There are claims that at least 11 genes confer resistance in barley (Hordeum vulgaris) to one or more components of the soilborne barley mosaic virus complex but, apart from the immunity conferred by the widely used gene rym4, little is known about their mode of action. This study used mechanical (sap) and plasmodiophorid vector-inoculation techniques combined with ELISA, RT-PCR, symptom development and virus transmission to investigate the response of different genotypes to Barley mild mosaic virus (BaMMV). Barley genotypes were grown at 20 and 12°C to test for temperature sensitivity. Plants with the genes rym3 or rym6 were fully susceptible to the virus, whereas those with genes rym1, Rym2, rym5 or rym11 appeared to be immune, as BaMMV was never detected in any tissue type nor was the virus transmitted from them to susceptible genotypes. The remaining genotypes could all be infected to some extent by BaMMV using one or both inoculation methods, and virus could be transmitted from their roots by the plasmodiophorid vector Polymyxa graminis. Plants with the rym7 gene had delayed symptoms compared to susceptible controls at 12°C. Plants with the rym8 gene could be infected by both inoculation methods, but there was no virus in the leaves at 12°C. Plants with the rym9 gene could be infected only by vector inoculation, and virus remained localized in the roots. Plants with the rym10 gene appeared susceptible by mechanical inoculation at both temperatures, but after vector inoculation virus moved to leaves only at 20°C. This suggests the operation of translocation resistance in plants with the rym8, rym9 or rym10 genes, which is temperature-sensitive in rym8 and rym10 and perhaps tissue-specific in rym9. No resistance to P. graminis was observed in any of the genotypes.
Barley mild mosaic virus (BaMMV) and Barley yellow mosaic virus (BaYMV) are the causal agents of the economically important soilborne barley mosaic disease of winter barley in North-western Europe and East Asia (Huth & Adams, 1990; Plumb, 2002). Both viruses are members of the genus Bymovirus (family Potyviridae) transmitted by the plasmodiophorid soilborne vector Polymyxa graminis. Infection with BaYMV, BaMMV, or both first becomes apparent as small, kite-shaped patches in the crop with a pale green or yellow appearance. On closer inspection, irregularly distributed chlorotic streaks can be seen on the youngest leaves, often associated with an upward rolling of leaf margins, giving the plants a ‘spiky’ appearance. Later yellowing and/or necrosis may result in death of the oldest leaves. If susceptible crops are grown in subsequent years, the patches enlarge and whole fields may become infested. Yield losses of 20–45% have been observed when levels of BaMMV in the field are high, and losses can be more severe when the winter has been particularly harsh (Adams, 2002).
The bymoviruses are transmitted by the obligate root-infecting parasite P. graminis (Plasmodiophorales) that produces resting spores able to persist in the soil for more than 10 years while remaining viruliferous (Usugi, 1988). Consequently, crop rotation is not a satisfactory means of limiting the disease, nor is chemical control, which is inefficient, expensive and ecologically damaging (Kusaba et al., 1971; Adams et al., 1993). Therefore, control of soilborne barley mosaic disease is based on the use of resistant varieties (Kanyuka et al., 2003).
A single recessive gene, rym4, originating from the Dalmatian landrace ‘Ragusa’, has been used extensively in breeding programmes in Europe where most of the resistant commercial cultivars carry this gene (Graner & Bauer, 1993). This gene confers complete immunity to both BaMMV and BaYMV. It has been shown that if virus inoculum is introduced artificially to the leaves of a resistant cultivar, it is rapidly degraded (Schenk et al., 1995) and P. graminis zoospores propagated from the roots of rym4 cultivars very rarely transmit the virus (Adams et al., 1987). Winter barley cultivars with rym4 resistance were grown successfully on infested land in Europe until the late 1980s, when a new pathotype of BaYMV, namely BaYMV-2, able to overcome rym4 resistance, was discovered in Germany and Britain and subsequently in other European countries (Huth, 1989; Hariri et al., 1990; Adams, 1991). The occurrence of this resistance-breaking strain forced breeders to investigate exotic barley genotypes for sources of novel resistance genes. Following experiments using both artificial inoculation and field trials on infested land, and work with molecular markers, there are now reports of at least 11 resistance genes (all of which, except Rym2, are recessive) that are claimed to be effective against one or more of the components of the disease complex (Ordon et al., 1996). The mechanisms of resistance conferred by these novel resistance genes are not known, and therefore the durability of the resistance remains difficult to estimate. However, field variants of both BaMMV and BaYMV capable of overcoming some of these resistance genes have been reported from France (Hariri et al., 2000), China (Chen et al., 1996) and Japan (Kashiwazaki et al., 1989). Most recently a pathotype of BaMMV has been reported from France that is able to break the resistance of cultivars carrying the gene rym5, which otherwise confers complete immunity to European isolates of both viruses, including BaYMV-2 (Graner et al., 1999a; Hariri et al., 2003). These developments have increased the need for research to understand how rym resistance genes work against the soilborne barley mosaic viruses.
This study aimed to investigate the mechanisms of resistance to BaMMV conferred by these novel resistance genes, as compared with the complete immunity conferred by rym4 in commercially available cultivars. By using different inoculation techniques and P. graminis transmission experiments, this report outlines the different mechanisms of resistance to BaMMV and suggests the importance of whole-plant testing when evaluating novel sources of virus resistance.
Materials and methods
Barley genotypes containing the rym resistance genes tested during this investigation are shown in Table 1. All seeds were pregerminated before planting. Except for sand cultures of the soilborne vector (Adams et al., 1986), plants were grown in peat compost. In all experiments, genotypes with rym4 and those with no known resistance gene were used as resistant and susceptible controls, respectively. All experiments were done in a glasshouse under ambient light. To ensure a 14 h day, supplementary lighting was provided when ambient light intensity was <130 µmol m−2 s−1 and switched off at 180 µmol m−2 s−1.
Table 1. Sources of barley genotypes used in testing for resistance to BaMMV
Seed provided by: aDr T. Konishi, Faculty of Agriculture, Kyushu University, Fukuoka-shi, 812-8581, Japan; bDr F. Ordon, Institute of Agronomy and Plant Breeding, Justus-Liebig University, 35390 Giessen, Germany; cCPB-Twyford Ltd, Thriplow, Royston SG8 7RE, UK; dRothamsted Research.
For mechanical inoculation, an isolate of BaMMV originally isolated in Streatley, Bedfordshire, UK and subsequently maintained by mechanical passage on susceptible cultivars was used. Plants were grown in seed boxes in rows of 10 seedlings, interspersed with susceptible and resistant controls in a randomized design, and were inoculated at the two- to three-leaf stage following the procedure of Adams et al. (1986). Each experiment tested two genotypes and the results from duplicate experiments were combined. No plants carrying Rym2 were tested by mechanical inoculation as there were few seeds of this genotype available for examination.
For vector inoculation, two cultures of viruliferous P. graminis were used, one from Streatley, UK and one from Göttingen, Germany. Cultures, prepared as a root powder containing P. graminis resting spores, were applied to the roots of seedlings. Six seedlings were inoculated per genotype, and each inoculation was repeated at least twice per temperature. Seedlings were grown in sand culture for 28 days. At this point zoospores were liberated, counted and inoculated onto a BaMMV susceptible cultivar, and grown as described by Adams et al. (1986). Subsequently plants were grown until 70 days after inoculation (dai), when leaf and root samples were taken. Roots containing resting spores were dried and ground using a sterile mortar and pestle, and the resulting powder was used to inoculate the roots of seedlings of a susceptible cultivar.
Evaluation of infection
Mechanically inoculated plants were visually assessed weekly for typical BaMMV symptoms until 28 or 42 dai for plants grown at 20 and 12°C, respectively. Leaf samples were also taken at 28 dai from plants grown at both temperatures for testing by ELISA and RT-PCR. Plants grown in sand culture following inoculation with root powder containing viruliferous P. graminis resting spores were grown until 70 dai, when leaf samples were tested by ELISA. Root samples were tested by ELISA and RT-PCR.
Transmission of BaMMV from barley genotypes via zoospores or root powder containing resting spores was visually assessed on a weekly basis. Plants needed to be cut back 1–2 cm from soil level at 28 dai to promote the development of virus symptoms.
Statistical analysis of P. graminis zoospore production on the barley genotypes was done by anova using genstat 5 (Payne et al., 1993), after a log10 transformation of the number of spores produced per mL.
BaMMV was detected by the indirect F(ab′)2 ELISA technique described by Barbara & Clark (1982) with the modifications of Adams (1991). Samples were prepared in a 1/10 (w/v) dilution of plant material to extraction buffer [136 mm NaCl, 15 mm KH2PO4, 9 mm Na2HPO4, 2·7 mm KCl containing 0·02% (v/v) Tween 20 + 2% PVP w/v, +1% full-cream milk powder (Farley's First Milk, HJ Heinz Co Ltd, Middlesex, UK)]. Absorbance (A405nm) readings in excess of twice that of the healthy control were taken as positive.
Total RNA was extracted using the TRIzol reagent (Invitrogen, Paisley, UK) following the manufacturer's instructions. Reverse transcription and PCR were performed using primers designed to regions of homology in RNA-1 of UK, German and Japanese isolates of BaMMV (Kashiwazaki et al., 1992; Foulds et al., 1993; Schlichter et al., 1993). Forward primer, M3 (5′-ACA GAG CAC GAG GAG GAA-3′) and reverse primer, M4 (5′-GCA TGA GAG ATC TAC CGG-3′) were used to amplify an 899 bp DNA fragment from the 3′ end of RNA-1, encompassing the coat protein. cDNA was synthesized using 10 µL RNA and 200 U Superscript II RT enzyme (Invitrogen) following the manufacturer's instructions. The PCR reaction was performed in a total reaction volume of 54 µL consisting of 5 µL 10 × PCR buffer, 1 mm MgCl2, 2·5 U Taq DNA polymerase (Invitrogen), 200 µm each of dATP, dTTP, dGTP, dCTP (Roche, Mannheim, Germany), 0·5 µm primers M3 and M4, 40·6 µL dH2O (Sigma-Aldrich, Poole, Dorset, UK) and 4 µL reverse transcription product. PCR conditions were as follows: 94°C for 2 min; 35 cycles of 94°C for 1 min, 50°C for 1 min, 72°C for 1 min; and a final extension step of 72°C for 5 min. PCR products were visualized on a 1·4% agarose gel, stained with ethidium bromide and photographed under UV light.
Susceptibility by mechanical inoculation
Chlorotic streaking and mosaic symptoms typical of BaMMV infection were observed on barley genotypes carrying rym3, rym6, rym7, rym10 or no resistance gene, grown at both 12 and 20°C. However, when grown at 12°C fewer plants with rym7 or rym10 had symptoms, which developed more slowly. Few plants containing rym8 developed symptoms, and then only at 20°C. No BaMMV symptoms were observed on plants carrying rym1, rym9 or rym11. ELISA and RT-PCR confirmed that symptomless plants did not contain detectable virus. Plants with rym4 or rym5 developed no symptoms at 12°C. However, a single rym4 plant and four rym5 plants had symptoms at 20°C (Table 2). ELISA absorbance readings of plants with rym8 or rym10 were lower than the susceptible control when grown at 20°C. Also, when grown at 12°C plants with rym7 or rym10 had ELISA readings lower than the susceptible control (results not shown).
Table 2. Reaction of barley lines carrying rym resistance genes following mechanical inoculation and growth at 12 or 20°C
Gene carried by barley line
Number of plants showing BaMMV symptoms/number of plants inoculated. Symptoms were scored at 42 dai (12°C) or 28 dai (20°C).
Inoculations with either the UK or German viruliferous isolate of P. graminis produced similar results and were combined. Following resting spore inoculation of barley genotypes containing different resistance genes, the subsequent production of P. graminis zoospores was not significantly different between the genotypes grown at either temperature (20°C, P > 0·05, 25 df, F = 0·171; 12°C, P > 0·05, 25 df, F = 0·455) (Table 3). BaMMV could be detected by ELISA in the leaves of a proportion of plants with rym3, rym7 or no resistance gene at either temperature. Virus was also detected in the leaves of plants with rym8 or rym10, when grown at 20°C. No virus was detected in the leaves of plants carrying any of the other rym genes. ELISA detected BaMMV in the roots of some plants with rym3, rym6, rym7, rym10 or no resistance gene at both temperatures. BaMMV was also detected in the roots of plants containing rym8 or rym9, when grown at 20°C (Table 4). BaMMV coat-protein region was amplified by RT-PCR from the roots of plants carrying rym3, rym6, rym7, rym8, rym9, rym10 or plants known to contain no rym resistance genes, when grown at either temperature. No virus was detected in the roots of plants carrying any of the other rym resistance genes (Fig. 1).
Table 3. Zoospore production from barley varieties carrying different rym resistance genes 28 dai with viruliferous Polymyxa graminis resting spores and growth at either 12 or 20°C
Gene carried by barley line
9·12 × 105
1·15 × 106
rym1 + rym5
6·46 × 105
1·17 × 106
9·33 × 105
1·20 × 106
8·32 × 105
3·31 × 106
1·15 × 106
1·62 × 106
6·03 × 105
2·45 × 106
8·32 × 105
2·00 × 106
8·71 × 105
2·00 × 106
8·13 × 105
2·24 × 106
7·41 × 105
2·00 × 106
1·02 × 106
2·34 × 106
1·05 × 105
2·24 × 106
9·33 × 105
2·63 × 106
SED (25 df)
Table 4. Detection of BaMMV by ELISA in leaf and root tissue of barley lines with rym resistance genes 70 dai with viruliferous Polymyxa graminis resting spores, and growth at 12 or 20°C
Gene carried by barley line
Number of plants with positive ELISA readings/number of plants tested.
Zoospores of P. graminis zoospores collected from plants carrying the different rym resistance genes were back-inoculated to a susceptible cultivar to investigate transmission of BaMMV. Zoospores propagated on plants grown at either 12 or 20°C, carrying rym3, rym6, rym7, rym8, rym9, rym10 or no rym genes, transmitted BaMMV to susceptible barley cultivars as judged by the appearance of typical virus symptoms at 112 dai. No virus symptoms were seen on susceptible cultivars inoculated with zoospores from plants with any of the other resistance genes grown at either temperature. There was also no BaMMV transmission by resting spores produced on any of these genotypes. On one occasion only, two plants inoculated with zoospores from a rym4 variety grown at 20°C developed symptoms, but no symptoms were observed on plants inoculated with resting spores collected from the same rym4 source (Table 5). Zoospores of P. graminis or root powder containing fungal resting spores collected from a rym4 variety grown at 12°C did not transmit BaMMV when inoculated to a susceptible variety.
Table 5. Transmission of BaMMV to a susceptible barley variety by Polymyxa graminis zoospores or resting spores, multiplied on barley lines carrying rym resistance genes that were grown at either 12 or 20°C
Gene carried by barley line
Number of plants showing virus symptoms/number of plants inoculated. Symptoms were scored 112 dai (zoospore inoculation) or 98 dai (resting spore inoculation).
Several resistance genes investigated in this study, rym1, Rym2, rym5 and rym11, all seem to confer immunity to BaMMV, similar to the resistance of barley with rym4 (Adams et al., 1986). This is the first time that the resistance conferred by these genes has been shown to function as a form of immunity, particularly as P. graminis collected from plants carrying one or more of these genes was no longer able to transmit BaMMV. Very low levels of BaMMV symptoms (<2%) were observed on plants with rym4 or rym5 following mechanical inoculation and growth at 20°C, which was unexpected. Cultivars in Europe containing either one of these genes have previously been reported to be resistant to BaMMV (Werner et al., 2000). Similar results in the past have usually been attributed to low-level seed contamination (Huth, 1984; Proeseler et al., 1999), and this seems the most likely explanation. A very low level of transmission from zoospores propagated on plants with rym4 was observed (<4%), and this has also been reported previously (Adams et al., 1987). This low level of transmission from zoospores propagated on rym4 plants could also be due to seed contamination but it is possible that when zoospores were liberated from the roots of immune plants at 28 dai, not all the resting spores used for the original inoculation had germinated. If these germinated during the zoospore liberation process they would have released zoospores that had not been produced from the immune variety and were therefore still viruliferous. Because resting spores produced on the rym4 plants at 70 dai did not transmit BaMMV, the latter is considered the most likely explanation.
Immunity has been reported for other viruses, where it is often termed ‘extreme resistance’. Such resistance, where no symptoms are seen on inoculated plants and the virus is not detected, has for example been observed for a number of genes conferring resistance to Turnip mosaic virus (TuMV, genus Potyvirus). Most of these genes are dominant and confer resistance to specific TuMV isolates only. For example, TuRBO1 is effective against all pathotype 1 isolates of TuMV, whereas TuRBO3 confers extreme resistance to the pathotype 4 isolate of TuMV CDN1 (Walsh & Jenner, 2002).
The remaining genotypes could all be infected to some extent by BaMMV using one or both inoculation methods, and virus could be transmitted from them by viruliferous P. graminis multiplying in their roots. Plants containing the gene rym3 were fully susceptible to BaMMV, but are known to be resistant to some strains of BaYMV (Konishi et al., 2002). Among the other genotypes there is a trend towards forms of partial resistance to BaMMV. Adams (1994) first observed partial resistance to BaMMV in the cultivar Sprite. In this variety, symptoms took longer to appear and were milder, but zoospores of the vector propagated from their roots transmitted virus. Plants with rym7 behaved as a susceptible variety when grown at 20°C, but when grown at 12°C fewer plants became infected. These results are similar to those of Graner et al. (1999a) and Proeseler et al. (1999), who suggested that rym7 confers partial resistance based on a reduction of virus titre and a delay in the development of BaMMV symptoms. Previous reports have also suggested that rym8 confers partial resistance, as seen for rym7 (Bauer et al., 1997). However, in the experiments reported here BaMMV multiplied in the leaves of rym8 plants at 20°C but not at 12°C. Furthermore, the virus could be detected in the roots at either temperature following resting spore inoculation but only moved into the leaves at 20°C. Consequently the resistance of rym8 has a temperature-sensitive element relating to BaMMV multiplication in the leaves. There was also evidence of temperature-sensitive resistance in rym10, despite the suggestion that this gene confers no form of resistance to BaMMV (Hariri et al., 2000). BaMMV was detectable in the leaves of rym10 plants following mechanical inoculation and subsequent growth at either temperature, but after root-powder inoculation the virus multiplied in the roots at both temperatures but was only detected in the leaves at 20°C. In plants with rym8 or rym10 there appears to be some form of blocking mechanism preventing the virus from moving from root to shoot (‘translocation resistance’).
A resistance mechanism based on blocking systemic virus movement has been proposed for other soilborne viruses transmitted by P. graminis. Field trials for resistance in wheat to Soil-borne wheat mosaic virus (SBWMV, genus Furovirus) demonstrated that, although some cultivars had no foliar symptoms, and virus could not be detected in the leaves by ELISA, virus was detectable by ELISA in the roots of these genotypes (Budge et al., 2002). Driskel et al. (2002) proposed that the resistance to SBWMV in some wheat varieties is directed against systemic virus movement. Furthermore, earlier research on SBWMV demonstrated that resistance of hard red winter wheat cultivars could be overcome by growing plants at 23°C after either fungal or mechanical inoculation, rather than at 15°C (Myers et al., 1993). Similarly, resistance in soft white winter wheat to Wheat spindle streak mosaic virus (WSSMV, genus Bymovirus) has been suggested to function through a mechanism that also restricts upward virus movement (Carroll et al., 2002).
The resistance of plants carrying rym9 may function in a similar way to the virus transport-blocking mechanism as in rym8- or rym10-carrying plants. However, BaMMV was never detected in the leaves of rym9 plants irrespective of the inoculation technique used. It is therefore plausible that rym9-controlled resistance could be tissue-specific, preventing leaf, but not root, infection. Tissue-specific resistance has been reported for the aphid-transmitted Watermelon mosaic virus (WMV, genus Potyvirus). Resistance against this virus in cucumber is controlled by two independent factors. A single resistance recessive gene, wmv-2, confers resistance to WMV that was shown to be expressed only in the cotyledons, whereas resistance controlled by the recessive resistance gene, wmv-3, plus either a dominant or recessive gene, Wmv-4/wmv-4, is expressed only in the leaves (Wai & Grumet, 1995).
Although it is not clear whether rym9 resistance is tissue-specific or not, this evidence emphasizes the need to test all plant tissues when evaluating genotypes for virus resistance. Whenever BaMMV was detected in root tissue of plants with any of the rym resistance genes, the virus could be acquired and transmitted by its vector. Plant breeders, when investigating novel resistance sources, should therefore use simple diagnostics such as ELISA and RT-PCR to test the possibility of any localization of resistance. Carroll et al. (2002) suggested that plants with resistance genes that reduce upward virus movement could contribute less to the WSSMV inoculum potential of the soil if planted more regularly than susceptible varieties. However, the field inoculum would increase, albeit more slowly.
Plants with the gene rym6 were susceptible to BaMMV following mechanical inoculation at either temperature, but following vector-borne inoculation virus could be detected only in roots. This could therefore be translocation resistance, but without the temperature sensitivity of rym8 and rym10. However the cultivar carrying rym6, Miho Golden, is a spring type which grew rapidly under the glasshouse conditions, and it is possible that virus simply did not move rapidly enough to reach the leaves. Field trials and mechanical inoculation of cultivars containing rym6 have generally concluded that this gene does not confer resistance to BaMMV (Nomura et al., 1996; Hariri et al., 2000) although it is effective against one strain of BaYMV in Japan (Iida et al., 1999). Moreover, Plumb et al. (1986) showed that spring barley cultivars are susceptible to the barley mosaic viruses but avoid infection because of unfavourable conditions for P. graminis infection. This suggests that near-isogenic lines would be preferable when investigating the reaction of resistance genes, so that any effect observed can be attributed solely to the gene of interest.
Resistance to P. graminis, based on zoospore production, was not observed in any of the rym genotypes, a result that supports those of Adams et al. (1986) and Adams & Jacquier (1994). So far, resistance against P. graminis has been reported to exist only in other Hordeum species such as H. bulbosum, where plants remained P. graminis-free for up to 6 months after inoculation (Kastirr, 2001). The mechanism of this resistance and its mode of inheritance have still to be investigated. Introgression of such a resistance into elite barley genotypes is also likely to be very difficult due to their diverse genetic backgrounds. Furthermore, the mechanism of resistance would need to prevent penetration by zoospores rather than later development if virus transmission was to be avoided. Peanut clump virus (genus Pecluvirus) is transmitted by P. graminis and causes a destructive disease on groundnuts, but the vector multiplies efficiently only on cereals. In groundnuts, resting spores do not form but zoospores penetrate sufficiently to transmit virus very effectively (Thouvenel et al., 1988).
These experiments have shown the existence of different mechanisms for resistance to BaMMV. It is doubtful whether genes conferring partial resistance (such as rym7), or where virus is confined to the roots (rym8, rym9, rym10), should be regarded as useful. Polymyxa graminis, collected from all the plants with genes conferring partial resistance or translocation resistance, transmitted the virus to susceptible cultivars. In such cases virus inoculum levels in the field would not decrease, as is likely with immune plants, but might even increase. The only genotype known to be resistant to all European and Japanese strains of BaMMV and BaYMV is Mokusekko 3 (Hariri et al., 2000; Konishi et al., 2002). This genotype has two resistance genes, rym1 and rym5, both of which have been shown to confer immunity to BaMMV in this study. Combining two or more genes that confer immunity to the disease by gene ‘pyramiding’ (Liu et al., 2000) may be the best way to obtain durable field resistance. Furthermore, combining rym genes from different chromosomes, as in Mokusekko 3, may increase the durability of resistance. However, to combine such rym genes into an elite barley cultivar requires the use of molecular markers. Suitable markers are available for rym4, rym5, rym9 and rym11 (Bauer et al., 1997; Graner et al., 1999b; Ordon et al., 1999), and breeding programmes are currently under way to construct pyramids of two or more of these genes in elite lines (Werner et al., 2000; Ordon et al., 2003). The combinations rym4 × rym11 and rym5 × rym11 appear the most promising, as both would combine rym genes that confer immunity, at least to BaMMV, and because the genes in the cross are located on different chromosomes. Moreover, Okada et al. (2003) suggested that rym11 might be allelic to rym1, so it is possible that the resistance of a cultivar containing both rym5 and rym11 would have the same field durability as Mokusekko 3.
Research into the mechanisms of resistance conferred by rym genes to the other barley viruses transmitted by P. graminis, BaYMV and BaYMV-2, using the methods described here, will further aid decisions as to which rym genes in combination will be most likely to lead to durable field resistance to the virus complex. However, to be able to use genes such as rym1 and Rym2, further efforts are required to map their chromosome location and produce suitable molecular markers to aid the introgression progress.
We thank Dr Frank Ordon, Institute of Agronomy and Plant Breeding, Justus-Leibig University, Germany and Dr Taeko Konishi, Faculty of Agriculture, Kyushu University, Japan for their kind gifts of barley seed carrying resistance genes, used in this work. We also thank Drs Kim Hammond-Kosack, Kostya Kanyuka and Hans Cools for critical reading of the manuscript and Dr Simon Hodge for statistical assistance. Graham McGrann is a PhD student registered at the University of Nottingham, UK. This work was partly funded by CPB-Twyford Ltd, UK. Rothamsted Research receives support from the Biotechnology and Biological Sciences Research Council of the UK.