Sher MUHAMMAD,Muhammad SAJJAD,Sultan Habibullah KHAN,Muhammad SHAHlD,Muhammad ZUBAlR,Faisal Saeed AWAN,Azeem lqbal KHAN,Muhammad Salman MUBARAK,Ayesha TAHlR,Muhammad UMER,Rumana KEYANl,Muhammad lnam AFZAL,lrfan MANZOOR,Javed lqbal WATTOO,Aziz-ur REHMAN

1 Department of Bioinformatics and Biotechnology,GC University,Faisalabad 38090,Pakistan

2 Department of Biosciences,COMSATS University Islamabad (CUI),Islamabad 45550,Pakistan

3 Centre of Agricultural Biochemistry and Biotechnology (CABB),University of Agriculture,Faisalabad 38000,Pakistan

4 Department of Plant Breeding and Genetics,University of Agriculture,Faisalabad 38000,Pakistan

5 Department of Biotechnology,Faculty of Life Sciences,University of Central Punjab,Lahore 54000,Pakistan

6 Wheat Research Institute,Ayub Agricultural Research Institute (AARI),Faisalabad 38950,Pakistan

Abstract Stripe rust is a continuous threat to wheat crop all over the world.It causes considerable yield losses in wheat crop every year.Continuous deployment of adult plant resistance (APR) genes in newly developing wheat cultivars is the most judicious strategy to combat this disease.Herein,we dissected the genetics underpinning stripe rust resistance in Pakistani wheat germplasm.An association panel of 94 spring wheat genotypes was phenotyped for two years to score the infestation of stripe rust on each accession and was scanned with 203 polymorphic SSRs.Based on D´ measure,linkage disequilibrium(LD) exhibited between loci distant up to 45 cM.Marker-trait associations (MTAs) were determined using mixed linear model (MLM).Total 31 quantitative trait loci (QTLs) were observed on all 21 wheat chromosomes.Twelve QTLs were newly discovered as well as 19 QTLs and 35 previously reported Yr genes were validated in Pakistani wheat germplasm.The major QTLs were QYr.uaf.2AL and QYr.uaf.3BS (PVE,11.9%).Dissection of genes from the newly observed QTLs can provide new APR genes to improve genetic resources for APR resistance in wheat crop.

Keywords:wheat,Puccinia striiformis,LD,GWAS,MTA,PCoA

1.lntroduction

Stripe rust (or yellow rust) caused byPuccinia striiformisf.sp.tritici(Pst),is a worldwide foliar disease of wheat crop(Rehmanet al.2013).In Pakistan,13 stripe rust epidemics in different time periods have incurred huge financial losses.In 2005,a new race of yellow rust (virulent toYr27gene incv.Inqlab-91) hit the northern area of Pakistan and devastated widely grown Inqlab-91 wheat cultivar (Rehmanet al.2013).The chemical control of rusts leads to environmental pollution,revival of the disease under conducive conditions and accumulation of chemical residues in grains and wheat straw (Luo 2009).However,genetic control of rusts is durable,safe for environment,human and animals health.Therefore,breeding wheat varieties with resistance to stripe rust (YR)races is the most economical,effective and clean approach.

Genetic control of rusts can be of vertical (also called race specific) or horizontal (also called race non-specific type,slow rusting or adult plant resistance).Race-specific resistance is underpinned by one or few major genes.The field life of varieties with this type of resistance is shorter due to emergence of new pathotypes which overcomes the resistance conferred by one or few major genes(Muhammadet al.2018).On the other hand,adult plant resistance (APR) is underpinned by several minor genes plus one or few major genes which usually hinders the progress of pathogens infestation on adult plants.Wheat varieties with APR resistance possess durable resistance for longer time period with almost same level of reaction against new pathotypes in changing climatic conditions.For example,Lyalpur-73 (a Pakistani wheat variety released in 1970’s) still exhibits high level of resistance to currentPstpathotypes due to the combination of minor APR genes and a major geneYr18(Rehmanet al.2013).

The ability ofPstpathotypes to evolve rapidly with increased virulence and epidemicity has overcome major resistance genes and even APR genes present in wheat varieties (Nsabiyeraet al.2018;Sajjadet al.2018).Up till now,more than 70Yrgenes have been formally registered and about 100 provisionally identified asYrgenes or quantitative trait loci (QTL) in cultivated and wild wheat species(Rosewarneet al.2013).Susceptibility of wheat varieties to new pathotypes is a major concern worldwide as about US$ 1 billion are lost annually due to the damage caused byPst(Beddowet al.2015).Pyramiding of effective APR genes in wheat varieties is essential to confer resistance against emergingPstpathotypes to deal with growing threat of YR and minimize financial losses.Identification of effective APR genes in local wheat germplasm is prerequisite for initiating an APR gene pyramiding program.

Genome-wide association analysis (GWAS) has been widely used in various wheat populations genotyped with SSRs to validate and identify novel QTLs/genes.Wheat populations of un-related individuals and SSR markers have been widely used to map QTLs for various traits related to yield (Yuet al.2011;Sajjadet al.2014),quality (Reifet al.2011;Shahzadet al.2016),biotic (Zhaoet al.2014;Liuet al.2017;Muhammadet al.2018) and abiotic stress tolerance(Ainet al.2015).Herein,we validated previously reported YR-related QTLs and identified novel APR QTLs for YR resistance in a panel of 94 diverse Pakistani wheat accessions.The panel was phenotyped for YR disease intensity during two crop years and genotyped with 203 genome wide SSR markers.

2.Materials and methods

2.1.Plant materials and phenotypic data recording

The association panel of 94 accessions (Appendix A) was sown under recommended farm practices at University of Agriculture,Faisalabad (coordinates:31°15´0´´N,73°03´0´´E;soil:silt loam;climate:semi-arid) during cropping seasons,2014–2016.Each entry was planted on an area of 1.5 m with row to row distance of 30 cm and plant to plant distance of 8 cm following augmented block design.All the recommended agronomic practices were applied to raise the crop.Prevalence and severity of stripe rust abrasions and varietal response was recorded following modified Cobb’s scale (Petersonet al.1948).The mixture of pathotypes was collected from stripe rust prone areas(Peshawar,Noshehra,Islamabad and Murree) of Pakistan.Plants were inoculated at alternate days from 23rd January to 7th February with urediniospores of pathotypes after mixing with talcum at a ratio of 1:24.Morocco was used as border,Punjab 76 (80S) was used as susceptible and Lyp 73 (TR) as resistant check.Rust data were recorded after 7 days interval from mid-February to the end of March.The data were pooled over the years for analysis.The stripe rust infestation was scored asabsence of abrasion(0),resistant (R);moderately resistant (MR);moderately susceptible (MS);moderately resistant to moderately susceptible (MRMS) and susceptible (S) under stripe rust stress (Gulnazet al.2019).

2.2.Genotyping

The DNA from each accession was extracted from fresh leaves following CTAB method (Doyle and Doyle 1987) and quantified (60 ng μL–1).Simple sequence repeats (SSR)markers were selected from GrainGenes Database (http://wheat.pw.usda.gov).The categories of selected SSRs were:CsLv34 (1);CsSr2 (1);cfb (1);gdm (2);mag (6);gpw (7);cfd(8);barc (12);sun (13);wmc (34);and gwm (118).The PCR reaction volume was used as 10 μL;dNTPs,0.4;MgCl2buffer,0.75 and 1.25 μmol L–1of each of forward and reverse primer and 0.04 μL×0.2 UTaqpolymerase.

The thermal profile for each SSR was adjusted as10 min at 95°C;followed by 30 cycles with touchdown profile,of 30 s at 60°C and 72°C for 30 s.The annealing temperature for each SSR was used as given at GrainGenes Database(http://wheat.pw.usda.gov) with a final extension of 5 min at 72°C.After amplification,2 μL of the PCR product was run and scored on 2.5% agarosegel.The remaining PCR product was separated using polyacrylamide gel electrophoresis (PAGE) by mixing M700 with 1–48 samples and M800 with 49–96 samples.The genotypes PIRSABAK 2005 and ZARGOON 79 were repeated in each set of PAGE (1–48 and 49–96) to confirm the accuracy of samples loading and scoring.The PAGE was prepared by mixing of 20 mL of TB (Tris buffer);150 μL of APS;6.5% Gel Matrix and 25 μL of TEMED (tetra methyl ethylene diamine).The gel apparatus was placed on LICOR instrument with bottom of gel immersed in the buffer tank which was filled with 1×TBE running buffer.After successful pre-run,samples were loaded and run using SAGA Software.

2.3.Statistical analysis

The marker-trait association test was performed based on genotypic and phenotypic data by using the R Software.The principal coordinate analysis (PcoA) was conducted using the Statistica v6 Software.The extent of marker-trait association was detected significant atP<0.01 and effect of QTL was estimated byR2.

GWAS was performed using mixed linear model (MLM) to find marker trait associations (Arief 2010;Hickeyet al.2011).

The extent of linkage disequilibrium was evaluated usingD´ andR2(Devlin and Risch 1995).The methods of detection of digenic interactions and QTLs detection were same as described by Zhaoet al.(2014) and Qiet al.(2016).

3.Results and discussion

3.1.Stripe rust phenotyping

To assess the infection type of YR on our association pane,score was recorded for each entry for the years 2014–2016.Disease prevalence was of fourteen categories from 0 to TS(Fig.1).Each of the 0 and TS type infection was observed on 15% genotypes while TR type infection was observed on 7% genotypes.In a panel of 44 Pakistani landraces,eight infection types of stripe rust were observed (Gulnazet al.2019).The 0 and TS infection types were 69.4 and 7.8%landraces,respectively.The TR infection was observed in 5% Pakistani landraces (Gulnazet al.2019).The variety of infection types in the present association panel reflected broad spectrum APR genes in the panel which resulted in several marker-trait associations (MTAs).

3.2.Principle co-ordinate analysis

To measure genetic structure among genotypes in the association panel,principle co-ordinate analysis (PCoA) was performed using SSR data.PCoA didn’t divide association panel in distinct groups,suggesting that the panel was fit for genome wide association study (Fig.2).The PCo1 and PCo2 explained 4.9 and 4.6% molecular diversity in the association panel.The grouping of genotypes in PCo1 and PCo2 were consistent with that in cluster analysis(Appendix B).

3.3.Linkage disequilibrium (LD)

Fig.1 Prevalence of different stripe rust infection types in the association panel:absence of abrasion (0),moderately susceptible to susceptible (MSS),totally to moderately resistant(TMR),totally to moderately susceptible (TMS),totally resistant(TR),totally susceptible (TS) and susceptible (S).The number before abbreviations are representing the scale from 1 to 100.

Fig.2 Low dimensional representation of the genetic distance among 94 genotypes using a principle co-ordinate analysis(PCoA).

To display inter-and intra-chromosomal linkage disequilibrium,chromosomalD´andR2values for loci in significant LD atP<0.01 were used.The mean value ofR2for all pairs was 0.45.The significant decay of LD with the increase of genetic distance was observed.Maccaferriet al.(2005) defined four classes of marker pairs in wheat:i) strongly linked (distance<10 cM);ii) moderately linked(10–20 cM);iii) loosely linked (20–50 cM);and iv) unlinked pairs (>50 cM).Based onD´,LD was observed upto the distance of 45 cM.Weaker LD was observed between loci which were 20 cM apart (Appendix C).A review of previous reports on LD extent concluded that LD extent varied with varying wheat populations (Chaoet al.2007;Sajjadet al.2012).Genome wide LD decayed to 0.2 within a distance of 10–20 cM in a panel of 43 USA cultivars.This LD extent was higher compared to that reported previously(Breseghello and Sorrells 2006).The LD in 44 varieties and 240 RILs were 0.5 and 30 cM,respectively,when assayed with 91 SSRs and STS markers (Tommasiniet al.2007).This significant difference in LD proves the usefulness of germplasm of unrelated varieties over bi-parental populations.The LD extent up to～80 cM (R2=0.015) was observed on chromosome 3A in 94 spring wheat genotypes surveyed with 23 SSRs (Sajjadet al.2014).The LD byR2atP<0.001 by using 660 DArT markers were also estimated(Mulkiet al.2013).On a genome-wide level,almost 25%of all pairs of loci were in significant LD.These reports on LD structure in hexaploid wheat revealed that LD structure varies with populations,genomic regions and type of marker.The changes in LD along chromosomes were indicative of areas of genome that were under selection pressure(Somerset al.2007).

3.4.Marker trait associations

According to the consensus map constructed by the information given by Somerset al.(2004) and Maccaferriet al.(2015),a total of 71 SSRs were found to be associated with stripe rust resistance (Fig.3).In the present study,many of the SSR markers associated with stripe rust were in the previously reported chromosomal regions whereYrgenes or QTLs had been reported.However,some novel genomic regions associated withYrresistance were observed on chromosomes 1AS,2AL,2DL,3DS,5BL,5DS,5DL,6DL,7AS and 7DL (Fig.3).Chromosomes IDS,4AL,4BL,5AS,5BS,6AL,6BS,and 6DS had the largest number of significantly associated loci for stripe rust.Most of these loci were previously reported as tightly linked withYrgenes namely,Yr1,Yr17,Yr32,Yr34,Yr48,Yr56,Yr60,Yr63andYr39.

Developments in plant molecular research have provided a tool to relate genotypic and phenotypic variation in sets of germplasm.Many studies have involved in associations of markers and traits among genetic materials with known pedigree and documented genes (Crossaet al.2007;Tommasiniet al.2007;Maccaferriet al.2010;Bansalet al.2011;Sajjadet al.2014;Shahzadet al.2016).MTA analysis detected 31 genomic regions across all 21 chromosomes underpinning stripe rust resistance (Fig.3).Out of 203 polymorphic SSRs,71 were found associated with stripe rust resistance configuring 31 QTLs (Table 1).On chromosome 1A,two QTLs were detected containing unknown genes.The marker interval ofQYr.uaf.1ASis longer,therefore,more SSRs should be used to narrow down the QTL interval(Fig.3).These QTLs can be focused to isolate newYrresistance genes in Pakistani germplasm.The long arm of chromosome 1B is rich inYrgenes.Three QTLs on 1BLQYr.uaf.1BL.1,QYr.uaf.1BL.2andQYr.uaf.1BL.3were found spanningYr9,Yr10,Yr64,Yr65,YrH52andYr29,respectively.These three QTLs were stable across various breeding germplasms and were previously reported in different breeding populations (Liuet al.2008;Singhet al.2009;Chenget al.2014).A new QTL with unknown gene was found on 1 DS between wmc432 and cfd15.Two previously reported QTLs on chromosome 2A were validated in Pakistani germplasm.The QTLs on 2AS and 2AL containedYr17,Yr56andYr32genes,respectively (Helgueraet al.2003;Eriksenet al.2004).TheQYr.uaf.2ALhad higher PVE value of 11.9%(Table 1).The two QTLs,QYr.uaf.2BL.1andQYr.uaf.2BL.2spanningYrC51(Zhenget al.2014),Yr5,Yr43andYr54(Murphyet al.2009) were validated with PVE values of 7.9 and 10.1%,respectively.At the distal end of 2D chromosome,a newYrlocus linked with barc124 (PVE,7.3%) was identified in Pakistani germplasm.CloningYrgene from this QTL can add a useful APR genetic resource for wheat breeding programs.TheQYr.uaf.2DLvalidated the prevalence ofYr54andYr55(Basnetet al.2014) in Pakistani germplasm.A new minor QTL with PVE value of 4.7% was observed on chromosome 3AS.On 3BS,a major QTL with PVE value of 11.9% was confirmed with three known genesYr4,Yr30,Yms-1b(Borneret al.2000;Suenagaet al.2003;Zhouet al.2014).Recently,Nsabiyeraet al.(2018) clonedYr80genes in 3BL between markers gwm108 and gwm376 but no MTA was detected for this gene in present study.Two new QTLs with unknown genes were found on 3DS (Table 1) and on 3DL QTL containingYr45also reported earlier (Liet al.2011).Two linked genesYr51andYr60between wmc776–gpw356 on 4AL (Randhawaet al.2014;Herrera-Foesselet al.2015)were found to be functional in Pakistani wheat germplasm.A major new QTL ‘QYr.uaf.4BS’ with unknown gene was discovered on 4BS with PVE value of 11.4%.Two QTLs on 4DS were detected which were previously reported to containYr28,YrS2388andYr46(Herrera-Foesselet al.2011;Liet al.2011).The two QTLs on 5A and 5B were reported to carry threeYrgenes namelyYr34,Yr48andYr47,respectively(Bansalet al.2011;Loweet al.2011).Two new QTLs on 5D were found in present study with PVE value of 6.5 and 5.6%,respectively.Two new minor QTLs one on each 6AS and 6DL were observed with PVE values of 4.3 and 3.2%,respectively.Two linked genesYr35andYr36were isolated from theQYr.uaf.6BSobserved herein (Uauyet al.2005;Dadkhodaieet al.2011).The two QTLs on 7BL contained fiveYrgenes namelyYr39,Yr52,Yr59,YrC59andYrZH8(Lin and Chen 2007;Renet al.2012;Zhouet al.2014).The QTL on 7DS spannedYr18gene (Lagudahet al.2009) and were validated in Pakistani germaplsm.

Fig.3 A microsatellite consensus map of chromosomes of wheat showing marker trait associations.

Table 1 Quantitative trait locus/loci for stripe rust resistance (YR) by GWAS in the Pakistani germplasm1)

4.Conclusion

The study validated 19 QTLs and 35Yrgenes previously reported in various breeding populations and discovered 12 new QTLs underpinning yellow rust resistance across all 21 chromosomes.The variety of stripe rust infections and PcoA suggested the fitness of the panel for genome wide association study.The LD structure in the association panel was found to be extended up to 45 cM.The novel QTLs reported,herein,can be focused for isolating newYrgenes to strengthen APR genetic resources in wheat crop.

Acknowledgements

The authors acknowledge Higher Education Commission(HEC),Islamabad,Pakistan for providing for funding (21-179/SRGP/R&D/HEC/2014) and the International Research Support Initiative Program (IRSIP) Scholarship for this research work.The authors are thankful to Dr.Harbans Singh Bariana and Dr.Urmil Bansal for giving an opportunity to conduct molecular part of this research work at their lab in Plant Breeding Institute,the University of Sydney,Australia.The authors are also grateful to Dr.Vivi N.Arief (School of Agriculture and Food Sciences,University of Queensland,Brisbane,Australia) for helping us in statistical analysis.

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm