Development of white-grain pre-harvest sprouting tolerant and pyramided protein-rich leaf rust resistant wheats using molecular breeding

The present study was undertaken for developing pre-harvest sprouting tolerant (PHST) wheat genotypes using marker-assisted backcross breeding (MABB). A major QTL for PHST was introgressed into an elite Indian wheat cv. Lok1 that is PHS susceptible. These PHST lines were also pyramided with one gene each for high grain protein content (Gpc-B1) and leaf rust resistance (Lr24). For introgression of PHST QTL, initially Lok1 was separately crossed with each of the two donors (PHS tolerant white-grained AUS1408 and CN19055). Backcrossing in each generation was followed by foreground and background selections using SSR markers. In advanced lines, KASP assay was also carried out for the candidate gene TaMKK3-A underlying the PHST QTL. The MAS derived lines homozygous for PHST QTL were screened for PHS using simulated rain chambers resulting in the selection of 10 PHST lines. For pyramiding of three QTL/genes (PHST QTL, Gpc-B1, and Lr24), MABB derived BC 4 F 2 plants (from the cross Lok1/CN19055) were crossed with a MAS derived BC 2 F 5 line [Lok1 (Gpc-B1 + Lr24)] developed earlier by us in the same background of Lok1. After foreground MAS followed by PHS screening, four advanced lines carrying all the three QTL/genes in homozygous condition were selected. These lines exhibited high level of PHST (PHS score 2–3) associated with signicant improvement in GPC with no yield penalty and resistance against leaf rust under articial epiphytotic conditions.


Introduction
Pre-harvest sprouting (PHS) refers to precocious germination of grains within physiologically mature spikes on prolonged rains before harvesting (Groos et al. 2002). Seed dormancy (SD) has also been used as a surrogate trait for PHS tolerance (PHST) because PHST is known to depend on seed dormancy and the two traits exhibit high correlation (Li et al. 2004). PHS leads to yield losses and signi cant reduction in grain quality due to activation of enzymes like lipases, amylases and proteases in developing grains leading to degradation of lipids, starch, and proteins (Andreoli et al. 2006;Simsek et al. 2014). PHS also leads to reduction in market value of wheat grain, by 50%, ultimately resulting in a global estimated loss of US $1 billion annually (Ali et al. 2019). PHS problem is common in all major wheat growing areas of the world, including Canada, Europe, Australia, China, USA, Japan, and also some parts of India (Rajjou et al. 2012;Zhou et al. 2018).
PHS is a complex trait controlled by a large number of genes and is in uenced by the environment. PHST is often associated with red grain color (DePauw and McCaig, 1983;Groos et al. 2002;Fofana et al. 2009), so that grain color and seed dormancy both are used as a markers for PHST (Flintham, 2000;Mares et al. 2005). However, PHS tolerant white grain wheat genotypes are also known, which include AUS1408 (Mares, 1987) and CN19055 (Hucl and Matus-Cadiz, 2002). Following parameters are generally used for estimating the level of PHST: PHS index, falling number, germination index and alpha amylase activity (Nörnberg et al. 2015;Shorinola et al. 2016). Among cereals, in wheat alone, ~50 QTL interval mapping studies involving ~40 mapping populations have already been conducted leading to identi cation of as many as >250 QTLs for PHST, which also included hexaploid synthetic wheats, tetraploid durum wheats and diploid einkorn wheats (see review by Gupta et al. 2020). Relatively, more important QTL for PHST and seed dormancy are known to be carried by chromosomes 3A, 3B, 3D, and 4A, although all the 21 chromosomes carry QTLs for this trait (see Gupta et al. 2020). MetaQTLs have also been identi ed for PHST in wheat (Tyagi and Gupta, 2012). Genome wide association studies (GWAS) have also been conducted for PHST and related traits leading to identi cation of ~250 MTAs (for details see review by Gupta et al. 2020). A number of these MTAs were located in genomic regions carrying QTLs earlier identi ed through interval mapping. Among known QTLs, 30 QTLs have also been shown to be stable over environments; these have been considered potentially useful for improvement of PHST in wheat. A major QTL on chromosome 3A, associated with red grain color (QPhs.ccsu-3A.1; ) and another QTL Phs-A1, located on chromosome arm 4AL are considered to be relatively more important for breeding (Mori et al. 2005;Mares et al. 2005;Liu et al. 2008). Phs-A1 has also been cloned and characterized (Barrero et al. 2015;Shorinola et al. 2017). Two genes, namely PM19-A1 and PM19-A2 (Plasma Membrane 19 genes) were proposed as the main candidates for the QTL Phs-A1 (Barrero et al. 2015), although on the basis of comparative analysis of synteny, as many as 10 orthologous genes were proposed as candidate genes. The QTL Phs-A1 also overlaps the gene TaMKK3-A for seed dormancy (Torada et al. 2016). Recently, markers for QTL Phs-A1 and the gene TaMKK3-A, were also developed (Shorinola et al. 2017). It was also shown that TaMKK3-A alleles are diagnostic for sprouting resistance in parents of mapping populations used for identi cation of the QTL Phs-A1.
It has been recognized that initially during green revolution, improvement in yield was a major concern, and that during the last three decades, GPC has also attracted the attention of wheat breeders to address the problem of nutritional security in developing countries (Joshi et al. 2007). Therefore, among important traits for wheat breeders, grain protein content (GPC) and resistance to three rusts including leaf rust are also important. Although, several loci affecting GPC have been identi ed in wheat, the most important gene affecting GPC is Gpc-B1, which was initially identi ed in wild emmer wheat (Triticum turgidum L. ssp. dicoccoides) and was subsequently cloned following map-based cloning strategy . Gpc-B1 encodes a NAC-domain transcription factor and is located on the short arm of chromosome 6B, tightly linked with genes for accelerated senescence and high grain Fe/Zn concentration (Distelfeld et al. 2007; reviewed in Tabbita et al. 2017).
Suitable RFLP, CAPS, and SSR markers are available for Gpc-B1 (Mes n et al. 1999;Khan et al. 2000;Olmos et al. 2003;Distelfeld et al. 2006;Uauy et al. 2006). During the last decade, Gpc-B1 has been extensively deployed in durum and common wheats for improvement of GPC, so that most modern wheat varieties all over the world now carry Gpc-B1 Hagenblad et al. 2012;Brevis and Dubcovsky 2010;Kumar et al. 2011;Balyan et al. 2013;Vishwakarma et al. 2014;Vishwakarma et al. 2016;Bokore et al. 2019;Gautam et al. 2020). However, it has been a challenge to develop wheat varieties with high GPC without yield penalty due to several factors including the following: negative correlation between the grain yield and GPC (rage of correlation = −0.30 and −0.60), large environmental effects, genotype-by-environment interactions, and a complex genetic system governing GPC (Simmonds, 1995;DePauw et al. 2007;Klindworth et al. 2009;Brevis and Dubcovsky 2010). The backcross-derived lines containing Gpc-B1 have also been found to have reduced kernel weight and earlier senescence, both adversely affecting grain yield (Sherman et al. 2008).
In our own research extending over two decades, we also observed that PHS tolerant wheats improved for GPC are often susceptible to leaf rust, which causes signi cant losses (~7-30%) in wheat yield worldwide (Bolton et al. 2008). Leaf rust resistance genes have been regularly discovered, so that ~80 Lr genes are now known in wheat (McIntosh et al. 2017). Among these Lr genes, Lr24 is an important gene, which has been widely deployed in wheat cultivars because it provides resistance at all developmental stages. The gene Lr24 is also linked with Sr24 so that Lr24 has become a gene of choice in wheat breeding programmes (Bhardwaj et al. 2019).
In view of the above, the development of PHS tolerant, protein-enriched leaf rust resistant wheats using molecular breeding became an important goal in our laboratory, since markers for all the three targeted traits are available. This goal has now been successfully achieved using wheat cv. Lok1, which has been widely grown in central India. The results of this study, extending almost a decade, are presented in this communication.

Materials And Method
Material Plant materials used in the present study included the following: (i) an Indian wheat cv. Lok1, (ii) two wheat genotypes, namely AUS1408 and CN19055, each carrying a major PHST QTL located on chromosome arm 4AL (Mares et al., 2005;Ogbonnaya et al. 2008); these two genotypes were used as two independent donors; (iii) a MAS derived line in the background of cv. Lok1 containing Gpc-B1+Lr24, where Gpc-B1 is a locus for high grain protein content, which is in close linkage with a stripe rust resistance gene (Yr36); this improved line was used as a donor for pyramiding of genes for GPC and resistance to leaf rust and on to PHST locus carried by the improved BC 4 F 2 plants. The Lok1 cultivar is a high yielding wheat cultivar, released in 1982 for late sown irrigated conditions of central India, where it has been the most preferred wheat variety for cultivation, due to its good bread making quality and superior agronomic performance. The choice of Lok1 as the recipient cultivar was based on the fact that Lok1 was the rst among the best 10 varieties, selected from 150 varieties for breeder's seed production across the country for seven consecutive years between 2004 and 2010. At present, Lok1 (63.91 Mt) has third rank after PBW343 (73.42 Mt) and HD2967 (93.38 Mt) in terms of breeder's seed indent for the year 2019-20; it is still grown in an estimated 1 mha wheat area in central India (https://www.lokbharti.org/Wheat-Research). Lok1 is also highly resistant to the most virulent rust pathotype Ug-99.

Molecular markers
The details of the markers linked with QTL/genes introgressed and pyramided during the present study are presented in Table 1. For the recovery of the recurrent parent genome (RPG) in the segregating backcross progenies during background selection, a total of ~130 polymorphic SSR markers (selected from 453 markers) covering all the 21 chromosomes were used in the following two crosses: Lok1/AUS1408 and Lok1/CN19055. In each backcross generation, the proportion of the genome from the recipient parent was estimated using the following formula; G = [(X+1/2Y) x 100]/N, where, X is the number of markers showing homozygosity for recipient parent alleles; Y is the number of markers showing heterozygosity involving the two parental alleles and N is the total number of polymorphic markers.
DNA extraction, primer synthesis and PCR analysis DNA isolation from 30-40 days old leaves was carried out using the CTAB method following Saghai-Maroof et al. (1984), with some modi cations. DNA was quanti ed on 1% agarose gel using agarose gel electrophoresis. Primers were synthesized by Euro ns Genomics India Pvt Ltd., Bangaluru, India.
Polymerase chain reaction (PCR) was carried out in 20 μL reaction mixture containing the following component: ~50 ng of template DNA, 1 U Taq polymerase (Sigma-Aldrich, USA), 0.02 mM of each dNTPs (Sigma-Aldrich, USA), 0.2 μM primers (forward and reverse), 1X Taq buffer in each case. PCR was performed using Veriti TM 96-well Thermal Cycler of Applied Biosystem, on following PCR pro le: initial denaturation for 4 min at 94 • C, followed by 35 cycles each with 30 sec at 94 • C, 30 sec at 50−60 • C (speci c for different primers), 1 min at 72 • C, and 7 min at 72 • C for nal extension. Ampli cation products for marker SCS73 719 were resolved on 2% agarose gel and visualized on UV trans-illuminator; products of other markers were resolved on 10% PAGE gel and visualized by silver staining.

KASP assay
Kompetitive allele speci c PCR (KASP) assay was also conducted to con rm the presence of PHST QTL in the improved lines developed during this study by using the primers earlier developed and used by Shorinola et al (2017). The KASP assay was carried out as described in Bansal et al (2020) Breeding scheme for introgression and pyramiding of QTL/genes Marker assisted backcross breeding (MABB) scheme was followed to transfer the PHST QTL from each of the two donor genotypes (AUS1408 and CN19055) independently into the genetic background of cv. Lok1 (Fig. 1). Recurrent parent Lok1 was used as female and crossed with each donor as male parent. F 1 plants were used for four generations of backcrossing leading to BC 4 F 1 . Both foreground and background selections were exercised. The MABB scheme followed in the present study is shown in Fig.  1. Foreground MAS in BC 4 F 2 and BC 4 F 3 was carried out for selecting plants homozygous for the PHST QTL; these plants were then screened in simulated rain chambers for PHST (more details are provided in the next section). Seed of the selected PHS tolerant plants was multiplied for conducting station yield trials for two consecutive years for evaluation of agronomic traits.
The BC 4 F 2 plants (improved for PHST) derived from only one of the two crosses namely Lok1/CN19055 was utilized for pyramiding of genes for high GPC and leaf rust resistance in the background of cv. Lok1. Scheme of the breeding program for pyramiding is also depicted as a part of Fig. 1. For the purpose of pyramiding, the improved PHST plants were crossed as female parent with MAS derived BC 2 F 5 line Lok1 (Gpc-B1+Lr24) earlier developed by us in a separate programme. The initial crossing was followed by forward breeding for a number of generations using plant to progeny scheme. The F 4 progenies were screened for rust resistance under arti cial epiphytotic conditions (for details see below). The progenies in F 4 to F 6 generations were also screened for PHS tolerance under arti cial rains (for detail see below).
Seed of the selected progenies was multiplied for station yield trials for two consecutive years (for details see below).

Evaluation for PHS and seed dormancy
For the PHS screening, ve randomly selected physiologically mature spikes (when 75% spikes in a plot turned yellow) were harvested from each selected progenies and immediately kept in specialized simulated rain chambers at room temperature providing arti cial rain every 6 h for seven consecutive days, when data on PHS was scored on a scale of 1-9; score of 1 represented no visible sprouting and a score of 9 represented complete sprouting (Kumar et al. 2010). Mean PHS scores were used for selection of PHS tolerant genotypes.
Data was also recorded on seed dormancy of MAS derived lines following Knox et al (2012). In brief 50 seed were placed in a Petri dish with the crease of each seed facing down to make it easier to observe the germination. Seed with white radicle or plumule protruding with a visible tear in the seed coat were treated as germinated seeds and removed from the plate. Seed counts were recorded daily for 21 days and germination index (GI) was calculated.

Screening of MAS-derived lines for rust resistance
Screening of the pyramided lines (containing QPhs.dpi.vic.4A.2, Gpc-B1 and Lr24) were carried out against a total of 16 pathotypes of the three rusts reported from different regions of India (for the details of pathotypes used see Table 2). The procedure for screening of rust resistance was described in an earlier study (Gautam et al. 2020). In brief, seedlings were assayed for the infection against each pathotype at 15 days post-inoculation (dpi) following Stakman et al (1962).

Yield trials of MAS-derived lines
The yield trials included the following 18 lines; (i)10 lines with only PHST QTL, (ii) 4 lines containing QPhs.dpi.vic.4A.2+Gpc-B1+Lr24 and (iii) one recipient and three donor parental genotypes [Lok1, Lok1 (Gpc-B1+Lr24), AUS1408, and CN19055]. Randomized block design with two replications was used at Meerut during 2018-19 and 2019-20 crop seasons. Each line in a replication was evaluated in 6 m 2 plots, each having six rows of six meter with a row-to-row distance of 20 cm. Phenotypic data were recorded on the following agronomic and grain quality traits: plant height (PH in cm), tillers per meter (TPM), grain number per spike (GNPS), 1000-grain weight (TGW in g), grain yield (GY in g), grain protein content (GPC in %), iron (Fe in ppm), and zinc (Zn in ppm) contents in grains. The GPC (%) at 12% grain moisture content was estimated using FOSS Infratec TM 1241 Grain Analyzer (Denmark).

Statistical analysis
For statistical analyses, mean values of the two replications for each trait in each trial were used. The descriptive statistics were obtained using Excel. The analysis of variance (ANOVA) for the eight traits was carried out using the SPSS Statistics for Windows, version 16.0 (SPSS Inc., Chicago, Ill., USA); signi cance in ANOVA was tested for the following sources of variation: replications, genotypes, environments, genotype-by-environment interactions.

Results And Discussion
MABB for PHST involving two donors (AUS1408/CN19055): For introgression of PHST QTL separately from each of the two donors, namely AUS1408 and CN19055 into Lok1, two separate MABB programmes were followed. In each case, F 1 plants were subjected to four backcrosses to recipient parent cv. Lok1 involving foreground and background MAS. For foreground selection of PHST QTL, separate pairs of polymorphic SSR markers were used as follows: (i) Xgwm397 and Xbarc170 in case of AUS1408, and (ii) Xgwm637 and Xgwm894 in case of CN19055. Similarly, for background selection a set ~130 polymorphic SSR markers were used in each cross. The ampli cation pro les of markers used for foreground selection in two crosses are shown in Fig. 2 (for details of SSRs used for background selection see Supplementary Table 1). In BC 4 F 2/3 , homozygous plants were selected and were then subjected to phenotypic selection for PHS in subsequent generations (BC 4 F 4 -BC 4 F 5 ). KASP assay involving SNPs associated with PHST was also carried out in BC 4 F 6 lines to con rm the presence of PHST QTL and the seed dormancy gene TaMKK-3 (Fig. 3). Only 62 of the 115 progenies carried the SNP allele associated with PHST in homozygous condition. It appears that during the foreground selection for PHST QTL, recombination occurred in the large intervals (14.2 cM in case of AUS1402 and 17.8 cM in case of CN19055) between the markers anking the PHST QTL, leading to the loss of PHST QTL allele in some progenies or to the selection of heterozygous allele at the PHS locus. Therefore, in future, closely linked markers need to be used. Eventually, 10 lines (8 involving AUS1408 and 2 involving CN19055 were selected, each carrying the marker associated with the desired QTL in homozygous condition, associated with 90-96% RPG, PHS score <3 and germination index (G17 at 7 days for testing seed dormancy) ranging from 0.2-0.6. RPG recovery in the two crosses in the present study did not differ from several earlier wheat breeding programmes where MABB was used for introgression of genes/QTL for biotic and abiotic stresses in wheat (Yadav et al. 2015;Mallick et al. 2015;Vishwakarma et al. 2016;Rai et al. 2018;Randhawa et al. 2019;Todkar et al. 2020: Gautam et al. 2020a). This con rmed the utility of MABB for effective introgression of genes/QTL into the agronomically superior genotypes associated with recovery of recurrent parent genome in a relatively short time as also reported earlier in wheat (Mallick et al. 2015), rice (Divya et al. 2015;Ellur et al. 2016;Ramalingam et al. 2020;Jamaloddin et al. 2020), and maize (Sureshkumar et al. 2014;Kaur R et al. 2020).
Pyramiding of QTL/genes for PHST, high GPC and resistance to leaf rust MAS derived BC 4 F 2 plants containing the PHST QTL from the donor CN19055 in the background of cv. Lok1 were used for the pyramiding of Gpc-B1, and Lr24, on to QPhs.dpivic.4A.2 (already introgressed in BC 4 F 2 plants). Markers utilized during pyramiding are listed in Table 1. In F 2 generations, foreground selection was carried out for three QTL/genes (Gpc-B1, Lr24 and QPhs.dpi.vic.4A.2) and plants homozygous/heterozygous for the three QTL/genes were selected (Fig. 4). Further foreground selection and phenotypic selection was exercised for PHS and four advanced F 5 lines were selected, which carried the above three QTL/genes in homozygous conditions and PHS score 2-3 (Fig. 5). These lines were subjected to rust screening and evaluation in yield trials.

Screening of the pyramided lines for rust resistance
The original cv. Lok1 was only moderately resistant to most of the pathotypes of the three rusts except one pathotype for each (Table 2). However, the MAS-derived parental genotype Lok1 (Gpc-B1+Lr24) used for gene pyramiding, was resistant against three pathotypes (238S119, 78S84, and 47S103) of stripe rust, all the ve pathotypes (43232, 77-5, 77-8, 77-9, and 104-2) of leaf rust, and four pathotypes (21A-2, 34-1, 40-1 and 117-6) of stem rust (Table 2). Similar was the resistance pattern of the pyramided lines containing Gpc-B1+Lr24+PHST QTL QPhs.dpivic.4A.2 (Table 2). Therefore, it is obvious that the pyramiding approach used in the present study was successful to produce improved version of cv. Lok1 with improved grain quality and rust resistance. In India, Lr24 (derived from Agropyron elongatum) continues to provide effective resistance, although virulence against Lr24 has been reported in South America, North America, and South Africa (Tomar et al. 2014). Therefore, introgression of Lr24 using MAS has also been reported in several earlier studies in India (Samsampour et al. 2009;Revathi et al. 2010;Chhuneja et al. 2011;Kaur et al. 2020). However, for maximum protection, it is recommended that Lr24 should be used in combination with other Lr genes (Bolton et al. 2008).
The Lr24 gene is also linked with Sr24, a stem rust resistance gene effective even against the deadliest race TTKSK (Ug99) (Singh et al. 2006). According to some estimates, the wheat cultivars carrying Lr24/Sr24 occupied a total of about 15 Mha over a period of 20 years (Tomar et al. 2014). In addition to the linked genes Lr24/Sr24, use of other resistance genes, namely Sr31 and Sr36 for stem rust and Lr9, Lr19, and Lr34 for leaf rust has also been suggested for breeding resistant varieties (Bhardwaj et al. 2019). In India, some of the important genes conferring resistance against the prevalent pathotypes of three rusts include the following: Lr24, Lr37, Lr76 for leaf rust; Sr24, Sr38 for stripe rust; and Yr17, Yr36, Yr70 for yellow rust . However in recent years, new virulent races have emerged for some of these genes, of which 77 is the most important race of leaf rust and pathotype 77-9 is the most prevalent. In leaf rust samples of cultivated wheat collected from India and Nepal, pathotype 77-9 was identi ed in 149 (51.1%), 77-13 in 20.2% and 77-5 in 15.1% rust samples; Lr24 provided resistance against the pathotype 77-9. Using gene pair Lr24/Sr24, 18 wheat varieties for different ecological conditions ranging from moisture stress to irrigated environments, and timely sown to late sown conditions have been released for commercial cultivation in India (Tomar et al. 2014). The presence of this gene pair in our pyramided lines was also con rmed using two stem rust pathotypes (34-1 and 40-1), where all the derived pyramided lines were found to be completely resistant to the above two pathotypes ( Table 2).

Performance of MAS-derived lines
The results of ANOVA for two yield trials involving 14 MAS derived lines along with parents [Lok1, Lok1 (Gpc-B1+Lr24), AUS1408, and CN19055] are presented in Table 3. Differences between genotypes were signi cant for ve (TGW, GY, GPC, Fe, and Zn) of the eight traits. Environment effects were also signi cant for four of the six traits (except TGW and GY). Genotype-by-environment (g × e) interactions were signi cant for only three (PH, GY, and GPC) traits suggesting signi cant effect of environment on the expression of these three traits.
Phenotypic evaluation of the PHS tolerant lines (a) Progenies containing only PHST QTL The ten (10) improved wheat lines obtained during the present study following MABB involving two crosses (8 from Lok1/AUS1408 and 2 from Lok1/CN19055) and subsequent phenotypic selection for PHST were also evaluated for important agronomic traits. All the 10 derived lines were similar to the recurrent parent cv. Lok1 for four of the eight traits (PH, TPM, GNPS, and GPC), possibly as a result of the 90-96% recovery of the RPG in these lines following four backcrosses along with background selection (Table 4). However only one line (CCSU-78) showed higher TGW, and only three lines  showed higher GY relative to the recipient cv. Lok1. Interestingly, seven lines also had improved grain Fe content (36.5-52.6 ppm) and three lines also had higher grain Zn content (28.7-35.0 ppm) relative to the grain Fe/Zn contents of the recurrent parent and each of the two donors (Table 4). Since there is no known relationship between the PHS tolerance and grain mineral contents, we speculate that the observed bioforti cation may be due to the possible accumulation of complementary genes for high grain Fe and Zn distributed in the recipient parent cv. Lok1 and each of the two donor genotypes (AUS1408 and CN19055). Future studies may be planned to understand the basis of associated bioforti cation in PHS tolerant lines.
(b) Pyramided progenies containing genes/QTLs for GPC, leaf rust, and PHST Four (4) improved Lok1 lines with QPhs.dpivic.4A.2+Gpc-B1+Lr24 (PHS score=2-3) were also evaluated for ve agronomic traits and three grain quality traits. These lines resembled the recipient cv. Lok1 for all the ve agronomic traits, but carried higher GPC (12.3% to 12.5%), and higher Fe; one line (CCSU-35) also had higher Zn (Table 4). Interestingly, the higher GPC in these lines had no yield penalty despite the known negative association of GPC with grain yield and the effect of genotype, environment and epistatic interactions on GPC (Davies et al. 2006;Brevis and Dubcovsky 2010;Kumar et al. 2011). Wheat lines with high GPC andno yield penalty were also developed earlier through introgression of Gpc-B1 gene following MAS (Kade et al. 2005;Brevis and Dubcovsky 2010;Kumar et al. 2011: Vishwkarma et al. 2016). Using the above Gpc-B1 gene, two commercial bread wheat cv. Lassik and Farnum were also developed earlier (Brevis and Dubcovsky 2010;http://variety.wsu .edu/extensionpubs/Farnumtrifold.pdf,http://uvdavis.edu/ les2/57360.pdf; for review see Balyan et al. 2013). The high GPC gene Gpc-B1 has also been exploited in breeding (mostly following MAS) for improvement of grain Fe and Zn associated with GPC in wheat (for reviews see Tabbita et al. 2017;Gupta et al. 2020). These results suggest that while improving PHS tolerance, pyramiding with Gpc-B1 can be used not only for protein enrichment but also for bioforti cation with Fe and Zn through MAS associated with judicious phenotypic selection.
In summary, a combination of marker-assisted foreground/background and phenotypic selection is a successful strategy for development of improved lines for the trait like PHST. The present study is yet another study of its kind, where MAS was utilized for the improvement of PHS in wheat. The two earlier studies were also conducted in our own laboratory for improvement of wheat cultivars for PHST (Kumar et al. 2010;Tyagi et al. 2014). However in both these earlier studies, we used a major PHS QTL located on 3A which was associated with red grain color, which is not preferred by the consumers. During this study, we not only developed PHS tolerant wheat lines with white grains, but also pyramided the PHS tolerance with high GPC and rust resistance. These improved lines should serve as new genetic resources for breeders in future wheat breeding programs.    Ampli cation pro les of markers associated with PHST QTL in parental lines.

Figure 3
KlusterCaller output view of segregation of KASP marker associated with PHST gene TaMKK3-A in BC4F6 population. Co-dominant KASP marker segregating for FAM tailed TaMKK3-A-snp1-res_FAM (blue colour) and VIC tailed TaMKK3-A-snpA1-sus_VIC (red colour) alleles on X-and Y-axes of the plot, respectively, and heterozygous individuals on mid-axis (pink colour).

Figure 4
Ampli cation pro les of markers associated with (a) Gpc-B1 and (b) Lr24 in parental lines and F2 plants.