Volume 65, Issue 5 p. 734-743
Original Article
Free Access

Resistance in Australian barley (Hordeum vulgare) germplasm to the exotic pathogen Puccinia striiformis f. sp. hordei, causal agent of stripe rust

P. M. Dracatos

Corresponding Author

P. M. Dracatos

Plant Breeding Institute Cobbitty, The University of Sydney, Private Bag 4011, Narellan, Sydney, NSW, 2567 Australia

E-mail: [email protected]Search for more papers by this author
M. S. Khatkar

M. S. Khatkar

Faculty of Veterinary Science, University of Sydney, 425 Werombi Road, Camden, NSW, 2570 Australia

Search for more papers by this author
D. Singh

D. Singh

Plant Breeding Institute Cobbitty, The University of Sydney, Private Bag 4011, Narellan, Sydney, NSW, 2567 Australia

Search for more papers by this author
F. Stefanato

F. Stefanato

National Institute of Agricultural Botany, Huntingdon Road, Cambridge, CB3 0LE UK

Search for more papers by this author
R. F. Park

R. F. Park

Plant Breeding Institute Cobbitty, The University of Sydney, Private Bag 4011, Narellan, Sydney, NSW, 2567 Australia

Search for more papers by this author
L. A. Boyd

L. A. Boyd

National Institute of Agricultural Botany, Huntingdon Road, Cambridge, CB3 0LE UK

Search for more papers by this author
First published: 29 August 2015
Citations: 13

Abstract

Eighty-eight Australian and 10 international barley cultivars were assessed for resistance to the barley stripe (yellow) rust pathogen, Puccinia striiformis f. sp. hordei (Psh). All cultivars were tested for seedling resistance to two UK-derived isolates of Psh (11.01 and 83.39) that were shown to differ in virulence based on responses on 16 differential barley genotypes. The 98 barley cultivars differed substantially in stripe rust response; 45% were susceptible to Psh 11.01, 53% to Psh 83.39 and 44% to both isolates. The observed diverse infection types (ITs) suggest the presence of both known and uncharacterized resistance. However, further multipathotype tests are required for accurate gene postulation. The Yerong × Franklin (Y×F) doubled haploid (DH) population was phenotypically assessed as seedlings using both Psh isolates. Yerong and Franklin were immune and highly resistant, respectively, to both isolates used in this study. Marker-trait and QTL mapping identified a major effect on the long arm of chromosome 7H contributed by Franklin in response to all isolates. The resistance of Yerong was mapped to 113·96 and 169·38 cM on chromosome 5HL in response to Psh 11.01 and 83.39, respectively. The Psh resistance sources identified in this study can be used for further genetic analysis and introgression for varietal improvement.

Introduction

In Australia, barley (Hordeum vulgare) is the second most important cereal crop, with an estimated annual production of seven million tonnes. With a reputation for consistent, contaminant-free, high-yielding barley varieties, Australia currently has the highest malting selection rate of all barley exporting countries. However, barley is susceptible to foliar diseases, including the Puccinia rust pathogens. Currently in Australia there are two formae speciales of the stripe rust pathogen Puccinia striiformis: P. striiformis f. sp. tritici (Pst, causal agent of wheat stripe rust) and Pstriiformis f. sp. pseudo-hordei (Psph, causal agent of barley grass stripe rust). Generally, Pst does not infect barley in Australia, while Psph can cause low levels of infection on a few commercial barley varieties (Wellings et al., 2000a). A previous study estimated a 10% yield loss in susceptible barley to Psph (Wellings et al., 2000a). Barley stripe rust, caused by Puccinia striiformis f. sp. hordei (Psh), is a potential threat to the barley industry in many production areas of the world, especially Australia, where it is not currently present (Wellings, 2007; Derevnina et al., 2015).

Rust pathogens have the ability to travel large distances via wind-aided migration, to rapidly increase population size, and to mutate and acquire virulence for resistance genes, especially where cereal monocultures predominate (Watson & de Sousa, 1983; Hovmøller et al., 2002). In Australia, several incursions of new rust pathogens (e.g. Pst) and of new pathotypes of existing rust pathogens, including wheat stem rust pathogen P. graminis f. sp. tritici (Pgt) and the wheat leaf rust pathogen P. triticina (Pt), have had serious implications for Australia's cereal industries. Nevertheless, stripe rust has been slower in its global spread compared with other cereal rust pathogens such as Pgt and Pt. For example, while Pgt and Pt have been present in Australia for some 200 years, since at least European colonization, Pst was first found in Australia in 1979 and in New Zealand in 1982 (O'Brien et al., 1980; Wellings, 2007). Psh has also been slow to migrate into new barley-growing regions. Predominately of western European, southern Asian and eastern African descent, Psh was first identified in South America near Bogota, Colombia in 1975. Following this, it spread quickly south into most barley-growing regions of South America (Argentina, Bolivia, Chile, Ecuador and Peru) where it has caused significant yield losses (Dubin & Stubbs, 1986). Stripe rust has caused yield losses in South America, southern Asia, Africa and the USA. Race analysis has detected over 70 different pathotypes of Psh in the USA, some of which are virulent on many of the known sources of resistance to the disease (X. Chen, Washington State University, USA, personal communication).

Psh was first reported in Mexico in 1987, and samples collected from Los Valles Altos were determined to be race PSH24 by Dr R. Stubbs (University of Wageningen, the Netherlands, personal communication). Psh was not detected in the USA until 1991, where it has since been reported to cause yield losses of up to 70% (Brown et al., 2003). Preliminary studies determined that 70% of Australian barley varieties were susceptible when screened in CIMMYT nurseries in Mexico with PSH24 race, suggesting that, if it were introduced into Australia, Psh would cause significant production losses (Wellings et al., 2000b; Wellings, 2007). Pre-emptive breeding strategies for resistance to Psh are therefore challenging, given the absence of the disease (Cakir et al., 2003; Wellings, 2007). This study was undertaken to determine the vulnerability of Australian barley germplasm to two European-derived Psh isolates. It was aimed to identify and map novel resistance sources in Australian barley germplasm that would enable pre-emptive breeding for Psh resistance in Australian barley in preparation for future incursions.

Materials and methods

Plant materials

Australian barley germplasm was sourced from a collection at the Plant Breeding Institute, University of Sydney in 2013 and pedigree information for each cultivar can be found in Table 1. A subset (70 lines) of the doubled haploid (DH) population of 92 lines derived from a cross between the Australian cultivars Yerong (six-row feed barley) and Franklin (two-row malting barley) (Y×F) was also used for genetic analysis of resistance to Psh.

Table 1. Pedigree information and infection type response of 88 Australian and 10 exotic barley cultivars phenotypically assessed with two UK-derived isolates of the barley stripe rust pathogen Puccinia striiformis f. sp. hordei (Psh)
Cultivar Psh isolate Groupa Pedigree
83.39 11.01
Arapiles 2+3Cb 2+3 D Noyep/Proctor//CI3576/Union/4/Kenia/3/Research/2/Noyep/Proctor/5/Domen
Barque 2CN 2+C D Triumph/Galleon
Baudin 2+C/23C ;0;/1C E Stirling/Franklin
Beecherc 33+ 33+ A Atlas/Vaughan
Binalong 23C 1 = C/2 E Blenheim//Skiff/O'Connor
Bandulla 3+ 2++C/3 B Unknown
Bowmanc 3+ 3+ A Klages//Fergus/Nordic/3/ND-1156/4/Hector
Buloke 2–C 2–C D Franklin/VB9104//VB9104
Bussell 12C 1–C/2C D Prior/Ymer
Capstan 33+ 1+C/3 B Waveney/WI2875//Chariot/Chebec
Caskc 3/3+4 1+CN B SCRI-8313/Fleet//Regatta
Chebec 33+ 12+C B (USA) Orge Matin/2 Clipper(86)//Schooner (Aust.) (Orge Martin Clipper#2)/86 Schooner
Chevronc 3+ 3 A (S)LV-Switzerland
Chieftain ;C ;C D Brittania/Prisma
Clipper 33+ 2C/23 B Proctor/Prior A
Corvette 2C/3 2–C/3 D Bonus/CI3576
Cosmicc 1–C 1 = C D Unknown
Cowabbie 2C/3–C 1+CC E (AB6/Franklin//Franklin-early)/3/(Rubin/Skiff-early)
Cutter 3+ 3+ A Proctor/Prior A
Dash 3+ 2++3 A Chad/Joline//Cask
Dhow 2++3 2+/2+C D WI2808//Skiff/Haruna Nijo 9
Dictator 1+C 1–C D Reselection of USDA accession CI2204
Doolup 2++3C 1+CN E (XBVT210)/3/(B6729)Prior/Lenta(75S:323)/(MndS,74S:314)Dampier//(A14)Prior/Ymer/
Empress 3+ 4 A H1006·3/HE902
Fitzgerald 1+C/3C 1C/2C D Onslow/Tas 85-466
Fitzroy 3/2+C 2+C/3C D WI2808/Alexis
Flagship ;C/3 4 A Chieftan/Barque//Manley/VB9104
Fleet 2+C 23 D Mundah/Keel//Barque
Franklin 0;C ;C E Shannon/Triumph
Gairdner 2C 1C E Onslow/Tas 83-587
Galaxy ;/3+ ;NN E 24719DB/Robin SIB
Galleon 3+C 3-C A Clipper/Hiproly//3 Proctor/CI3576
Gilbert 3+ 2+C B Reselection of Mx(Q21517)
Grimmett 1CN 1CN D Bussel/Zepher
Grout 3+ 3+ A Cameo/Arupo 31-04
Gusc 3+4 3+4 A Unknown
Hamelin 3– 3+ A Stirling/Harrington
Hannan 3++ 3+ A WABAR2023//Windich/Morex
Harringtonc 3+ 1+C B Klages/3/Gazelle/Betzes//Centennial
Haruna Nijoc 4 4 A Unknown
Heartlandc 3+ 3+ A Klondike/BT-146
Henley 1CC 1++C D Unknown
Kaputar 3++C 3+ A 5604/1025/3/Emir/Shabet//CM67/4/F3 Bulk Hip
Keel 3+ 3+ A CPI18197/Clipper//WI2645
Ketch 3+ 3+ A Noyep/Lenta
Lara 12C 12C D Research/Lenta
Lindwall ;C ;C/2C D Triumph/Grimmett
Lockyer 3+C 3+ A Tantangara/VB9104
Lofty Nijo 3C 3+ A Unknown
Mackay 3+ 3+ A Cameo/Koru
Malebo ;C ;C D Selection from CPI11083(Palladium WWB 18)
Maritime ;CN 3++CN E Unknown
Milloy 3+ 3+ A Unknown
Molloy 3+ 3+ A Golden Promise/WI2395(WARI2-38)/4/(72S:267)XBVT210/3/(66S08-4)Atlas57//(A14)Prior/
Moondyne 3+C 3+C A Dampier//(A14)Prior/Ymer/3/Kristina/(70S20-20)/4/(73S13)Clipper/Tenn-65-117
Morexc 3+ 3+ A Cree/Bonanza
Morrell 3 3+ A Wum-221/P-23822/5/Forrest/4/Psaknon/Dampier//M-19/3/Zephyr
Mundah 3+ 1+C B Unknown
Noyep 2+3C 2+CN D (S) Prior/(S) Chevalier
O'Connor 2++N/3 3N D Proctor/CI3576(WI2231)/3/(XBVT212)Atlas 57//(A14)Prior/Ymer
Onslow 1CN/2C 1C E Forrest/Aapo
Osprey ;CN ;CN D 24719-DB/Robin
Oxford 3++ 2-C B Unknown
Pacific Ranger 3+ 2+3 A PC11/AC Rosser [PC11 is a CIMMYT selection with resistance to stem rust race QCCJ]
Parwan 3+ 3+ A Plumage Archer/Prior//Lenta/3/Research/Lenta
Picola 3+ 3+ A 75031/Elgina(75031 =  Noyep/Prior//CI3576/Union/Kenia/4/Research/Noyep/Prior)
Prior 3+ 3+ A Selected from either Archer or Chevalier
Quasar 3+ 3 A Chalice/NFC breeding line
Quickstar 3+ 3C A Unknown
Research 3C 3C A Plumage-Archer/Prior
Roe 2++C 3C D Doolup//Windich/Morex
Schooner 3+ 3+ A Proctor/Prior A//Proctor/CI3576
Scope 33C 3C A Unknown
Shannon 2++C 2 = C E Proctor 4/Ethiopian line CI3208-1
Shepherd 3 2 B Baronesse/Cheri
Skiff 3+C 3 A Abed Deba/3/Proctor/CI-3576//CPI-18197/Beka/4/Clipper/Diamant/Proctor/CI-3576
Sloop 3 3 A RL1577/84/Schooner
Sloop SA 3 3 A CCN6-3/Sloop3
Starmalt ;C ;CN E Unknown
Stirling 3C 3C A Dampier/Prior/Ymer/3/Piroline
Tallon 0; 0; D Triumph/Grimmett
Tantangara 3C 0; B AB6/Skiff (AB6 is Hordeum spontaneum CPI71283/4*Clipper)
Tilga ;CNN ;CNN D Forrest/Cantala
Torrens 1C/23C 1+C D Galleon/CIMMYT 42002
Tulla 3C 3+C A Skiff/FM437
Ulandra 3++C 3+C A Warboys/Alpha
Unicorn 2++C 2+3C D 54C25/51C38
Urambie 23C 1C E Unknown
Vadac 4 4 A Hordeum laevigatum/Svalofs-Gull
Vertess 1C 0;C D Franklin/Cooper
Vlamingh 3C 2C,;CN D WABAR0570/TR118
Weeah 3C 3C A Prior/Research
Westminister ;C/1C 1C D Unknown
Windich 3+ 3+ A Atlas 57//(A16)Prior/Ymer(68S17-75)/3(B6729)Prior/Lenta//Noyep/Lenta
Wyalong 2C,3 2 D Schooner/Stirling
Yagan 2++C 3–C D Unknown, tested as IB/286, WUM143
Yambla ;C 3 C Skiff/FM437
Yerong 0;C 0; D M22/Malebo
  • a Barley cultivars were grouped based on infection type in response to Psh isolates 83.39 and 11.01. Group A had high ITs to both isolates, group B had high ITs to 83.39 and low ITs to 11.01, group C had a high IT to 11.01 and low IT to 83.39, group D showed the same low IT response to both isolates and group E were resistant to both isolates with differing low ITs possibly due to the presence of additional resistance genes.
  • b Disease response was assessed using a 0–4 infection type (IT) scale as described for P. striiformis f. sp. tritici by McIntosh et al. (1995), where ‘0’ = no visible symptoms, ‘;’ = necrotic flecks, ‘;N’ = necrotic areas without sporulation, ‘1’ = necrotic and chlorotic areas with restricted sporulation; ‘2’ = moderate sporulation with necrosis and chlorosis, ‘3’ = sporulation with chlorosis, ‘4’ = abundant sporulation without chlorosis. Variations of the ITs were indicated by use of ‘−’ (less than average for the class), ‘+’ (more than average for the class), ‘C’ (chlorosis) and/or ‘N’ (necrosis). Where two predominant ITs are observed both ITs are recorded and separated with a comma (e.g. IT=‘1,1+’ and IT=‘3,3+’). Infection types of ‘3’ or higher were considered to indicate a compatible response (susceptibility in the host).
  • c Exotic barley cultivars from Canada (Harrington and Heartland), Netherlands (Vada), USA (Beecher, Bowman, Chevron, Gus, Morex), UK (Cask and Corvette).

Pathogen cultures

Rust pathogen cultures were sourced from the Psh survey collection maintained at the National Institute of Agricultural Botany (NIAB), Cambridge, UK. Isolate nomenclature is defined first by the year of collection, followed by sample number for that particular year e.g. 83.39.

Phenotypic analysis of resistance to Psh

Seedlings were grown in 77-well plastic trays containing a peat-based potting mix. Differential lines were sown as clumps (six seeds per line). Plants were grown in a Conviron growth cabinet with a day cycle of 16 h at 17°C (light intensity > 300 μE m−2) and 8 h at 14°C (in darkness), being watered every 2 days. Assessment of resistance to Psh in both Australian barley varieties and the Y×F DH population was performed at the seedling stage using two UK Psh isolates, 83.39 and 11.01. Pathotypes were discriminated based on a collection of 16 barley lines assembled by Dr Colin Wellings (University of Sydney, Australia).

Inoculation procedures

Inoculations were performed on 12-day-old seedlings. Urediniospores, stored in lyophilized 50 mg ampoules, were mixed with talc (1:10 ratio) and deposited onto plants using a compressor in a sealed glass inoculation chamber. To prevent contamination between successive inoculations, the spray equipment was washed in 96% ethanol that was then ignited to burn off any ethanol residue and rinsed in running tap water. Incubation following inoculation of barley seedlings was performed in a dew chamber, essentially as described by Dracatos et al. (2010), to enable the urediniospores to germinate and allow rust infection to take place. After incubation, the seedling trays were removed from the dew chambers and placed back in the Conviron growth cabinet using the normal growth conditions specified above.

Disease assessment

For seedling assessment, disease response was assessed 12 days after inoculation using a modified 0–4 scale, as described by McIntosh et al. (1995). Variations of the infection types (ITs) were indicated by the use of ‘–’ (less than average for the class), ‘+’ (more than average for the class), ‘C’ (chlorosis) and ‘N’ (necrosis). A comma separating different ITs was used to indicate heterogeneity within a given test host genotype. When two different ITs were observed on a single leaf, they were written together without a comma. For inheritance studies, the ITs for each genotype of the Y×F DH population were converted to binary features, whereby ITs < 3 were deemed resistant and ITs >3 were susceptible. The presence of chlorosis or necrosis were incorporated into the quantitative (0–4) score described by McIntosh et al. (1995), for example ITs of 1C = 1 and 2N = 2, while modifiers to the IT such as ‘+’ and ‘−’ were incorporated by the addition or subtraction of 0·25 (i.e. ‘+’) and 0·5 (i.e. ‘++’). DH barley lines that exhibited ITs ‘0;C’ and ‘;’ scored as a 0·5.

DArT-Seq genotyping and genetic map construction in the Y×F DH population

Genomic DNA was extracted from the leaf tissue of a single plant from 92 different DH lines and the parental genotypes of the Y×F population using the CTAB method, essentially as described by Fulton et al. (1995), and subsequently diluted to 100 ng μL−1. The Y×F DH population was genotyped using the DArT-Seq platform commercialized by Diversity Arrays Technology (Canberra, Australia). The DArT-Seq presence–absence markers (PAMs) were subsequently curated as described by Li et al. (2015). The physical positions of all DArT sequences were determined based on the Hordeum vulgare ‘Bowman’ genome assembly and consensus map (Mayer et al., 2012). Linkage map construction for the Y×F DH population was performed on a fee-for-service basis by Diversity Arrays Technology Pty. Ltd using purpose-built software, as described by Li et al. (2015).

Marker-trait association (MTA) analysis for resistance to Psh

The MTA analysis was conducted on the Y×F population using DArT-Seq marker genotypes and phenotypic scores for resistance to each Psh pathotype. A realized additive relationship matrix (K) (Endelman & Jannick, 2012) was computed to estimate molecular kinship among all DH lines. The association between a DArT-Seq marker and each resistance phenotype (Psh pathotype) was tested with single marker regression while adjusting for line similarity using the following linear mixed model:
urn:x-wiley:00320862:media:ppa12448:ppa12448-math-0001
where y is the vector of phenotypic value (trait), X is incidence matrix incorporating mean and the DArT-Seq marker genotype; β vector representing coefficients of the fixed effect, Z is an incidence matrix mapping phenotype records to the lines, u is a vector of polygenic genetic effects. The model was fitted using ASReml (Gilmour et al., 2009) with quantitative and binary resistance score traits fitted using family argument in ASReml-R as Gaussian and binary respectively. Genomewide false discovery rate was computed using the q-value package of R/bioconductor (http://www.bioconductor.org/). The –log10 of P values were plotted against the positions of each DArT-Seq marker on the physical Bowman genome assembly (Mayer et al., 2012) by means of a genomewide (Manhattan) plot.

QTL analysis for resistance to Psh

QTL analysis was performed on a subset of 70 of the total 92 Y×F DH lines selected for DArT-Seq genotypic analysis and subsequent linkage map construction (Dracatos et al., 2014). DArT-Seq marker loci were selected for QTL analysis every 5 cM to provide even coverage of the genome (Fig. S1). Single marker regression (SMR) analysis was used initially to identify significant associations with selected genetic markers. Composite interval mapping (CIM) methods were used to identify and confirm the presence of QTL using Windows QTL Cartographer v. 2.5 (Wang et al., 2012). The maximum logarithm of the odds ratio (LOD score) of association between the genotype and trait data was calculated for CIM, and QTL location predictions were accepted for CIM for values greater than a threshold value of 2·5. Permutation analysis (1000 iterations) was used to establish an experiment-wise significance value at the 0·05 confidence level, defined as a minimum LOD threshold for each trait in CIM (Churchill & Doerge, 1994; Doerge & Churchill, 1996).

Results

Pathotypic discrimination using differential lines for Psh resistance

Sixteen barley genotypes, including 9 out of the 12 USA Psh differential barley cultivars, were used to assess the virulence of the two UK Psh isolates (83.39 and 11.01; Table 2). Barley differentials Bigo (Rps1.b), Berac, Keg and Varunda (rpsVa1, rpsVa2) differentiated 83.39 and 11.01 as different pathotypes (Table 2). Psh 83.39 was virulent on plants with resistance genes Rps4, rpsAst, Rps1b, rpsTr1, rpsTr2, rpsVa1 and rpsVa2 as well as other genotypes including Berac, Biosaline-19, Fong Tien, Keg, Sultan and the susceptible control Topper (no resistance genes). Psh 11.01 was a more avirulent isolate in comparison to 83.39; however, virulence was identified against Rps4, rpsAst, rpsTr1, rpsTr2, and on Biosaline-19, Fong Tien and Topper. The main difference between the two pathotypes was avirulence/virulence on Bigo (Rps1.b) and Varunda (rpsVa1 and rpsVa2), and both Berac and Keg, which carry uncharacterized potential resistance for Psh 11.01 (Table 2). Both isolates were virulent on Astrix, Atem, Biosaline-19, Fong Tien and Topper, and avirulent on Agio, Emir, Triumph and Trumf. Different low ITs observed on Emir may indicate that both 11.01 and 83.39 were probably avirulent to different resistance genes in Emir (either rpsEm1 or rpsEm2). However, 83.39 generated an intermediate IT on Emir and may not confer as strong an avirulent reaction to rpsEm1, or alternatively could indicate the presence of an unidentified R-gene–avirulence interaction. Atem, Biosaline-19 and Fong Tien were more susceptible than Topper to both UK isolates used in this study and therefore Biosaline-19 was used as a susceptible control for experiments with the UK-derived Psh isolates on the Y×F DH population.

Table 2. Seedling response of selected barley genotypes to two UK-derived pathotypes of Puccinia striiformis f. sp. hordei (Psh)
Differential line Gene Psh 83.39 Psh 11.01
Topper Nil 3+ S 2+ S
Emir rpsEm1, rpsEm2 2+C R ;CN R
Astrix Rps4, rpsAst 3+ S 3+ S
Trumf rpsTr1, rpsTr2 1C R ;C R
Mazurka Rps1.C No data 1C R
Bigo Rps1.b 3– S 1C R
Agio Not known 2CN R 2CN R
Cambrinus Rps4 No data 2+/3+ S
Sultan Not known 3– S No data
Berac Not known 3+ S 2C R
Keg Not known 3+C S 2C R
Varunda rpsVa1, rpsVa2 3C S ;C R
Atem rpsTr1, rpsTr2 3+ S 33C S
Triumph Not known ;1C/2C R 1C/3 R
Biosaline-19 Not known 3+4 S 4 S
Fong Tien Not known 3 S 3+ S
Yerong RpsYer1, RpsYer2 0; R 0; R
Franklin RpsFra 0;C R ;C R
  • S, susceptible; R, resistant.
  • Disease response was assessed using a 0–4 infection type (IT) scale as described for P. striiformis f. sp. tritici by McIntosh et al. (1995), where ‘0’= no visible symptoms, ‘;’= necrotic flecks, ‘;N’= necrotic areas without sporulation, ‘1’= necrotic and chlorotic areas with restricted sporulation; ‘2’= moderate sporulation with necrosis and chlorosis, ‘3’= sporulation with chlorosis, ‘4’= abundant sporulation without chlorosis. Variations of the ITs were indicated by use of ‘−’ (less than average for the class), ‘+’ (more than average for the class), ‘C’ (chlorosis) and/or ‘N’ (necrosis). Where two predominant ITs are observed both ITs are recorded and separated with a comma (e.g. IT=‘1,1 + ’ and IT=‘3,3 + ’). Infection types of ‘3’ or higher were = considered to indicate a compatible response (susceptibility in the host).

Resistance to Psh in Australian barley germplasm

Eighty-eight Australian and 10 exotic barley varieties (Table 1) were assessed phenotypically at the seedling growth stage with Psh 83.39 and 11.01 (Table 1). The cultivars were separated into five groups (A–E) based on IT responses in independent seedling tests with each isolate. Group A comprised 43 (44%) cultivars (7 exotic and 36 Australian cultivars), which displayed high ITs of 3C to 4 with both pathotypes (Table 1). Group B consisted of 11 (11%) cultivars that had high ITs to Psh 83.39 (ITs >3) and lower ITs (ranging from 0; to 3) to Psh 11.01 (Table 1). Group C consisted of one (1%) cultivar (Yambla) that had a high IT to 11.01 (IT >3) and a low IT to 83.39. Group D consisted of 31 (32%) cultivars that showed the same low IT response to both isolates (Table 1). Group E consisted of 12 (12%) cultivars that had different low ITs in response to both isolates, possibly due to the presence of additional resistance genes.

Nearly half of the Australian barley cultivars tested were susceptible (IT greater or equal to 3) to Psh 83.39 and/or 11.01 (53 and 44%, respectively), while 60% of the exotic cultivars tested were highly susceptible to both isolates (Table 1). A range of resistance ITs were observed in Australian barley germplasm, which suggests the presence of different resistance genes (Fig. 1). Fewer cultivars were more resistant to 83.39 than 11.01, consistent with the wider virulence of 83.39 than 11.01 on the differential tester lines examined. Only one cultivar (Yambla) was resistant to 83.39 and susceptible to 11.01, whilst 10 cultivars (Bandulla, Capstan, Cask, Chebec, Clipper, Gilbert, Mundah, Shepherd, Tantangara and Oxford) were resistant to 11.01 and susceptible to 83.39. Five of the group B cultivars (Capstan, Cask, Chebec, Harrington and Mundah) had similar ITs, suggesting that they may carry the same resistance gene (Table 1).

Details are in the caption following the image
Seedling leaves of Australian barley cultivars (left to right): Tilga, Galaxy, Baudin, Cosmic and the susceptible control Biosaline-19 infected with Puccinia striiformis f. sp. hordei isolates (a) 83.39 and (b) 11.01.

The cultivars Baudin, Chiefton, Galaxy, Grimmet, Malebo, Osprey, Starmalt, Tallon and Yerong were highly resistant to one or both pathotypes (Table 1; Fig. 1). Whilst still resistant, Maritime, O'Connor and Tilga had higher ITs, and, in the case of O'Connor, low levels of sporulation (Table 1; Fig. 1). A subset of four Australian barley cultivars (Baudin, Cosmic, Galaxy and Tilga) were compared for their contrasting resistance phenotypes (ITs ranging from 00; to 3+) in response to both isolates (Fig. 1). Tilga gave a distinct IT compared to the other Australian barley cultivars, with resistance against both UK-derived isolates (Fig. 1). Galaxy was immune to 83.39, but displayed both chlorosis and necrosis towards 11.01. In contrast, Baudin was immune to 11.01, but displayed an intermediate IT (2–C) to 83.39. These lines were postulated to carry more than one resistance gene to Psh. The susceptible barley research line Biosaline-19 was used as a control and in all experiments was highly susceptible (IT 3+ to 4; Fig. 1).

Inheritance of Psh resistance in the Y×F DH population

A previously developed DH barley population derived from Yerong and Franklin (Y×F) was assessed at the seedling stage with Psh 11.01 and 83.39. Both Yerong (IT 0; and 0;C) and Franklin (0;C and ;C) were resistant in response to the UK Psh isolates; however, subtle variations in ITs were observed between isolates (Table 1). Segregation was observed in the progeny of the Y×F population, where approximately 25% of the DH lines were highly susceptible compared to the highly resistant phenotypes observed in the parents towards both isolates (Fig. 2a; Table S1). The Y×F DH population segregated in a 3 resistant (R): 1 susceptible (S) ratio in seedling tests with Psh 11.01 (observed segregation 48R: 21S; χ2 = 1.08, > 0·2) and Psh 83.39 (observed segregation 54R: 15S; χ2 = 0·39, > 0·5). These segregation patterns, using binary phenotypic data suggested the presence of two resistance genes, one from each parent, effective against each UK isolate. As observed for the chi-squared analysis of binary phenotypic data, the frequency distributions of quantitative infection scores in the Y×F DH population suggests that resistance in response to both Psh isolates was polygenically inherited (Fig. 2b).

Details are in the caption following the image
(a) Seedling leaves of the parental genotypes of the Yerong × Franklin (Y×F) doubled haploid (DH) mapping population and selected DH lines inoculated with Puccinia striiformis f. sp. hordei isolate 11.01 (left to right): Yerong, Franklin, Y×F5, Y×F21 and Y×F16, and the susceptible barley selection Biosaline-19. (b) Frequency distribution of seedling quantitative infections scores in response to isolates 11.01 (black) and 83.39 (diagonal stripes).

Mapping resistance to Psh in the Y×F population

Both QTL and marker-trait analysis (MTA) was performed in the Y×F DH population using 8564 DArT-Seq markers and phenotypic data derived from independent seedling tests with Psh 11.01 and 83.39. Phenotypic infection responses were scored both as quantitative characters and binary features. The MTA results for both scoring methods largely overlapped; however, for Psh 83.39, although there were significant (< 0·001) associations, the q-values exceeding 0·2 suggested that, for binary scores, the sample size was too small to accurately predict marker-trait associations without false discovery (data not shown).

Both forms of analysis using quantitative phenotypic data determined that DArT-Seq marker alleles contributed by the Franklin parent on the long arm of chromosome 7H (7HL, 120·4–126·42 cM) were significantly (< 0·0001) associated with seedling resistance in response to both isolates (Tables 3, 4; Fig. 3). This suggests that the resistance in Franklin is effective against both Psh isolates. Using MTA analysis, a second significant marker association contributed by Franklin was identified on chromosome 2HL (94·87 cM) in response to both isolates. The resistance in Yerong in response to both Psh 83.39 and 11.01 was significantly associated with DArT-Seq markers on chromosome 5HL at 169·38 cM and 113·96 cM, respectively, which probably represent independent loci (Fig. 3; Tables 3, 4). Both SMR and CIM analysis also identified QTL contributed by Yerong on chromosome 5HL at the same positions in response to both isolates. Weak marker trait associations contributed by the Yerong parent were also observed on chromosome 1H (103·82 cM) in response to 11.01 (Fig. 3a; Table 3).

Table 3. Marker trait analysis in the Yerong × Franklin doubled haploid population showing the most significantly associated DArT-Seq markers with seedling responses to Puccinia striiformis f. sp. hordei (Psh) isolates 83.39 and 11.01
Psh isolate Marker Chromosome Position (cM) P Effect gwas scorea q-value Parent
83.39 3259480¦F¦0-20:A>G-20:A>G 2H 94·87 0·00328 0·57 2·48 0·2224 Franklin
3396692¦F¦0 5H 169·38 0·00199 0·54 2·70 0·1684 Yerong
3267165¦F¦0 7H 126·42 3·05E−06 0·74 5·52 0·0260 Franklin
11.01 3263737¦F¦0 1H 103·82 0·00016 −0·73 3·80 0·0081 Yerong
3259480¦F¦0-20:A>G-20:A>G 2H 94·87 4·81E−06 0·97 5·32 0·0005 Franklin
3274035¦F¦0 5H 113·96 0·00127 −0·71 2·89 0·0547 Yerong
3266011¦F¦0 7H 120·4 6·98E−12 1·13 11·16 0·0000 Franklin
  • a gwas = –log10(P value).
Table 4. Summary of composite interval mapping (CIM) analysis for barley stripe rust resistance in the Yerong × Franklin doubled haploid population
Isolate Chromosome SMR < 0·01a (cM) Most closely linked marker CIM Parent
Maximum LOD score Position (cM) A b R c
11.01 2H 63·53 3258146¦F¦0 2·15 64·64 −0·42 0·067 Franklin
5H 112·4–124·1 3274035¦F¦0 2·75 115·41 0·48 0·093 Yerong
7H 126·56–138·22 3397945¦F¦0 6·01 129·36 −0·74 0·244 Franklin
83.39 4H 37·96 3398823¦F¦0-28:T>C-28:T>C 2·94 36·98 0·43 0·103 Yerong
5H 165·97 3396934¦F¦0 3·35 160·43 0·48 0·129 Yerong
7H 126·56–132·22 3397945¦F¦0 3·004 127·36 −0·43 0·1108 Franklin
  • a Position of significant (< 0·01) association identified by single marker regression (SMR) analysis.
  • b Additive effect of substituting alternative marker alleles at marker locus.
  • c Proportion of variance explained by QTL.
Details are in the caption following the image
Marker trait association analysis scans for resistance to Puccinia striiformis f. sp. hordei isolates (a) 11.01 and (b) 83.39 in the Yerong × Franklin doubled haploid population screened with 8564 DArT-Seq markers. The vertical axis represents –log10 of the P value of each marker trait association. Marker traits with values above the minimum threshold of 2 (= 0·03) can be considered significantly associated. Different colours were used to differentiate between chromosomes (1H to 7H).

Discussion

Unlike most cereal rust pathosystems, research into the physiological specialization and race structure of Psh is poorly defined, primarily due to the lack of characterized genetic stocks for resistance in barley. Researchers in the USA have used a set of 12 barley differential lines, while in Europe other differentials with uncharacterized resistance genes are used, including Agio, Berac, Keg, Sultan and Triumph. In this study, two Psh isolates collected in the UK were identified as different pathotypes using a collection of 16 barley genotypes. More virulence factors were identified in Psh 83.39 compared to 11.01 using these 16 barley differential genotypes. The broader virulence of 83.39 was also reflected in its ability to render a marginally higher number of Australian barley cultivars susceptible, when compared to 11.01. While only one Australian cultivar (Yambla) was susceptible to 11.01 but resistant to 83.39, 10 were susceptible to 83.39 and yet resistant to 11.01 and therefore may carry Rps1b or the Varunda resistance (rpsVa1, rpsVa2). Preliminary field data suggests that approximately 70% of Australian barley cultivars are susceptible to the CIMMYT Psh race PSH24 as adult plants (C. Wellings, University of Sydney, Australia, personal communication) compared to <50% observed in this seedling study, indicating that PSH24 was more virulent than the UK-derived Psh isolates.

In contrast to the wheat stripe rust pathosystem, relatively few studies have effectively characterized and designated resistance genes to Psh in international barley collections. Only two dominant and three recessive genes conferring resistance to Psh have been previously described, whilst QTL have been mapped to all barley chromosomes (1H–7H) (Nover & Scholz, 1969; Bakshi & Luthra, 1971; Lehmann et al., 1975; Chen et al., 1994; Toojinda et al., 2000; Castro et al., 2002a,b, 2003). In the present study, nearly half of the Australian barley cultivars phenotypically assessed as seedlings were resistant to both UK isolates; therefore, it is probable that at least some of the cultivars carry genes that confer uncharacterized resistance to a wide array of Psh pathotypes. Australian barley cultivars Yerong and Franklin were both highly resistant at the seedling stage in response to both UK isolates. Therefore, to map the observed resistance in both parents the Y×F DH population was assessed at the seedling stage with both isolates. The identification of highly susceptible (IT 4) DH lines indicated the presence of different seedling resistance in both parents. This mapping result was in agreement with the chi-squared analysis. However, while in most cases the genetic analysis of the phenotypic data fitted a two-gene model, there were often additional MT associations and QTLs detected in response to both isolates.

MTA and QTL analysis identified two distinct QTLs on chromosome 5HL contributed by the Yerong parent that were tentatively designated RpsYer1 and RpsYer2 in response to Psh 11.01 and 83.39, respectively. This suggests the presence of independent pathotype-specific resistance loci (RpsYer1 and RpsYer2) on chromosome 5HL in response to each isolate used in this study. Conversely MTA and QTL mapping results suggest that the same major effect on chromosome 7HL from Franklin, tentatively designated RpsFra, confers resistance to both Psh isolates used in the study. A gene for qualitative Psh resistance on 7HL (Rpsx), effective against four North American Psh pathotypes (PSH-13, PSH-14, PSH-21 and PSH-31) was reported by Castro et al. (2003). CIM results in this study determined that RpsFra on 7HL accounted for 24·4% phenotypic variation in response to Psh 11.01. Previous studies have also reported a very strong reduction in disease severity in lines carrying the Rpsx resistance allele (Castro et al., 2003). This suggests that if RpsFra has the same specificity as Rpsx, it may be effective against both North American and European Psh pathotypes.

Two recent studies reported a major seedling resistance QTL for Psph on 7HL in the same position as the Psh QTL reported here. In the first study, Golegaonkar et al. (2013) reported on the mapping and genetic characterization of a recessive seedling resistance gene on 7HL derived from an Algerian landrace Sahara3771. In a second study, Derevnina et al. (2015) mapped two QTL for resistance to Psph in the Y×F DH population on 5HL and 7HL. According to Sui et al. (2010) a single dominant gene (YrpstY1), conferring resistance to the non-adapted Pst pathotype CYR32 in the Chinese barley line Y12, also mapped to the long arm of 7H, 2 cM away from the QTL identified by Derevnina et al. (2015). This suggests that these loci on chromosome 7HL are probably the same gene, or that they are very closely linked genes effective to different formae speciales of P. striiformis. Furthermore, adult plant screening with Psh race PSH24 in Mexico determined that Franklin was highly resistant (5R–10MR), while Yerong was moderately susceptible (70–80S), over multiple seasons (D. Singh, unpublished data). Derevnina et al. (2015) reported an adult plant resistance (APR) gene for Psph on 5HS contributed by Franklin and mapped in the Y×F population. As in the case of the large effect QTL identified in both Yerong and Franklin as seedlings, this APR may also be effective to Psh. Taken together, the data presented in both previous and the present study suggest that, due to Australian quarantine regulations and a ban on importing exotic rust cultures, Psph could be used for barley pre-emptive breeding purposes if the identified resistance is closely linked or the same. However, this approach must be taken with caution, given that barley is an intermediate host to Psph and only 10% of barley cultivars are susceptible (Golegaonkar et al., 2013; Derevnina et al., 2015). This suggests that although some cultivars may share resistance effective to both Psh and Psph, many more cultivars are resistant to Psph than to Psh.

The absence of the pathogen in Australia has previously complicated all stages of the pre-emptive breeding approach including: phenotypic assessment, identification, selection and integration of resistance into high-yielding cultivars adapted to Australian conditions. Despite over a decade of field assessment in Mexico using Psh race PSH24, in the absence of multipathotypic analysis, the presence of APR cannot be differentiated from seedling responses. This study represents the first attempt to assess seedling resistance in Australian barley cultivars in response to representative European isolates. Further studies and pre-emptive breeding strategies will therefore require a coordinated approach involving multipathotype testing in international screening facilities using pathotypes with a broad virulence spectrum, including race PSH24. This would then permit gene postulation and the retrospective use of historical field data sets to determine the presence of APR in Australian barley cultivars. Once both seedling and APR genes have been postulated in Australian barley cultivars, select mapping populations will be phenotypically assessed to characterize the inheritance and to map and develop molecular markers for assisted selection and varietal improvement.

Acknowledgements

The authors acknowledge Dr Sarah Holdgate and Amelia Hubbard for the provision of viable barley stripe rust pathogen cultures. The authors thank the Australian Government for funding the project and Dr Dracatos as an Endeavour Executive Award Fellow. The authors thank Dr Andrzej Kilian at Diversity Arrays Technology for the genotypic analysis of the Y×F DH barley population and linkage map construction. The authors thank Dr Colin Wellings for the assembly and provision of seed for the 16 differential barley lines used in this study. The authors have no conflict of interest in regard to any data presented in this manuscript.