A Novel QTL Controlling Flag Leaf Width Located on Chromosome Arm 7AS in Bread Wheat (Triticum Aestivum L.)


 Background: Wheat is an important cereal crop and improving wheat production is essential for meeting the food demand from the growing population worldwide. Flag leaf width (FLW) is an important trait affecting plant architecture and contributing to grain yield. To detect loci conferring FLW, we assessed a population of recombinant inbred lines (RILs) from a cross of EGA Wylie/Sumai 3 in different environments.Results: A total of six QTL were detected from the population. Two of them located on chromosome 2B and the other four located on chromosomes 2D, 4B, 7A, and 7B, respectively. The percentage of phenotypic variation (PEV) explained by these loci ranged from 14.6% to 33.8%, with LOD scores varying from 3.01 to 7.81. Of them, the locus located on chromosome arm 7AS is likely novel. Significant effects of this locus were detected in multiple trials conducted and the PEV explained by this QTL varied from 14.6% to 19.8% among the different trials. An orthologous analysis based on rice and Arabidopsis identified 3 putative genes underlying this potentially novel locus.Conclusion: This study identified a stable potentially novel QTL in multiple environments and predicted three candidate genes of it, which laid the foundation for further fine-mapping and cloning the gene(s) underlying QFlw.WS-7A with the contribution to grain yield.

Conventional breeding methods are time-consuming, while marker-assisted selection (MAS) is faster and more e cient. MAS, as an optimal method in wheat breeding, mainly depends on the genes/QTL and their linkage markers. QTL mapping lay the foundation for MAS. QTL mapping as a popular and e cient method has a long story. Sax (1923) [22] rst suggested the basic idea for studying QTL through linkage markers, and the idea was put into practice by Thoday (1961) [23]. Paterson et al. (1988) [24] proved for the rst time that the QTL mapping using molecular markers worked well, which opened the door to QTL mapping in many traits. Afterward, many researchers have contributed to the improvement of QTL mapping methods for different conditions by different models or algorithms [25][26][27][28][29][30]. Now, QTL mapping has become a powerful tool and it is widely used in many species [31][32][33][34][35]. Many QTL associated with every aspect of traits were detected using QTL mapping in wheat [36][37][38][39][40][41][42][43][44][45][46][47][48]. The availability of the high throughput molecular markers [49][50][51][52][53][54] and the high-quality genome reference IWGSC RefSeq v1.0 [55] have resulted in more precise identi cation of QTL via the use of dense genetic maps.
Following the identi cation of QTL conferring FLW using a RIL population with an existing genetic map consisting of Diversity Arrays Technology (DArT) markers, we identi ed candidate genes for a novel locus through orthologous analysis. These results are reported in this publication.

Results
Phenotypic variation of ag leaf width in the mapping population FLW of Sumai 3 was signi cant wider than that of EGA Wylie (Fig. 1). FLW was measured against the RIL population under for different environments. Signi cant correlations were detected for results from these trials, with correlation coe cients ranging from 0.352 to 0.861 (p < 0.01) ( Table 1). Two of these are conducted in the eld environments at CSIRO Research Station in Queensland in 2018, one was located at 27°32'16.4"S 152°20'14.6"E (designated as E1), and the other at 27°33'56.9"S 152°19'49.4"E (designated as E2). FLW ranged from 1.17 to 2.08 in E1 and from 0.99 to 1.73 in E2. The mean value was signi cantly higher in E1 (1.66) than that in E2 (1.34), while the phenotypic diversity index was signi cantly higher in E2 (0.93) than that in E1 (0.88). The other two trials were conducted in glasshouses at the Queensland Bioscience Precinct (QBP) in Brisbane, Australia, one in 2017 and the other in 2018 (designated as E3 and E4, respectively). FLW varied from 0.95 to 2.10 in E3 and from 1.00 to 1.80 in E4. The mean values and the phenotypic diversity indexes of FLW in E4 were higher than that in E3. BLUP values of FLW from these trials ranged from 1.20 to 1.75 with a mean of 1.44 and the phenotypic diversity index being high as 0.96.
The estimated h 2 for FLW from these trials was 0.96 ( Table 2), suggesting that genetic effects were the major determinant for the phenotypic variance of this trait. The numbers of RILs for different FLW followed the normal distribution in all the four trials (Fig. 2). The identi cation of QTL for ag leaf width To identify QTL for FLW in the RIL population, the trait was evaluated in each of the four trials and BLUP values from these trial results were also obtained and used. A total of 6 QTL were detected, two of them located on chromosome 2B, and the other four located on chromosomes 2D, 4B, 7A, and 7B, respectively (  (Fig. 4). According to the Chinese Spring reference RefSeq v1.0 (IWGSC) and RefSeq Annotation v1.1 [55], the 31 high-con dence genes in the region of QFlw.WS-7A were selected for collinearity analysis with Arabidopsis and rice. Based on the function of their orthologous genes, we predicted three candidate genes that may be associated with FLW. They were TraesCS7A02G050900, TraesCS7A02G051200, and TraesCS7A02G052000, which were involved in melatonin degradation, substances and energy metabolism, and leaf development that affected the leaf width directly and indirectly.

QTL analysis
The population used in this study was derived from EGA Wylie and Sumai 3. Both parents were commercial varieties of great values and they performed well in agronomical and morphological traits. Identifying and utilizing the elite FLW QTL from these varieties could be an effective way to contribute to grain yield. We identi ed six QTL associated with FLW by assessing this population. These loci were all detected in multiple conditions except for QFlw.WS-2B.  [19]. The physical location of Xwmc139 was around 19.90 Mb, thus these two loci were separated by a physical interval of around 4.47 Mb. We analyzed the RIL population and their parents using Xwmc139 and found that the marker was not polymorphic. We thus believe that QFlw.WS-7A was a different locus from those loci reported previously.
Many QTL have been reported but few have been utilized in breeding programs. One of the important factors determining the useful of a locus was its hereditary. In this study, the detected h 2 for FLW was high, which could lay the foundation for the identi cation of major and stable QTL. Based on the presence/absence of the two different alleles at QFlw.WS-7A, the population was classi ed as two groups, one with EGA Wylie allele and the other Sumai 3 allele. The FLW of Sumai 3 type was signi cantly higher than that of Wylie type in each of the trials (Fig. 5). The presence of the Sumai 3 allele increased the FLW by an average of 16.6% based on the results from the two eld trials, by an average of 16.1% based on results from the two glasshouse trials, and by an average of 18.6% based on results from the BLUP values. These results suggested that the allele of QFlw.WS-7A from Sumai 3 is a stable and effective QTL for breeding programs.
The analysis of candidate genes underlying QFlw.WS-7A Three candidate genes were identi ed for QFlw.WS-7A. They were TraesCS7A02G050900, TraesCS7A02G051200, and TraesCS7A02G052000. These genes were involved in melatonin degradation, substances and energy metabolism, and leaf development that affected the leaf width directly and indirectly. Both TraesCS7A02G050900 and TraesCS7A02G051200 are homologous to rice gene 2ODD11 (2oxoglutarate-dependent dioxygenase 11) that was involved in melatonin degradation [58]. According to the reports of Arnao and Hernández-Ruiz (2015) [59], the absorption of melatonin can increase leaf size. It has also been reported that 2ODD was involved in several structural modi cations in the biosynthesis of gibberellins and it played a key role in many growth and developmental processes including leaf expansion [60]. Function analysis also showed that 2ODD11 participated in the dioxygenase activity, Lascorbic acid binding, and metal ion binding. TraesCS7A02G052000 is aligned with rice gene SDH8A (Succinate dehydrogenase subunit 8A). GO annotation analysis showed that this gene was involved in the pathway of tricarboxylic acid cycle, which is the important hub of substances and energy metabolism. The carbohydrate metabolism and photosynthesis involved in the tricarboxylic acid cycle provide the substances and energy for leaf development, which affecting FLW indirectly. Based on the report of Zhao et al. (2015) [61], amino acids and organic acids participated in the tricarboxylic acid cycle were signi cantly different in metabolite levels between two rice accessions with different leaf width. There could be a relationship between leaf width and the tricarboxylic acid cycle [61-62].
Orthologs for these genes were also found in Arabidopsis. TraesCS7A02G050900 and TraesCS7A02G051200 were orthologous with Arabidopsis gene ANS and SRG1, respectively. Both genes belong to the iron/ascorbate-dependent oxidoreductase family. ANS was involved in the avonoid biosynthetic process and regulation of jasmonic acid mediated signaling pathway [63]. SRG1 was involved in leaf senescence [64].

Conclusion
To detect QTL for FLW, a RIL population consisting of RIL was assessed in different environments. We identi ed 6 QTL in these trials. They were located on chromosomes 2B, 2B, 2D, 4B, 7A, and 7B, respectively.
Of them, QFlw.WS-4B was a stable major QTL with a PVE of up to 33.8%. Compared with the location of previously reported QTL associated with FLW, QFlw.WS-7A was potentially novel. Based on an orthologous analysis with rice and Arabidopsis, we identi ed three candidate genes for this locus. They were involved in the regulation of plant growth, substances and energy metabolism, and leaf development. The QTL mapping and the candidate genes prediction laid the foundation for cloning the gene(s) underlying QFlw.WS-7A.

Plant materials
A population of 92 RILs derived from a cross of EGA Wylie/Sumai 3 [44] was used in this study to detect QTL for FLW. EGA Wylie is a commercial variety widely grown in Australia and Sumai 3 was a variety The genotyping and QTL mapping The genetic map for the population used in this study was generated using DArT Pty Ltd as described in an earlier study [44]. The QTL analysis was carried out using software MapQTL 5.0 through MQM mapping [69]. For each trial, the signi cant LOD threshold was determined by a test of 1000 permutations with the whole genome scanning of 0.05 level. QTL detected in multiple trials was considered as a stable QTL. The genetic maps were drawn with MapChart 2.32 (https://www.wur.nl/en/show/MapChart-2.32.htm). Comparison of QTL detected in this study and those reported earlier was conducted based on the Chinese Spring reference RefSeq v1.0 [55].
The analysis of candidate genes Identi cation of candidate genes for the novel QTL was conducted based on an orthologous analysis. Firstly, we delineated the physical intervals of novel QTL based on the wheat reference genome RefSeq v1.0 of Chinese Spring [55]. Gene sequences in the targeted interval were extracted using TBtools [70]and used to blast against the wild emmer (T. turgidum ssp dicoccoides) genome [71] and protein database SWISS-PROT [72] using TBtools [70] with default parameters. Functions of the candidate genes were extracted from their homologs in Oryza sativa and Arabidopsis. Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Figure 1
The ag leaf width of parents. The Wylie and Sumai 3 are maternal and paternal, respectively  The physical position for QFlw.WS-7A and their predicted genes