Pyramiding of High Grain Weight With Stripe Rust and Leaf Rust Resistance in Elite Indian Wheat Cultivar Using a Combination of Marker Assisted and Phenotypic Selection

Wheat (Triticum aestivum L.) is an important cereal crop globally as well as in India and yield improvement programs encounter a strong impediment from ever-evolving rust pathogens. Hence, durable rust resistance is always a priority trait for wheat breeders globally. Grain weight, represented as thousand grain weight (TGW), is the most important yield-contributing trait in wheat. In the present study high TGW has been transferred into two elite Indian wheat cultivars PBW343 and PBW550 from a high TGW genotype, Rye selection 111, selected from local germplasm. In the background of PBW343 and PBW550, an increase in TGW upto 27.34 and 18% was observed, respectively (with respect to recipient parents), through conventional backcross breeding with phenotypic selections in 3 years replicated RBD trials. Resistance to leaf rust and stripe rust has been incorporated in the high TGW version of PBW550 through marker assisted pyramiding of stripe rust resistance gene Yr15 using marker Xuhw302, and a pair of linked leaf rust and stripe rust resistance genes Lr57-Yr40 using marker Ta5DS-2754099_kasp23. Improved versions of PBW550 with increased TGW ranging from 45.0 to 46.2 g (up to a 9% increase) and stacked genes for stripe and leaf rust resistance have been developed. This study serves as proof of utilizing conventional breeding and phenotypic selection combined with modern marker assisted selection in improvement of important wheat cultivars as a symbiont of conventional and moderan techniques.


INTRODUCTION
Wheat (Triticum aestivum L.) is an important cereal crop in India, ranking second after rice for the area (29.31 million hectares) and production (103.6 million metrics) with the state of Punjab sharing 18% of production. With global per capita consumption of 67.4 kg/year, wheat is the most widely consumed food grain (Djanaguiraman et al., 2019). In India only, the population is projected to cross the 1.70 billion mark by 2050 with a domestic demand of wheat exceeding 140 million tons (Nagarajan, 2005) 1 . A consistent increase in the wheat yields is a primary goal for food security of the growing population (Singh et al., 2007;Ye and Smith, 2008). Crop yield is a complex quantitative trait determined by different parameters of tiller number, grain number, grain weight, etc. Grain or kernel weight (1,000 grain weight in g, TGW), consisting of grain length, width, and area has high heritability (> 0.68) (Giura and Saulescu, 1996;Kuchel et al., 2007b) which not only translates into higher yields but also has a favorable effect over flour yield (Gegas et al., 2010). Moreover, uniform and larger sized grains are visually appealing and fetch higher market prices.
As rusts are major diseases in India, keeping pace of increasing yields with rust resistance is one of the most challenging tasks. The three species of rust viz., stripe rust caused by Puccinia striiformis, leaf rust caused by Puccinia triticina, and stem rust caused by Puccinia graminis are severely affecting wheat yield. An estimated loss of 200 million rupees occur every year due to rusts (Mehta, 1950). Leaf rust is prevalent in all the wheat growing zones of India and its widespread occurrence was observed during periods of 1971-1973, 1993-1994Nayar et al., 1997). The 70% of the total area under wheat cultivation in cooler parts Northern India is under constant threat of ever evolving epidemic stripe rust pathogenic races. It occurred almost every year from 1967 to 1974, with high incidence recorded in 2001 and 2011 (Nayar et al., 1997;Prashar et al., 2007;Pannu et al., 2010;Tomar et al., 2014;Pal et al., 2015). An approximate loss of rupees 236 crore have been recorded in Punjab state during the epidemic year of 2009-2010 (Jindal et al., 2012). Stem rust on other hand is important central and peninsular India since its first epidemic reported in 1786AD (Nagarajan and Joshi, 1975).
Despite the devastating nature and continuous occurrence of rusts, wheat production is growing linearly due to the continuous addition of new rust resistance genes in the wheat gene pool. Use of resistant wheat cultivars is not only effective and economical but also environment friendly (Peng et al., 2003;Singh et al., 2020). Due to pathogen evolution, combination of resistance genes is being pursued by the crop breeders world over for increasing durability of resistance. Resistance gene pyramiding has been found to increase the life of each gene though the synergistic effect of pyramided genes (Klymiuk et al., 2018;Mundt, 2018). One of the greatest successes in the history of resistance breeding is with the pyramiding of resistance genes in controlling stem rust . The Ug99 race group of pathotypes of stem rust defeated rust resistance genes Sr31 and Sr38 but their differential pyramiding combinations of different resistant genes Sr22, Sr25, Sr26, Sr33, Sr35, Sr45, and Sr50 were found to be effective. For leaf rust, pyramiding of different genes in combination with leaf rust resistance genes provided long lasting resistance (Kolmer, 1996;Bhawar et al., 2011;Aboukhaddour et al., 2020;Babu et al., 2020). Resistance to stripe rust has also been significantly improved 1 https://www.thehindubusinessline.com; https://www.eastasiaforum.org by pyramiding of different resistance genes (Zheng et al., 2017;Randhawa et al., 2019).
The present study reports the introgression of high thousand grain weight (TGW) to wheat varieties PBW343 and PBW550 from a local selection named "Rye Selection 111." Pyramiding of three rust resistance genes viz Yr15 and linked Lr57-Yr40 into improved TGW version of PBW550 was done using "transfer first and assemble later" strategy (Ishii and Yonezawa, 2007a,b). Wheat cultivar PBW343, released in India in 1995, was the most widely grown cultivar in the country and soon was recognized as the new miracle genotype with wider adaptability and yield potential (Gupta et al., 2018). PBW550 was released in 2008 and farmers' receptivity to this short duration variety with bold grains has accelerated its spread in the first few years of its release (Gupta et al., 2018). However, both of these hexaploid elite cultivars succumbed to emerging races of YR thus it became imperative to restructure the genetic makeup of these cultivars for improved yield and rust resistance.

Introgression of High Grain Weight in PBW343 and PBW550
For transfer of high TGW to wheat variety PBW343, cross was made with RyeSel 111 as male parent and the phenotypic selection was done in subsequent backcross generations (Figure 1a). The BC 2 F 5 progenies of cross PBW343-RyeSel 111 with high TGW will be referred to as HGW343 hereon. Selected HGW343 plants were used as donor to transfer high TGW to wheat variety PBW550 and BC 2 F 5 progenies with high TGW (Figure 1b) thus obtained will be termed as HGW550 hereon. All the crosses and selections were done at PAU, Ludhiana, India.

Phenotyping and Experimental Design
In both the above crosses, traits were recorded on single plants in BC 2 -F 1 /F 2 generations while in BC 2 -F 3 /F 4 /F 5 progenies, traits were recorded on 10 single plants of each progeny as well as bulk of progeny. Single plants from uniform and promising progenies were selected and carried forward. Evaluation of each of the BC 2 -F 3 /F 4 /F 5 progenies along with parental lines were done in three replications (1.5 m paired rows with the plant to plant distance of 10 cm and row to row distance of 20 cm) in randomized block design. Wherever single plant selections were done, the weight of 300-500 grains (as available per plant) was recorded and converted into TGW, while the weight of thousand grains was recorded for progeny bulks. Selections were done for plant/progenies having TGW higher than recurrent parent. Similarly other phenotypic traits of plant height, tiller number, spikelets per spike, spike length, and grains per spike were also recorded and plants/progenies having either similar value to or higher than recurrent parent were selected (data not given).
For each progeny the mean values were calculated for recorded traits in each replication and adjusted means of replications were finalized for each year. Three years data was again used to calculate adjusted means across the years for final selections.

Combining High Grain Weight and Rust Resistance in PBW550
For pyramiding HGW with stripe rust resistance gene Yr15, first the selected HGW550 progenies obtained from the above cross were crossed with NIL PBW550 + Yr15. The HGW550 + Yr15 thus obtained was again crossed with another NIL, PBW550 + Lr57-Yr40, to develop the HGW550 + Yr15 + Lr57-Yr40 line. Selections were done for high TGW and rust resistance as for single plant in F 1 /F 2 generations, while in F 3 /F 4 /F 5 progenies 10 single plants along with progeny bulks recorded to select single plants from best performing progenies. Since both the parental lines were from the same background, no backcrossing was done.
Marker Assisted Selection for Yr15 and Lr57-Yr40 Genes Selection for rust resistance was done in a combined phenotypicgenotypic manner. For marker assisted selection, DNA from parental lines and segregating progenies was extracted using CTAB method (Saghai-Maroof et al., 1984) with the small modification of reducing incubation time to 30 min at 65 • C with CTAB buffer and to 20 min at room temperature for solvent extraction using chloroform: isoamyl alcohol (24:1). Yr15 positive plants were selected by amplifying gene specific marker Xuhw302 (Klymiuk et al., 2018). For the selection of Lr57-Yr40 gene, linked Kasp marker Ta5DS-2754099_kasp23 (Bansal et al., 2020) and caps marker Lr57-Yr40_caps16 (Toor et al., 2016) were used. The PCR reaction for gel based marker Xuhw302 and Lr57-Yr40_caps16 was done in 10 µl reaction volume (60 ng DNA, 5 µl of 2× EmeraldAmp R GT PCR Master Mix, 0.75 µl of 5 µM each primer) in 384 well microtiter plate in an Applied Biosciences 384 thermal cyclers. The PCR products were resolved using 2.5% agarose gel electrophoresis and visualized and photographed using gel documentation system. The scoring was done identifying the presence of gene specific amplicon as per the positive control. The Kasp marker Ta5DS-2754099_kasp23 was amplified in 4 µl reaction (20 ng DNA, 1.944 µl of 2X KASP V3.0 master mix from LGC, Biosearch Technologies, 0.056 µl of primer mix in ratio of 12:12:30:46 Allele specific Primer I:Allele specific II primer: Common primer: Water) in a 384 well microtiter plate. The amplicons were identified by measuring allele specific flourescence in a high throughput TECAN infinite F200 PRO plate reader. Kluster Caller software was used to view the allele specific calls in an x-y plot to identify the positive and negative allels against the positive control.

Screening Against Stripe Rust and Leaf Rust
Screening for stripe rust and leaf rust was also done in the field along with marker selection to validate the effectiveness of gene pyramiding. F 3 and F 4 progenies of HGW550 X PBW550 + Yr15 were screened against stripe rust at the adult plant stage in the field during season 2016-2017 and 2017-2018. Similarly, F 3 and F 4 progenies of HGW550 + Yr15 X PBW550 + Lr57-Yr40 were screened both against stripe rust and leaf rust at the adult plant stage in the field during season 2018-2019 and 2019-2020. For screening, artificial epiphytotic conditions for rust were created by spraying the urediniospores mixed and diluted in water, containing Tween-20, of Pst pathotypes (100S119, 78S84), and Pt pathotypes (77-1, 77-2, 77-5, 104-2) and a mixture of leaf rust and stripe rust inoculum collected from farmers' fields. For the uniform spread of disease, highly susceptible cultivar WL711 was planted as spreader rows all around the field and after every 20 rows. Rust data was recorded when WL711 showed complete susceptibility. Rust was recorded using Cobbs scale, as illustrated in McIntosh et al. (1995) where types of spores were recorded as zero (immune); TR (traces of severity); MR (moderately resistant), MS (moderately susceptible); and S (susceptible), and numeric numbers associated with these scores signified the percentage of leaf area covered by rust. The TR score is given to a plant when there are no lesions, MR when there are no visual postules of spores on the leaf but apoptotic lesions are visible, MS when there are minute visual postules of spores on the leaf with apoptotic lesions while S score is given to a plant when advanced postules of spores are visible to the leaf.  Figure 2). Six BC 2 F 5 plants of progenies with significantly high TGW, namely HGW343-3 (TGW-45.36 g, 24.13% increment), HGW343-6 (TGW-45.54 g, 24.63% increment), HGW343-8 (TGW-46.08 g, 26.11% increment), HGW343-60 (TGW-46.48 g, 27.20% increment), HGW343-61 (TGW-46.71 g, 27.83% increment), and HGW343-11 (TGW-46.81 g, 28.10% increment) were  selected (Figure 3). The popularity of PBW343 in the late 1990s and early 2000s led to its widespread sowing which facilitated the selection of virulence for a super aggressive stripe rust race 78S84 (Prashar et al., 2007), breaking down resistance of its predominant gene Yr27, crumbling the wheat production in India. For restructuring the resistant version of PBW343, the high TGW progenies in PBW343 background were selected and disease resistance gene pyramiding carried out under a separate breeding program (not included in the present manuscript).

Marker Assisted Pyramiding of HGW550 and Yr15
Three selected HGW550 plants; HGW550-8, HGW550-6, and HGW550-63 were crossed with NIL PBW550 + Yr15 (Figure 1c). F 1 s thus obtained were selfed to develop 798 F 2 plants and 439 stripe rust resistant plants with TGW of more than 42 g were selected. In F 3 , plants having a TGW equal to or more than 45.0 g were selected from promising uniform stripe rust resistant progenies. The selected plants were screened with molecular marker Xuhw302, and 83 plants, homozygous for Yr15 were selected (Supplementary Figure S1A). These 83 plants were further sown in plant to progeny rows in F 4 and 23 single F 4 plants with TGW ranging between 45.0 and 48.0 g and positive for Yr15 were selected (Figure 4). These plants were also completely resistant to stripe rust in the field.
Marker Assisted Pyramiding of HGW550 + Yr15 and Lr57-Yr40 from superior progenies with TGW between 45.0 and 46.2 g and complete resistance to stripe rust and leaf rust have been selected (Figure 4).

DISCUSSION
The stacking of genes governing multiple traits enhances the value of breeding material besides improving its durability. Yield is a trait of foremost priority for the commercial success of a variety, but combining improved grain yield with disease resistance, and high grain quality is a need of the changing time and environment. The introgression and pyramiding of major genes/QTL for different traits through marker-assisted selection (MAS) has been reported in wheat (Gupta et al., 2010;Kumar et al., 2010;Tyagi et al., 2014;Gautam et al., 2020). We used RyeSel 111 as the donor to transfer the high TGW trait to elite cultivar PBW343 and PBW550 through conventional phenotypic selection. Though QTLs for high grain weight have been reported in the RyeSel 111-Chinese Spring RIL population (Kumar et al., 2006), the reported linked markers were found to be non-polymorphic between the receipient and donor combinations used in the present study. Thus a conventional route is followed through phenotypic selection for high TGW. Mapping and transfer of QTLs for yield related traits often sound more logical theoretically, but these complex traits are controlled by many QTLs of large and small effects and pyramiding all these QTLs in one background through MAS to retrieve the donor effect is not always achieved. This is more often pronounced in cases where the minor QTLs additively contribute toward the trait along with major ones. Thus phenotypic selection especially for an easily quantified trait like grain weight, gives more realistic results. Wheat yield is controlled by traits of yield per area (includes grains per spike, grain weight, and spikes per area) and yield per spike (spikelet number per spike, grain number, and grain size) (Slafer et al., 2014). Several QTLs related to grain yield related traits have been mapped on different chromosomes but none have been cloned or effectively utilized for marker assisted selection. From different factors influencing wheat yield, grain weight was found to be one of the most stably inherited (Sidwell et al., 1976;Kuchel et al., 2007a,b;Aisawi et al., 2015) suggesting selection for heavier grains lead to effective yield improvement. Raising plant yield by conventional breeding methods ) has remained very successful in the past, though new emerging technologies are creating a different niche. High TGW was introgressed effectively into two popular wheat cultivars cv. PBW343 and PBW550 through the phenotypic selection, and improved progenies with an increase of TGW by 27.34 and 18% were obtained, respectively. Initially, direct crosses of RyeSel 111 were done with PBW343 then using improved HGW343 progenies as the donor, HGW introgressed into PBW 550. Similarly improvement in grain yield by improving TGW has been reported by several studies (Giunta et al., 2007;Beche et al., 2014;Qin et al., 2015;Zhang et al., 2016;Gao et al., 2017).
PBW343 has been one of the most popular wheat varieties and was grown in about 25% of the 27 million hectares under the wheat cultivation in the country and contribute roughly 55% of the total wheat output in the country (Pavithra et al., 2017). Due to its wider adaptability to a range of environments, it has been a variety of choice for many improvement programs and was selected in this program for improving grain weight. After the PBW343 succumbed to 78S84 pathotype of stripe rust in 2004 and so did the whole breeding pipeline at PAU, it became mandatory to quickly mobilize some known stripe rust resistance genes to PBW343 and other advance germplasms to have a rust resistant variety for the farmers of the region. It was then that the development of rust resistant version of PBW343 became the main mandate of the wheat breeders, and reintroduction of its rust resistant version led to the shift of this work to a separate program.
PBW550, on the other hand, has been recognized as a good quality cultivar with bold grains and was spread to whole wheat growing regions of the country in the few years since its release in the year 2008. This variety was tested and also recommended for cultivation in Eastern and Central India for processing the commercial wheat flour under the PAU-ITC (Indian Tobacco Company) agreement. However, the evolution of the 78S84 pathogen was so targeted and led to the development of a very aggressive strain which not only rendered PBW343 susceptible to stripe rust but also slowly wrapped up PBW550 including all the newly released varieties during that time. As a result, PBW550 too became susceptible to yellow rust after 6-7 years of release despite having multi-pathogen resistance gene loci Lr34/Yr18/Sr57. The short-lived resistance makes it of utmost importance to introgress rust resistance in addition to yield improvement in these cultivars.
Though HGW has been introduced with backcross breeding only while the rust resistance was facilitated through MAS. Increased adoption of combining conventional breeding with the MAS approach (Gupta et al., 2010) in recent years has given multifold advantages, the major one being accelerated mobilization of the desirable gene(s) and their efficient stacking in elite backgrounds. Two different NILs carrying stripe rust resistance gene Yr15 and linked leaf rust-stripe rust resistance gene Lr57-Yr40 have been developed by PAU. NIL PBW550 + Yr15 has also been released as a variety for cultivation in Punjab under timely sown irrigated conditions as Unnat PBW550. Yr15 mapped near distal Nor-B1 on the short arm of chromosome 1B was first discovered in the 1980s in wild emmer wheat T. turgidum var. dicoccoides. Avocet + Yr15 was initially used as a donor to transfer this gene in PBW550 background. As per the Global Rust Reference Centre 2 , Yr15 has been providing a complete and broad spectrum resistance (Klymiuk et al., 2018). Linked Lr57-Yr40 (on chromosome 5DS) has been transferred from an introgression line INGR15046, developed in the wheat wide hybridization program of PAU by transfer of leaf rust and stripe rust resistance from Aegilopes geniculata to wheat cultivar WL711 (Kuraparthy et al., 2007). This linked gene provides complete resistance to leaf rust and partial resistance (20MR) to stripe rust in field. 2 https://agro.au.dk/forskning/internationale-platforme/wheatrust/ For pyramiding, the "transfer first and assemble later" approach (Ishii and Yonezawa, 2007a,b) already given above was followed where PBW550 + Yr15 and PBW550 + Lr57-Yr40, NILs were used as donors of respective genes. Since both the parental genotypes were in the PBW550 background, backcrossing was not required. Similarly, HGW550 + Yr15 and PBW550 + Lr57-Yr40 were crossed, and selfing generations F 2 , F 3 , and F 4 were evaluated for HGW and two resistances to leaf and stripe rust leading to the selection of 11 HGW550 + Yr15 + Lr57-Yr40 progenies with three rust resistance genes and high TGW. Gene pyramiding of disease resistance genes reported improving the durability of resistance, though the combined effect depends upon the nature of the individual genes and their synergism. Resistance breeding by marker led pyramid in the last decade is being used in wheat programs globally. Moreover, the combination of major and minor genes was also found to have more significant effects. Although Yr15 has already succeeded in conferring stripe rust resistance for many years in different introgression lines around the world, marker assisted gene pyramiding of Yr15, and Lr57-Yr40 genes provides durable resistance to stripe rust in combination with leaf rust resistance. Advanced breeding lines of PBW550 with three rust resistance genes and high TGW can lead to the development of new varieties and also serve as valuable germplasm for breeders to be used in the varietal development program to aid future breeding programs.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s. The raw data for trials has been added as Supplementary Table S1.

AUTHOR CONTRIBUTIONS
SK conceived and designed the research, planned, and conducted the experiments, did rust screening, compiled the work, and produced the final draft of the manuscript. JK and RS conducted field experiments and did marker assisted selection. GM and AS conducted the field experiments. GD did statistical analysis and compilation of the manuscript. UD helped in field experiments. PC conceived and designed the research, and reviewed the manuscript. All authors contributed to the article and approved the submitted version.