517

Mapping QTL for grain yield, yield components, and spike features in a doubled haploid population of bread wheat

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Bahram Heidari, Badraldin Ebrahim Sayed-Tabatabaei, Ghodratollah Saeidi, Michael Kearsey, and Kazuhiro Suenaga

Abstract: A doubled haploid (DH) population derived from a cross between the Japanese cultivar ‘Fukuho-kumogi’ and the Israeli wheat line ‘Oligoculm’ was used to map genome regions involved in the expression of grain yield, yield components, and spike features in wheat (Triticum aestivum L). A total of 371 markers (RAPD, SSR, RFLP, AFLP, and two morphological traits) were used to construct the linkage map that covered 4190 cM of wheat genome including 28 linkage groups. The results of composite interval mapping for all studied traits showed that some of the quantitative trait loci (QTL) were stable over experiments conducted in 2004 and 2005. The major QTL located in the Hair–Xpsp2999 interval on chromosome 1A controlled the expression of grains/spike (R2 = 12.9% in 2004 and 22.4% in 2005), grain weight/spike (R2 = 21.4% in 2004 and 15.8% in 2005), and spike number (R2 = 15.6% in 2004 and 5.4% in 2005). The QTL for grain yield located on chromosomes 6A, 6B, and 6D totally accounted for 27.2% and 31.7% of total variation in this trait in 2004 and 2005, respectively. Alleles inherited from ‘Oligoculm’ increased the length of spikes and had decreasing effects on spike number. According to the data obtained in 2005, locus Xgwm261 was associated with a highly significant spike length QTL (R2 = 42.33%) and also the major QTL for spikelet compactness (R2 = 26.1%). Key words: doubled haploid, grain yield, QTL, spike features, wheat. Résumé : Une population de lignées haploïdes doublées (DH) dérivées d’un croisement entre le cultivar japonais ‘Fukuhokumogi’ et la lignée israélienne ‘Oligoculm’ a été employée pour identifier des régions du génome qui sont impliquées dans l’expression du rendement en grains, des composantes du rendement ainsi que des caractéristiques de l’épi chez le blé (Triticum aestivum L.). Au total, 371 marqueurs (RAPD, SSR, RFLP, AFLP et deux caractères morphologiques) ont été utilisés pour produire une carte génétique s’étendant sur 4190 cM répartis sur 28 groupes de liaison. Les résultats d’une analyse de cartographie d’intervalles composite pour tous les caractères à l’étude a montré que certains des locus de caractères quantitatifs (QTL) étaient stables pour les expériences réalisées en 2004 et 2005. Le QTL majeur situé dans l’intervalle Hair– Xpsp2999 sur le chromosome 1A contrôlait le nombre de grains/épi (R2 = 12,9 % en 2004 et 22,4 % en 2005), le poids en grains/épi (R2 = 21,4 % en 2004 et 15,8 % en 2005), ainsi que le nombre d’épis (R2 = 15,6 % en 2004 et 5,4 % en 2005). Les QTL pour le rendement en grains qui sont situés sur les chromosomes 6A, 6B et 6D expliquaient conjointement 27,2 % et 31,7 % de la variation pour ce caractère en 2004 et en 2005 respectivement. Les allèles du parent ‘Oligoculm’ contribuaient à augmenter la longueur des épis et à décroître le nombre d’épis. Sur la base des résultats obtenus en 2005, le locus Xgwm261 a été associé avec un QTL très hautement significatif (R2 = 42,33 %) pour la longueur des épis et avec un QTL majeur pour la densité des épillets (R2 = 26,1 %) Mots‐clés : haploïdes doublés, rendement en grains, QTL, caractéristiques de l’épi, blé. [Traduit par la Rédaction]

Introduction The progress in genetic improvement over the last century was based mainly on the use of single genes with relatively

clear-cut effects on the phenotype (Newbury 2003). As agronomic traits such as grain yield and its components show continuous variation and are controlled by quantitative genes, analysis of quantitative trait loci (QTL) is of importance for

Received 11 July 2010. Accepted 31 January 2011. Published at www.nrcresearchpress.com/gen on 02 June 2011. Corresponding Editor: Patrick Gulick. B. Heidari. Department of Crop Production and Plant Breeding, College of Agriculture, Shiraz University, Shiraz, 7144165186, Iran. B.E. Sayed-Tabatabaei. Department of Biotechnology, College of Agriculture, Isfahan University of Technology, Isfahan, 841568311, Iran. G. Saeidi. Department Plant Breeding, College of Agriculture, Isfahan University of Technology, Isfahan, 841568311, Iran. M. Kearsey. School of Biosciences, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom. K. Suenaga. Japan International Research Center for Agricultural Sciences, 1-1 Ohwashi, Tsukuba, Ibaraki, 305-8686, Japan. Corresponding author: Bahram Heidari (e-mail: [email protected]). Genome 54: 517–527 (2011)

doi:10.1139/G11-017

Published by NRC Research Press

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

518

plant breeders. Although it is possible to undertake breeding programs using only phenotypic selection, understanding the number and genomic location of genes controlling agronomically important traits can enhance the efficiency of selection (Kearsey and Pooni 1996). The use of molecular markers such as RAPD, RFLP, SSR, and AFLP to track QTL and locate them in genome regions in various crops is now widely applied in many laboratories (Bariana, et al. 2010; Hai et al. 2008; Lörz and Wenzel 2004; McIntyre et al. 2010; Sourdille et al. 2000). QTL can be mapped by following their cosegregation with molecular marker loci (Newbury 2003). The most widely used methods for identifying marker–trait associations (QTL analysis) are map construction, phenotyping and genotyping with molecular markers, or using segregation populations (Lörz and Wenzel 2004; van Eeuwijk et al. 2010). The most widely used segregating populations for identifying the QTL in crops are F2, backcross (BC1F1), recombinant inbred lines (RILs), and doubled haploid (DH) populations (Newbury 2003). DH populations are quicker to produce than RILs (Segman et al. 2006), when methods to produce DH are available. Statistical approaches developed for mapping QTL include interval mapping (Lander and Botstein 1989), marker regression (Kearsey and Hyne 1994), multiple regression (Haley and Knott 1992), and composite interval mapping (Zeng 1994; Jiang and Zeng 1995). In the composite interval mapping (CIM) approach, markers outside the region of the QTL are used as background controls or cofactors to eliminate the genetic variance caused by other QTL, and consequently they allow the residual variance to be reduced (Jansen 1996). Analysis of QTL has increased the knowledge of quantitative traits by associating genome information with phenotypic measurements. Kumar et al. (2007) used a set of RILs derived from different crosses to locate the QTL controlling grain yield and yield components in wheat. McCartney et al. (2005) used a DH population of wheat to study some economically important traits. They identified QTL for grain weight on chromosomes 2A, 3D, 4A, 4B, 4D, and 6D and QTL for grain yield on chromosomes 2A, 2B, 3D, 4A, and 4D. Using a linkage map consisting of 363 AFLP and 43 SSR markers, a number of QTL, 10, 16, and 14, respectively, were identified for grain yield, yield components, and spike morphology (Marza et al. 2006). In the study of Börner et al. (2002), two and three major QTL were detected for grain number/spike and grain weight, respectively. Suenaga et al. (2005) constructed a linkage map of wheat for QTL analysis and reported a major QTL controlling the spike length on chromosome 2D that explained 33.3% of the total phenotypic variation and was common in both greenhouse and field experiments. The intervarietal genetic map constructed using 131 RILs covered 3636.7 cM of the wheat genome, and 46 putative QTL were detected for grain yield, grain weight, spikelets/spike, and grains/spike (Li et al. 2007). In comparison, Balyan et al. (2005) found two QTL on chromosomes 7A and 2B controlling grain weight/spike, which were common in four environments. In the study of Hai et al. (2008), a total of 30 QTL were detected for grain yield and yield related parameters across four environments. Five of these QTL located on chromosomes 1A, 1B, 2B, 2D, and 7D exhibited pleiotropic effects for the traits studied.

Genome, Vol. 54, 2011

McIntyre et al. (2010) evaluated 194 wheat F7 sister lines and reported two putative grain yield QTL (log of odds (LOD) > 3) colocated with a putative QTL for increased stem carbohydrate content. The present study was conducted to evaluate 107 wheat DH lines along with their parents and to identify QTL for grain yield, yield components, and spike features in two growing seasons. The detection of stable QTL controlling yield related traits is expected to be useful in wheat breeding programs.

Materials and methods Mapping population A mapping population consisting of 107 DH lines was derived from the F1 of a cross between the Japanese cultivar ‘Fukuho-kumogi’ and an Israeli wheat line with “gigas” features, ‘Oligoculm’, (Atsmon and Jacobs 1977). The DH population was obtained by the method of crossing wheat to maize (Suenaga and Nakajima 1993). ‘Oligoculm’ has restricted tillering ability and large spike size; whereas ‘Fukuhokumogi’ has a normal spike size and tillering ability. These two parental varieties are very distant in their pedigree (Suenaga et al. 2005). Experimental design and field evaluations The parental lines along with 107 DH lines were evaluated under field conditions in a randomized complete block design with three replications in the growing seasons of 2004 and 2005 at the Research Farm of Isfahan University of Technology (32°32′N, 51°32′E), Isfahan, Iran. Each experimental plot in 2004 and 2005 consisted of four 2 m rows spaced 20 cm apart. Data for grains/spike, grain weight/spike (g), spike length (cm), spikelets/spike, and spikelet compactness (spikelets/cm spikes) were recorded from 10 randomly selected plants grown in the centre rows of each plot. Grain yield (g/m2) and fertile spikes/m2 were recorded from all plants harvested from the center parts of each plot. Also 1000-grain weight (g) was measured. Statistical analysis of field experimental data Data obtained from field evaluations in 2004 and 2005 were subjected to combined analysis using the GLM procedure of SAS software (SAS Institute Inc. 2000) to test the significance between years and genotypes (DH lines and the parents) by year interaction effects. Heritability of all the traits was estimated based on the variance components (Kearsey and Pooni 1996). Genetic correlation coefficients between grain yield related traits were calculated for data obtained in 2004 and 2005. The descriptive statistics of all traits measured in 2004 and 2005 and for combined data of 2 years were calculated. Molecular markers and map construction Twenty-seven detected AFLP markers were added to the existing framework map consisting of 344 RAPD, SSR, and RFLP markers, plus 2 morphological markers (glume colour and pubescence) created by Suenaga et al. (2005). For AFLP markers, a series of primer combinations of PstI + ANN and MseI + CNN were used to genotype 107 DH lines and their parents at the Biotechnology Laboratory of Isfahan UniverPublished by NRC Research Press

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Heidari et al.

519

sity of Technology, Isfahan, Iran. Those AFLP primer combinations that amplified polymorphic bands between the parents were selected to genotype the 107 lines of the DH population. The nomenclature of AFLP markers (polymorphic bands) was set according to their size and also related primer combination (p and m denote for PstI and MseI primers, respectively). All newly detected AFLP markers and 344 previously available markers (Suenaga et al. 2005) were used and analyzed with the JoinMap® 3.0 software (Van Ooijen and Voorrips 2001). Those marker loci that were determined as identical by the similarity of loci command in this software were excluded. Also, markers with high segregation distortion (Kammholz et al. 2001) were omitted from the analysis based on the c2 test (P < 0.001). The selected markers by the primary analysis of JoinMap® (Van Ooijen and Voorrips 2001) were further used to construct linkage groups using MAPMAKER/EXP 3.0b (Lander et al. 1987). First, the ANCHOR command was used to assign anchor loci to each chromosome. Subsequently, the assignment of other markers to chromosomes was based on a LOD of ≥3. Within each linkage group, the best order of markers was determined according to the maximum likelihood method, and the genetic distances (cM) were calculated using the Kosambi (1944) function.

ity estimates in 2005 were lower than in 2004 for most of the traits (Table 1). The genetic correlation coefficients between traits studied are shown in Table 3. The highest correlation was observed between grains/spike and grain weight/spike in 2004 (r = 0.77, P < 0.01) and in 2005 (r = 0.89, P < 0.01). Grain yield was significantly correlated with grains/spike (r = 0.21* in 2004 and 0.23* in 2005), grain weight (r = 0.33** in 2004 and 0.44** in 2005), and spikes/m2 (r = 0.40** in 2004 and 0.17 in 2005).

QTL detection For each trait evaluated in 2004 and 2005, QTL mapping was conducted using a CIM approach (Zeng 1994; Jiang and Zeng 1995) and Windows QTL Cartographer 2.5 software (Wang et al. 2007). A forward selection, backward elimination, and stepwise regression with model 6 was used to select cofactors for CIM. A 10 cM scan window was used for analysis and the LOD statistic was computed at a walk speed of 2 cM. To identify significant threshold values (LOD) and significant QTL, a permutation test (Churchill and Doerge 1994; Doerge and Churchill 1996) with 1000 random samples was performed, but LOD thresholds between 2 and 3 were also considered for detection of minor QTL over years. Also, because of the high number of detected QTL, the LOD plots were only presented for major and (or) stable QTL of some traits.

Grains/spike The results of permutation tests showed that the significant LOD scores for declaring the presence of a QTL for grains/ spike were 3.2 and 2.8 for the data of 2004 and 2005, respectively (Table 4). Based on these thresholds, a major QTL (R2 = 12.9%; LOD = 5.17) was detected in 2004 in the Hair–Xpsp2999 interval on chromosome 1A (Fig. 1a). This QTL was also detected at the same interval in 2005 (LOD = 7.34; Fig. 1b) and explained 22.4% of the phenotypic variation of grains/spike (Table 4). According to the estimates of additive effects (–5.04 in 2004 and –6.25 in 2005), ‘Oligoculm’ alleles had increasing effects on grains/ spike in the Hair–Xpsp2999 interval. The QTL for this trait identified on chromosome 2B was only expressed in 2004 and explained 12.2% of the total variation.

Results Agronomic performance of DH lines and their parents The analyses of variance (Table 1) showed that the effect of genotype and genotype by year interactions were significant for all traits. Phenotypic performance of DH lines and related descriptive statistics (Table 2) indicates that ‘Fukuhokumogi’ and ‘Oligoculm’ had significantly different performances for all traits. ‘Oligoculm’ had greater values for grains/ spike, grain weight/spike, 1000-grain weight, spike length, and spikelets/spike in 2 years; whereas the performance of ‘Fukuho-kumogi’ was much better than ‘Oligoculm’ for grain yield, spikes/m2, and spikelet compactness. The range of variation of the DH lines showed transgressive segregation (Falconer and Mackay 1996) in most of the traits investigated. The heritability estimates were highest for 1000-grain weight (93%) and grains/spike (92%), but were lowest for grain yield (63%) and spikes/m2 (68%) in 2004 (Table 1). The heritabil-

Linkage map A total of 371 loci, including 27 newly detected AFLP markers and 344 markers on the map by Suenaga et al. (2005), resulted in a map with 28 linkage groups constructed by MAPMAKER/EXP 3.0b (Lander et al. 1987). Three linkage groups were not assigned to any chromosome, and chromosomes 3A, 6A, 3D, and 7D were defined by two linkage groups. The total map length was 4190 cM including three unassigned linkage groups (92 cM). The mean interval between loci was 11.29 cM. Among the 371 loci mapped to the 28 linkage groups, only 93 loci (25%) were assigned to genome D compared with 120 (32%) and 147 (39%) loci assigned to genomes A and B, respectively. QTL identified for grain yield and its components

1000-grain weight Four and three putative QTL were detected in 2004 and 2005, respectively (Table 4). The putative QTL on chromosome 4B positioned in the Xgwm935b–Xgwm48c interval accounted for 21.5% and 15.2% of the total variation of 1000-grain weight in 2004 and 2005, respectively. Mapping QTL on chromosome 4B using data obtained in 2005 showed that there were two other QTL in the RhtB1– Xgwm935b and Xgwm495–Xgwm149 intervals that accounted for 12.9% and 10.8% of the total variation, respectively. Three significant QTL located on chromosomes 2B (R2 = 15.7% and 13.3%) and 2D (R2 = 12.1%) were also identified in 2004. Grain weight/spike A major and stable QTL (LOD = 7.44 in 2004 and 5.53 in 2005) was detected at the interval of Hair–Xpsp2999 on chromosome 1A. This QTL explained 21.4% of the total phenotypic variation of this trait in 2004 (Table 4). Although this Published by NRC Research Press

520

Genome, Vol. 54, 2011

Table 1. Pooled analyses of variance over years (2004 and 2005) and derived heritabilities for grain yield, yield components, and spike features. Mean squares

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Source of variation Year Rep (year) Genotype Genotype × year Error Heritability 2004 2005

df 1 4 108 108 432

Grains/ spike 593** 151.6 863.9** 146.7** 48.4

1000-grain weight (g) 536** 55.02 80.10** 14.83** 8.86

Grain weight/ spike (g) 1.29** 1.86 1.63** 0.27** 0.14

Spikes/m2 12 114 056** 16 040 72 504** 34 107** 20 288

Grain yield (g/m2) 23 801 453** 56 834 66 400** 44 830** 26 346

Spike length (cm) 6011** 3.3 18.9** 3.1** 1.2

Spikelets/ spike 119.7** 3.4 13.4** 2.29** 0.89

Spikelet compactness 67.5** 0.01 0.20** 0.04** 0.01

0.92 0.86

0.93 0.63

0.89 0.80

0.68 0.39

0.63 0.32

0.89 0.87

0.86 0.87

0.90 0.83

Note: **, significant at 1% level of probability; Rep (year), replication within year.

Table 2. The range of grain yield, yield components, and spike features for ‘Fukuho-kumogi’ (F) and ‘Oligoculm’ (O) and 107 doubled haploid (DH) lines in 2004 and 2005. Parentsa Trait Grains/spike

1000-grain weight (g)

Grain weight/spike (g)

Spikes/m2

Grain yield (g/m2)

Spike length (cm)

Spikelets/spike

Spikelet compactness

Experiment 2004 2005 M 2004 2005 M 2004 2005 M 2004 2005 M 2004 2005 M 2004 2005 M 2004 2005 M 2004 2005 M

F 58.4a 55.5a 56.9 37.0a 32.7a 34.9 2.29a 2.04a 2.16 848.7a 586.5a 717.6 1116.4a 651.0a 883.7 15.1a 9.5 12.3 18.1a 19.8a 19.0 1.20a 2.08a 1.64

DH lines O 81.2b 95.6b 88.4 41.8b 40.3b 41.0 3.62b 5.30b 4.46 490.3b 217.6b 354.0 1037.7a 625.16a 831.4 22.1b 19.0b 20.5 19.8a 22.9b 21.3 0.89b 1.20b 1.05

Min. 25.5 38.2 40.8 25.7 25.2 25.5 1.13 1.15 1.14 350.0 195.6 289.0 461.1 277.1 497.4 13.7 8.5 11.4 16.3 17.0 16.8 0.79 1.29 1.05

Max. 42.9 101.7 124.3 45.0 42.5 42.9 3.98 4.86 4.04 1012.0 1075.3 885.3 1385.8 785.4 1005.6 23.6 15.4 18.5 23.5 25.9 24.2 1.47 2.32 1.86

Mean 34.6 61.6 59.4 35.5 33.7 34.6 2.34 2.26 2.30 702.6 430.7 566.7 896.8 515.1 705.9 17.9 11.8 14.8 19.3 20.1 19.7 1.09 1.74 1.41

SD 3.6 12.6 12.8 4.3 3.4 3.6 0.48 0.54 0.46 146.3 116.0 107.9 162.2 102.5 103.9 2.1 1.5 1.6 1.5 1.6 1.4 0.13 0.24 0.18

CV (%) 20.5 21.6 19.4 12.3 10.3 10.4 20.8 24.0 20.3 20.8 26.9 19.0 18.0 19.9 14.7 11.7 12.7 11.3 8.0 8.2 7.5 12.5 14.2 12.7

Note: M, mean over years; SD, standard deviation of the mean; CV, coefficient of variation of the traits. Parental means with different letters (a and b) in each year have significant differences based on LSD test.

a

QTL was identified in 2005, its expression (R2 = 15.8%) was slightly lower than the experiment conducted in 2004. The second major QTL (R2 = 10.6%) detected in the vicinity of Xbarc124c on chromosome 2D in 2004 was not detected in 2005. The grain weight/spike was controlled by another two QTL located on chromosomes 5B (R2 = 7.7%) in 2004 and 4A (R2 = 7.6%) in 2005. At the location of the QTL identified on chromosomes 1A and 4A, the ‘Oligoculm’ allele increased grain weight/spike in the DH population, whereas the ‘Fukuho-kumogi’ allele of the QTL on chromosome 2D had an increasing effect.

Spikes/m2 The only major QTL expressed for spikes/m2 in 2004 was located in the Hair–Xpsp2999 interval on chromosome 1A (Table 4). This QTL accounted for 15.6% of the total variation of spikes/m2 and was the only stable QTL detected in 2005, although its expression was much lower (R2 = 5.4%) compared with the 1st year. The estimate of additive effect at the location of this QTL showed that ‘Oligoculm’ alleles decreased spikes/m2. Although all the QTL located on chromosomes 1A, 7A, and 2D did not meet the LOD threshold determined by the permutation test (LOD = 3.2), they nonePublished by NRC Research Press

Heidari et al.

521

Table 3. Genetic correlation between grain yield related traits in 2004 (above diagonal) and 2005 (under diagonal) in 107 wheat doubled haploid (DH) lines and their parents.

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Trait Grains/spike 1000-grain weight (g) Grain weight/spike (g) Spikes/m2 Grain yield (g/m2) Spike length (cm) Spikelets/spike Spikelet compactness

Grains/ spike 1 –0.03 0.89 –0.73 0.23 0.28 0.61 0.11

1000-grain weight (g) –0.51 1 0.37 –0.17 0.28 0.01 –0.17 –0.12

Grain weight/ spike (g) 0.77 0.12 1 –0.73 0.44 0.32 0.44 –0.03

Spikes/m2 –0.62 0.07 –0.71 1 0.17 –0.27 –0.39 0.08

Grain yield (g/m2) 0.21 0.18 0.33 0.40 1 0.07 0.12 0.06

Spike length (cm) 0.35 –0.32 0.26 –0.23 0.04 1 0.18 –0.78

Spikelets/ spike 0.64 –0.56 0.33 –0.35 0.05 0.21 1 0.43

Spikelet compactness 0.04 –0.03 –0.05 0.02 0.02 –0.85 0.39 1

Note: The absolute value of numbers greater than 0.25 and 0.19 are significantly different from 0 at 0.01 and 0.05, respectively.

theless explained 29.1% of the total phenotypic variation of spikes/m2 in 2005. Grain yield Although according to results of the permutation test (LOD = 3.1), the QTL detected close to the AFLP marker (Xp02m22-3) for grain yield was not significant (LOD = 1.18), and it explained 11.5% of its variation in 2004 (Table 4). The ‘Oligoculm’ alleles close to this QTL increased grain yield. There were two minor QTL located in the Xbarc173–Xgwm469 interval and close to the Xgwm169 that controlled expression of grain yield in 2004. According to the data of 2005, a major and significant QTL (R2 = 20.9%) located on the second linkage group of chromosome 6A was in the vicinity of Xp10m31-10. Alleles of this QTL inherited from the ‘Fukuho-kumogi’ cultivar increased grain yield. The second QTL identified in 2005 was positioned in the Xgwm469–Xwmc113b interval and contributed in 11.8% of the total phenotypic variation of grain yield. QTL for spike features Spike length Near the Xcfd53 locus on chromosome 2D (Fig. 2a) there was a major QTL (LOD = 10.1) that accounted for 37.5% of spike length variation in 2004. In the 2nd year of evaluation, this QTL was not detected exactly in the same location (Fig. 2b), but its position was in the Xgwm261–Xcfd53 interval and had the largest contribution (R2 = 42.33%) to spike length (Table 5). Three QTL were found on chromosome 6A2 that explained 8.5%, 8.0%, and 7.6% of spike length variation in 2004. The QTL positioned in the Xp10m31-13– Xgwm497d interval was the only common QTL detected over years. Chromosome 1A had a significantly expressed QTL (LOD = 4.10) in the Hair–Xpsp2999 interval in 2005. According to the estimates of additive effects (Table 5), alleles inherited from ‘Oligoculm’ increased spike length for all QTL detected in 2004 and 2005. Spikelets/spike Four significant QTL for spikelets/spike were detected in 2004 located on chromosomes 4B, 2A, and 1A, whereas the only significant QTL in 2005 was on chromosome 1A (Table 5). The QTL positioned in the Xta556–RhtB1 interval and close to the RhtB1, Xgwm71a, and Xwmc59 markers were the most important and significant QTL that explained 22.5%, 18.4%, 10%, and 9.6% of variation in 2004. Both pa-

rental cultivars contributed alleles of increasing effect on spikelets/spike. In the 2nd year of evaluation, the only major QTL (R2 = 10.8%) with a LOD score of 3.91 was located in the Hair–Xpsp2999 interval. Mapping QTL for data obtained in 2005 showed that there were three minor QTL (with 2 ≤ LOD ≤ 3) located at chromosomal regions that were common over years (Table 5). Spikelet compactness The highest numbers of stable QTL over years were detected for spikelet compactness. These common QTL were located on chromosomes 2A, 2D, and 6A2 (Table 5). On chromosome 2D there was a putative QTL positioned near Xcfd53 (R2 = 22.4%) in 2004 and in the Xgwm261–Xcfd53 interval (R2 = 26.1%) in 2005 (approximate distance between these loci was 3 cM, therefore, they are probably the same QTL). The Xwmc111–Xgwm296 interval was the second important region of chromosome 2D for expression of spikelet compactness. A major QTL detected in this interval was identified in the evaluations in both years. It was located in the Xgwm71a–Xgwm328 interval on chromosome 2A and had lower expression in 2005 (R2 = 8.0%) than 2004 (R2 = 13.5%). The linkage group assigned to chromosome 6A (or 6A2) carried two significant QTL (LOD > 4), one in the Xpsp3029b–Xwmc256 interval in both years and the other in the Xgwm497d–Xpsp3029b interval in 2005 alone. The QTL positioned on chromosome 4B was specific to the 2004 growing season. ‘Fukuho-kumogi’ alleles at all QTL increased spikelet compactness in the DH lines.

Discussion In the present study there were some stable QTL whose expression changed because of environmental conditions over years (2004 and 2005). In addition, some of QTL were only expressed in one of the years, probably because of genotype by environment interaction. Therefore, the results show the importance of trait evaluation in experiments replicated across environments. Paterson et al. (1991) also showed that the experiments conducted in a single environment (year or location) underestimate the number of QTL controlling a certain trait, whereas repeated experiments are useful in determining stable QTL over environments. Also, the effect of QTL may vary among different environments becasue of QTL by environment interactions. Grain yield is one of the most complex traits with its expression strongly influenced by environment. Therefore, it is Published by NRC Research Press

522

Genome, Vol. 54, 2011

Table 4. The results of composite interval mapping analyses for detection of QTL for grain yield and its components in ‘Fukuho-kumogi’ × ‘Oligoculm’ DH population in 2004 and 2005. Trait Grains/spike

Year

Chr.a

2004

1A 2B

2.01 99.51

Hair–Xpsp2999 Xgwm71b–Xgwm429

2005

1A

2.01

Hair–Xpsp2999

2004

4B 2B 2B 2D

71.11 140.51 127.11 64.21

Xgwm935b–Xgwm48.c Xgwm55 Xwmc344–Xac14 Xcfd53–Xbarc168

2005

4B 4B 4B

77.11 62.91 88.01

Xgwm935b–Xgwm48.c RhtB1–Xgwm935b Xgwm495–Xgwm149

2004

1A 2D 5B

2.01 0.01 197.41

Hair–Xpsp2999 Xbarc124c Xwmc28–Xgwm790a

2005

1A 4A

4.01 21.21

Hair–Xpsp2999 Xpsp3028

2004

1A

0.01

Hair–Xpsp2999

2005

7A 2D 1A

135.71 207.81 0.01

Xgwm282–Xwmc116 Xgwm349–Xgwm320 Hair–Xpsp2999

2004

6A2 6D 6A2

0.01 21.11 245.41

Xp02m22–3 Xbarc173–Xgwm469 Xgwm169

2005

6A2 6D

43.71 53.01

Xp10m31-10 Xgwm469–Xwmc113b

Position (cM)

Closest markers or interval

LOD

R2 (%)b

Additive effect

5.17 4.54 TR = 3.0c 7.34 TR = 2.8

12.9 12.2

–5.04 4.99

22.4

–6.25

8.62 6.56 5.50 4.05 TR = 3.3 4.97 4.00 3.36 TR = 3.0

21.5 15.7 13.3 12.1

–2.16 –1.76 –1.61 1.54

15.2 12.9 10.8

–1.46 –1.28 –1.17

7.44 4.13 3.02 TR = 3.0 5.53 2.83 TR = 3.0

21.4 10.6 7.7

–0.24 0.16 0.13

15.8 7.6

–0.22 –0.15

4.92 TR = 3.1 2.42 2.79 1.83 TR = 3.2

15.6

59.23

8.4 8.3 5.4

–34.05 –33.74 27.70

1.18 2.76 2.09 TR = 3.1 3.06 2.44 TR = 3.0

11.5 9.2 6.5

–56.11 –51.03 43.05

20.9 11.8

47.87 –36.69

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

1000-grain weight (g)

Grain weight/spike (g)

Spikes/m2

Grain yield (g/m2)

Note: Stable QTL over years are bolded. Chr. denotes the chromosomes. b Explained phenotypic variation by each QTL. c Threshold value (TR) declaring the presence of a QTL is significant based on permutation test. a

much better to study this trait by focusing on its components (Bezant et al. 1997). According to field evaluations of 2004 and 2005, the major QTL detected in the Hair-Xpsp2999 interval (on chromosome 1A) controlled the expression of both grains/spike and grain weight/spike. This coincidence may be due to either pleiotropy or genetic linkage. In this same interval there was also a major QTL for spikelets/spike identified in the 2005 trial. Apart from the QTL located on chromosome 1A, the other QTL detected in 2004 for grains/spike and grain weight/spike were not identified in 2005 because of environmental conditions. Studies reported by several authors identified QTL for grain yield and its components (Börner et al. 2002; Kato et al. 2000; Kumar et al. 2007;

Sourdille et al. 2003; McCartney et al. 2005; Marza et al. 2006; Li et al. 2007). Grain yield was the only trait measured in this study that had no stable QTL over 2004 and 2005. The QTL identified to control the expression of grain yield only accounted for 27.2% and 31.7% of total variation of this trait in 2004 and 2005, respectively. These results are in agreement with the results obtained by Kumar et al. (2007), Li et al. (2007), and Kato et al. (2000), indicating no stable QTL for grain yield in replicated experiments, but McIntyre et al. (2010) reported stable yield QTL on chromosomes 6D and 7A using wheat F7 sister lines evaluated in three locations. Identified QTL on homoeologous group 6 (chromosomes 6A, 6B, and 6D) Published by NRC Research Press

Heidari et al.

523

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Fig. 1. The LOD plots showing the location (arrows: peak of the curves) of major and stable QTL of grains/spike (a and b in the experiment 2004 and 2005, respectively) on chromosome 1A. Horizontal and vertical axes represent the length of linkage group (cM) and LOD values, respectively.

Published by NRC Research Press

524

Genome, Vol. 54, 2011

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Fig. 2. The LOD plots showing the location (arrows: peak of the curves) of major and stable QTL of spike length on chromosome 2D (a and b in the experiment 2004 and 2005, respectively). Horizontal and vertical axes represent the length of linkage group (cM) and LOD values, respectively.

Published by NRC Research Press

Heidari et al.

525

Table 5. The results of composite interval mapping analyses for detection of QTL of spike characters in ‘Fukuho-kumogi’ × ‘Oligoculm’ doubled haploid (DH) population in 2004 and 2005.

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Trait Spike length (cm)

Year

Chr.a

Position (cM)

Closest markers or interval

LOD

R2 (%)b

Additive effect

2004

2D 6A2 6A2 6A2

52.21 188.51 198.41 195.11

Xcfd53 Xp10m31-13–Xgwm497d Xpsp3029b Xgwm497d–Xpsp3029b

2005

2D 6A2 1A

39.31 188.51 4.01

Xgwm261–Xcfd53 Xp10m31-13–Xgwm497d Hair–Xpsp2999

10.15 2.91 3.16 3.16 TRc = 3.3 16.31 4.84 4.10 TR = 2.8

37.5 8.5 8.0 7.6

–1.30 –0.64 –0.64 –0.61

42.33 11.79 8.17

–1.01 –0.70 –0.45

2004

4B 4B 2A 1A 1A 2A

53.61 62.91 117.61 173.61 160.71 106.01

Xta556–RhtB1 RhtB1 Xgwm71a Xwmc59 Xgwm497b–Xgwm99a Xag8a–Xgwm339

6.70 6.48 3.89 3.77 2.77 2.65 TR = 3.1 3.91 2.15 2.19 2.03 2.21 TR = 3.0

22.5 18.4 10.0 9.6 8.8 7.6

0.75 0.69 0.50 –0.49 –0.47 0.43

2005

1A 1A 1A 2A 1A

0.01 179.61 162.71 123.61 169.31

Hair–Xpsp2999 Xwmc59 Xgwm497b–Xgwm99a Xgwm71a–Xgwm328 Xgwm99a–Xwmc59

10.8 7.3 7.1 6.3 6.3

–0.56 –0.45 –0.46 0.42 –0.43

2004

2D 2D 2A 6A2 4B

40.21 29.41 125.61 200.41 64.91

Xgwm261–Xcfd53 Xwmc111–Xgwm296 Xgwm71a–Xgwm328 Xpsp3029b–Xwmc256 RhtB1–Xgwm935b

9.89 6.15 5.94 4.38 3.97 TR = 3.0 10.45 7.41 4.61 4.07 3.47 TR = 3.0

22.4 15.9 13.5 9.7 8.3

0.06 0.05 0.05 0.04 0.04

2005

2D 2D 6A2 6A2 2A

39.31 27.41 202.41 195.11 123.61

Xgwm261–Xcfd53 Xwmc111–Xgwm296 Xpsp3029b–Xwmc256 Xgwm497d–Xpsp3029b Xgwm71a–Xgwm328

26.1 20.4 11.4 9.1 8.0

0.12 0.11 0.09 0.08 0.07

Spikelets/spike

Spikelet compactness

Note: Stable QTL over years are bolded. Chr. denotes the chromosomes. b Explained phenotypic variation by each QTL. c Threshold value (TR) declaring the presence of a QTL is significant based on permutation test. a

indicated the importance of these chromosomes in expression of grain yield in the present study. The linkage groups assigned to genome A (1A, 2A, and 7A) had the largest contribution to variation in spikelets/ spike. Based on evaluation of DH lines in 2004, 76.9% of phenotypic variation of spikelets/spike was determined by six QTL. In 2005, this amount was decreased to 49.3%. There were some QTL mapped in the Xta556–RhtB1, RhtB1–Xgwm935b, and Hair–Xpsp2999 intervals that simultaneously controlled spikelets/spike and also grains/spike. The gigas features (large spike and restricted tillering) of ‘Oligoculm’ were analyzed in both the 2004 and 2005 experiments. The prominent and stable QTL in the Hair– Xpsp2999 (chromosome 1A) and Xgwm261–Xcfd53 (chromosome 2D) intervals detected for spikes/m2 and spike length

were also identified in the study of Suenaga et al. (2005) using the same DH population as in the present study. Therefore, it can safely be concluded that these QTL are robust across environments. Richards (1988) also reported a gene for tiller inhibition (Tin) on chromosome 1A in wheat lines of Israeli origin. Atsmon and Jacobs (1978) also reported that in wheat, uniculm and ‘Oligoculm’ are under polygenic control. The Xgwm261 locus is known to be closely linked to the semi dwarf gene Rht8, which is possessed by ‘Fukuhokumogi’ (Worland et al. 1998). Also, the presence of a QTL (R2 = 7.7%) for spike length was reported near the Xgwm261 locus by Sourdille et al. (2003). Detection of a common QTL for spike number and spike length in the Hair–Xpsp2999 interval in the present study and also in the reports of Atsmon and Jacobs (1977) and Suenaga et Published by NRC Research Press

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

526

al. (2005) indicates the relationship between restricted tillering and large spikes in wheat. The size of the map constructed in the present study (4190 cM) is comparable to those obtained by Nelson et al. (1995) and Marino et al. (1996) with 3551 and 4110 cM, respectively. Based on the results of this study, the Hair–Xpsp2999 interval and Xgwm261 locus were most frequently associated with QTL of various traits. The CIM approach indicated the Hair–Xpsp2999 interval carried QTL for grains/spike, grain weight, spikes/m2, spike length, and spikelets/spike, whereas the Xgwm261 locus was close to major QTL for spike length and spikelet compactness. In theory, all markers linked to QTL are potentially valuable for an indirect selection program, but those markers that are codominant (SSR or RFLP) and tightly linked to more than one QTL are of more value in marker-assisted selection (Collard et al. 2005; Mohan et al. 1997). Therefore, according to the results of the CIM approach, SSR markers such as Xgwm261 and Xpsp2999, which were close to the QTL of most of the traits, would be more useful to increase efficiency of indirect selection in the breeding program of wheat.

Acknowledgment The authors appreciate the French National Institute of Agricultural Research (INRA) that provided the primer sequences of cfd markers and related information.

References Atsmon, D., and Jacobs, E. 1977. A newly bred ‘gigas’ form of bread wheat (Triticum aestivum L.): morphological features and thermophotoperiodic responses. Crop Sci. 17(1): 31–35. doi:10.2135/ cropsci1977.0011183X001700010010x. Atsmon, D., and Jacobs, E. 1978. Genetic and environmental factors affecting tillering and yield components in uni- and ‘Oligoculm’ bread wheats and in their crosses with standard varieties. In Proceedings of the 5th International Wheat Genetics Symposium, New Delhi, India, 23–28 February 1978. Edited by S. Ramanujam. Indian Society of Genetic and Plant Breeding, New Delhi. pp. 1002–1003. Balyan, H.S., Gupta, P.K., Kulwal, P.T., and Kumar, N. 2005. QTL analysis for three grain quality traits in bread wheat using intervarietal mapping population.Czech J. Genet. Plant. Breed. 41: 281–283. Bariana, H.S., Bansal, U.K., Schmidt, A., Lehmensiek, A., Kaur, J., Miah, H., et al. 2010. Molecular mapping of adult plant stripe rust resistance in wheat and identification of pyramided QTL genotypes. Euphytica, 176(2): 251–260. doi:10.1007/s10681010-0240-x. Bezant, J., Laurie, D., Pratchett, N., Chojecki, J., and Kearsey, M.J. 1997. Mapping QTLs controlling yield and yield components in barley (Hordeum vulgare L.) cross using marker regression. Mol. Breed. 3(1): 29–38. doi:10.1023/A:1009648220852. Börner, A., Schumann, E., Fürste, A., Cöster, H., Leithold, B., Röder, M.S., and Weber, W.E. 2002. Mapping of quantitative trait loci determining agronomic important characters in hexaploid wheat (Triticum aestivum L.). Theor. Appl. Genet. 105(6–7): 921–936. doi:10.1007/s00122-002-0994-1. PMID:12582918. Churchill, G.A., and Doerge, R.W. 1994. Empirical threshold values for quantitative trait mapping. Genetics, 138(3): 963–971. PMID: 7851788. Collard, B.C.Y., Jahufer, M.Z.Z., Brouwer, J.B., and Pang, E.C.K. 2005. An introduction to markers, quantitative trait loci (QTL)

Genome, Vol. 54, 2011 mapping and marker-assisted selection for crop improvement: the basic concept. Euphytica, 142(1–2): 169–196. doi:10.1007/ s10681-005-1681-5. Doerge, R.W., and Churchill, G.A. 1996. Permutation tests for multiple loci affecting a quantitative character. Genetics, 142(1): 285–294. PMID:8770605. van Eeuwijk, F.A., Bink, M.C., Chenu, K., and Chapman, S.C. 2010. Detection and use of QTL for complex traits in multiple environments. Curr. Opin. Plant Biol. 13(2): 193–205. doi:10. 1016/j.pbi.2010.01.001. PMID:20137999. Falconer, V.A., and Mackay, T.F.C. 1996. Introduction to quantitative genetics. Ronald Press, New York. Hai, L., Guo, H., Wagner, C., Shihe, X., and Friedt, W. 2008. Genomic regions for yield and yield parameters in Chinese winter wheat (Triticum aestivum L.) genotypes tested under varying environments correspond to QTL in widely different wheat materials. Plant Sci. 175(3): 226–232. doi:10.1016/j.plantsci. 2008.03.006. Haley, C.S., and Knott, S.A. 1992. A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity, 69(4): 315–324. PMID:16718932. Jansen, R.C. 1996. Complex plant traits: time for polygenic analysis. Trends Plant Sci. 1(3): 89–94. doi:10.1016/S1360-1385(96) 80040-9. Jiang, C., and Zeng, Z.B. 1995. Multiple-trait analysis of genetic mapping for quantitative trait loci. Genetics, 140(3): 1111–1127. PMID:7672582. Kammholz, S.J., Campbell, A.W., Sutherland, M.W., Hollamby, G.J., Martin, P.J., Eastwood, R.F., et al. 2001. Establishment and characterisation of wheat genetic mapping populations. Aust. J. Agric. Res. 52(12): 1079–1088. doi:10.1071/AR01043. Kato, K., Miura, H., and Sawada, S. 2000. Mapping QTLs controlling grain yield and its components on chromosome 5A of wheat. Theor. Appl. Genet. 101(7): 1114–1121. doi:10.1007/ s001220051587. Kearsey, M.J., and Hyne, V. 1994. QTL analysis: a simple marker regression approach. Theor. Appl. Genet. 89(6): 698–702. doi:10. 1007/BF00223708. Kearsey, M.J., and Pooni, H.S. 1996. The genetical analysis of quantitative traits. Chapman and Hall, London, UK. Kosambi, D.D. 1944. The estimation of map distances from recombination values. Ann. Eugen. 12: 172–175. Kumar, N., Kulwal, P.L., Balyan, H.S., and Gupta, P.K. 2007. QTL mapping for yield and yield contributing traits in two mapping population of bread wheat. Mol. Breed. 19(2): 163–177. doi:10. 1007/s11032-006-9056-8. Lander, E.S., and Botstein, D. 1989. Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics, 121(1): 185–199. PMID:2563713. Lander, E.S., Green, P., Abrahamson, J., Barlow, A., Daly, M.J., Lincoln, S.E., and Newberg, L. 1987. MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics, 1(2): 174–181. doi:10.1016/0888-7543(87)90010-3. PMID: 3692487. Li, S., Jia, J., Wei, X., Zhang, X., Li, L., Chen, H., et al. 2007. A intervarietal genetic map and QTL analysis for yield traits in wheat. Mol. Breed. 20(2): 167–178. doi:10.1007/s11032-0079080-3. Lörz, H., and Wenzel, G. (Editors) 2004. Molecular marker systems in plant breeding and crop improvement. Springer-Verlag, Berlin, Heidelberg, New York. Marino, C.L., Nelson, J.C., Lu, Y.H., Sorrells, M.E., Leroy, P., Tuleen, N.A., et al. 1996. Molecular genetic maps of the group 6 Published by NRC Research Press

Genome Downloaded from www.nrcresearchpress.com by Peking University on 06/03/13 For personal use only.

Heidari et al. chromosomes of hexaploid wheat (Triticum aestivum L. em. Thell.). Genome, 39(2): 359–366. doi:10.1139/g96-046. PMID: 18469899. Marza, F., Bai, G.H., Carver, B.F., and Zhou, W.C. 2006. Quantitative trait loci for yield and related traits in wheat population Ning7840 × Clark. Theor. Appl. Genet. 112(4): 688– 698. doi:10.1007/s00122-005-0172-3. PMID:16369760. McCartney, C.A., Somers, D.J., Humphreys, D.G., Lukow, O., Ames, N., Noll, J., et al. 2005. Mapping quantitative trait loci controlling agronomic traits in spring wheat cross RL4452 × ‘AC Domain’. Genome, 48(5): 870–883. doi:10.1139/g05-055. PMID:16391693. McIntyre, C.L., Mathews, K.L., Rattey, A., Chapman, S.C., Drenth, J., Ghaderi, M., et al. 2010. Molecular detection of genomic regions associated with grain yield and yield-related components in an elite bread wheat cross evaluated under irrigated and rainfed conditions. Theor. Appl. Genet. 120(3): 527–541. doi:10.1007/ s00122-009-1173-4. PMID:19865806. Mohan, M., Nair, S., Bhagwat, A., Krishna, T.G., Yano, M., Bhatia, C.R., and Sasaki, T. 1997. Genetic mapping, molecular markers and marker assisted selection in crop plants. Mol. Breed. 3(2): 87– 103. doi:10.1023/A:1009651919792. Nelson, J.C., Sorrells, M.E., van Deyze, A.E., Lu, Y.H., Atinkson, M., Bernard, M., Leory, P., Faris, J.D., and Anderson, J.A. 1995. Molecular mapping of wheat: major genes and rearrangements in homoeologous groups 4, 5, and 7. Genetics, 141(2): 721–731. PMID:8647405. Newbury, H.J. 2003. Plant molecular breeding. CRC Press, USA. Paterson, A.H., Damon, S., Hewitt, J.D., Zamir, D., Rabinowitch, H. D., Lincoln, S.E., et al. 1991. Mendelian factors underlying quantitative traits in tomato: comparison across species, generations, and environments. Genetics, 127(1): 181–197. PMID: 1673106. Richards, R.A. 1988. A tiller inhibitor gene in wheat and its effect on plant growth. Aust. J. Agric. Res. 39(5): 749–757. doi:10.1071/ AR9880749.

527 SAS Institute Inc. 2000. The SAS System for Windows. Release 8.01, SAS Institute Inc., Cary, NC. Segman, K., Bjørnstad, A., and Ndjiondjop, N. 2006. Principles, requirements and prospects of genetic mapping in plants. Afr. J. Biotechnol. 5(25): 2569–2587. Sourdille, P., Snape, J.W., Cadalen, T., Charmet, G., Nakata, N., Bernard, S., and Bernard, M. 2000. Detection of QTLs for heading time and photoperiod response in wheat using a doubled-haploid population. Genome, 43(3): 487–494. doi:10.1139/gen-43-3-487. PMID:10902713. Sourdille, P., Cadalen, T., Guyomarc’h, H., Snape, G.W., Perretnat, M.R., Charmet, G., et al. 2003. An update of Courtot × Chinese spring intervarietal molecular marker linkage map for the QTL detection of agronomic traits in wheat. Theor. Appl. Genet. 106(3): 530–538. doi:10.1007/s00122-002-1044-8. PMID:12589554. Suenaga, K., and Nakajima, K. 1993. Segregation of genetic markers among wheat doubled haploid lines derived from wheat × maize crosses. Euphytica, 65(2): 145–152. doi:10.1007/BF00022576. Suenaga, K., Khairallah, M., William, H.M., and Hoisington, D.A. 2005. A new intervarietal linkage map and its application for quantitative trait locus analysis of “gigas” features in bread wheat. Genome, 48(1): 65–75. doi:10.1139/g04-092. PMID:15729398. Van Ooijen, J.W., and Voorrips, R.E. 2001. JoinMap® 3.0, software for the calculation of genetic linkage maps. Plant Research International, Wageningen, the Netherlands. Wang, S., Basten, C.J., and Zeng, Z.B. 2007. Windows QTL Cartographer V2.5 user manual. Bioinformatic Research Center, North Carolina State University, Raleigh, NC, USA. Worland, A.J., Korzun, V., Röder, M.S., Ganal, M.W., and Law, C.N. 1998. Genetic analysis of the dwarfing gene Rht8 in wheat. Part II. The distribution and adaptive significance of allelic variants at the Rht8 locus of wheat as revealed by microsatellite screening. Theor. Appl. Genet. 96(8): 1110–1120. doi:10.1007/s001220050846. Zeng, Z.B. 1994. Precision mapping of quantitative trait loci. Genetics, 136(4): 1457–1468. PMID:8013918.

Published by NRC Research Press