Analysis of genetic control and QTL mapping of essential wheat grain quality traits in a recombinant inbred population

Wheat cultivars are genetically crossed to improve end-use quality for traits as per demands of baking industry and broad consumer preferences. The processing and baking qualities of bread wheat are influenced by a variety of genetic make-ups, environmental factors and their interactions. Two wheat cultivars, WL711 and C306, derived recombinant inbred lines (RILs) with a population of 206, were used for phenotyping of quality-related traits. The genetic analysis of quality traits showed considerable variation for measurable quality traits, with normal distribution and transgressive segregation across the years. From the 206 RILs, few RILs were found to be superior to those of the parental cultivars for key quality traits, indicating their potential use for the improvement of end-use quality and suggesting the probability of finding new alleles and allelic combinations from the RIL population. Mapping analysis identified 38 putative QTLs for 13 quality-related traits, with QTLs explaining 7.9–16.8% phenotypic variation spanning over 14 chromosomes, i.e., 1A, 1B, 1D, 2A, 2D, 3B, 3D, 4A, 4B, 4D, 5D, 6A, 7A and 7B. In-silico analysis based on homology to the annotated wheat genes present in database, identified six putative candidate genes within QTL for total grain protein content, qGPC.1B.1 region. Major QTL regions for other quality traits such as TKW have been identified on 1B, 2A, and 7A chromosomes in the studied RIL population. This study revealed the importance of the combination of stable QTLs with region-specific QTLs for better phenotyping, and the QTLs presented in our study will be useful for the improvement of wheat grain and bread-making quality.


Introduction
Recent research in wheat contributes to yield enhancement and disease resistance, but quality is lacking in today's status quo. However, wide consumer demand has forced wheat breeders to focus on wheat quality improvement as per consumer preferences and industrial demands. Bread wheat (Triticum aestivum L.) is a globally accepted food crop and is consumed mainly in the form of baked products. The end-use quality of wheat is governed by a plethora of gene networks that are majorly affected by environmental conditions. Further, the end use property of wheat is measured by its seed quality and rheological traits such as Grain protein content (GPC), Sedimentation rate (SDS), Hectolitre weight (HW), 1000-kernel weight (TKW), Seed diameter (SD), Wet gluten content (WGC), Dry gluten content (DGC), Flour water absorption (FWA), Dough development time (DDT), Dough stability time (DST), Mixing tolerance index (MTI), Break down time (BDT) and Kernel hardness (KH). Quantitative traits loci (QTLs) for quality traits including GPC [1][2][3], KH [4,5], and dough quality traits, namely, MTI, mixing time, dough extensibility and dough tenacity [6,7], have been mapped. Groos et al. (1) reported four QTLs for GPC on chromosomes 2A, 3A, 4D, and 7D.
Linkage mapping and subsequent QTL mapping is the prerequisite for applying a successful marker-assisted selection (MAS) programme for individual traits. Earlier, MAS was executed in hexaploid wheat for high GPC (Gpc-B1), which was mapped and introgressed from the wild tetraploid wheat T. turgidum var. Dicoccoides [8]. Further, the role of the QTLs (Gpc-B1) for increased GPC was confirmed in tetraploid and hexaploid wheat using near-isogenic lines (NILs) with distinct Gpc-B1 alleles [9]. In additions, two independent studies conducted by Kumar et al. [10] and Tabbita et al. [11] showed that GPC was increased in Indian and Argentine hexaploid wheat carrying Gpc-B1. However, the pleotropic effect of the QTL Gpc-B1 is associated with reduced grain size and grain yield that ultimately lead to a reduction in wheat production [11,12].
Dough rheological properties and KH strongly affect the end-use quality of wheat. Doughmaking properties are often used as indicators of food baking quality. Dough strength and starch pasting characteristics are reported as quantitative traits; therefore, their expression is governed by multiple genes [13]. Presently, no specific bread-making quality trait-controlling genes have been identified that have direct associations with end-product quality. Nonetheless, a few QTLs for end-product quality traits have been reported [14]. Wheat quality is affected by temperature and humidity, but their effect is specific to developmental growth stages. Nuttall et al. [15] have reported that high temperatures during grain filling were responsible for reduced dough strength. Further, Cavanagh et al. [16] identified additional traits, such as the percentage of unextractable polymeric protein (%UPP) and dough strength, which were directly affected by temperature during the grain filling stage. The KH of wheat grain is a major determinant of food end-product quality. KH refers to the texture of the grain (caryopsis) that represents physical hardness or softness of the endosperm. KH is predominantly controlled by the Puroindoline (Pin) genes Pin a and Pin b,which are part of only the D subgenome and are located on chromosome 5 at the Hardness (Ha) locus. Furthermore, different classes' grain textures have been determined by unique allelic blends of Pin genes (Pin a and Pinb) in wheat, with diverse end-use characteristics [17]. The key role of the Pin a and Pin b genes is to determine the structure of the proteins in wheat grain, as well as the possible antimicrobial effects [18]. Therefore, to develop a variety with the desired KH, pronounced understanding of the allelic composition of Pin genes in a diverse set of germplasms is of the utmost importance for the selection of parental donors.
In the present study, a total population of 206 RILs was used for phenotyping of qualityrelated traits in three different locations in India, namely, Delhi, Karnal and Indore. The aim of the present research was to unravel the genetic factors controlling bread-making qualityrelated traits by mapping wheat population grown in three different environmental conditions through mapping of QTLs associated with these quality traits.

Plant materials and experimental design
In the present study, a mapping population of 206 RILs (F 9:11 ) was genotyped and evaluated for different quality traits.The RIL population was developed by crossing two wheat cultivars WL711 (S308/Chris/Kalyansona) and C306 (RGB/CSL3//2/C591/3/C217/N14//C281) [19]. WL711 is known for low end product quality traits while C306 is well known for its impact on good bread and cahapati making quality.The grain samples were taken from three independent field experiments conducted at . These three regions are geographically located in the traditional wheat agro-ecosystems. RILs, along with parents, were sown in three environments in a randomized complete block design (RCBD) pattern in the field with three replications per experiment. Sowing was done in a plot containing 3 rows that were 1.5 m long; each row was equally spaced by 25 cm and in each row, a total of 30 seeds were planted. RILs were sown during mid November and harvesting was done in April at DL10 and KL08, while at IN09, they were sown in early November and harvesting was done in early March.

Quality traits analysis
RIL grain samples collected from each experimental location were analysed in the same year at Cereal Quality Laboratory, Division of Genetics, Indian Agriculture Research Institute, New Delhi, India (S1 Table). There were three replicates used from each experiment for each quality traits samples were hand-cleaned and air-aspirated to remove foreign material and shrivelled kernels. The estimation of GPC was done by near-infrared reflectance (NIR) (RACI-CCD, 2010) using a NIR instrument (Foss 6500, FOSS NIR Systems, Inc., Laurel, MD) [20]. Estimated sedimentation volume, represented in height (mm) of the sediment measured during the SDS sedimentation test, was estimated as gluten strength [21]. Wheat flour of 206 RILs and the two parental genotypes (WL711 and C306) used for quality analysis was produced by a Cyclotec Mill (Tecator AB, Sweden) fitted with a 1 mm sieve. Five flour quality traits, namely, BDT, DDT, DST, FWA, and MTI were recorded by a Farinograph (Brabender, Germany) according to AACC 2000 [20].
Clean samples of 20 g seeds with grain moisture content ranging between 10% and 11% were used for the analysis of KH, TKW and SD using the Single Kernel Characterization system (SKCS) 4100 (Perten Instruments, Australia) with the AACC method (2000). HW was measured as the volume of grain per unit. Further, grain protein gluten was measured as wet and dry gluten using Glutomatic 2200 (Perten Instruments) according to the AACC method (2000).

Statistical analysis of the traits
Statistical and genetic analysis for quality traits was performed by GenStat14 [22]. The analysis was conducted in two stages while taking account of experimental design factors, first spatial analysis [23], to find the best linear unbiased estimates (BLUEs). Analysis of variance (ANOVA) was conducted for all traits separately for estimating variance components for evaluation of the significance of genotypes and trial effects and their interactions in the WL711/ C306 RIL population. ANOVA was done using three factor factorial analysis of the statistical programme MSTAT-C, version 1.41, Michigan State University, USA.
The broad sense heritability (h B 2 ) value was calculated for each trait across environments as h 2 = σ 2 g /(σ 2 g + σ 2 gxe /e) whereσ 2 g = [MS RIL -MS RILxe )/e], σ 2 gxe = MS RILxe , Wheree is the number of environments, MS is the mean square and x is the sign of multiplication.

Genetic analysis of the traits
The information regarding the genotyping of the RIL population and linkage map was given in Shukla et al. [24].

In-silico identification of genes within QTL region
Total genes present within underlying QTLs were identified utilising NCBI blast to reference chromosome from wheat genome sequence. Markers flanking (gmw413 and cfd65) to the QTL qGPC.1B.1 was blast to 2A wheat chromosome sequence and genomic sequence of the QTL region was downloaded. Annotated CDS present within the QTL region was selected from the wheat genome annotated CDS present in EMBL database (ftp://ftp.ensemblgenomes. org/pub/plants/release-42/fasta/triticum_aestivum). Function of the genes was predicted using blast2go tool [25]. Genes having more than 70% of the similarity were selected.

Phenotypic data and correlation analysis
Experiments were conducted at three locations in three different years. Performance of both the parents was observed along with the RIL populations. WL711 showed a low quality score at KL08 compared to that of the other locations; however, C306 showed better performance for traits at the same location and years (Table 1). Measurable phenotypic variation was observed among both the parents for SDS, TKW, WGC, DGC, FWA, DST, MTI, BDT and KH. All quality-related traits significantly differed among the RILs and exhibited transgressive segregation (Table 1). A combined ANOVA was performed over all trials which indicated statistically significant main effects for genotypes (G), trials (T), GxT interactions for quality traits ( Table 2). Variance due to GxT interaction was substantially lower than variation due to genotype for all the traits. GPC and HW showed high broad sense heritability, while TKW and KH showed moderate heritability ( Table 2). A highly significant positive correlation was recorded between GPC, WGC, DGC and FWA; between SDS,WGC, DGC, DDT and DST; between DDT,BDT, and KH; and between DST and BDT. Highly significant but negative correlations were recorded between GPC and TKW; between WGC and DST; and between MTI, DST, and BDT (Table 3).

QTL x environment interactions and epistatic QTL
The effects of the QTL x environment interactions (QE) for quality-related traits were recorded and listed in Table 5. From the measured quality traits, two QQ interactions were detected for GPC and TKW. In addition, a few genomic regions identified in this study showed QE, QQ and QQE interactions, and their effects were less noticeable than the main additive effects (a). These results indicated that the additive effects were more significant than the epistatic effects in the studied quality traits. Epistatic QTLs showed QTL x QTL (QQ) and QTL x QTL x environment (QQE) interaction.

Gene identification within QTLregion
The genomic region within the flanking markers of QTL qGPC1B.1 was retrieved from the NCBI genome database. In-silico analysis showed total 346 genes were found within this QTLs region. Out of 346, 110 genes showed more that 70% functional similarity with the existing A positive value means that the parent-type effect is greater than the recombinant-type effect A negative value means that the parent-type effect is less than the recombinant-type effect a GPC-grain protein content, TKW-thousand kernel weight b QTL_i and QTL_j are a pair of QTL involved in epistasis c QQ, the epistatic main effect d QQE, the epistasis x environment interaction effects e R2 (QQ) %, Phenotypic variation explained by QQ effects protein in database. Among these 110 genes, 50 were enzymes, 7 transcription factor, 4 transporters, 7 ribosomal protein, 6 chloroplast, 5 motochondial subunits encoded by genome, 4 receptors and 27 were belongs to different function (S2 Table). Further analysis based on homology to the annotated wheat genes present in database showed that only six genes namely PGKY_Phosphoglycerate kinase, cytosolic, CBP2_Serine carboxypeptidase2, PALY_Phenylalanine ammonia-lyase, HBP1C_Transcription factor HBP-1b (c1), MT1_Metallothionein-like protein 1 and UBC2_Ubiquitin-conjugating enzyme belongs to Triticum aestivum (Table 6).

Phenotypic and genotypic variation in the parents and the RILs
Growing genotypes under well-adapted conditions with strong phenotypic expression can lead to overestimation of the genetic component, which could be avoided by including contrasting environments and seasons in which observations are made. In accordance with this notion, the experimental materials consisting of a population of 206 RILs that was developed from the cross WL711/ C306 were grown under three environmental conditions. A total of 38 QTLs were identified through CIM for thirteen quality-related traits across environments. Continuous phenotypic variation and transgressive segregation for all the traits observed in the RIL population revealed the quantitative inheritance of these traits. Further, both the parents contributed beneficial allele for quality traits strengthened usefulness of this population for QTL analysis and genetic interaction analysis between the alleles.

Genetic locus for quality traits GPC, TKW and KH
Increased GPC is a focus area of current wheat quality breeding programmes. Parent C306 and the RILs showed a significantly high mean GPC (above 15%) in the IN09 environment, where RILs were exposed to heat between post-anthesis and the grain filling stage. These results were in agreement with Maphosa et al. [26]. GPC showed a low value (below 12%) in the DL10 and KL08 experiments, when crops experienced cool and moist conditions. Li et al. [27] indicated that total GPC is linked to temperature and low humidity. A negative correlation between GPC and TKW was recorded in this population, which was reported in previous studies as well [28]. The QTLs related to GPC were reported earlier on the regions of several chromosomes, showing several loci controlling wheat GPC; those studies also suggested very fewer differences in GPC in the parental line, but QTLs were still detected [29,30]. In the present study, QTL analysis for GPC revealed six QTLs with PV ranging from 9.8-15.8% located on six different chromosomes, i.e., 1B, 1D, 3B, 3D, 5D and 7A. The chromosomes 3B and 7A were earlier also explored for the GPC content [31]. Although, the difference in protein content between the parents was lower, transgressive segregants were observed for GPC. These transgressive segregants for high GPC might be due to minor genes segregating in the population and the different GPC-controlling alleles in the parents, confirming the suitability of this Table 6. Genes senquence name and genomic position of identified genes within the flanking markers of QTL qGPC1B.1based on homology to the annotated wheat genes present in the database.

SeqName
Genomic position Description  [32]. A set of epistatic QTLs showed weak additive × additive × environment effects (AAE), andthe interactions suggested that the additive effects played an important role in wheat GPC. In this study, QTLs for GPC and SDS were mapped near the Glu-D1 region which is present on chromosome 1D. Similar results were observed in other studies as well [33,34]. In fact, the Glu-D1gene that codes HMW subunits (2 +12 and5+10) was also found to affect the protein quality in a ChSh population [35]. Furthermore, another wheat protein, triticin, which is encoded by Tri-D1, was reported to positively affect wheat dough bread-making quality, which was also present on the short arm of chromosome 1D [36]. The other two QTLs for GPC, on chromosomes 3B and 5D, had larger effects and can be used for further genetic improvement. TKW is one of the important yield components. Selection of TKW directly increases the grain yield [37]. Its correlation with quality parameters has been reported [38]. Selection for quality traits alone will not improve this trait. A pronounced and significant variation for TKW suggested several genes with major and minor effects that were involved in the phenotypic expression of this trait. TKW was controlled by 5 QTLs identified in our study, which were present on the chromosomes 1B, 2A, 2D, 6A and 7A.Sun et al. [39] also identified seven QTL regions on chromosomes 2A, 2D, 3B, 4A, 5D, 6A, 6B, and 7B in RIL population. In addition, Reif et al. [40] identified 12 putative QTLs on chromosomes 1A, 3A, 5A, 7A, 1B, 3B, 6B, 1D, 3D, 4D and 7D in a RIL population. In these studies, only one QTL(6B) was found similar, which suggested that many genes govern the trait TKW. Of the eight QTLs identified by Sun et al. [38] only two QTLs i. e. 2D and 6A, shared chromosomal location in the present study. Wheat chromosome 7A was earlier also endorsed for the study of QTLs for different agronomic traits and also for TKW as similar to our study [41]. Recently, MAS was used for the transfer of three garin weight QTL QGw.ccsu-1A.2, QGw.ccsu-1A. 3 and QGw.ccsu-1B.1 identified from NILs derived from Raj3765 and K9107 [42]. In this study, one epistatic QTL was identified with negative Additive × Environment (AE) or AAE interactions, which showed that an additive effect responsible for the main genetic variance of TKW.
KH played a major role in determining quality of bread wheat and end use properties. Additionally, the Ha locus is mainly known for affecting grain hardness in wheat. Several QTLs for KH that are distributed on all twenty-one wheat chromosomes except for 3D and 6A have been reported in different mapping populations [43]. Both parents contributed favourable alleles for KH, which confirmed the quantitative nature of the trait [44].

Identification of gene-rich regions/ QTL clusters
In wheat, associations of qualitatively inherited genes together represent gene-rich regions form the hot spots of recombination. QTL are usually spread over all the chromosomes, but clusters of QTLs in certain chromosomal regions have been observed. QTLs affecting several traits are common and may be due to pleiotropy or close linkage [34]. Since most of the QTL hotspots in this study were located in the short and long arm of the chromosomes, QTL colocation of yield QTLs has also been identified previously in wheat [1,37]. Similarly, 5 QTLs were mapped on 5D,5 QTLs on 7B and 4 QTLs on 1B, and some of them showed stability across the environments, which also suggested that the two QTL clusters might have pleiotropic effects. It is likely that the clusters represent similar gene/protein content. Several linked markers in the clusters suggest the usefulness of these markers for marker-assisted breeding of these QTLs to enhance the end-product quality of wheat.

Conclusions
Overall, 38 QTLs for 13 end product quality traits were mapped, explaining 7.9 (qSDS.4B.1) to 16.8% (qSDS.7A.1) of PV detected on total 14 chromosomesi.e.,1(ABD), 2(A, D), 3(B,D), 4 (ABD), 5D, 6A, 7A and 7B. The additive effect was found to be positive in 17 QTLs, contributed by WL711 while, 21 were negative and contributed by C306. Eight QTLs for three major quality traits affecting the bread-making quality, namely, SDS (5), DST (2) and DGC (1), were identified, with 9.6 to 16.8% PV. For SDS, five of the three alleles were contributed by WL711, and for DST and DGC, both were contributed by C306..For GPC, six QTLs were reported on chromosome 1B, 1D, 3B, 3D, 5D and 7A, showing 9.8-15.8% of PV for the trait, with positive alleles coming from WL71l at two QTLs (qGPC.3D.1 and qGPC.7A.1) and from C306 at four QTLs. The strongest effect for GPC (11.9), with 15.8% PV, was located on qGPC.5D.1, with the positive allele being contributed by C306. Six putative candidate genes have been identified by Insilico analysis of QTL qGPC.1B.1 region based on homology to the annotated wheat genes present in the database. This study revealed the importance of the combination of stable QTLs with region-specific QTLs for better phenotyping, and the QTLs presented in our study will be useful in MAS efforts after validation for the improvement of wheat grain and bread-making quality.
Supporting information S1 Table. Mean value of the measured traits data of a 206 RILs population along with parents recored from three independent experiments conducted at three different locations (KL08, IN09, DL10) in an independent year. (XLSX) S2 Table. List of genes present within flanking markers of QTL qGPC1B.1 were identifed form EMBL database and function was pridicted based on homology by Blast2go tool. (XLSX) S1 Data.