Introduction

Wheat (Triticum aestivum) is Australia’s most important crop, covering an area of nearly 13 M ha, and producing on average 20–25 M t annually (Pink 2008) with a value of AU$4.7 billion (Murray and Brennan 2009). The root-lesion nematode Pratylenchus thornei is a migratory endoparasite that feeds and reproduces in the cortex of wheat roots and is found in all major Australian wheat-growing regions (Thompson et al. 2008; Hollaway et al. 2008; Vanstone et al. 2008). P. thornei can reduce yield by as much as 65% in intolerant wheat cultivars (Thompson et al. 1999) and has been estimated to cost Australian wheat growers AU$50 M (wheat priced at AU$239/t) annually (Murray and Brennan 2009). Most of these losses occur in the northern grains region, where P. thornei is the dominant plant-parasitic nematode, being present in 67% of fields (Thompson et al. 2010).

A resistant (R) plant is one in which nematodes reproduce poorly and a tolerant plant is one that shows little injury, even under attack by large populations of nematodes (Rhode 1972). Incorporating partial resistance to P. thornei into wheat cultivars can increase yields by up to 17% compared to tolerant commercial cultivars (Thompson et al. 2001). Accordingly, genetic resistance is preferred over cultural, biological and chemical control methods. When resistance is coupled with tolerance, the crop will yield well while reducing nematode densities in the soil to non-damaging levels for subsequent crops, thereby allowing shorter and more flexible rotations (Roberts 2002). ‘Putting the technology into the seed’ through genetic resistance and tolerance requires no additional management costs to growers. Despite the development of partially R advanced breeding lines that have increased yield on a nematode infested site (Thompson et al. 1999, 2001), there remains a need to identify higher levels of resistance from genetically compatible species to facilitate the development of R parent lines and ultimately commercial cultivars.

Wheat is an allohexaploid comprising three genetically related genomes (A, B and D) that originated as a hybrid of emmer wheat (Triticum turgidum ssp. dicoccon, BBA u A u) and Aegilops tauschii (DD) (Mukai et al. 1993). Emmer wheat is a hybrid of the diploid grasses Aegilops speltoides (SS, putative progenitor of the B-genome) and Triticum urartu (A u A u) (Valkoun 2001). One of the richest sources of new genes for disease resistance for wheat improvement has been the progenitor and related species that have genomes homeologous to those of wheat, including the diploid species A. tauschii, Triticum monococcum (A m A m), T. urartu and A. speltoides and the tetraploid species Triticum tugidum (BBA u A u) and Triticum timopheevii (GGA u A u) (Cox 1991).

Pratylenchus thornei is endemic to the region of the Middle East (Di Vito et al. 1994; Greco et al. 1988; Orion et al. 1979; Pourjam et al. 1999) where modern wheat and its progenitors have evolved (Feldman and Sears 1981). Much of the diversity among plant species is believed to reflect natural selection imposed by biotic stress, causing the evolution of a new resistance character that reduces pathogen attack (Rausher 2001). Therefore, it is conceivable that, through co-evolution, resistance to P. thornei might have developed in some of these species. For example, 39 of 244 accessions of A. tauschii originating from Iran and Azerbaijan were R to P. thornei (Thompson and Haak 1997) including three accessions that had previously been found to be highly R to cereal cyst nematode (CCN; Heterodera avenae) (Eastwood et al. 1991).

Until now, the resistance to P. thornei of diploid and tetraploid relatives of wheat carrying the A, B and related genomes has not been investigated. This paper establishes that P. thornei resistance is located on the A-genome, confirms resistance on the B-genome and identifies many accessions that could be valuable in developing wheat cultivars with resistance to P. thornei.

Materials and methods

Germplasm

Two glasshouse experiments were conducted to determine the resistance to P. thornei of 241 unique accessions of principally wild relatives of wheat compared to 10 reference treatments. The following species of wild relatives and number of accessions of each in this study were: 21 A. speltoides, 21 T. urartu, 23 T. monococcum, 2 A. peregrina (S p S p U p U p; syn A. variabilis), 25 T. timopheevii ssp. armeniacum, 30 T. carthlicum (=T. turgidium ssp. carthlicum), 25 T. dicoccoides (=T. turgidium ssp. dicoccoides) and 1 T. turanicum (=T. turgidium ssp. turanicum, BBA u A u). Up to 100 hexaploid wheats including 15 hard red winter wheat cultivars with disease resistance introgressed from A. tauschii at Kansas State University (KSU) (http://www.k-state.edu/wgrc/Germplasm/grmplsm.html), four lines derived from the French wheat cv. Lutin and an accession of A. peregrina (Rivoal et al. 2001) carrying two CCN-resistance genes designated CreX and CreY (Barloy et al. 2007), 18 selections from the Iraqi landrace AUS4930 (syn. Iraq 48) putatively R to CCN pathotype Ha13 (syn. Heterodera australis, Subbotin et al. 2002) and 43 doubled haploid lines derived from the moderately R bread wheat GS50a were also examined in each experiment. The reference treatments comprised four susceptible Australian wheat cultivars (Batavia, Cunningham, Gatcher and Janz), three partially R wheat accessions (GS50a, QT8343 and QT9048), one partially R durum cultivar (Yallaroi), one partially R canarygrass cultivar (Phalaris canariensis cv. Moroccan) and one inoculated unplanted treatment to simulate fallow.

Glasshouse procedures

In both experiments, entries were replicated three times in a randomised block design with resistance determined by counting the final number of P. thornei in the roots and soil of each entry after 16 weeks of growth. Soil temperature was maintained at 22°C, the optimum temperature for P. thornei reproduction (Thompson et al. 1999), by under-bench heating. Air temperature was maintained between 20 and 25°C by the use of shade cloth (as required) and evaporative coolers. Plants were sprayed as required with 1 ml/l of Milcurb® (125 g/l dimethirimol) to control powdery mildew and with 20 ml/l of Yates Pyrethrum Insecticide® (16 g/l piperonyl butoxide, 4 g/l pyrethrins) or 1.5 g/l of Pirimor WG Aphicide® (500 g/kg pirimicarb) to control aphids.

Experiment 1 followed the procedure described by Thompson and Haak (1997). A polythene bag (100 μm thickness) was filled with 1 kg (oven-dry equivalent) of vertosolic soil of the Irving clay soil association (Thompson and Beckman 1959) pasteurised using aerated-steam at 70°C for 45 min (Thompson 1990). Pre-weighed soil and roots from open pot cultures supplying 2,500 P. thornei/kg was added to each bag of soil and mixed thoroughly by shaking. Nutrient solutions providing nitrogen (200 mg/kg), phosphorus (25 mg/kg), zinc (5 mg/kg), potassium (88 mg/kg) and sulphur (36 mg/kg) were then added and mixed thoroughly. The plastic bags and their contents were placed into 15 cm-diameter plastic pots and moved onto benches in the glasshouse. Three seeds per pot were laid on the soil surface and pressed 2 cm into the soil, covered and watered to a gravimetric content of 0.56 g/g (equivalent of pF2). After emergence, extra plants were removed to leave one plant per pot. Pots were watered as required to return the soil to 0.56 g/g moisture content.

Experiment 2 followed the procedure described by Sheedy and Thompson (2009). Plants in 70 mm-square (150 mm high) plastic pots containing 330 g (oven-dry equivalent) of soil fertilized with 1 g of Osmocote® Native Gardens plus micronutrients (17–1.6–8.7 NPK) slow-release fertilizer, (Scotts Australia Pty Ltd., Baulkham Hills, NSW, Australia) were grown on benches fitted with a bottom-watering system regulated by a float valve set to a water tension of 2 cm. Three seeds of each entry were placed on a base layer of soil (80% of total weight) and inoculated with 3,300 P. thornei (equivalent to 10,000/kg oven-dry soil) suspended in 15 ml of water pipetted around the seeds. The remaining soil was placed over the seeds as a cap. Extra plants were removed after emergence to leave one plant per pot.

In both experiments, plant growth stage (Zadoks et al. 1974), tiller number and height were recorded at 16 weeks after sowing. Then, the plant and soil were removed from the pot and a vertical section (Experiment 1) or a horizontal section (Experiment 2) removing 50% of the soil and roots was collected and stored at 3°C. Plants were re-potted with fresh pasteurised soil, staked and returned to the glasshouse, watered and grown on for seed.

Nematode extraction and enumeration

The sample of soil and roots was thoroughly mixed and the roots cut into ≈1 cm lengths. A 100 g subsample of the processed soil and roots was dried at 105°C for 48 h to determine gravimetric moisture content. A 150 g subsample was extracted at 22°C for 48 h using the Whitehead tray method (Whitehead and Hemming 1965) and nematodes were collected on a 20 μm sieve. Samples were stored in 30 ml vials at 3°C. Nematodes extracted from soil and roots were counted using a 1 ml Hawksley slide under a compound microscope (40×).

Statistical analysis

An analysis across years was conducted for ln (P. thornei/kg soil + 1) counts, where the ln transformation was applied to address variance heterogeneity. The data were analysed in a linear mixed model framework, with genotype by year effects fitted as random terms with an unstructured form for the genetic co-variance matrix across years. Replicate effects were fitted as random, and a separate residual variance was fitted for each experiment. Estimates of variance parameters were obtained using residual maximum likelihood (REML) (Patterson and Thompson 1977) and best linear unbiased predictions (BLUPs) (Robinson 1991; Piepho 1998) were obtained for nematode counts of genotypes. The program ASReml-R (Butler et al. 2009) was used to fit the linear model and form predictions. The heritability of each experiment was calculated using the approach of Cullis et al. (2006), and probabilities of genotype effects being more R than GS50a were calculated from the full matrix of standard errors of differences. Additionally, the significance of the ploidy and species structure within the genotype set was tested as a fixed effect in the linear mixed model providing best linear unbiased estimates (BLUEs).

Average reproduction factor (RF) for each accession was calculated by dividing back-transformed final by initial P. thornei numbers in soil and roots for each experiment and then used to classify accessions according to the Australian national disease rating and management guide for nematode resistance (http://www.nvtonline.com.au/) into one of nine categories ranging from R to very susceptible (VS).

Results

Twenty-nine accessions, including 11 (52%) A. speltoides, 10 (43%) T. monococcum, 5 (24%) T. urartu, 1 (4%) T. dicoccoides and 2 (2%) T. aestivum were found to be R with a RF ≤ 1. A further 59 accessions, 10 A. speltoides, 11 T. monococcum, 13 T. urartu, 2 A. peregrina, 8 T. dicoccoides, 7 T. timopheevii and 8 T. aestivum were moderately R (Table 1). The probabilities calculated from the pair-wise comparison of each genotype with GS50a identified 18 accessions and the inoculated unplanted treatment in Experiment 1, and nine accessions and the inoculated unplanted treatment in Experiment 2, that produced significantly lower final P. thornei populations than GS50a. The nine accessions more R than GS50a in both years include three T. urartu (A u-genome; AUS26935, AUS26978, AUS26979), two T. monococcum (A m-genome; AUS27037, AUS27049), three A. speltoides (S-genome; AUS26952, AUS26983, AUS26984) and one T. dicoccoides (BA u-genomes; AUS27025). All of the A. tauschii derived germplasm from Kansas State University was susceptible, as were Lutin and the A. peregrina derived lines.

Table 1 Final P. thornei population and RF of 251 accessions of wheat and related species tested in two glasshouse experiments
Full size table

The strong genetic correlation between years (0.90) indicates that ranked performance for resistance to P. thornei was relatively stable across the two experiments. While Experiment 2 had a greater genetic variance, it also had a higher level of residual variance, resulting in an overall lower heritability than Experiment 1. Hence Experiment 1 gave slightly better discrimination between genotypes (Table 2).

Table 2 Variance components and heritability estimated from the linear mixed model for two experiments
Full size table

Analysing accessions by ploidy level (Table 3) showed that the diploid accessions were significantly more R than the tetraploid accessions, which were significantly more R than the hexaploid accessions. The proportion of R accessions (RF < 1) decreased from 41 to 1 or 2% as ploidy increased from diploid to tetraploid or hexaploid, respectively. The trend of susceptibility increasing as the level of ploidy increased remained evident when a less stringent level of selection (RF < 4) was applied with 92% of diploid accessions, but only 21 and 13% of tetraploid and hexaploid accessions, being moderately R or better.

Table 3 Diploid accessions were more R to P. thornei than tetraploid and hexaploid accessions
Full size table

A total of 115 accessions in the species A. speltoides, T. urartu, T. monococcum, T. dicoccoides and T. timopheevii had passport data identifying country of origin. These had been collected from Armenia (3), Azerbaijan (2), Iran (2), Iraq (41), Israel (17), Italy (1), Lebanon (11), Syria (1) and Turkey (37). Overall, accessions from Lebanon had the lowest average RF, while accessions from Israel had the highest. Israel, Turkey and Lebanon had the lowest average RF for A. speltoides, T. monococcum and T. urartu accessions, respectively, with accessions from Italy, Iran and Turkey the highest. Armenia and Syria had the most R T. timopheevii and T. dicoccoides accessions while Azerbaijan and Israel had the most susceptible.

Discussion

A high strategic priority for practical cereal improvement worldwide is to enrich the cultivated gene pools by incorporating favourable alleles, genes or gene complexes from wild relatives (Feuillet et al. 2007). The P. thornei-R diploid and tetraploid species identified in this study belong mostly to the primary gene pool of common wheat, and gene transfer could be achieved by direct hybridization with adapted durum and bread wheat cultivars, homologous recombination, backcrossing and selection (Friebe et al. 1996). Several authors report that crosses between A. speltoides (Bijral et al. 1997; Valkoun 2001), T. urartu (Johnson and Dhaliwal 1976; Valkoun 2001), or T. monococcum (Johnson and Dhaliwal 1976; Ma and Hughes 1993) and common or durum wheats produced viable F1s that were completely male sterile and partially female sterile. F1s produced using a durum parent generally recovered full fertility after the first backcross (Valkoun 2001), whereas F1s derived from common wheat have generally required several additional backcrosses before the progeny were self-fertile (Bijral et al. 1997; Schmolke et al. 2011; Qiu et al. 2005). The A and B-genome chromosomes of T. dicoccoides show high homology with those of durum and bread wheats (Valkoun 2001) and readily hybridize. Suppression of the resistance gene(s) or dilution of its products may result in a reduction of expressed disease resistance when transferred from a species of a lower level of ploidy to one of a higher level (Cox 1991; Gill et al. 1986; Kerber and Green 1980; Potgieter et al. 1991). However, many genes conferring resistance to diseases and pests have been transferred using direct hybridization (Cox 1991, 1998; Friebe et al. 1996), including CCN resistance from T. monococcum to durum and bread wheat cultivars (Singh et al. 2010).

Previous studies on moderately P. thornei R hexaploid wheats from the West Asia and North Africa (WANA) region have used single marker regression and QTL analysis to identify resistance loci on chromosomes 2B, 3B, 6D and 7A (Schmidt et al. 2005), 1B, 2B and 6D (Toktay et al. 2006) and a susceptibility locus on 1B (Schmidt et al. 2005). Similar studies on synthetic hexaploid derived populations have identified P. thornei resistance loci on chromosomes 6DS and 6DL (Zwart et al. 2005), 1B and 3B (Toktay et al. 2006) and 2BS, 6DS and 6DL (Zwart et al. 2006, 2010). It is not surprising that virtually all the resistance loci have been identified on the B and D-genomes, considering all of the A. speltoides accessions examined in this study were at least moderately R to P. thornei, as were 16% of A. tauschii accessions screened by Thompson and Haak (1997). However, prior to this study there was only limited evidence for P. thornei resistance on the A-genome of wheat. The P. neglectus resistance gene Rlnn1 has been mapped to chromosome 7A (Williams et al. 2002), but the only P. thornei resistance locus identified on the A-genome (also 7A) was detected only in single marker regression (Schmidt et al. 2005). The results of this study clearly establish that resistance to P. thornei exists on the A-genomes of both T. urartu and T. monococcum.

Since amphiploids are usually highly fertile and true-breeding (Cox 1998) they could be an alternative or concurrent method of exploiting these novel resistances. Synthetic hexaploids have been used widely by breeding programs (Feuillet et al. 2007) to introduce disease resistance into wheat cultivars (Ogbonnaya et al. 2008), and could be used to combine the P. thornei resistances of T. dicoccoides and A. tauschii. Although A. speltoides carries genes that support homoeologous chromosome pairing (Schneider et al. 2008), certain accessions, particularly from subspecies aucheri and ligustica, may also introduce gametocidal genes to wide cross progeny thereby reducing female fertility by 50% or more (Marais and Pretorius 1996). However, amphiploids between A. speltoides and T. monococcum have been successfully used to transfer disease resistance to hexaploid wheat (Kerber and Dyck 1990), suggesting that substitution of T. monococcum with T. urartu may also be effective. If successfully developed, the synthetic tetraploids could be hybridized with P. thornei R accessions of A. tauschii to produce synthetic hexaploids with P. thornei resistance on each of their genomes.

Previous experience in amphiploid production has shown that when partially P. thornei-R durums and A. tauschii accessions were combined in synthetic hexaploids, the resultant level of P. thornei resistance was not diluted and was equal to that of the parents and better than synthetic hexaploids derived from only one R parent (Thompson 2008). This was likely due to P. thornei resistance being polygenic and additive (Zwart et al. 2004; Thompson and Seymour 2011). As a result, a proportionate number of additive resistance genes may be required in hexaploids to limit P. thornei multiplication to levels observed in species of lower ploidy (Thompson 2008).

Although our research identifies P. thornei R accessions, the number of genes controlling the resistance remains unknown. Since 92% of the diploid accessions evaluated in this study were R or moderately so to P. thornei, it could be concluded that resistance is controlled by a few dominant genes of large effect. This would be comparable to the Cre genes that individually confer complete or partial resistance to pathotypes of CCN. Conversely, several studies have shown that combinations of two to five minor resistance genes can considerably reduce disease incidence or even result in near-immune reactions in wheat cultivars (Miedaner et al. 2006; Singh et al. 2011). Chromosome addition lines allow the study of the genetic effects of individual alien chromosomes in the background of hexaploid wheat and fortunately one such population has been developed for the P. thornei R A. speltoides accession AUS26955 (synonyms: TA2780, TAM9212-2) (Friebe et al. 2000). A preliminary study of the disomic chromosome addition lines indicated that P. thornei resistance in A. speltoides could be controlled by up to five genes of varying strength (Sheedy et al. 2008a). Interestingly the strongest P. thornei resistances, slightly weaker but statistically equal to A. speltoides, were observed in the DA6S and DA5S addition lines (Sheedy et al. 2008a). Neither of the corresponding B-genome chromosomes has been identified in QTL analysis of P. thornei resistance in landrace bread wheats and synthetic hexaploid wheats. The addition lines DA2S, DA3S and DA4S had an intermediate level of resistance, more susceptible than A. speltoides but more R than Chinese Spring (Sheedy et al. 2008a), with chromosome 2B and/or 3B commonly identified in QTL analysis of hexaploid wheat (Schmidt et al. 2005; Toktay et al. 2006; Zwart et al. 2006, 2010). Chromosome 1B has been reported to have both susceptibility (Schmidt et al. 2005) and resistance (Toktay et al. 2006) loci but the DA1S addition line was more susceptible than Chinese Spring (Sheedy et al. 2008a). Given there may be up to five genes controlling the P. thornei resistance in A. speltoides (Sheedy et al. 2008a) and that many studies of hexaploid wheat have concluded that P. thornei resistance is polygenic with additive gene action (Schmidt et al. 2005; Thompson and Seymour 2011; Toktay et al. 2006; Zwart et al. 2004, 2005, 2006, 2010), it seems reasonable to expect that P. thornei resistance in the diploid relatives of wheat may also be polygenic. Further evaluation of this set of addition lines and an associated set of disomic chromosome substitution lines (Friebe et al. 2000) would greatly aid our understanding of the chromosome locations and strengths of the various resistance genes.

If a proportionate number of additive resistance genes are required in hexaploids to produce P. thornei resistance levels similar to R diploids and the resistance of A. speltoides (and possibly T. urartu and T. monococcum) is controlled by up to five resistance genes, does this mean we would need to recover up to ten genes in tetraploids and 15 genes in hexaploids? Since it appears that the genes have varying effects, recovery of the genes with the largest effects from each of the A, B and D-genomes would likely confer levels of P. thornei resistance in hexaploids similar to the levels in the diploid species. A. speltoides appears to have two genes of large effect (Sheedy et al. 2008a) as does A. tauschii (Zwart et al. 2006, 2010) and if T. urartu or T. monococcum were similar there may only be a need to recover six genes to maintain the diploid level of resistance in a hexaploid background. That would be a similar position to some current breeding efforts that aim to recover up to six genes to transfer partial P. thornei resistance from landraces to adapted cultivars (Thompson and Seymour 2011).

Evaluation of wheat and durum landraces and A. tauschii accessions originating from the WANA region where P. thornei is widely distributed has proven successful in identifying R or moderately R accessions (Sheedy and Thompson 2009; Thompson and Haak 1997; Thompson et al. 2009). Many of the R wild relatives identified in this research, particularly A. speltoides, T. monococcum, T. urartu and T. dicoccoides, originated from the countries occupying the Levantine coast of the eastern Mediterranean where P. thornei is endemic (Di Vito et al. 1994; Greco et al. 1988; Orion et al. 1979). This supports the notion that germplasm collections selected from areas where a disease is endemic are more likely to contain R accessions.

The A. peregrina selections (AUS90633SH, AUS90633LH) classified as moderately R to P. thornei in this study are also CCN-R, but we found all of the CCN-R lines developed from A. peregrina (Barloy et al. 2007) were susceptible to P. thornei like their cultivated T. aestivum parent ‘Lutin’. Evaluation of germplasm with the CCN resistance genes Cre2 (Nombela and Romero 1999), Cre3 (Thompson 2008), Cre7 (Nombela and Romero 2001) and CreR (Sheedy et al. 2008b) also showed that they did not confer resistance to P. thornei. Similarly, the KSU germplasm with resistances to Hessian fly (Mayetiola destructor), greenbug (Schizaphis graminum), leaf rust (Puccinia triticina), powdery mildew (Blumeria graminis f.sp. tritici), soilborne mosaic virus and/or wheat spindle streak mosaic virus introgresssed from A. tauschii did not confer resistance to P. thornei. These results reinforce the need for targeted selection of R germplasm because serendipitous selection for P. thornei resistance is unlikely to be successful even if the germplasm is R to other nematodes, insects or diseases.

The wild relatives of common wheat appear to be very valuable sources of resistance to P. thornei, providing the opportunity to introduce one or more resistance genes into each of the three wheat genomes. Particular emphasis should be given to the novel A-genome resistances that so far have not been detected in hexaploid breeding lines. Direct hybridization with durum and bread wheat and the production of amphiploids seem the most efficient methods of utilizing the P. thornei R accessions identified here.