Yang Liu,Hui Chen,Chunxin Li,c,Lirong Zhang,d,Mingqin Shao,Yuhui Pang,e,Xiangyang Xu,Guihua Bai,g,*

a Institute of South Subtropical Crops,Chinese Academy of Tropical Agricultural Sciences,Zhanjiang 524013,Guangdong,China

b Department of Agronomy,Kansas State University,Manhattan,KS 66506,USA

c Henan Academy of Agricultural Science,Zhengzhou 450002,Henan,China

d College of Plant Protection,Hebei Agricultural University,Baoding 071000,Hebei,China

e College of Agriculture,Henan University of Science and Technology,Luoyang 471023,Henan,China

f USDA-ARS Wheat,Peanut,and Other Field Crop Research Unit,Stillwater,OK 74075,USA

g USDA-ARS,Hard Winter Wheat Genetic Research Unit,Manhattan,KS 66506,USA

Keywords:

A B S T R A C T Wheat leaf rust is a prevalent foliar disease in wheat worldwide.Growing resistant cultivars is an effective strategy to minimize the impact of leaf rust on yield and grain quality.Lr42 is a leaf rust resistance gene identified from Aegilops tauschii and is still effective against current predominant leaf rust races in the United States and many other countries.In this study,we developed diagnostic DNA markers for Lr42 using the sequence polymorphisms of a differentially expressed gene(TaRPM1)encoding a putative NBARC protein in the Lr42 candidate region identified by RNA-sequencing of two near-isogenic lines contrasting in Lr42 alleles.Markers were designed based on a deletion mutation and a single nucleotide polymorphism(SNP)in the gene.Haplotype analyses of the newly developed markers in the three diversity panels demonstrated that they are diagnostic for Lr42,and superior to previously used markers in selection accuracy.These markers have the advantages of low cost and easy assay,and they are suitable for marker-assisted selection in breeding programs with either high-or low-throughput marker screening facilities.

1.Introduction

Wheat(Triticum aestivumL.)is one of the most important cereal crops in the world[1].Diseases and insects are major biotic constraints threatening wheat production and food safety worldwide.Leaf rust,caused byPuccinia triticinaErikss.,is one of the most serious wheat foliar diseases[2].Infected leaves are usually covered with rust pustules,which significantly limits photosynthesis by reducing green tissue area,results in insufficient grain filling,and thus decreases grain weight.Severe rust epidemics can cause more than 40%losses in wheat yield[3].Recent global warming provides more conducive environments for the rust pathogen reproduction and spread,which may result in wheat leaf rust epidemics being more frequent and severe than before[4].

Growing resistant varieties has proven to be one of the most economical and eco-friendly approaches to manage the disease.Two types of wheat leaf rust resistance have been reported,seedling resistance and adult plant resistance.A plant with a seedling resistance gene typically shows high resistance at all wheat growth stages.However,seedling resistance genes can be easily overcome by a newly emerged rust race,which evades race-specific recognition by the resistance gene.Adult plant resistance is usually controlled by race-nonspecific genes,which are presumably not dependent on pathogen recognition,and therefore is more durable[5].One type of adult plant resistance is slow rusting,which usually has a longer latent period and fewer small pustules with fewer spores than in susceptible plants after infection.This type of resistance can last as long as several decades[6].To date,more than 100 wheat leaf rust resistance genes and quantitative trait loci(QTL)have been reported with 79 genes formally named(Lr1–Lr79)[7–10].Most of the named genes are seedling resistance genes.Only a small portion of them are adult-plant resistance genes[11],includingLr12,Lr13,Lr22a,Lr22b,Lr34,Lr35,Lr37,Lr46,Lr48,Lr49,Lr67,andLr68.Among them,Lr34[12],Lr46[13],Lr67

[14],andLr68[15]demonstrate slow rusting resistance.

To develop durable leaf rust resistant cultivars,two main strategies were proposed:accumulating multiple partial resistance genes for adult plant resistance and pyramiding multiple racespecific resistance genes or both race-specific and adult plant resistance genes in a cultivar.Using the new cultivars with multiple resistance genes will reduce the formation rate of new races and the resistance is not easily overcome by the new rust races[16,17].Therefore,stacking leaf rust resistance genes from different sources can enhance resistance gene diversity and improve the durability of rust resistance in new cultivars.

Lr42was identified from anAegilops tauschiiCoss.accession TA2450[18]and integrated into a hard winter wheat cultivar‘Century’through backcrossing[19].Lr42showed partial resistance with an infection type of 2 in both seedling and adult plant stages[19].The derived wheat line KS93U50 withLr42has been successfully used in several breeding programs[20,21].Lr42was reported to play a role in increasing wheat yield and kernel weight in leafrust infected fields[22].

Lr42was previously mapped on the distal end of chromosome 1DS using simple sequence repeat(SSR)markers[23].Liu et al.[24]mappedLr42between XWMC432 and XGDM33 using an F2population of KS93U50×Morocco.High false-positive rates of these SSR markers were observed in the wheat lines from the US Hard Winter Wheat Regional Performance Nurseries(RPN),which suggests a low selection accuracy of these markers in markerassisted breeding.Gill et al.[25]further evaluated a recombinant inbred line(RIL)population from the same cross(KS93U50×Morocco)for rust resistance and definedLr42to a 3.7 cM genetic interval between markers TC387992 and WMC432.They developed two Kompetitive allele-specific polymerase chain reaction(KASP)markers(SNP113325 and TC387992)for marker-assisted selection.However,these markers have not been validated in diverse germplasm.The objectives of this study were to 1)develop breeder-friendly diagnostic markers that can be used in breeding programs with different throughput marker screening facilities and 2)validate the usefulness of these markers in diverse breeding populations.

2.Materials and methods

2.1.Plant materials

The resistant accessions with theLr42resistance allele used in this study are oneAe.tauschiiaccession TA2450 and one wheat accession KS93U50.The susceptible accessions with the susceptibility allelelr42include oneAe.tauschiiaccession TA10132 and one wheat accession Morocco.Two sets of near-isogenic lines,includingLr42-NIL-R1 andLr42-NIL-R3 as the resistant NILs andLr42-NIL-S1,andLr42-NIL-S3 as the susceptible NILs,were selected from the two(KS93U50×Morocco)F4:5heterogeneous inbred families(HIFs)[26].The eight accessions were used for initial marker screening for polymorphisms.A population of 140 F5:6plants that were selected from the three selfed F4:5HIFs(RIL-14,RIL-80 and RIL-125)of KS93U50×Morocco was genotyped using the polymorphic markers in theLr42region.Several other populations from different sources were used to validate the newly developed markers from this study.These included a population of 56 NILs with known phenotypes derived from KS93U50×OK92G205[23],and a panel of 67 wheat breeding lines and parents that were selected from the crosses with anLr42resistant parent,which was kindly provided by Dr.Ravi Singh from International Maize and Wheat Improvement Center(CIMMYT,Mexico).We also tested a set of 73 breeding lines including 68 randomly selected breeding lines from 2018 US Hard Winter Wheat Regional Germplasm Observation Nursery(RGON)and four breeding lines from the 2018 Hard Winter Wheat RPN(https://www.ars.usda.gov/plainsarea/lincoln-ne/wheat-sorghum-and-forage-research/docs/hardwinter-wheat-regional-nursery-program/research/) and one breeding line selected from 2015 RPN that carries SNP113325 alleles associated withLr42resistance.Selected KASP markers were also screened in 2020 RPN and RGON(https://www.ars.usda.gov/plains-area/lincoln-ne/wheat-sorghum-and-forage-research/docs/hard-winter-wheat-regional-nursery-program/research/).

2.2.Evaluation of leaf rust resistance

A total of 149 F5:6plants derived from three HIFs of(KS93U50×Morocco)F4:5RILs,56 F5:6NILs derived from KS93U50×OK92G205[23],and 64 selected breeding lines from CIMMYT were inoculated with a rust isolate PNMRJ at Kansas State University.PNMRJ is avirulent toLr42but virulent toLr24[23].All the plant materials were planted in 72-cell plastic flats and the seedlings were inoculated with the isolate PNMRJ at the two-leaf stage.Before inoculation,rust spores from a liquid nitrogen tank were heat-shocked in a water bath at 42°C for 5 min to break dormancy.The spores were suspended in Soltrol 170 light oil(Chevron Phillips Chemical Company,Bellevue,WA,USA)and misted uniformly over the seedling leaves using a pressure sprayer.After inoculation,plants were incubated in a dew chamber at 20 °C and 100% relative humidity for 18–24 h before being moved to a greenhouse bench for rust establishment.The inoculated plants were grown at 20±3°C supplemented with 10 h daylight.The rust infection type(IT)was recorded at 10 d after inoculation using a 1–4 scale[27]and rechecked two days later for possible error.Based on the reaction of KS93U50 to the rust isolate PNMRJ infection,plants with IT≤2 were classified as resistant and IT＞2 as susceptible.

2.3.DNA extraction and primer design

Two pieces of 2.0 cm-long leaf tissues per sample were collected into a 96-deepwell plate at the three-leaf stage,dried in a freeze dryer(Thermo Savant,Holbrook,NY,USA)for 48 h,and then ground into a fine powder using a Mixer Mill(MM 400,Retsch,Germany).DNA was extracted using a modified cetyltrimethylammonium bromide(CTAB)protocol[28].Based on the reference genome sequence of Chinese Spring(GenBank:KU641029),genespecific primers for two STS markers and two KASP markers(Table S1)were designed using software Primer Premier v.5.0(Premier Biosoft International,Palo Alto,CA,USA).Primer sequences for other markers were obtained from a previous publication[25].

2.4.Differentially expressed genes(DEGs)in Lr42 region identified by RNA-seq

Leaf tissues were collected from inoculated plants of two sets of NILs(Lr42-NIL-R1/S1 andLr42-NIL-R3/S3)at 0,24,and 36 h postinoculation(HPI)with the rust isolate PNMRJ.Total RNA was extracted using TRIzol from Invitrogen(ThermoFisher Scientific,Grand Island,NY,USA)following the manufacturer’s instructions.The quantity and quality of total RNA was assessed using a Bioanalyzer 2100 system(Agilent Technologies,Santa Clara,CA,USA).The cDNA libraries were constructed and sequenced using an Illumina HiSeq 2500 sequencer(Illumina,San Diego,CA,USA)in the Genomics Sciences Laboratory-North Carolina State University.After low-quality reads were removed,all trimmed reads were aligned onto IWGSC RefSeq v1.0 assembly[29].The reads with high-quality scores were annotated using IWGSC RefSeq v1.0 annotation(https://urgi.versailles.inra.fr/download/iwgsc/IWGSC_RefSeq_Annotations/v1.0/).To identify candidate genes forLr42,all DEGs in theLr42interval that were defined by Gill et al.[25]were selected for further sequencing to isolate full-length genes for marker development.

2.5.Cloning of full-length genomic and coding sequences of TaRPM1

The full-length genomic DNA sequence of the selected DEG was amplified from the two parents and one pair of the NILs with contracting alleles atLr42by PCR using the gene-specific primers(Table S1).To amplify the full-length coding sequence(CDS)of the DEG,total RNA was extracted from the NILs and 1 μg of total RNA was treated with DNase I from Invitrogen to remove residual DNA and then reverse-transcribed to cDNA using SuperScript IV First-Strand Synthesis System and an oligo(dT)20primer following the manufacturer’s instruction(Invitrogen,USA).The full-length CDS of the DEG was amplified from the NILs by PCR using the gene-specific primers(Table S1).The PCR products of the full-length genomic and coding sequences of the DEG from the NILs were cloned into the pCR-XL-2-TOPO vector from Invitrogen(ThermoFisher Scientific,Grand Island,NY,USA).The inserted fragments in the construct were verified by sequencing in an ABI 3730 DNA Sequencer(ThermoFisher Scientific).The sequence data were assembled and aligned using BioEdit(http://www.mbio.ncsu.edu/Bioedit/bioedit.html)and DNASTAR software(http://www.dnastar.com/).Gene structure was predicted using the SnapGene Viewer program(https://www.snapgene.com).

2.6.Genotyping assays

To develop a gel-based marker, a 13-μL PCR mix consisted of 150 nmol L-1each of the two primers, ～60 ng DNA, 3.0 U of Taq,and a 5× PCR master mix (New England Biolabs, Inc., Beverly,MA, USA). PCR was started at 94 °C for 3.0 min, followed by 35 cycles of 30 s at 94 °C, 30 s at 60 °C, 30 s at 72 °C, and ended with a final extension at 72 °C for 10 min. PCR products were separated in a 2.0% (w/v) agarose gel. For the capillary-electrophoresis-based marker, a 13 μL PCR mix contained 1× ASB buffer, 2.5 mmol L-1of MgCl2, 200 μmol L-1of dNTP, 30 nmol L-1of the fluorescencelabeled M13 primer (5′-ACGACGTTGTAAAACGAC-3′), 50 nmol L-1of a forward primer with the M13-tail sequence added to its 5′-end, and 80 nmol L-1of a reverse primer, 3 U Taq polymerase,and ～60 ng DNA. All PCRs were carried out in a C1000 Touch Thermal Cycler (BioRad Laboratory Inc. Hercules, CA, USA) using a touchdown program starting at 94 °C for 5 min, followed by 10 cycles of 94 °C for 30 s, annealing at 68 °C for 30 s, with a decrease of 0.8 °C in each subsequent cycle, and extension at 72 °C for 30 s;then PCR went through an additional 30 cycles of 94 °C for 30 s,55 °C for 30 s and 72 °C for 30 s, and ended with a final extension at 72 °C for 10 min. The PCR products labeled with fluorescent dyes were pooled and analyzed in an ABI 3730 DNA Sequencer, and scored using GeneMarker v1.75 (SoftGenetics LLC, State College,PA, USA).

The KASP assays were performed using a 4 μL reaction mix including 1.93 μL 2×KASP master mix,0.07 μL KASP primer mix and 2 μL DNA(～20 ng μL-1)followed the manufacturer’s instruction from LGC Genomics(https://www.lgcgenomics.com).PCR products were analyzed in an ABI 7900HT fast real-time PCR system(Applied Biosystems)and KASP SNPs were called using Klustercaller v3.4.1.39(LGC Genomics).

3.Results

3.1.RNA-seq identified a unique NB-ARC gene differentially expressed in Lr42 interval

A previous study mappedLr42to a 3.7 cM region flanked by the markers TC387992 and WMC432(Fig.1A)[25].Alignment of the primer sequences of the two markers to the Chinese Spring wheat reference genome[29]identified a 600 kb region corresponding to theL42interval that includes 19 high confidence genes(Fig.1B),whereas the corresponding physical interval in theAe.tauschiireference genome[30]is 1.03 Mb that includes 20 high-confidence genes(Fig.1C).To identify possible candidate genes for marker development,RNA-sequencing was conducted using the two pairs of NILs carrying the contrastingLr42alleles.Five genes in theL42interval showed differential expression(Fig.1B),with four putative protein kinase genes(TraesCS1D01G017700,TraesCS1D01G017800,TraesCS1D01G017900,andTraesCS1D01G018000)and one NBSLRR disease resistance gene(TraesCS1D01G018700)(Table S2)based on the wheat reference genome[29].However,only two genes showed differential expression based on the reference sequence ofAe.tauschii:a putative protein kinase gene(AET1Gv20040700,a homolog ofTraesCS1D01G018000in wheat)and a NBS-LRR disease resistance gene(AET1Gv20042900,a homolog ofTraesCS1D01G018700in wheat)(Fig.1C;Table S2).BecauseLr42was fromAe.tauschii,we selected the two genes fromAe.tauschiias the putative candidates.Since most cloned rust resistance genes in wheat are NBS-LRR genes,we selected

AET1Gv20042900fromAe.tauschii,designatedTaRPM1(resistance toPseudomonas syringaeinArabidopsis thaliana),as the final candidate for marker development.Another two NBS-LRR genes(TraesCS1D01G018800andTraesCS1D01G018900)were identified next toTaRPM1in the Chinese Spring reference genome;however,none of them was expressed in eitherLr42-NIL-R orLr42-NIL-S lines at 0,24,36 h post-inoculation,and onlyTraesCS1D01G018900has a homolog in theAe.tauschiireference genome.

Sequencing the full-length genomic DNA ofTaRPM1from the NILs using the primersTaRPM1-F andTaRPM1-R(Table S1)that were designed based on Chinese Spring reference genome sequence(GenBank:FN564434.1)identified a 3447 bp sequence including a 371 bp 5′-UTR,an open reading frame(ORF)of 2714 bp and 362 bp 3′-UTR in the susceptible NIL and parent(Fig.2),but 3520 bp including the same(371 bp)5′-UTR and a much longer(1613 bp)3′-UTR in the resistant NIL and parent.Sequencing the full-length coding sequences(CDS)ofTaRPM1from the two NILs confirmed the gene prediction.Comparison of the full-length sequences ofTaRPM1between the contrasting NILs and between the two parents identified 23 SNPs and two indels.A 31-bp insertion was found at the end of the predicted ORF from the resistant NIL and resistant parent,which caused early termination of protein translation(Fig.2).

3.2.Gene-specific markers for TaRPM1

One pair of the gene-specific sequence-tagged site(STS)primers(Table S1)was designed for a gel-based assay,namedTaRPM1-GSM,to capture the 31 bp insertion polymorphism between the NILs(Fig.3A).This co-dominant marker amplified a 339 bp band in the four susceptible genotypes(Morocco,TA10132,Lr42-NILS1,andLr42-NIL-S3)and a 370 bp band in the resistant genotype(KS93U50,TA2450,Lr42-NIL-R1,andLr42-NIL-R3)in an agarose gel(Fig.4A).The marker polymorphism between the NILs suggests that theTaRPM1-GSM marker is specific for chromosome 1DS.

Fig.1.High-confidence genes(HCs)and differentially expressed genes(DEGs)in the 0.60 Mb of Lr42 candidate region on chromosome arm 1DS of Chinese Spring(CS)reference genome and Ae.tauschii.(A)Genetic map and markers location in a KS93U50×Morocco F5 recombinant inbred line(RIL)population[25];HCs and DEGs(highlighted in green)annotated in(B)CS reference genome and(C)Ae.tauschii chromosome.

To improve assay throughput,the gel-based STS marker was converted to a capillary-electrophoresis-based STS marker,named asTaRPM1-ABI,to be genotyped in an ABI 3730 DNA Sequencer by adding a fluorescence-labeled M13-tail sequence to the 5′-end of the forward primer(Fig.3A;Table S1).BothTaRPM1-GSM andTaRPM1-ABI were assayed in the secondary RIL population of 140 F5:6plants developed from the three HIFs(RIL-14,RIL-80,and RIL-125).As expected,TaRPM1-ABI showed the identical codominant segregation pattern asTaRPM1-GSM and both cosegregated with rust resistance in the population(Table S3).In addition,the marker clearly separated homozygotes(Lr42andlr42)from heterozygotes(Fig.4B).

To develop KASP marker,primers for markerTaRPM1-KASP1(Table S1)were designed to capture the same 31 bp indel inTaRPM1(Fig.3A).Because the indel is not a typical SNP,we designed two forward primers in different sequence positions and one common reverse primer.Screening of the KASP marker in the 140 F5:6plants from KS93U50×Morocco showed identical genotypic result as those from theTaRPM1-GSM andTaRPM1-ABI markers(Fig.4C;Table S3).Meanwhile,another KASP marker,TaRPM1-KASP2,was designed based on an SNP between the NILs(Fig.3B;Table S1),and this marker produced identical genotypic results as those from the markersTaRPM1-GSM,TaRPM1-ABI andTaRPM1-KASP1 in the 140 F5:6plants from KS93U50×Morocco(Table S3).All these markers were codominant and are useful for marker-assisted selection ofLr42in early segregating generations.

3.3.Marker validation in different populations

To validate the newly developed markers,they were analyzed in a panel of 67 CIMMYT breeding lines from different crosses withLr42in one of their parents and some of their parents.The panel contains bothLr42resistant and susceptible lines with Quaiu as theLr42donor(Table 1).The fourTaRPM1markers produced identical results for all accessions evaluated and generated two haplotypes,Hap I(TaRPM1-1a)and Hap II(TaRPM1-1b),in this panel.All the accessions with Hap I carry the KS93U50 allele,whereas all the accessions with Hap II carry the Morocco allele.The leaf rust data matched with the haplotype classification of most tested lines with the exception of four breeding lines(Table 1),indicating that both STS and KASP markers ofTaRPM1are near diagnostic forLr42in the panel.The fourTaRPM1markers were also validated in a population of 56 NILs derived from the cross between a resistant germplasm line KS93U50(Lr42)and a susceptible breeding line OK92G205(Non-Lr42)used by Sun et al.[23].As expected,the four markers clearly separated the two contrasting genotypes,and the genotypic groups matched with their rust data(data not shown).

Fig.2.Gene and protein structures of TaRPM1.(A)Structures of the two contrasting alleles of TaRPM1.Solid blue bars denote exons.Red bars refer to sequence insertions in the ORF region of TaRPM1-1a.Numbers are fragment lengths in bp.(B)Analysis of TaRPM1 protein structure using NCBI Conserved Domain Database(CDD).Rx_N,resistance protein N-terminal domain;NB-ARC,nucleotide-binding adaptor shared by APAF-1,R protein,and CED-4;LRR,leucine-rich domain.

The newly developed markers were compared with the previously used KASP markers forLr42(Table S1),SNP113325[25]and PB12(S.Z.Liu,personal communication),in the two wheat populations.For the 56 NILs from KS93U50×OK92G205,the results were the same among the six tested markers(data not shown).In the 67 CIMMYT accessions,however,inconsistent results were found for samples 15 and 50(Table 1).The sample 15 showed leaf rust resistance,but the marker SNP113325 amplified a susceptibility allele;the sample 50 was leaf rust susceptible,but the markerPB12amplified a resistance allele.Therefore,both SNP113325 andPB12incorrectly identify one sample each in the CIMMYT accessions,suggesting that these four newly developed markers are better markers than the two previously developed SNP markers.

A set of 73 US hard winter wheat breeding lines were also evaluated with theTaRPM1KASP marker and the data were compared with a previously reported marker SNP113325.SNP113325 identified eight lines with the allele associated withLr42resistance,which represented all positive lines identified after screening 392 lines of 2018 RGON and 95 lines of 2018 RPN.The KASP markers forTaRPM1,however,only identified two lines asLr42positive(Table S4).BothTaRPM1-KASP1 andTaRPM1-KASP2 markers were also screened in 2020 RPN(91 accessions)and RGON(375 accessions)and only one line(TX16A001289)was identified to carry theLr42marker allele.Phenotyping the line for rust resistance confirmed that the line is rust resistant(data not shown).

4.Discussion

To date,Lr42is still effective against predominant rust races in the USA and many other countries[3].However,Lr42has not been widely used in wheat breeding programs due to a lack of diagnostic markers for selectingLr42in breeding.Therefore,the development and validation of molecular markers forLr42will facilitate its deployment by marker-assisted selection.

Gill et al.[25]mapped theLr42to a 3.7 cM region.Blasting the flanking marker sequences from Gill et al.[25]identified a 0.6 Mb(7,874,340–8,475,810 bp)physical interval in the Chinese Spring genome[29]and a 1.03 Mb(8,765,910–9,797,282 bp)interval inAe.tauschiigenome[30]that harbor 19 and 20 high confidence genes,respectively.RNA-seq analysis using theLr42NILs identified five DEGs among 11 expressed genes in the interval based on the Chinese Spring wheat reference genome and two DEGs among eight expressed genes in the interval based onAe.tauschiireference genome.Among the five DEGs,onlyTraesCS1D01G018700(TaRPM1)is an NBS-LRR disease resistance gene that was differentially expressed and mapped on both reference genomes of the Chinese Spring wheat andAe.tauschii.To date,most cloned rust resistance genes in wheat encode NBS-LRR-type proteins[11,31–37];therefore,we selectedTaRPM1as the candidate gene for further marker development.In addition,two genes encoding NBSLRR proteins(TraesCS1D01G018800andTraesCS1D01G018900)were mapped next toTaRPM1in theLr42interval based on the Chinese Spring reference genome[29];however,none of them expressed in either NILs at any of the three time points of inoculation.

Comparison ofTaRPM1genomic sequences between the NILs revealed two InDels and 23 SNPs in the coding region ofTaRPM1,suggesting a high level of sequence polymorphisms in the coding region ofTaRPM1between the two NILs.In the previous report,TaRPM1was not included in the list of 13 candidate NBS-LRR resistance genes[25];however,further analysis ofTaRPM1protein structure using the NCBI Conserved Domain Database(CDD)(https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) found thatTaRPM1encodes a protein with an Rx N-terminal domain(a coiled-coil domain in many plant resistance proteins),an NB-ARC domain(a novel signaling motif shared by plant resistance gene products)and a LRR domain involving in protein–protein interactions.Therefore,TaRPM1is a gene coding for a typical NBS-LRR resistance protein in wheat.TaRPM1contains the RPM domain in the leaf rust-susceptible genotype,while the leaf rust-resistant genotype produced a truncated TaRPM1 protein that was prematurely terminated due to the 31 bp fragment insertion.It remains unknown ifTaRPM1is the causal gene forLr42based on the data available in this study.Further studies need to be conducted to validate the function ofTaRPM1on rust resistance.However,even ifTaRPM1is not the causal gene,it tightly links toLr42,thus it was selected for developing the diagnostic markers.

Fig.3.Sequence alignment among different alleles of TaRPM1 from KS93U50 and Morocco,and from homeologous genes of TaRPM1 on chromosomes 1A and 1B in Chinese Spring to identify the conserved sequence for development of Lr42 diagnostic markers.(A)Primer sequences and locations for TaRPM1-GSM,TaRPM1-ABI,and TaRPM1-KASP1.(B)Primer sequence and location for TaRPM1-KASP2.CS-Lr42-1A,CS-Lr42-1B,KS93U50-Lr42-1D,and Morocoo-Lr42-1D represent TaRPM1 sequences in chromosome 1A of Chinese Spring,chromosome 1B of Chinese Spring,chromosome 1D of KS93U50 and chromosome 1D of Morocco,respectively.Different colored arrows labeled the locations of marker primer sequences in TaRPM1 gene.

In this study,the 31 bp insertion and one SNP site inTaRPM1were used as the targets for developing four markers of different types,TaRPM1-GSM,TaRPM1-ABI,TaRPM1-KASP1,andTaRPM1-KASP2,to fit different marker genotyping platforms.These new markers provide more accurate prediction ofLr42in different genetic backgrounds compared to previously developed markers.Several markers have been previously available forLr42,including TC387992,WMC432,KASP113325,and PB12.These markers are all close toLr42(Fig.1)and are usable in certain genetic backgrounds.However,none of them have been validated in diversity panels.Both TC387992 and WMC432 markers produced too many falsepositive signals when they were assayed in a panel of US hard winter wheat breeding lines.KASP113325 is closer toLr42and has been used in the US Small Grain Genotyping Laboratories to screenLr42for several years,but it is still about 3.2 cM distal toLr42[25].PB12was a gene-based marker developed fromTraesCS1D01G018200that is also close toLr42.Compared to theTaRPM1-based markers,both PB12 and KASP113325 showed inconsistency between genotypic and phenotypic data in one of the 67 tested lines from the CIMMYT breeding panel(Table 1).For the new markers developed in this study,their genotypes matched well with the phenotypic data of these CIMMYT lines with four exceptions.The four unmatched lines(lines 12,19,51,and 61)were rust-resistant but carry the susceptibility allele at all sixLr42markers tested,suggesting that resistance in the four lines might be contributed by other leaf rust resistance genes.Another possibility is thatTaRPM1may not be the causal gene forLr42,thus,we cannot rule out the possibility of false negative signals of the KASP markers in the four lines.In addition,in the US hard winter wheat breeding lines,out of 392 tested accessions from the 2018 RGON and 95 accessions from RPN screened with KASP113325,only eight accessions showedLr42positive(https://www.ars.usda.gov/plains-area/lincoln-ne/wheat-sorghum-andforage-research/docs/hard-winter-wheat-regional-nursery-program/research/).When the eight KASP113325 positive lines were screened withTaRPM1-KASP1,only two were positive(Table 1),suggesting much lower false positive ofTaRPM1-KASP1 than KASP113325.In 2020,both RPN(91 accessions)and RGON(327 accessions)were screened with both markersTaRPM1-KASP1 andTaRPM1-KASP2,only the accession TX16A001289 was identified to carryLr42resistance allele(https://www.ars.usda.gov/plainsarea/lincoln-ne/wheat-sorghum-and-forage-research/docs/hardwinter-wheat-regional-nursery-program/research/).This line was confirmed to be resistant after phenotyping at Kansas State University.Based on our knowledge,Lr42was not widely used in the US hard winter wheat breeding programs,therefore its extremely low frequency in the US hard winter wheat breeding programs is expected.Only three positive lines identified from＞800 lines using these markers suggesting these markers are highly diagnostic.

Fig.4.Genotypic analysis of two selected wheat accessions,two Ae.tauschii accessions and four near-isogenic lines(NILs)contrasting in Lr42 alleles using the gene-specificmarkers.(A)A gel image of TaRPM1-GSM marker,from left are DNA ladder,KS93U50,TA2450,Lr42 resistant near-isogenic lines NIL-R1 and NIL-R3,Morocco,TA10132,and lr42 susceptible near-isogenic lines NIL-S1 and NIL-S3.(B)TaRPM1-STS peaks for KS93U50(Lr42-1a),Morocco(Lr42-1b),and a heterozygous F5:6 line of KS93U50×Morocco analyzed in an ABI DNA sequencer.(C)Segregation of KASP markers TaRPM1-KASP1 and TaRPM1-KASP2 in the(KS93U50×Morocco)F5:6 population showing co-dominant segregation of the markers TaRPM1-KASP1(left)and TaRPM1-KASP2(right).Blue dots are for the resistant homozygotes,red dots for heterozygotes and green dots for susceptible homozygotes.

Table 1Haplotypes of six markers linked to Lr42 in 67 wheat breeding lines and parents from CIMMYT.

Table 1(continued)

All the markers are co-dominant and can clearly distinguishLr42homozygotes from heterozygotes,which enable breeders to make the selection in early generations of breeding.Although KASP113325 and PB12 are useful for marker-assisted selection,theTaRPM1-based markers are better markers for predictingLr42and meet all three criteria of good markers[38]:co-segregating withLr42phenotypes in multiple populations,absence of theLr42positive alleles in non-Lr42lines and flexible assays to fit various types of laboratory settings.Therefore,they can be used for marker-assisted breeding to speed up the screening process,reduce the screening cost,and improve the selection efficiency.

In summary,we developed four markers of different types to fit various screening platforms based on a disease resistance candidate gene in theLr42region identified by RNA-sequencing of theLr42NILs.These markers have been validated to be diagnostic forLr42in different diversity populations and proved to be superior to previously reported markers with improved accuracy and high flexibility.Therefore,these markers should be useful for markerassisted selection ofLr42to improve wheat leaf rust resistance in wheat breeding programs.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Guihua Baiconceived the study;Yang Liu,Hui Chen,Chunxin Li,Lirong Zhang,Xiangyang Xu,Yuhui Pang,and Mingqin Shaoconducted experiments;Yang Liu,Hui Chen,and Chunxin Lianalyzed data;Yang Liu,Hui Chen,and Guihua Bai

wrote the manuscript.All authors revised and approved the final manuscript.

Acknowledgments

This is contribution number 21-071-J from the Kansas Agricultural Experiment Station.The authors thank Dr.Sanzhen Liu from Kansas State University for providing PB12 marker sequence,Dr.Robert Bowden from USDA helped with leaf rust phenotyping,and Dr.Ravi Singh from International Maize and Wheat Improvement Center(CIMMYT)for providingLr42breeding materials for marker validation.This project is partly funded by the National Research Initiative Competitive Grants(2017-67007-25939)from US Department of Agriculture,the National Institute of Food and Agriculture.Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.02.012.