Powdery mildew disease resistance and markerassisted screening at the Pm60 locus in wild diploid wheat Triticum urartu

Fuki Zho, Yinghui Li,Boju Yng, Hongo Yun,c,Cong Jin, Lixun Zhou,Hongcui Pei,c, Lifng Zho,Yiwen Li, Yilin Zhou, Jinkun Xie,*, Qin-Hu Shen,c,**

aCollege of Life Sciences,Jiangxi Normal University, Nanchang 330022,Jiangxi,China

bState Key Laboratory of Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences,Beijing 100101,China

cUniversity of Chinese Academy of Sciences,Beijing 100049,China

dState Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences,Beijing 100193,China

A B S T R A C T

1.Introduction

Powdery mildew fungi infect nearly 10,000 plant species,including important cereal crops such as wheat and barley[1].Powdery mildew fungi are obligate biotrophs.Wheat powdery mildew, caused by Blumeria graminis f. sp. tritici (Bgt), is a worldwide fungal disease that can cause yield losses of 5%-30% in an epidemic year [2]. The Bgt fungus over-winters as chasmothecia on cereal residues where is undergoes sexual reproduction. In early summer the haploid conidiospores are asexually reproduced, easily dispersed by wind and can lead to disease epidemics[3-5].

Deployment of effective powdery mildew resistance (Pm)genes in wheat varieties is an economical and environmentally friendly approach to reduce disease losses.More than 60 powdery mildew resistance loci have been identified and/or mapped in wheat and their wild relatives; they are named from Pm1 to Pm64[6-8].Partly due to the complexity of wheat genome,only five Pm genes have been cloned,including Pm2,Pm3, Pm8, Pm21, and Pm60, and interestingly, all of them encode nucleotide-binding domain and leucine-rich repeat(NLR) immune receptors [9-14]. In addition, two multipathogen resistance genes in wheat, Pm38 (same as Lr34/Yr18/Sr57, encoding an ABC transporter) and Pm46 (same as Lr67/Yr46/Sr55, encoding a hexose transporter), have also been cloned [15,16]. Some Pm gene loci consist of multiple alleles; for example, the well characterized Pm3 gene harbors 17 alleles,each encoding a different NLR receptor [17].

Closely related wild relatives as well as domesticated wheat species have been used as sources of powdery mildew resistance in wheat improvement[18];for example,Pm6 from

Triticum timopheevii, Pm8 from rye (Secale cereale) and Pm21 from Haynaldia villasa [19-22]. T. urartu, as a nondomesticated, diploid relative and the A genome donor of polyploid wheat, represents a valuable genetic resource for wheat improvement [23-25]. A draft genome and a highquality genome sequence of T. urartu have been published[26,27].These will facilitate the isolation of disease resistance genes from this species.Gene Pm60 was recently cloned from T. urartu accession PI428309; it encodes a typical CC-NB-LRR type of NLR immune receptor [14]. In addition, three alleles were identified at the Pm60 locus, and designated as Pm60,Pm60a, and Pm60b. Interestingly, the three PM60 variants differ significantly in the LRR regions, with two LRR motifs involving 80 amino acids deleted from PM60a but inserted in PM60b, as compared to PM60 [14]. Whether there are further unidentified Pm60 alleles and how to effectively utilize these alleles in wheat breeding are still not fully addressed.

In the present study, we systematically characterized the interaction between T.urartu and Bgt at microscopic level.We screened 227 T.urartu accessions for reaction to Bgt isolate E09 and categorized different levels of powdery mildew resistance in the collection. Through a marker-assisted screen targeting the Pm60 locus we identified a non-functional Pm60a-like allele, named as Pm60a′, in multiple susceptible accessions.Based on the sequence variation between Pm60a and Pm60a′,two new molecular markers were developed to distinguish different functional Pm60 alleles from the non-functional Pm60a′. Our study reveals the complexity and diversity of powdery mildew resistance in T. urartu and provides new insights into the Pm60 locus.

2. Materials and methods

2.1. Plant materials

The 227 T. urartu accessions used in this study (Table S1)were collected from different countries in the Fertile Crescent.The information of their origins is available from the Genesys database (https://www.genesys-pgr.org/). All accessions were grown in a climate-controlled greenhouse.

2.2. Bgt inoculation and evaluation of responses to Bgt

Bgt isolate E09 was maintained on highly susceptible common wheat ‘Xuezao' in a growth cabinet at 70% relative humidity,22 °C light/20 °C darkness and 12 h light/12 h photoperiod.Inoculations was carried out when first seedling leaves were fully expanded [28]. Xuezao and T. urartu accession G1812 were used as the susceptible controls. Ten well-developed seedlings from each accession were evaluated for powdery mildew response at 7-10 days post inoculation(dpi)when the susceptible controls were fully sporulating. Infection types(IT) were classified into six classes in accordance with a previous study with IT 0, 0, 1, 2, 3, and 4 representing no visible symptoms, necrotic flecks, highly resistant, resistant,susceptible, and highly susceptible reactions, respectively[28]. They were classified into two groups, resistant (IT 0-2)and susceptible(IT 3 and 4)[28].

2.3. Histological staining of Bgt

Histological staining of Bgt structures was carried out as described previously [29,30]. In brief, Bgt-infected T. urartu leaves were collected at 0, 0.5, 4, 8, 15, 24, 30, and 48 h post inoculation (hpi), and at 3 and 5 dpi. Leaves were separately put into 10 mL centrifuge tubes filled with 6 mL of KOH(1.0 mol L-1)plus a droplet of Triton-100 and incubated at 37°C for 24 h. The leaves were then washed three times with Tris-HCl (50 mmol L-1, pH 7.5) for neutralization before staining with 20 μg mL-1WGA-FIFC solution for 24 h at 4 °C. After staining, the leaves were gently washed with double distilled H2O and examined using a Nikon A1 confocal laser-scanning microscope using a 488 nm laser beam.

2.4. Genomic DNA isolation and marker-based analysis

Total genomic DNA was extracted from T.urartu leaf tissues by following a cetyltrimethylammonium bromide (CTAB)method [31]. Each PCR mixture (20 μL) contained 50-100 ng of template DNA,0.4 μmol L-1of each of forward and reverse primers, 10 μL of 2× Taq PCR StarMix (Genstar, A012-B100).PCR amplification with marker M-Pm60 was programmed as 94°C for 5 min;35 cycles of 94°C for 35 s,57°C for 35 s,and 72 °C for 1 min 45 s, terminated after an extension at 72 °C for 10 min [14]. PCR products were separated on a 1.2%agarose gel.

2.5. Cloning and sequencing of Pm60 alleles

The fragment of Pm60a′ was amplified and sequenced with the primers of marker M-Pm60. The full-length ORFs of different Pm60 alleles were amplified using the primers of marker Pm60-ORF that was developed based on the Pm60 sequence (Table S2). PCR mixture (50 μL) contained 50-100 ng of template DNA, 0.3 μmol L-1of each of forward and reverse primers, 25 μL of 2× KAPA HiFi HotStart ReadyMix PCR(KAPABIOSYSTEMS, KK2601). PCR amplification with marker Pm60-ORF was programmed as 95 °C for 3 min; 35 cycles of 98 °C for 20 s, 61 °C for 15 s, and 72 °C for 2 min 40 s, terminated after an extension at 72 °C for 10 min. PCR products were subcloned into a vector (TA/Blunt-Zero Cloning Kit, VazymE,C601-00) by following the manufacturer's recommendations for further sequencing. Sequencing of PCR products was performed by the Ruibio BioTech Company, Beijing.

2.6. Sequence alignment and development of new markers

Sequence alignment was performed using DNAMAN software. Two independent markers (M-Pm60-S1 and M-Pm60-S2)were designed to differentiate Pm60a′ from Pm60a, and the two primers for each marker were located in the LRR regions that contained multiple SNPs and 1 aa deletion between the two allele sequences (Table S2). PCR mixtures were prepared as described above.PCR amplification with M-Pm60-S1 and MPm60-S2 was programmed as 94°C for 5 min;35 cycles of 94°C for 35 s,57°C for 35 s,and 72°C for 1 min,terminated after an extension at 72°C for 10 min.PCR products were separated in 1.2%agarose gels.

3. Results

3.1. Bgt development in T.urartu leaves

Although the reaction of T. urartu to Bgt was studied histologically, the Bgt and T. urartu interaction had not been systematically investigated at a microscopic level. We inoculated the susceptible line G1812 with Bgt isolate E09,using the resistant line PI428215 that contains a functional Pm60b allele as a control [14]. The fungal structures and intracellular growth on T. urartu leaves were visualized using wheat germ agglutinin conjugated with a fluorophore. In the compatible interaction with G1812 Bgt developed primary (PGT) and secondary (SGT) germ tubes at 0.5 and 4 hpi, respectively.The appressorial germ tube (AGT) and penetration peg (P)were formed at 8 hpi and 15 hpi, respectively; the immature haustorium(H)and mature haustorium were developed at 24 and 30 hpi,respectively.Secondary hyphae(SH)developed at 48 hpi and the epiphytic mycelium appeared at 3 days post inoculation (dpi), and the conidiophores appeared and produced large number of conidia from 3 to 5 dpi (Fig. 1, left panels).The infection process of Bgt in T.urartu was in general similar to that of Bgh in barley, nevertheless, we noticed that the formation of Bgt haustorium in T. urartu was delayed by 4-6 h compared to Bgh infection in barley[32].

In the resistant line PI428215 that harbored the Pm60b allele [14] Bgt spores generated PGT or SGT at 0.5 or 4.0 hpi,similar to G1812.A second cell instead of AGT that emerged at 8 hpi was twisted and elongated from 15 to 48 hpi.No further fungal development was evident(Fig.1,right panels).We also investigated Bgt development in resistant accessions PI428306 and PI662269 containing Pm60 and Pm60a, respectively [14].Likewise, the fungus could not complete its asexual life cycle in these lines (Fig. S1). These results indicated that the three different Pm60 alleles in PI428306, PI662269, and PI428215 triggered immunity to Bgt isolate E09.

3.2. Reaction of T. urartu accessions to Bgt isolate E09

The 227 T. urartu accessions were collected from the Fertile Crescent and neighboring countries, including 83 from Lebanon,77 from Turkey,39 from Syria(Fig.2A).Forty one percent of accessions (93/227) were resistant to Bgt E09, including 59 immune (IT = 0 or 0;), 19 highly resistant (IT = 1), and 15 moderately resistant (IT = 2) accessions. The majority of resistant accessions were from three countries, i.e. Lebanon(57 accessions,61.3%),Syria(19 accessions,20.4%)and Turkey(12 accessions, 12.9%), and only 5.4% were from other countries(Fig.2C).About 69%of the accessions from Lebanon,49% from Syria and 16% from Turkey were resistant. These results indicated that different types of disease resistance were present in the collection and that resistance was more prevalent in accessions from Lebanon and Syria than in those from Turkey.

3.3. Marker-assisted identification of Pm60 allele in T. urartu accessions

The Pm60 gene was first isolated from T. urartu accession PI428309, and two more functional alleles were further identified, i.e. Pm60a and Pm60b from PI428210 and PI428215,respectively. The three Pm60 alleles differ in length of LRR coding region (Fig. S2). To isolate more Pm60 alleles from T.urartu, we screened the 227 T. urartu accessions using the reported marker M-Pm60 (Tables S1, S2). M-Pm60 is a PCRbased molecular marker that covers the LRR regions and can differentiate the three alleles by the size of amplified products[14].

As shown in Fig. 3, the M-Pm60 marker amplified no product from the susceptible accession G1812 but amplified distinct products from resistant accessions, PI428306(1551 bp),PI428215(1791 bp),and PI662269(1311 bp).Sequencing confirmed that these were the fragments of Pm60, Pm60b,and Pm60a, respectively (Fig. S2). Unexpectedly, M-Pm60 also amplified a Pm60a-like fragment from the susceptible line PI662227 (Figs. 3A, S3). Confirmed by sequencing this Pm60alike fragment differed from the Pm60a (PI662269) by many SNPs and a 3 nt InDel in the LRR region.Therefore,this Pm60alike allele from PI662227 is most likely non-functional; we named it Pm60a′ (Figs. 3B, S4). From this preliminary screening, we concluded that marker M-Pm60 can be used to screen Pm60 alleles but it cannot distinguish Pm60a from Pm60a′.

Fig.1-Bgt development in susceptible or resistant T.urartu accessions.Conidiospores of Bgt E09 were inoculated on leaves of T. urartu accessions G1812 and PI428215,and fungal development was observed at indicated time points after staining with WGA-FIFC solution(see M&M).G1812 is a susceptible line while PI428215 contains the Pm60b allele thus is resistant to Bgt E09;conidium(Con),primary germ tube(PGT), secondary germ tube(SGT),appressorial germ tube(AGT),penetration peg(P),haustorium(H),and secondary hyphae(SH)are indicated by arrows.Bar,10 μm(0-30 hpi),20 μm(48 hpi),50 μm(3 dpi),and 100 μm(5 dpi).hpi,hours post inoculation;dpi,days post inoculation.

By using M-Pm60,we screened 52 resistant accessions and identified a Pm60-like sequence in 46 accessions and a Pm60blike sequence in six accessions (Table S3). By comparing the entire open reading frame (ORF) of Pm60-like or Pm60b-like sequence in multiple accessions,we found no difference from the published Pm60 (MF996807.1) or Pm60b (MF996808.1)sequences, respectively. We further identified a Pm60a-like allele in 15 accessions, among which six were resistant,containing the Pm60a (MF996806.1) sequence and the other nine were susceptible accessions containing the Pm60a′sequence identical to that from PI662227(Fig.3, Table S3, Fig.S3).The Pm60a′sequence was 98.52%identical to Pm60a,with 58 SNPs and a 3 nt deletion(Fig.S3).In summary,we identified the mutant allele Pm60a′ in multiple susceptible T. urartu accessions. We presumed it is a non-functional allele at the Pm60 locus.

3.4. Development of two specific markers for differentiating Pm60a alleles in T.urartu

Since the marker M-Pm60 was unable to distinguish Pm60a from Pm60a′ we developed new markers specifically for amplifying the functional Pm60a but not the Pm60a′ allele.M-Pm60-S1 and M-Pm60-S2 were designed based on the nucleotide polymorphisms and 3 nt InDel difference within the LRR coding region of Pm60a and Pm60a′(Figs.S5,S6).These two markers also covered the LRR region containing the 240 bp deletion or insertion, therefore they should have distinguished the three reported Pm60 alleles by the length of PCR products (Figs. 3, S6). First, we tested M-Pm60-S1, MPm60-S2, and M-Pm60 in a panel of T. urartu accessions containing different Pm60 alleles (Fig. 4). From resistant accessions that contained Pm60a, Pm60, or Pm60b, these markers could each amplify a product with distinct size, i.e.M-Pm60-S1 could amplify products of 591, 831, or 1071 bp; MPm60-S2 could amplify products of 516,756,or 996 bp;and MPm60 could amplify products of 1311, 1551, or 1791 bp,respectively, thus each marker was able to differentiate the three functional alleles (Fig. 4). However, in susceptible accessions containing the Pm60a′ allele, M-Pm60 also amplified a fragment whereas M-Pm60-S1 and M-Pm60-S2 amplified no products (Fig. 4). These data demonstrated that new markers M-Pm60-S1 and M-Pm60-S2 could distinguish the three functional Pm60 alleles; importantly, they could discriminate Pm60a from the non-functional Pm60a′ allele.

Fig.2- The country of origin and infection type of T. urartu accessions.(A)Origins of 227 T. urartu accessions.(B) Different infection types(IT)at 10 dpi shown by T.urartu accessions following infection with Bgt E09;IT 0(PI428306), 0;(PI428303), 1(PI428299),2(PI428300),3(PI428209),4(PI428278)(top panel).(C)Distribution and country of origion of 93 resistant accessions(IT 0-2).LBN,Lebanon;TUR,Turkey;SYR,Syria;JOR,Jordan;IRN,Iran;ARM,Armenia;IRQ,Iraq;USSR,former Union of Soviet Socialist Republics;GEO,Georgia.(D)Overall distribution of accessions with each infection type.

Fig.3- Marker-assisted screening of the Pm60 locus in T. urartu accessions and infection types of different accessions.(A)Agarose gel electrophoresis of PCR products amplified by marker M-Pm60 from five T.urartu accessions;M(DNA ladder,M121 GenStar).(B)Infection types of T. urartu accessions at 10 days post inoculation.G1812(no Pm60 allele),PI428306(Pm60),PI428215(Pm60b), PI662269(Pm60a),PI662227(Pm60a′).

Fig.4-Amplification pattern of the three Pm60 markers from different T.urartu accessions.The three Pm60 markers,M-Pm60(top panel),M-Pm60-S1(middle panel),and M-Pm60-S2(bottom panel)were used for testing T.urartu accessions containing Pm60a(1-5,PI662269,PI662251,PI428219, PI428208,01C0104213-3);Pm60a′(6-9,PI538727,PI662229,PI662274,PI662227);Pm60(10-14,PI428203,PI428215,PI428204,PI662230,PI428315); or Pm60,(15-24 PI428309,PI428310,PI538737,PI428276,PI428306, PI538740,PI428296,PI428288,PI428325,PI428324), respectively.M, DNA ladder,M121 GenStar.

To test these two new markers, we generated several segregating populations using resistant lines containing individual Pm60 alleles. Among them, three F2populations segregating for Pm60 (CITR17664/PI428186, PI538737/PI554599,and PI538751/G1812),and three F2populations segregating for Pm60b (PI428315/PI428241, PI428215/PI428274, and PI428196/PI428186). Marker Pm60-S1 was tested in these populations and identified the functional Pm60 alleles in the resistant but not the susceptible plants (Fig. S7), and marker Pm60-S2 cosegregated with Pm60-S1(Fig.4).These results suggested that markers Pm60-S1 and Pm60-S2 could be used for identifying the Pm60 alleles in segregating populations and therefore likely for marker-assisted selection in breeding programs.

4. Discussion

An analysis of the global gene co-expression network has revealed that different gene networks are associated with the two different types of Bgt disease resistance in T. urartu, i.e.immunity and HR resistance. Moreover, three candidate NBLRR genes were predicted to likely contribute to the resistance[33]. The Pm60 gene isolated from T. urartu encodes a typical CC-NB-LRR immune receptor that confers resistance associated with HR cell death [14]. Previously, different types of R genes associated with powdery mildew resistance were predicted or identified in T. urartu, including NB-LRR genes,wall-associated receptor genes and others [14,27,33,34]. Sequencing projects also revealed that the T. urartu genome contains more NB-LRR type genes than barley, rice, maize or sorghum[26,27,35].By screening 227 T.urartu accessions with Bgt isolate E09, we found at least three types of defense responses,including immune(IT=0,0;),highly resistant(IT=1)and moderate resistant(IT=2)(Fig.2B).We believe that the accessions with moderate resistance may contain different resistance genes from the NLR type. Accessions that are immune or highly resistant may contain novel NB-LRR genes.Our screen helps to better understand the Bgt resistance in T.urartu and provides a valuable resource for the identification and utilization of novel powdery mildew resistance genes.

As summarized in Table S3, different Pm60 alleles are distributed in accessions collected or originating from different countries. Among 67 accessions harboring Pm60 alleles, Pm60 is the most prevalent one and was present in accessions from LBN (29), SYR (14), and TUR (3). The other alleles were less prevalent, i.e. Pm60b was present in accessions from TUR (4) and LBN (2), Pm60a in accessions from TUR (3) and SYR (2), and Pm60a′ in accessions mainly from TUR(3)and SYR(4)(Table S3).It is noteworthy that the accessions from TUR contain all forms of Pm60 alleles, and that the functional Pm60a allele coexists with the nonfunctional Pm60a′in accessions from TUR and SYR.Further analysis is required for a better understanding of the evolution of Pm60 alleles.

It is known that some NB-LRR gene loci have been subjected to diversified selection and have evolved multiple alleles, with each allele detecting an isolate-specific AVR effector protein [36-40]. For instance, the barley Mla and the wheat Pm3 powdery mildew resistance loci harbor 30 and 17 alleles, respectively [17,39]. By contrast, three Pm60 alleles,Pm60, Pm60a and Pm60b, have so far been identified in T.urartu,each conferring similar levels of resistance against Bgt isolate E09 (Figs. 1, 3, S1). In this study, we identified 93 resistant accessions of T. urartu but no new functional Pm60 allele was found among them (Fig. 2C, Table S3). Since this collection contains a relatively large representation of accessions from different countries in the Fertile Crescent, it is tempting to speculate that only three functional alleles have been selected at the Pm60 locus. Nevertheless, it would be necessary to screen more accessions, as well as related wild diploid wheat species, and tetraploid wild wheat to identify new Pm60 alleles or homologs.

We identified allele Pm60a′ in nine susceptible T. urartu accessions. This allele shared 98.52% sequence identity with Pm60a (Fig. S3). Sequence alignment revealed that Pm60a′differed from Pm60a by 58 SNPs and one 3-nucleotide deletion (Fig. S3), resulting in an overall 37 amino acid (aa)substitutions and a 1 aa deletion in Pm60a′ (Fig. S4). Only four aa substitutions occurred in the N-terminal CC-NB domain whereas the remaining 33 aa substitutions and 1 aa deletion were all located in the LRR region (Fig. S4). While the aa substitutions in the CC-NB domains might be responsible for the loss of function of PM60a′, it is equally possible that some of the aa substitutions and/or the deletion in the LRR region of PM60a′ contribute to loss-offunction or alteration of recognition specificity,considering that a primary role of the LRR is in determining recognition specificity [41-44]. In this regard, it is surprising that deletion/insertion of two LRR motifs in PM60a/PM60b does not alter the recognition specificity to Bgt E09 [14]. We envisage that the newly isolated Pm60a′ allele can help to better understand the structural and functional properties of the PM60 immune receptors.

Based on the sequence variations in the LRR regions between Pm60a and Pm60a′ we developed molecular markers M-Pm60-S1 and M-Pm60-S2 that can differentiate the nonfunctional Pm60a′ from the functional Pm60 alleles, in addition to the three functional alleles(Fig.S5,Fig.4).These two markers should be useful for screening of new Pm60 alleles, and for marker-assisted breeding and utilizing of the Pm60 gene in common wheat.

5. Conclusions

In this study,we systematically characterized the interaction between T. urartu and Bgt at microscopic level for the first time.We screened 227 accessions of T.urartu and categorized different levels of disease resistance indicative of different resistance genes. Marker-assisted screening targeting the Pm60 locus identified a nonfunctional allele named Pm60a′.We developed two molecular markers based on sequence differences between Pm60a and Pm60a′ that can differentiate these two alleles in addition to the three functional Pm60 alleles identified in T.urartu accessions.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2019.09.007.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by the National Key Research and Development Program of China (2016YFD0100602), National Natural Science Foundation of China (31530061), the Ministry of Agriculture of China, the National GMO project(2016ZX08009-003-001),and the Innovation Fund for Graduate Students of Jiangxi Province of China (YJS2017057). We thank Chaojie Xie,China Agricultural University,for providing seeds of wheat cultivar Xuezao.