Application of Genomics Tools in Wheat Breeding to Attain Durable Rust Resistance

Abstract

Wheat is an important source of dietary protein and calories for the majority of the world’s population. It is one of the largest grown cereal in the world occupying over 215 M ha. Wheat production globally is challenged by biotic stresses such as pests and diseases. Of the 50 diseases of wheat that are of economic importance, the three rust diseases are the most ubiquitous causing significant yield losses in the majority of wheat production environments. Under severe epidemics they can lead to food insecurity threats amid the continuous evolution of new races of the pathogens, shifts in population dynamics and their virulence patterns, thereby rendering several effective resistance genes deployed in wheat breeding programs vulnerable. This emphasizes the need to identify, characterize, and deploy effective rust-resistant genes from diverse sources into pre-breeding lines and future wheat varieties. The use of genetic resistance has been marked as eco-friendly and to curb the further evolution of rust pathogens. Deployment of multiple rust resistance genes including major and minor genes in wheat lines could enhance the durability of resistance thereby reducing pathogen evolution. Advances in next-generation sequencing (NGS) platforms and associated bioinformatics tools have revolutionized wheat genomics. The sequence alignment of the wheat genome is the most important landmark which will enable genomics to identify marker-trait associations, candidate genes and enhanced breeding values in genomic selection (GS) studies. High throughput genotyping platforms have demonstrated their role in the estimation of genetic diversity, construction of the high-density genetic maps, dissecting polygenic traits, and better understanding their interactions through GWAS (genome-wide association studies) and QTL mapping, and isolation of R genes. Application of breeder’s friendly KASP assays in the wheat breeding program has expedited the identification and pyramiding of rust resistance alleles/genes in elite lines. The present review covers the evolutionary trends of the rust pathogen and contemporary wheat varieties, and how these research strategies galvanized to control the wheat killer genus Puccinia. It will also highlight the outcome and research impact of cost-effective NGS technologies and cloning of rust resistance genes amid the public availability of common and tetraploid wheat reference genomes.

Introduction

During the evolution of agriculture, intense changes in the biology of a wider range of traits of cereals have been documented (Preece et al., 2017). These changes assisted in the domestication of wheat and its cultivation in the Fertile Crescent, 10,000 years ago (Nevo et al., 2013). Wheat is a segmental allohexaploid (2n = AABBDD = 42) accommodating three inter-related diploid genomes (Marcussen et al., 2014). Around seven million years ago, diploid progenitors of the AA genome (Triticum urartu) and BB genome (closely related to Aegilops speltoides) diverged from a common ancestor (Marcussen et al., 2014), and during evolution, both of these genomes combined through a polyploidization event which led to the evolution of emmer wheat (T. turgidum), an allotetraploid (AABB) around half a million years ago (El Baidouri et al., 2017). Around five million years ago, DD genome (Ae. tauschii) evolved from these two diploid progenitors through homoploid speciation (Sandve et al., 2015; El Baidouri et al., 2017; Pont and Salse, 2017). Common wheat (AABBDD) evolved 8,000 to 10,000 years ago through hybridization between T. turgidum and Ae. tauschii. Further, domestication, cultivation, and selection by mankind have refined the agronomical features of common wheat (Kihara, 1944; McFadden and Sears, 1946). Wheat is an extensively cultivated cereal across the globe covering a wide range of altitudes except for lowland tropical regions (Feldman, 1995). As per global grain production statistics, around 734 million tonnes of wheat was produced in the crop year 2018 at 215 million ha around the world which puts wheat second to maize (Zea mays L.) in production among the cereal crops (FAOSTAT, 2020). Common wheat covers around 95% of the total wheat acreage, whereas durum wheat and others occupy the remaining 5% (Shewry, 2009). The world population is projected to reach 9.7 billion by 2050 (Elferink and Schierhorn, 2016), and around 70% in additional production is needed to fulfil the global wheat demand.

Among many biotic stresses which hinder wheat production, rust pathogens are considered as most pressing threats with large economical losses worldwide (Ellis et al., 2014). Stem rust and stripe rust can cause 100% loss, whereas leaf rust can result in 50% loss under severe rust epidemics (Bhardwaj et al., 2019). To minimize the loss incurred by the rust pathogens, there are two effective approaches: 1. chemical control via fungicidal spray and 2. genetic control via breeding for rust resistance (Asad et al., 2012). The International Maize and Wheat Improvement Centre (CIMMYT), Mexico has been rigorously engaged in improving wheat for rust resistance since its inception in the 1940s, and rich dividends up to 27:1 benefit-to-cost ratio have been attributed to the development of resistant cultivars through genetic approach (Kolmer et al., 2009). The worldwide estimated losses caused by wheat rusts were as high as USD 170 million for stripe (Pakistan) (Hussain et al., 1980), AUD 100–200 million for stem (Australia) (McIntosh et al., 1995) 83 and USD 100 million for leaf rust (Pakistan) (Duveiller et al., 2007), and hence the economic value of rust diseases has been a driving factor for research funding. Breeding for rust resistance is an environmentally safe approach adapted by wheat breeders permitting a reduction in the use of pesticides and greater yield stability (Bariana et al., 2013).

Diverse accessions including diploid progenitors, landraces, old cultivars, and synthetic wheat have an inherent ability to restrict the infection induced by the rust pathogens. Such class of resistance is termed as host resistance, and it can be classified into two broad categories, All Stage Resistance (ASR) and Adult Plant Resistance (APR). ASR is active from the seedling to adult plant stages and generally is only effective against specific races of the pathogen. APR is controlled by non-race specific genes with exceptions where some genes provide race-specific resistance such as Lr13 (Ellis et al., 2014), while Lr48 confers hypersensitive response (Bansal et al., 2008). These genes express resistance at post seedling stages when adult plants are close to flag leaf or leaf emergence at boot stage. The effect of each APR gene is partial, not enough to cope up high disease pressure; however, combining APR genes provides high levels of resistance (Singh, 1992). The resistance effect of ASR is neutralized due to selection pressure on the pathogen to evolve or mutate to acquire virulence against resistance genes. Such genes when deployed alone result in boom and bust cycles. The successful breeding for rust resistance requires a deep understanding of evolutionary nature of the pathogen, identification of diverse resistance sources, and judicious deployment of resistance genes in combinations (McIntosh et al., 2001).

Development of gene-specific DNA markers is desirable to expedite pyramiding of rust resistance genes in future wheat varieties. Enormous genome size, complexity, repetitive sequences, and extreme level of similarity between homoeologous genome sequences (95–99% in coding regions), and a higher proportion of repetitive elements (over 80%) have posed a challenge to develop robust DNA markers and draft wheat genome sequences (Borrill et al., 2015). In the last decade, polymerase chain reaction-based markers, including Simple Sequence Repeat (SSR), Sequence Tagged Site (STS), and Cleaved Amplified Polymorphic Sequences (CAPS) were largely explored as promising tools for wheat improvement (Somers et al., 2004; Rosewarne et al., 2013). Wheat genomics has been a witness to the advancement in Next-Generation Sequencing technologies (NGS) from whole-Genome Shotgun (WGS) approach (Brenchley et al., 2012), flow-sorting technology (Vrána et al., 2012), chromosome arm-specific sequencing (IWGSC, 2014; Mayer et al., 2014), long-range sequencing via the WGS approach (Chapman et al., 2015), and de novo whole-genome assembly (Appels et al., 2018). Development of SNP-based breeder-friendly KASP (Kompetitive Allele-Specific PCR) markers is the outcome of these technologies. These cost-effective technologies have been utilized to isolate rust resistance genes via positional cloning and consequently the development of perfect markers for effective ASR and APR genes (Nsabiyera et al., 2020). In this review, we attempt to revisit the nature of rust pathogen resistance sources and available genomic tools useful in gene-specific marker development.

Wheat Rust Diseases and Their Impact on Productivity

Three rust diseases namely leaf rust caused by Puccinia triticina Eriks (Pt), stem rust caused by P. graminis Pers. f. sp. tritici Eriks. & Henn. (Pgt), and stripe rust caused by P. striiformis Westend. f. sp. tritici (Pst) afflict wheat. Rust fungi are obligate plant parasites which include over 7,000 species having agricultural and economic importance (Aime et al., 2018). The genus Puccinia belongs to the Pucciniale order. The genus evolved around 235 million years ago and is characterized by vast genome size ~380Mb (Tavares et al., 2014). Since the evidence of rust spores dating back to 700BC, rust diseases have been a formidable challenge to wheat cultivation. Roman farmers offered dogs, sheep, incense, and wine to the rust god “Robigus” to be propitious (Zadoks, 1985). Individual species of Puccinia require a conducive environment regimes for the proliferation of rust inoculum followed by their dispersal via wind (Singh and Rajaram, 1993). Stem rust is usually favored under warm and humid (≤30°C) conditions. Leaf rust pathogen also prefers humidity but comparatively lower temperatures (20–24°C) Stripe rust proliferates under cooler temperatures (12–20°C) (Singh et al., 2011).

The Puccinia spp. inciting rusts on wheat are obligate parasites having five spore stages viz., spermatia, aeciospores, urediniospores, teliospores, and basidiospores in its complex life cycle (Money, 2016). Wheat is a primary host for rust fungi. Berberis species and Mahonia species are alternate/secondary hosts of Pgt and Pst, while Thalictrum, Anchusa, Isopyrum, and Clematis are the alternate hosts of Pt (Jin et al., 2010; Wang et al., 2015). Teliosori containing teliospores/teleutospores replace urediniosori as wheat crop matures. The thick-walled, bicelled teliospores can resist overwintering on stubbles. During the onset of spring, the teliospores produce a promycelium which bear four uninucleate haploid basidiospores from each cell. These basidiospores infect the Berberis leaves to form pycnial cups/spermagonia containing spermatia/pycniospores and receptive hyphae. Transfer of spermatia (sexual gametes) to receptive hyphae between oppositely charged (+ or −) spermatia and receptive hyphae on barberry leaves with the help of insects completes the sexual reproduction. The resultant dikaryotic spores (aeciospores) are formed on the lower side of the Berberis leaves. Aeciospores are dispersed through the wind to fall on wheat leaves to penetrate the stomata by the formation of an appressorium (Money, 2016). Further, infectious mycelia transform into haustorium that initiates further infection and spread in wheat without affecting its plasma membrane. After a week of infection on wheat, urediniospores are produced in millions on a susceptible variety. During the crop season, rust pathogens can produce numerous cycles of urediniospore production, its dispersal takes place through wind, and lead to build-up of epidemic over a vast region (Aime et al., 2018). In countries like India, alternate hosts are not functional and urediniospores act as repeating spores. The rusts survive on off-season wheat crop or grasses and become active in the season to cause infection on wheat through urediniospores (Bhardwaj et al., 2019).

Among rust diseases, leaf rust is the most common wheat disease worldwide (Huerta-Espino et al., 2011). Due to its wide adaptation, leaf rust fungi can initiate their infection and spread under mild temperatures and moist conditions. Although, yield losses caused by this disease display temporal and geographical variation; the economic significance of the disease is substantial (Figueroa et al., 2016). Devastating nature of the rust pathogen became more apparent after the susceptible response of a North American wheat variety “Thatcher” against leaf rust pathotypes (Pt; Anderson, 1963). This epidemic devastated wheat crop on a large area and forced researchers and policymakers to consider Pt as an alarming pathogen in the USA, former USSR, and China. Higher rate of reproduction, the ability to travel long-distance, the potential to survive on various host species and diverse environments increased aggressiveness and genetic diversity of the pathogen (Watson, 1981; Walter et al., 2016). In India, severe leaf rust epidemics occurred due to susceptibility of green revolution wheat varieties “Kalyansona” and “Sonalika” and, contemporary cultivars K68 and Pbc 306 in the Northwest region of the country during 1971–73. Over one-million-ton yield losses were estimated (Joshi et al., 1975). Similarly, another leaf rust epidemic was evident due to susceptibility of Mexican mega-wheat variety “Jupateco 73” in Northwestern Mexico. It occupied 75 per cent of the wheat acreage in the southern Sonora region within four years of its release (Dubin and Torres, 1981). Another leaf rust epidemic was reported in the eastern Prairies due to gained additional virulence against Lr13 and Lr21 in 1999 (Aboukhaddour et al., 2020).

Significant yield losses in areas with cool and high-altitude climates are commonly inflicted by stripe rust (Roelfs et al., 1992; Singh et al., 2000). In the majority of wheat-growing areas stripe rust have shown regular regional crop losses of up to 0.1–5% and in few exceptions, the losses extend up to 5–25% (Wellings, 2011) besides, severe epidemics due to the breakdown of genes in larger areas may inflict up to 70% yield losses. One-billion USD annual loss is estimated due to the severity and around ninety per cent of the world’s wheat crop is vulnerable to stripe rust (Schwessinger, 2017). Stripe rust infection at early growth stages can cause complete yield losses through a reduction in grain quality, seed vigor and poor germination in susceptible cultivars (Chen, 2005). In Asia, around 46% of yield losses have been due to stripe rust epidemics (Singh et al., 2004a). Tajikistan and Uzbekistan also witnessed widespread epidemics of stripe rust in 2010 (Sobhy Draz, 2019). In India, stripe rust is a serious threat in over ten million hectares of Northern India. Restriction on the cultivation of a mega wheat variety of this region namely PBW343 was imposed due to acquired virulence of Yr9 during 2008–09. The disease pressure was recorded up to 80%, which necessitated the cultivation of stripe rust-resistant wheat varieties (Aggarwal et al., 2018; Bhardwaj et al., 2019).

Stem rust can cause greater damage when susceptible wheat cultivars are grown over wide geographical regions and early infections may lead to large scale epidemics in a short period (Bhavani et al., 2019). Severe stem rust pressure can affect tillering (Rowell and Roelfs, 1976), plant growth by reducing the photosynthetic area, nutrients and water supply, and their transportation in the plant system (Roelfs, 1985). The two most severe epidemics of Europe occurred in Scandinavia (1951) and central Europe (1932) causing nationwide yield losses up to 20–33% (Zadoks and Bouwman, 1985). During, 1948–56 a few severe epidemics were evident in China due to higher than average temperatures and frequent rains which created a conducive environment for the infections. Other historical events of wheat losses due to this disease occurred in Australia during 1973 (AUD 100–200 million loss, McIntosh et al., 1995), United States (North Dakota) during 1919–1923 (1.95 million tons loss, Roelfs, 1978) and 1950–1955 (3.74 million tons, Roelfs, 1978), New Zealand in 1980 (Beresford, 1982), continental Europe, and the Indian subcontinent (Chaves et al., 2013; Singh et al., 2015). The last major stem rust epidemics occurred in Ethiopia in 1993 and 1994 (Shank, 1994). In a major outbreak in 1998, a new race (Ug99) of stem rust pathogen virulent to Sr31 was first reported in Uganda (Pretorius et al., 2000), and its lineages (TKTTF) have caused 100% yield losses on susceptible wheat cultivars during 2013–14 in Ethiopia (Singh et al., 2015). It is still considered to be a potential threat for Kenya, Ethiopia, Yemen, the Middle East, and South Asia. The overall global crop losses associated with Pgt alone were estimated to be approximately USD 53.7 billion per annum (Pardey et al., 2013).

Identification and Mapping of Rust Resistance Genes

Breeding for genetic resistance assists in developing rust-resistant wheat varieties. Genetic resistance can be categorized in three groups: 1. monogenic resistance (resistance governed by a single gene), 2. oligogenic resistance (resistance governed by few genes) and 3. polygenic resistance (resistance governed by many genes with additive effects) (Russel, 1978). Based on the resistance exhibited, individual rust resistance genes have been named as major genes and minor genes (Russel, 1978). In pathological terms, rust resistance could be distinguished as horizontal resistance and vertical resistance. Horizontal resistance is race non-specific resistance and effective against genetic variants of a rust pathogen. Vertical resistance is effective against certain pathogen races, so it is also called race-specific resistance (Van der Plank, 1963). The term durable resistance is frequently used instead of non-race specific or horizontal resistance to describe long-lasting resistance. The durability does not mean that resistance is active against all genetic variants of a pathogen but simply that the resistance was effective for a long time (Johnson and Law, 1975). Durable resistance may be conferred by a single gene or by combinations of many genes (McIntosh and Watson, 1982). The theory of gene-for-gene relationship implies that for each resistance gene in the host, there is a corresponding and specific avirulence gene in the pathogen (Flor, 1956). This relationship shows two reactions: 1. Incompatible reaction and 2. Compatible reaction. The incompatible reaction can be evident when a pathogen carries avirulence alleles corresponding to the host’s resistant alleles which results in a resistance response. Compatible reaction reveals a susceptible response when a pathogen carries virulence alleles to a susceptible host. The pathogen could carry many avirulence genes; each one corresponding to a resistance gene of a cultivar (Person, 1959). The gene-for-gene hypothesis defines existing diversity in physiological races of a pathogen and resistant genes among the relevant host species.

Wild progenitor and related species of wheat constitute an immense reservoir of genetic variation important for several economic traits including resistance to biotic stresses. This had resulted in an opportunity to breeders for transferring resistance genes from wild species to modern wheat cultivars through inter-specific hybridization (Bansal et al., 2017). Over two-hundred twenty rust resistance genes comprising 60 stem rust resistance, 79 leaf rust resistance genes and 83 stripe rust resistance genes have been formally named and catalogued using bi-parental populations in wheat (McIntosh et al., 2017; Li et al., 2020a). During the last decade, discovery and precise mapping of rust resistance genes expedited, and robust breeder-friendly linked-markers have been reported for several genes ( Tables 1A, B). This was largely possible due to the availability of high throughput genotyping platforms including DArTSeq (www.diversityarrays.com), i-select 90K SNP Infinium array (Wang et al., 2014) and genotyping-by-sequencing approaches (Poland and Rife, 2012).

Table 1A.

Seedling rust-resistant genes and their linked markers.

Seedling rust-resistant genes	Location	Donor	Linked Marker(s)	References
Yr15	1BS	A WEW accession G25	uhw264 and uhw259	Klymiuk et al., 2018
Yr51	4AL	A wheat landrace AUS27858	owm45F3R3 and sun104	Randhawa et al., 2014
Yr53	2BL	A durum wheat accession PI 480148	wmc441 and LRRrev/NLRRrev350	Xu et al., 2013
Yr60	4AL	A wheat line Almop	wmc219 and wmc313	Herrera-Foessel et al., 2014a
Yr61	7AS	A Chinese cultivar Pindong34	STS5467 and STS5765b	Zhou et al., 2014a
Yr63	7BS	A wheat landrace AUS 27955	IWB33120 and IWB52844	McIntosh et al., 2017
Yr70/Lr76	5DS	An introgressed line IL393-4 carrying gene from Ae. umbellulata	gwm190	Bansal et al., 2017
Yr73	3DL	An AUS cv. Avocet R (YrA resistance was inherited by two complementary dominant genes Yr73 and Yr74)	DArT markers 116719[F]0	Dracatos et al., 2016
Yr74	5BL	An AUS cv. Avocet R (YrA resistance was inherited by two complementary dominant genes Yr73 and Yr74)	DArT markers 1091695[F]0	Dracatos et al., 2016
Yr81	6AS	Wheat landrace AUS27430	KASP_3077 and gwm459	Gessese et al., 2019
Yr82	3BL	Wheat landrace AUS27969	sun KASP_300 and sunKASP_8775	Pakeerathan et al., 2019
Lr16/Sr23	2BS	A Canadian spring wheat line BW278 (AC Domain*2/Sumai3)	2BS_5175914_kwm847	Kassa et al., 2017
Lr23	2BS	A common wheat genotype BT-Schomburgk selection and durum wheat cultivar Tamaroi	sunKASP_16 and sunKASP_47	Chhetri et al., 2017
Lr24/Sr24	3DL	A wheat variety Agent carrying a translocation from Agropyron elongatum.	Sr24#12-F and Sr24#12-R	Mago et al., 2005
Lr37/Sr38/Yr17	2AS#2NS	T. ventricosum accession #10	VENTRIUP and LN2	Helguera et al., 2003
Lr52/Yr47	5BS	A wheat landrace Aus28183	sun180	Qureshi et al., 2016
Lr65	2AS	A swiss spelt wheat ‘Altgold Rotkorn’	gwm614 and barc124	Mohler et al., 2012
Lr71	1B	A swiss spelt wheat cv. Altgold Rotkorn	gwm18 and barc187	Singh et al., 2001
Lr72	7BS	A Mexican durum wheat cultivar Atil C2000	wmc606	Herrera-Foessel et al., 2013
Lr73	2BS	A wheat genotype Morocco	wPt-4453 and wPt-8235	Park et al., 2014
Lr79	3BL	A durum wheat landrace Aus26582	sun786	Qureshi et al., 2018c
Sr13	6AL	An emmer wheat Khapli and Ethiopian accession ST464	EX24785	Zhang et al., 2017
Sr26	6AL	A wheat genotype Avocet ‘S’ carrying an alien segment 6Ae#1L from Ag. elongatum	sunKASP_224 and sunKASP_225	Qureshi et al., 2018b
Sr52	T6AS·6V#3L	A wheat-Dasypyrum villosum disomic addition line DA6V#3	BF201083 and BE490365	Qi et al., 2011
Sr53	5DL	TA5599 (T5DL-5MgL·5MgS) carrying an alien segment from Aegilops geniculata	BE443021/MboI and BE442814/HaeIII,	Liu et al., 2011
Sr54	2D	A winter wheat cultivar Norin40	barc228 and wmc181	Ghazvini et al., 2012
Sr60	5A	T.monococcum accession PI 306540	ucw530 and ucw540	Chen S. et al., 2020

Table 1B.

Adult plant rust-resistant genes in wheat and their linked markers.

Adult plant rust resistance genes	Location	Donor	Linked Marker(s)	References
Yr48	5AL	Synthetic wheat PI610750	wmc727 and cfa2149	Lowe et al., 2011
Yr52	7BL	A wheat germplasm PI 183527 carrying HTAP resistance	barc182 and wgp5258	Ren et al., 2012
Yr54	2DL	A wheat line Quaiu3	wpt-667162, wpt-667054 and gwm301	Basnet et al., 2014
Yr56	2AS	A landrace AUS 91575 and Wollaroi (AUS 99174)	sun167 and sun168	McIntosh et al., 2017
Yr59	7BL	Iraqi germplasm PI 178759	wgp5175 and barc32	Zhou et al., 2014b
Yr62	4BL	Spring wheat germplasm PI 192252	gwm192 and gwm251	Lu et al., 2014
Yr68	4BL	Undesignated International Nursery ex New Zealand 03.25 (AGG91587WHEA)	IWB74301 and IWA4640	McIntosh et al., 2017
Yr71	3DL	An Australian wheat variety Sunco	gwm114b, KASP_16434, KASP_17207 and KASP_20836	Bariana et al., 2016
Yr75	7AL	An Australian wheat variety Axe	IWB34640 and cfa2040	McIntosh et al., 2017
Yr77	6DS	A common wheat PI322118	IWA167 and barc54	McIntosh et al., 2017
Yr78	6BS	A common wheat PI519805	IWA7257 and IWA4408	Dong et al., 2017
Yr79	7BL	A common wheat PI182103 carrying HTAP resistance	barc72 and wmc335	Feng et al., 2018
Yr80	3BL	A wheat landrace AUS27284	sunKASP_65624 and sunKASP_53113	Nsabiyera et al., 2018
Lr68	7BL	A common wheat cultivar Parula	cs7BLNLRR-F/cs7BLNLRR-R	Herrera-Foessel et al., 2012
Lr74	3BL	A common wheat genotype BT-Schomburgk selection	GBS2256311 and IWB69699	McIntosh et al., 2017
Lr75	1BL	A Swiss winter wheat cultivar ‘Forno’	B2g38480 and B2g38640	Singla et al., 2017
Sr56	5BL	A Swiss winter wheat Arina	sun209 and sun320	Bansal et al., 2014
Sr55/Lr67/Yr46	4DL	A PAK landrace PI 250413	TM 4 and TM10	Moore et al., 2015
Sr57/Lr34/Yr18	7DS	A wheat genotype Parula (PI58548)	csLV34	Lagudah et al., 2009
Sr58/Lr46/Yr29	1BL	A wheat cultivar Pavon76	SNP1_Lr46G22	Lagudah, personnel communication
Sr2/Lr27/Yr30	3BS	A wheat cultivar Hope	csSr2	Mago et al., 2011

Pleiotropic Adult Plant Resistances

Four APRs Lr27/Yr30/Sr2, Lr34/Yr18/Sr57, Lr46/Yr29/Sr58, and Lr67/Yr46/Sr55 harboring multiple pathogen resistance have been identified (McIntosh et al., 1995; Table 1B ). These genes resist devastation incurred by three rusts and powdery mildew caused by Blumeria graminis f. sp. tritici. The response of Lr34, Lr46, and Lr67 is well characterized by their partial resistance to each of these pathogens and a phenotypic marker, leaf tip necrosis (Ltn) (Lillemo et al., 2008; Herrera-Foessel et al., 2014b). The pleiotropic APR Lr34 has a comparatively higher level of rust resistance than other APRs (Lillemo et al., 2008; Yang et al., 2013). The APRs Lr34/Yr18/Sr57/Pm38/Ltn1 and Lr46/Yr29/Sr58/Pm39/Ltn2 have expressed additive gene interaction in combination and is frequently used in CIMMYT breeding programs (Lillemo et al., 2008; Lillemo et al., 2013). Unlike ASRs, the class of APR does not rely on a gene-for-gene hypothesis and partial resistance against a range of pathotypes (Dodds et al., 2010). Cloning of Lr34/Yr18/Sr57/Pm38/Ltn1 and Lr67/Yr46/Sr55/Pm46/Ltn3 (Krattinger et al., 2009; Krattinger et al., 2011; Moore et al., 2015) has demonstrated their role in conferring disease resistance in cereals like Sorghum (Schnippenkoetter et al., 2017), Durum wheat (Rinaldo et al., 2017), Rice (Krattinger et al., 2016) and Barley (Risk et al., 2013; Milne et al., 2019). It is also indicated that the resistance response of Lr34res and Lr67res occurs due to gain of resistance function through alteration in two amino acids in each of ATP-Binding Cassette (ABC) transporter and Hexose (sugar) Transporter Protein (STP), respectively (Krattinger et al., 2011; Moore et al., 2015). The molecular mechanism of few of these genes is not well understood as compared to NLR genes, and the mode of action of Lr34 transporter and molecules it transports is completely unknown (Rinaldo et al., 2017). There are three possible scenarios which may explain the mode of action of Lr67res; First, is altered carbon partitioning caused by Lr67res to limit nutrients to the pathogen, secondly, altered hexose/sucrose ratio triggering defense response and thirdly reduced pathogen growth due to unknown mechanism triggered by altered function of Lr67res (Milne et al., 2019).

Bi-Parental Mapping vs Genome-Wide Association Studies (GWAS)

Gene postulation is a classical approach to identify new genes phenotypically, using rust infection type response, multi pathotype test for gene postulation and differential sets. There are two alternative approaches to identify new rust resistance genes. The first approach includes the identification of donor parents expressing a high level of rust resistance and developing bi-parental mapping populations by crossing donor lines with susceptible lines. Genetic studies at F₂ and F₃ generations help us to estimate the number of genetic factors controlling the trait and their interactions in controlling resistance. Selective genotyping at the early segregating generations and/or advanced stages like F₅ and F₆ generations assists in identifying genomic region(s) underpinning rust resistance in the population (Bansal and Bariana, 2017). Bulk segregant analysis (BSA) is the most commonly used, efficient, and rapid method to identify markers closely linked to a specific gene or any genomic regions (Michelmore et al., 1991) by genotyping a pool of DNA from selected individuals with two extreme phenotypes and the contrasting parents. Then the allelic frequency differences between the bulks will be analyzed and compared with the phenotype to ascertain the possible linkage between the markers and trait of interest (Kthiri et al., 2018). Earlier, in the decade, SSR markers evenly spanning the wheat chromosomes were very frequently used to screen for polymorphic markers but, it was time-consuming and laborious as it was done manually. In recent times, the integration of next-generation sequencing-based 90K SNP arrays with BSA approach has quickened the identification of polymorphic markers and chromosomal location of major genes or the QTLs (Abe et al., 2012; Ramirez-Gonzalez et al., 2015; Qureshi et al., 2016; Kthiri et al., 2018; Wu et al., 2018; Pakeerathan et al., 2019). The BSA approach has also been combined with RNA-Seq to identify gene associated SNPs, NLR exome capture and sequencing (also called as MapRenSeq) to define NLRs associated with resistant genes (Narang et al., 2020) and NGS-based SLAF-seq to locate the gene of interest/QTL (Yin et al., 2018). However, the mapping of disease resistance genes and quantitative trait loci (QTL) through traditional bi-parental mapping approaches is restricted only to the limited diversity present in the parents and the less number of recombination events which results in low map resolution.

Alternatively, the second approach, Genome-wide association studies (GWAS) is primarily focused on quickly screening germplasm collection for genes of interest and to capture historical recombinations to achieve high-resolution mapping at a specific locus. As an advantage over mapping in bi-parental populations, this approach makes use of already existing natural populations, accounts for greater allelic diversity among the diverse individuals at a given locus and uses the recombination events that occur throughout the evolutionary process of germplasm (Yao et al., 2019). To excel in GWAS analysis, as a prerequisite, this approach would need a diversity panel constituting accessions and/or elite lines with similar photoperiod requirements and adaptation (Yu and Buckler, 2006). It has played a remarkable role in dissecting various complex traits in cereals (Yu et al., 2011; Juliana et al., 2015), and several of major and minor quantitative trait loci (QTL) relying on significant marker-trait associations (MTAs) for rust resistance have been identified through this approach (Rosewarne et al., 2012; Li et al., 2014; Zhang et al., 2014; Maccaferri et al., 2015; Juliana et al., 2018) but, none of the named rust genes has been identified through this approach until now. Moreover, several of these novel QTLs have not been validated and functionally characterized which have limited their use in breeding programs (Juliana et al., 2018). Though advances in statistical analysis and genotyping platforms have facilitated in ruling out the false-positive MTAs and rapid discovery of candidate genes using genome references (Bradbury et al., 2007; Zhang et al., 2010; Tang et al., 2016), population structure is still a critical concern and can be overcome by using mixed models to integrate genetic relatedness.

Application of Genomics Tools in Mining and Isolation of Rust Resistance Genes

DNA Marker Technology in Wheat

DNA markers reveal allelic variations among individual organisms or species. In wheat research, DNA marker technology has been used in tracing targeted genes. This technique is consistent, mostly neutral/free from environmental variations and does not affect the developmental stage of the plant as it is based on variations in DNA sequences (Collard et al., 2005). Some morphological and biochemical markers are also being used to identify lines carrying durable genes, for instance, pseudo-black chaff and high-temperature-induced seedling chlorosis linked with Sr2 (Slinkard and McIntosh, 1998), red-glume gene (Rg1) linked with Yr10 and leaf-tip necrosis (Ltn) linked with Lr34, Lr46, Lr67, and Lr68 (Lagudah et al., 2006; Herrera-Foessel et al., 2012; Herrera-Foessel et al., 2014b). The DNA marker technology offers a major advantage in plant breeding by exploring modern genomics tools. The Polymerase chain reaction (PCR)-based markers have been designed and applied to detect marker-trait association across seven groups of wheat chromosomes. The DNA markers namely Restriction fragment length polymorphism (RFLP), Sequence Characterized Amplified Region (SCAR), Sequence Tagged Site (STS), Simple Sequence Repeat (SSR), and Cleaved Amplified Polymorphic Sequences (CAPS) were widely explored as valuable tools for wheat improvement. Gene mapping in wheat was initiated using conventional genetic markers RFLPs (Anderson et al., 1992) and SSRs (Röder et al., 1998). Later, additional PCR-based markers like RGAP, CAPS, EST, STS, AFLP, and RAPD have been implemented in the mapping of disease resistance genes (Bariana et al., 2001). The consensus map of 2004 SSRs has been used to integrate conventional markers to some extent for drafting an integrated map of the identified gene/loci using cMap (Somers et al., 2004; Rosewarne et al., 2013). Sanger sequencing accelerated the identification of the single base pair variations (Wang et al., 1998) and helped to replace gel-based marker systems with robust and bi-allelic single nucleotide polymorphisms (SNPs). This SNP-based DNA marker has enhanced its implementation to detect allelic variation in targeted DNA fragments. The advantage of MAS (Marker Assisted Selection) is not only rapid detection assays for the presence of the gene but also enables gene pyramiding strategies wherein it is difficult to detect gene combinations due to epistatic/masking effect if race-specific genes.

Next-Generation Sequencing (NGS) Platforms

The NGS platforms have revolutionized genetics and genomics studies in crop plants. Several NGS-based approaches played a significant role in allele discovery and genotype-by-sequencing (GBS) through thousands of markers simultaneously to precisely cover the entire crop genome (Davey et al., 2011; Liu H. et al., 2014). In wheat, NGS technology enables detecting the abundance of DNA markers within a short period. The size of the wheat genome is huge (~16GB—around five times bigger than the human genome), more complex, and contains a high percentage of repetitive elements. Several marker development methods have been reported, including RRLs (reduced-representation libraries), CRoPS (complexity reduction of polymorphic sequences), RAD-seq (restriction-site associated DNA sequencing), SBP (sequence-based polymorphic marker technology), MSG (low coverage multiplexed shotgun genotyping), and GBS (genotyping-by-sequencing) to reduce the complexity (Yang et al., 2012). Additionally, GBS pipeline was also developed to reduce the complexity of the wheat genome. The RAD-seq and GBS were explored to target cutting sites after restriction digestion to avoid further redundancy in the genome (Poland et al., 2012). High throughput genotyping platforms including DArT-Seq, SNPs, GBS markers, and population-specific tGBS (targeted genotyping-by-sequencing) have expedited the precise mapping of genomic regions underpinning rust resistance (Qureshi et al., 2018b; Nsabiyera et al., 2020).

Discovery of SNPs Using a Genotyping Array

Discovery of single nucleotide polymorphisms (SNPs) in wheat has been expedited through the advances in genomics resources. Cavanagh et al. (2013) developed a high-throughput SNP genotyping array called i-select 9K SNP Infinium Bead chip Assay using 2,994 worldwide wheat accessions. Later, a high-density SNP map was developed after incorporating the first map and customized as i-select 90K SNP Infinium array covering seven groups of wheat chromosomes (Wang et al., 2014). These linked SNPs would facilitate to overcome, a major bottleneck associated with the application of DNA markers to characterize a wide range of germplasm (Yang et al., 2012). This high-density SNP map comprises nearly 90,000 gene-linked SNPs to characterize genetic variability in tetraploid and hexaploid wheat (Wang et al., 2014). Similarly, the GBS approach was utilized to characterize over 40,000 CIMMYT wheat germplasms as part of Seeds of Discovery (SeeD) initiative (Li et al., 2015). A consensus map of GBS markers was constructed to detect potentially new APR linked with stripe rust, leaf rust, and stem rust resistance and validate known genes/QTL through aligning them on wheat genome reference (IWGSC Refseq v 1.0; Appels et al., 2018). Identified trait-linked SNPs can be converted into Kompetitive allele-specific PCR assay (KASP) using PolyMarker, an automated bioinformatics pipeline (http://www.polymarker.info). The same SNPs are available in different wheat chromosomes at various positions due to its evolutionary nature and diploidization phenomenon. PolyMarker can design SNP-based KASP of the known genome-specificity using IWGSC RefSeq v1.0 (Ramirez-Gonzalez et al., 2015). Developing KASP markers cannot be guaranteed their utility in marker-assisted selection. Mendelian inheritance of individual KASP in a bi-parental mapping population and linkage with phenotypic data can ensure their application in the wheat breeding program.

Challenges to Draft Consensus Maps of Rust Resistance Genes

Several new resistance loci and robust markers of reported genes were developed and explored in gene pyramiding (Bariana et al., 2007; Wessels and Botes, 2014; Randhawa et al., 2019). High level of redundancy in these identified loci and common MTAs was evident among different mapping studies. A consensus map using public database was drafted for stripe rust resistance on wheat chromosomes with the help of flanking markers to highlight around 140 stripe rust resistance loci and project redundant loci identified in over 30 bi-parental studies (Rosewarne et al., 2013). This study reduced the number of genomic regions conferring stripe rust resistance and laid a platform for further works. Later, Maccaferri et al. (2015) developed an integrated map of stripe rust resistance regions after incorporating significant QTLs identified in ten experiment-wise GWAS studies, 56 YR genes and 169 mapped QTLs. An “iterative maps compilation” tool in Biomercator v4.2 was explored to integrate the consensus map of SSRs (Somers et al., 2004), Synthetic × Opata ITMI BARC SSR map (Song et al., 2005), i-select 90K SNP Infinium array, Synthetic × Opata DH GBS map (Saintenac et al., 2013a) and Diversity Array Technology (DArT) map (http://www.diversityarrays.com) to create an integrated map (Maccaferri et al., 2015). This consensus map facilitated in the integration of different marker systems like RFLP, DArT, SSR along with flanking robust SNPs and align them together using published genetic map. This information is available on the GrainGenes webpage (https://wheat.pw.usda.gov/GG3/). Similar approaches have been adapted to develop stem rust resistance loci consensus map of detected in 24 bi-parental mapping populations, three association panels, two backcross populations (Yu et al., 2014), and QTL conferring leaf rust and powdery mildew resistance in wheat (Li et al., 2014).

Several genomic loci are well known to harbor more than one gene, but these integrated maps did not allow further delineation of these regions. For instance, wheat chromosome 5A carries two genes Yr34 and Yr48 which were later confirmed as allelic or the same (Qureshi et al., 2018a). Gao et al. (2015) attempted to enrich region of SrCad (source- Canadian wheat cultivar AC Cadillac) and Sr42 (Norin 40) using GBS and 90K SNP array approach which placed both these genes at the same locus on chromosome 6DS. Moreover, both genes shared a common donor, BW553 and resistance to TTKSK isolate which also suggested that both genes are the same or allelic. Differential expression of AC Cadillac to non-TTKSK Pgt races compared to Norin 40 is due to the existence of background genes like Lr34. Cloning of Rph15 in barley confirmed that Rph15 and Rph16 are allelic or the same (Chen C. et al., 2020).

Integrating wheat genome reference (IWGSC RefSeq v 1.0) with the previously published genetic map of rust resistance loci was a major challenge; however, the new genetic map viewing software, Pretzel, was designed to fix the issues. This software is currently being explored to align trait-linked SNPs, DArTseq, and GBS markers with multiple SNP platforms of available wheat and related genomic resources like Triticum dicoccoides, T. turgidum, Aegilops tauschii, T. urartu, Brachypodium_distachyon, Hordeum vulgare, and Oryza sativa (Keeble-Gagnere et al., 2019; http://plantinformatics.io/mapview). It assists in the positioning of various genomic regions of the different populations using IWGSC RefSeq v1.0. It would be a milestone achievement in differentiating rust resistance genes/MTAs underlying in similar genomic regions.

Rapid Rust Resistance Gene Cloning Methods

In plants, most of the disease resistance genes are race-specific and contain the Nucleotide-binding site (NBS) and LRR (Leucine-rich repeat) domains. These genes are believed to be regulated by NBS domains through signal transduction, and specific sites of corresponding pathogen virulence genes are recognized by LRR domains (Gill et al., 2015). Six stripe rust resistance genes, Yr5/YrSP, Yr7, Yr10, Yr15, Yr36, and YrAS2388R (Fu et al., 2009; Liu H. et al., 2014; Klymiuk et al., 2018; Marchal et al., 2018; Zhang et al., 2019), four-leaf rust resistance genes Lr1, Lr10, Lr21, and Lr22a (Feuillet et al., 2003; Huang et al., 2003; Cloutier et al., 2007; Thind et al., 2017), and nine major stem rust resistance genes, Sr33, Sr35, Sr50, Sr22, Sr45, Sr13, Sr46, Sr21, and Sr60 (Periyannan et al., 2013; Saintenac et al., 2013b; Mago et al., 2015; Steuernagel et al., 2016; Zhang et al., 2017; Arora et al., 2018; Chen et al., 2018; Chen S. et al., 2020), and two pleiotropic APR (Lr34/Yr18/Sr57 and Lr67/Yr46/Sr55) (Krattinger et al., 2009; Moore et al., 2015) have been cloned ( Table 2 ).

Table 2.

List of cloned rust resistance genes in wheat.

Gene/Locus	Chromosome	Approach	Major/Minor gene	Protein encoded	References
Yr5/YrSP	2B	MutRenSeq	Major	NBS-LRR	Marchal et al., 2018
Yr7	2B	MutRenSeq	Major	NBS-LRR	Marchal et al., 2018
Yr10	1BS	Map-based cloning	Major	NBS-LRR	Liu W. et al., 2014
Yr15	1BS	Map-based cloning	Major	A wheat tandem kinase 1	Klymiuk et al., 2018
YrAS2388R	4DS	Map-based cloning	Major	CC-NBS-LRR	Zhang et al., 2019
Lr1	5DL	Map-based cloning	Major	CC-NBS-LRR	Cloutier et al., 2007
Lr10	1AS	Map-based cloning	Major	CC-NBS-LRR	Feuillet et al., 2003
Lr21	1DS	Map-based cloning	Major	CC-NBS-LRR	Huang et al., 2003
Sr13	6AL	map-based cloning	Major	CC-NBS-LRR	Zhang et al. (2017)
Sr21	2AmL	map-based cloning	Major	CC-NBS-LRR	Chen et al. (2018)
Sr22	7A	MutRenSeq	Major	CC-NBS-LRR	Steuernagel et al., 2016
Sr33	1D	Map-based cloning	Major	CC-NBS-LRR	Periyannan et al., 2013
Sr35	3AL	Map-based cloning	Major	CC-NBS-LRR	Saintenac et al. (2013b)
Sr45	1D	MutRenSeq	Major	CC-NBS-LRR	Steuernagel et al., 2016
Sr 46	2DS	AgRenSeq	Major	coiled-coil NLR protein	Arora et al., 2018
Sr50	1RS	physical mapping, mutation and complementation	Major	CC-NBS-LRR	Mago et al. (2015)
Sr60	5AmS of T. monococcum	Map-based cloning	Major	A wheat tandem kinase 2	Chen S. et al., 2020
Yr36	6BS	Map-based cloning	Minor	Kinase-START gene	Fu et al., 2009
Lr22a	2DS	TACCA	Minor	CC-NBS-LRR	Thind et al., 2017
Lr34/Yr18/Sr57	7DS	Map-based cloning	Minor	ABC transporter	Krattinger et al., 2009
Lr67/Yr46/Sr55	4DL	Map-based cloning	Minor	Hexose transporter	Moore et al., 2015

Map-based cloning includes fine mapping of an individual gene, discovery of candidate genes followed by validation of these genes using mutants and available genomics resources (Jander et al., 2002; Thind et al., 2017; Zhang et al., 2019). The first study towards the map-based cloning of Rpg1 and Rpg4 using rice as an inter-genomic mapping vehicle was reported in Barley (Kilian et al., 1997). Many genes of economic importance in cereal crops, most notably, chloronerva, a gene involved in the uptake of iron in higher plants encoding Nicotianamine synthase (Ling et al., 1999), beta and old-gold (og) color mutations in tomato (Ronen et al., 2000), Rf-1 in rice (Komori et al., 2004), Soybean Maturity Locus E3 (Watanabe et al., 2009), TTG1 homolog in Brassica (Zhang et al., 2009), and AvrLm6 avirulence gene in Leptosphaeria maculans (Fudal et al., 2007), Lr10 (Feuillet et al., 2003), Lr21 (Huang et al., 2003), and Lr1 (Ling et al., 2003) were isolated using this strategy. The main problem associated with map-based cloning is lack of sufficient DNA markers to create a fine map in different crops (Drenkard et al., 2002). Besides, it was essential to construct a physical map, identify markers, and reach close to the gene through chromosome walking. It was then followed by isolation, transformation, complementation, and determination of the sequence of the entire region of interest in the absence of any reference sequence of the wild-type DNA (Jander et al., 2002).

In crops with large genomes like wheat, complexity reduction is very important to clone target genes more quickly and efficiently. A recently developed molecular tool, MutChromSeq (Mutgenesis Chromosome flow sorting and short-read Sequencing), works based on mutagenesis followed by flow sorting of chromosomes and their sequencing to identify the possible induced mutations. This approach does not require a similar gene in bait library and hence enables reference-free forward genetics to open up the pan-genome to functional genomics (Sanchez-Martin et al., 2016; Zhang et al., 2020). This rapid cloning approach which would need 18–24 months for mutagenesis and screening, approximately one month for chromosome flow sorting, sequencing and bioinformatics were successfully described for cloning of barley Eceriferum-q gene and wheat Pm2 gene (Sanchez-Martin et al., 2016). The first leaf rust resistance gene Rph1 (Rph1.a) on chromosome 2H from cultivated barley (Hordeum vulgare) cv Sudan was also rapidly cloned by combining this approach with genetic mapping. The successful application of this approach would need prior information on, chromosome location of the target gene (Dracatos et al., 2019; Zhang et al., 2020), large mutant populations to identify the target genes among five to six mutants (Mago et al., 2017; Dinh et al., 2020) and isolation of individual chromosomes (Periyannan, 2018).

As an alternate approach to MutChromSeq, TACCA makes use of flow sorting of chromosomes, next-generation sequencing, and cultivar specific de-novo chromosome assembly. Using this approach a broad-spectrum leaf-rust resistance locus Lr22a (encodes an intracellular immune receptor homologous to the Arabidopsis thaliana RPM1 protein) was cloned with molecular marker information and ethyl methanesulfonate (EMS) mutants (Thind et al., 2017).

There are other cloning strategies like MutMap (mutational mapping) which involve mutagenesis, sequencing, and mapping to identify SNPs between wild-type and homozygous mutants and then to narrow down to the region having gene of interest (Li et al., 2020b). Until recently, this approach was considered to be applicable only in crops with small genomes like rice (Abe et al., 2012), but it was demonstrated to successfully map and clone Ms1 from allohexaploid bread wheat using F₂ plants derived from the progeny of heterozygous ms1e mutants (Wang et al., 2017). Still, this is a better choice when the genome is small, and high-quality reference genome is available.

A three-step pipeline including chemical mutagenesis (Mut), exome capture and sequencing (RenSeq) techniques was developed for the rapid isolation of genes (Steuernagel et al., 2016). Here, ‘Mut’ confers mutagenesis that is a technique to produce random mutations or nucleotide substitutions in the DNA sequence by treating the parent material with mutagens like EMS (Ethyl Methane Sulfonate CH₃SO₃C₂H₅) or Sodium Azide (NaN₃). This substitution results in GC to AT transition and can cause loss of function of the resistance gene. An initial kill-curve analysis is used to decide the optimum dose of mutagen to render 50% mortality and reduced growth (Periyannan et al., 2013). The ‘RenSeq’ technique relies on the sequence information of known resistance genes and uses as bait to capture similar gene fragments in closely related crops and species (Jupe et al., 2013; Steuernagel et al., 2016). In plants, most of the known disease resistance genes belong to NLRs, kinases (receptor-like kinase, receptor-like proteins, wall-associated kinase etc) and/or transporters families (Klymiuk et al., 2018). Two stem rust resistance genes (Sr22 and Sr45) were cloned using this technique. In MutRenSeq, NLR gene families are targeted, as it is the major class of R gene family conferring disease resistance in plants. The procedure includes: (1) Bait library construction containing RNA probes with more sequence identity with predicted NLR genes, (2) Preparation of template DNA from the accessions or germplasm conferring disease resistance, (3) Gene capture/hybridization using bait library in PCR based hybridization reaction, (4) Next-generation sequencing of captured and enriched fragments to predict NLR gene structures, (5) Assembly of NBS-LRR reads and their comparison and (6) Validation of candidate genes. MutRenSeq is a rapid and accessible to several organisms including wheat (Steuernagel et al., 2016). The major limitations of this technique are its restriction to discover only the NLR gene, a reference pan-genome and vital changes due to mutagenesis (Dinh et al., 2020).

AgRenSeq (Association genetics with R-gene enrichment sequencing) or speed cloning has been developed to align GWAS platform (to utilize diverse genome-wide natural variation) and the RenSeq. Unlike map-based cloning and MutRenSeq, this approach does not rely on bi-parental mapping populations or mutagenesis. For instance, AgRenSeq approach was successfully demonstrated to clone R gene, Sr46 and to identify the candidate gene sequence for SrTA1662 using a diverse panel of Ae. tauschii ssp.strangulata (Arora et al., 2018). It explores a pool of diverse wild relatives carrying many resistance genes as a result enabling the cloning of multiple genes at the same time (Dinh et al., 2020).

Breeding Strategies to Develop Rust-Resistant Wheat Varieties

Relentless efforts to enhance modern wheat cultivars for rust resistance is of primary importance to many breeding programs which come with a challenge of, not to compromise with yield and quality traits. A rust-resistant hard wheat variety, “Marquillo” developed in 1918 (Hayes et al., 1920) had considerable resistance to rust, but it lacked the flour quality, so it could not become popular in North Dakota (Stoa, 1945). Similarly, a highly rust-resistant line named, “Hope” (McFadden, 1930) which had poor yield potential under drought and heat stress conditions. However, the popular rust-resistant wheat cultivar, “Thatcher” released in 1934 (Hayes et al., 1936) carried multiple stem rust resistance genes and provided useful resistance against stem rust in North America (Singh et al., 2004b) occupying 60% of the area during that period under hard red spring wheat in North Dakota. Being a most preferred strategy of rust control, systematic breeding for genetic resistance in wheat started when the inheritance of yellow rust resistance was demonstrated following Mendelian principles of heredity (Biffen, 1905). The resistance controlled by a single locus was considered to be sufficient for complete control of the rust pathogens, until the discoveries of Flor (1956), when the incompatibility of host and pathogens was entirely said to be based on corresponding gene present in the organism and the resistance governed by a single gene need not be always durable. In the 20^th century, though, breeders were aware of possible single mutations in pathogens’ Avr genes, could result in new virulent pathotypes, still most of the breeding programs adopted single gene pyramiding into cultivars as major “breeding strategy”. This could be due to ease in the screening of breeding materials at the seedling stage and high heritability and clear phenotype of the resistance expressed by major genes (Pink and Hand, 2002).

There are multiple approaches, conventional and/or molecular, singly or in combination, being considered to develop wheat varieties with durable rust resistance around the globe. Traditionally, using conventional approaches, the available germplasm and breeding materials are screened at seedling and adult plant stage, to phenotypically identify resistant wheat lines and then to involve them in hybridization programs as one of the parents. Till today, there are several breeding teams with no access to advanced MAS tools have successfully adopted this approach to breeding future rust-resistant wheat varieties. This makes the best use of shuttle breeding, wherein, the breeding lines are grown in one or more off-season locations with congenial environments for all the three individual rust diseases. Due to the cultivation of host crops throughout the year, these offseason environments become the evolution hubs for the new races, which helps them to prolifer and multiply continuously. Besides, these environments will also assist breeder to carefully select wheat lines which show consistent resistance across the years and locations with a large number of naturally occurring pathotype mixtures in the environment.

Breeding strategies to improve wheat for seedling and adult plant resistance may vary usually due to their mode of inheritance which is either qualitative and/or quantitative. Among the many breeding methodologies available to improve a particular trait in crop plants, pureline selection, modified pedigree method, bulk breeding, recurrent selection, backcross breeding, mutation and single seed descent methods of breeding are applied to breed wheat for rusts resistance. Pureline selection improved the agronomical feature of several landraces resulting into popular wheat varieties namely Sharbati, White Pissi, Chandausi and Lal Kanak in India. The wheat variety Marquis was developed as a selection from hybridization between an early ripening Indian wheat “Hard Red Calcutta” and “Common Fife” in 1909. Due to early ripening, it escaped the injury caused by heat and rust to become a prominent variety for several years in Western North Dakota (Stoa, 1945). Breeding through a modified pedigree method would be more successful when resistance is governed by a major gene, while it does not give much improvements in combining minor genes in the population. While, single back cross-selected bulk breeding approach (Singh and Trethowan, 2007) can be used to transfer multiple minor resistance genes into a well-adapted and higher-yielding backgrounds (Singh et al., 2014). We have enlisted a few of popular Indian wheat varieties developed through conventional introgression of resistance genes for all three kinds of rusts ( Table 3 ).

Table 3.

List of popular Indian wheat varieties and their known rust resistant source.

Variety	Pedigree	Gene combination (s)
		Leaf rust (Lr)	Stripe rust (Yr)	Stem rust (Sr)
LOK 1	S308/S331	Lr13	Yr2ks	Sr2+Sr9b+Sr11
PBW502	WH485/PBW343//RAJ1482	Lr26+	Yr9+	Sr2+ Sr5+Sr31
PBW343	ND/VG9144//KAL/BB/3/YCO”S”/4/VEE#S “S”	Lr26+	Yr9+Yr27	Sr2+ Sr5+Sr31
GW273	CPAN2084/VW205	Lr13+Lr10	-	Sr8b+Sr11
GW322	PBW173/GW196	Lr13	-	Sr2+Sr11
RAJ3765	HD2402/VL639	Lr13+Lr10	Yr2ks+	Sr2+
WH711	ALD’S’HUAC//HD2285/3/HFW-17	Lr13	Yr2ks	Sr2+Sr7b
DBW17	CMH79A.95/3*CNO79//RAJ3777	Lr26+Lr23	Yr9+	Sr31
PBW550	WH 594/RAJ 3856//W 485	Lr26+R	Yr9+Yr18	Sr31
PBW373	ND/VG9144//KAL/BB/3/YCO”S”/4/VEE#S “S”	Lr26+	Yr9+Yr27	Sr2+ Sr5+Sr31
RAJ4037	DL788-2/RAJ3717	Lr24	-	Sr2+Sr24
HD2189	HD2963/HD1931	Lr13+Lr34	Sr2+Sr11	Yr2+Yr18
WH147	E4870/C303//S339/PV18	Lr13+Lr34	Yr2ks+Yr18	Sr7a+Sr11
UP2338	UP368/VL421//UP262	Lr26+Lr34	Yr9+Yr18	Sr2+Sr31
WH542	JUPATECO/BLUE JAY//URES	Lr26+Lr23	Yr9+Yr18	Sr31
HUW234	HUW12*2/CPAN1666//HUW12	Lr14a	Yr2ks+	Sr9b+Sr11
HD2329	HD1962/E4870/3/K65/3/HD1553/UP262	Lr13+Lr10+Lr34	Yr2+Yr18	Sr8b+Sr9b+Sr11
C306	REGENT1974/3*CHZ//*2C591/3/C217/N14/C281	Lr34	Yr2ks+Yr18	-
HD2285	HD1912/HD1592//HD1962/E4870/3/K65/4/HD2160/HD2186	Lr23	Yr2+	Sr9b+Sr11
SONALIKA	II53.388/AN//YT54/N10B/3/LR/4/B4946.A.4.18.2IY/Y53//3*Y50	Lr13	Yr2+	Sr2+Sr11
UP262	S308/BJ66	Lr23+Lr34	Yr2+Yr18	Sr2+Sr11
PBW226	WG138/JUSTIN//CHIRS/HD1941	Lr23	Yr2+	Sr2+
MACS2496	A selection from SERI ‘S’	Lr26+Lr23+Lr1+	Yr9+	Sr2+Sr31
WH283	HD1981/RAJ821	Lr10+	Yr2+	Sr8b+
GW190	VEE/3/BB ‘S’/SKA//ARJUN	Lr26+Lr23+Lr1+	Yr9+	Sr2+Sr31
HS 295	CQT/AZ//IAS55/ALD/3/ALD/NAFN/4/PJN/PEL SL 1276.69	Lr23+Lr34	Yr2ks+Yr18	Sr2+Sr8b+Sr11
HD2967	ALD/CUC//URES/HD2160M/HD2278	Lr23+	Yr2+	Sr8a+Sr11+Sr2
HD3086	DBW14/HD2733//HUW468	Lr13+Lr10+	Yr2+	Sr7b+Sr2
WH1105	MILAN/S87230//BABAX	Lr13+	Yr2+	R+

Nayar et al. (2001); Bhardwaj (2011); Singh et al. (2011).

Genomics Assisted Improvement of Rust Resistance

Most of the breeding programs with rust resistance as a major focus, work on identification, mapping, isolation, and introgression of diverse resistant genes into agronomically potential but, rust susceptible wheat lines. Though conventional breeding methods are effective, they are very time consuming, need laborious efforts to develop improved versions of crop plants for any target trait/s. In the last two decades, several genomic interventions have become possible which provide tremendous capabilities to characterize and manipulate wheat genome and to accelerate breeding. Though rust resistance has clear-cut phenotype to select for, the advent of molecular marker systems like SNPs and high-throughput genotyping platforms like SNP arrays have enabled precise estimation of marker-trait associations and hence, to select for the trait of interest without the greater involvement of environmental effects and to combine multiple genes at the same time.

To identify and map novel rust resistance genes, the biparental and genome-wide association based mapping are most widely followed, with the latter having the ability to localize multiple genes and/or QTLs. Rust resistance based on a single dominant gene is generally considered to be leading to genetic changes in pathogen virulence (Wilcoxon, 1981) and when incorporated singly into cultivars, it may become ineffective within a few years of cultivation. Hence, it is always recommended to use combinations of major and minor genes (Roelfs et al., 1992) which could confer broad-spectrum resistance. Here, MAS may be incorporated into backcross breeding to simultaneously introgress several rust resistant genes against a wider range of pathogens. Rapid generation advance protocols like doubled haploids (DH) and Speed breeding combined with marker-assisted breeding, would also assist in developing resistant cultivars in a very short period. Although MAS is a very efficient breeding strategy, its implementation is limited due to unavailability of tightly liked markers and can only identify few large effect QTLs.

In recent years, GS, an advanced version of MAS, is being largely used to increase the selection efficiency and to dissect complex polygenic traits with lower heritability. GS has the power to capture small individual effects of thousands of genes through whole genome prediction models, to estimate breeding values for the complex traits and hence overcome the problems associated with the MAS. Breeding for race nonspecific minor gene resistance is one way to minimize genotype and environment interaction of resistance, which in turn leads to stability in resistance and yield (Bekele et al., 2019). Breeders have already preferred to go for many minor gene based resistance instead of major genes, where GS is considered to be the best molecular breeding approach in place of MAS. The integration of GS with selected bulk, recurrent selection, and back cross breeding is being explored to maximize the genetic gain for quantitative traits like APR through selection. Transgene-free genetic engineering through genome editing with sequence specific nucleases has paved a way for manipulating any specific genomic sequence superseding random mutagenesis (Wang et al., 2018). Among the four types of nucleases used in genome editing, CRISPR/Cas9 has a higher success rate in gene modification and requires less know-how. The application of genome editing for disease resistance would be much feasible; due to the better understanding of the molecular mechanisms underlying resistance, modification of a single gene may lead to resistance phenotype, and targeted mutagenesis is well applicable to knockout susceptible genes (Borrelli et al., 2018).

Deployment of Adult Plant Genes Delaying the Evolution of Pathotypes

Race-specific APRs which can be screened at the greenhouse level at two or third-leaf stage like Lr13 and Yr49 (McIntosh et al., 1995) are more vulnerable to succumb matching virulence in rust pathogens. Race-specific APR Lr13 expresses at the seedling stage (two-leaf stage) in a greenhouse as well as at field conditions against avirulent pathotypes. Due to their expression at an early stage, some researchers have cited these genes as seedling resistance genes (Chen and Kang, 2017). APR Lr22a confers leaf rust resistance under field conditions and rendered resistance against 22 leaf rust pathotypes at flag leaf stage (Sawhney et al., 1982). Lr34 was reported to be partially effective against leaf rust pathotypes in India (Sawhney and Sharma, 1990) and it expresses enhanced resistance along with other resistance genes (German and Kolmer, 1992). APR Lr34 carrying genotype reduced the yield losses up to 18% and genotype lacking Lr34 revealed 60 to 84% of losses (Singh et al., 1994). Slow rusting characters of Lr34 and its interaction with other genes have made it useful for rust resistance (Nayar et al., 1999). Similarly, Lr46/Yr29/Pm39 gene complex was identified in “Pavon76” and conferred a slow rusting to leaf rust, stripe rusts, and powdery mildew in wheat (William et al., 2003; Lillemo et al., 2008; Singh et al., 2013). The leaf rust resistance genes Lr48 and Lr49 were reported in wheat cultivars CSP44 and VL404, respectively as hypersensitive adult plant resistance genes (Saini et al., 2002). Another gene Lr67/Yr46/Sr55 also confers slow rusting to leaf and stripe rusts (Hiebert et al., 2010; Herrera-Foessel et al., 2011). An Australian Pst pathotype 239 E237 A- 17 + 33+ has changed the rust rating of Australian cultivar “Axe” carrying Yr75 (Cuddy and Hollway, 2018). Previous studies showed the susceptible nature of Yr49 against Chinese Pst races (Ellis et al., 2014). Several stripe rust APRs have been overcome by newly detected Pst races in Europe and North America, so those APRs were tentatively classified as race-specific APRs (Hao et al., 2011; Sthapit et al., 2012; Sørensen et al., 2014). An accession HS628 expressed broad-spectrum resistance to Pt, Pst, and Pgt pathotypes in India (Pal et al., 2018). Developing wheat varieties possessing these diverse gene combinations could be a feasible approach to delay the evolution of pathotypes with matching virulence genes.

Concluding Remarks

Breeding for rust resistance is an integral part of wheat improvement. APRs Sr2, Lr34, Lr46, Lr67, and Lr68 have been found durable and non-race specific. These APRs can offer only partial resistance which is insufficient to address food insecurity threat. Deployment of both durable rust resistance genes along with major R genes has been reported as a sound breeding strategy to avoid rust epidemics worldwide. Development of gene-specific DNA markers is essential to introgress the rust resistance gene in the desired wheat background and avoid linkage drag. Positional-cloning and MutRenSeq approach along with available genomics resources were explored to develop gene-specific marker of around twenty rust resistance genes including four APRs (Lr34, Lr67, Yr36 and Lr22a). Stacking of five cloned resistance genes in modern wheat variety is underway via cis-genic approach. In future, this approach will be fruitful to develop resistant wheat against three rusts. Rust pathogen has been continuously evolving to shrink the number of effective rust resistance genes. The high-quality reference genome of over ten wheat varieties (10+ genome project) will assist in the cloning of additional major genes and APR. The major future challenge would be to identify effective resistance genes and their deployment to attain durable rust resistance.

Author Contributions

PB, DB, DW, HB, and AS wrote the manuscript. All authors contributed to the article edits and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.