Multiple Disease Resistance Loci And Their Relationship To Agronomic And Quality Loci In A Spring Barley Population

Patrick Hayes1*, Doris Prehn2, Hugo Vivar3, Tom Blake4, Andre Comeau5, Isabelle Henry6, Mareike Johnston7, Berne Jones8, Brian Steffenson9, C.A. St. Pierre10, and Fuqiang Chen11

*Corresponding author.


Genetic resistance is an effective means of disease control. Molecular markers can be employed to directly map loci showing simple inheritance and QTL strategies can be used to locate those showing quantitative inheritance. Because disease resistance loci are often present in unadapted backgrounds, linkage drag and negative pleiotropic effects are of concern. Mapping in a reference population phenotyped for multiple traits allows for detection of coupling and repulsion linkages and coincident QTL, which could either indicate pleiotropy or tight linkage. We mapped determinants of resistance to the PAV strain of barley yellow dwarf virus (BYDV), leaf rust (Puccinia hordei), stripe rust (Puccinia striiformis), and scald (Rhynchosporium secalis) in a population of doubled haploid lines and related these to effects detected for agronomic and malting quality traits. Determinants of resistance to leaf rust, BYDV, and scald mapped as single loci to chromosomes 1 and 3. Determinants of resistance to stripe rust were mapped as QTL to chromosomes 4 and 7 in greenhouse and field tests, respectively. Agronomic and malting quality trait QTL were mapped using data from field environments where these diseases were absent. There were no repulsion linkages, or coincident QTL in a repulsion phase, of resistance loci with agronomic or malting quality traits. QTL for grain yield and two malting quality traits coincided with stripe rust resistance QTL, but the resistant parent contributed favorable alleles. Thus, intra-population manipulation of disease resistance loci can proceed without negative effects on performance. The association of these resistance loci with agronomic and quality traits in other germplasm remains to be determined.


Genetic resistance is a cost effective and sound approach to disease control. However, disease resistance genes are often found in unadapted germplasm. Transfer of these genes to adapted germplasm can be a laborious proposition, particularly when they show quantitative inheritance. Furthermore, introgression may be complicated by linkage drag. Marker-based strategies can be used to detect these loci as QTL and markers can be used to expedite introgression, minimize linkage drag, and maximize recovery of the target genome. QTL analysis also offers a simple approach to determining the presence of linked or coincident QTL for target and non-target traits. Although linkage and pleiotropy may be indistinguishable at the level of resolution afforded by common population sizes and marker densities, information regarding the frequency of coupling vs. repulsion relationships can be invaluable in developing a breeding strategy.

QTL detection and manipulation have been generally considered separate steps. Indeed, most QTL reports to date have been of an essentially descriptive nature. In this context, exhaustive phenotyping for non-target traits may not be warranted, if the objective of the experiment is map-based cloning or detection of a locus and its transfer, via sexual hybridization and monitoring of tightly linked markers, to unrelated germplasm. If QTL detection and variety development are simultaneous events, such as in the breeding scheme proposed by Tanksley and Nelson (1995), then more extensive phenotyping for a whole range of characters would be justified.

A range of diseases can limit barley yield and quality. In the experiments leading to this report, we considered three fungal pathogens and one viral pathogen. Stripe rust (incited by Puccinia striiformis f. sp. hordei) has caused serious yield losses in Europe, in the Asian subcontinent, and in South America (Dubin and Stubbs, 1985). The progression of this disease toward the US , its impacts, and what little is known regarding the genetics of resistance were reviewed by Chen et al. (1994). In the US, the disease was first reported from Texas, in 1991. Roelfs and Huerta-Espino (1994) described seedling reactions to isolates from Texas, while Marshall and Sutton (1995) described the epidemiology, virulence, and yield loss in the region. Salient points are that by 1995 the disease was reported in every state of the western US, all spring barley cultivars grown in the region are susceptible, the genetics of stripe rust resistance in barley are poorly characterized , and there is considerable variation in stripe rust isolates collected in the region (R. Line, personal communication). Whether this variation is related to virulence is not known. Barley leaf rust (incited by Puccinia hordei) has been studied more extensively than stripe rust. This disease is found in many barley producing parts of the world. In the US, this disease has become a serious problem in the mid-Atlantic region of the US because isolates with virulence for the widely-used Rph7 resistance gene have become firmly established (Jin and Steffenson, 1994). The map locations of some Rph genes were reported by Jin et al., (1993). Scald (Rhynchosporium secalis) is a disease of worldwide importance and is common in cooler, humid climates. Barua et al. (1993) reported marker/phenotype linkages for scald resistance on chromosome 3, which may correspond to the Rh, Rh3, and/or Rh4 loci described by Takahashi (1983). BYDV, a luteovirus, is one of the most widespread and important virus diseases of barley. The disease has been extensively studied in terms of its vector, serology, and molecular biology, as reviewed by Collins et al. (1996). These authors provided a detailed map of the centromeric region of chromosome 3 showing markers tightly linked to the Ryd2 locus. The worldwide distribution and epidemiology of BYDV were reviewed by D'Arcy & Burnett (1995). Progress in breeding for resistance to this disease was reviewed by Burnett et al. (1995). The availability of markers linked to Ryd2 may be useful in developing resistant varieties.

In view of the threat to western US barley production posed by barley stripe rust (P. striiformis f. sp. hordei) we initiated a mapping project to locate the determinants of quantitative stripe rust resistance present in 'Calicuchima sib', a six-rowed germplasm line developed by Dr. Hugo Vivar (ICARDA/CIMMYT). We reported that resistance to the stripe rust inoculum encountered under field conditions in Mexico in 1991 and 1992 was controlled by one large-effect and one small-effect QTL (Chen et al. 1994). Calicuchima-sib is resistant to the spectrum of leaf rust (Puccinia hordei) and scald (Rhynchosporium secalis) virulence currently encountered in the Andean region of South America. The stripe rust susceptible parent we mated with Calicuchima-sib was a two-rowed germplasm line derived from backcrosses to 'Bowman', putatively carrying the Ryd2 locus conferring resistance to the PAV strain of BYDV.

Thus, our initial stripe rust resistance mapping efforts grew to include mapping of the determinants of resistance to leaf rust, BYDV and scald. At the same time, we generated phenotype data sets for agronomic performance and malting quality. Our objectives in this study were to (i) map determinants of resistance to stripe rust, leaf rust, scald, and BYDV (ii) provide colleagues with a catalog of markers for working with these mapped genes, and (iii) determine relationships of disease resistance loci with non-target agronomic and quality trait loci.

Materials and Methods

Germplasm: One hundred and ten doubled haploid (DH ) lines were derived from the F1 of the cross LBIran/UNA8271//Gloria/ Come x Bowman (BC). LBIran/UNA8271//Gloria/ Come, tested in the ICARDA/CIMMYT program as Calicuchima-sib, will be referred to as "LUGC" in the remainder of this report. The Bowman backcross (BOWBC) derivative putatively carries the Ryd2 gene for resistance to barley yellow dwarf virus (BYDV) and was kindly provided by Dr. Jerry Franckowiak of North Dakota State University.

Genotypes: Fifty two loci have been genotyped in this population . Mapping protocols were presented by Chen et al. (1994). Genome regions were mapped based on selective genotyping revealing likely marker/QTL associations, or a priori information regarding probable location of target loci. A subset of 33 loci was selected to provide relatively even spacing and unequivocal map ordering for QTL analysis. Linkage maps were constructed with GMendel (Holloway and Knapp, 1994) and define portions of all seven chromosomes (Figure 1). The Simple Interval Mapping (SIM) and Simple Composite Interval Mapping (sCIM) techniques available in MQTL (Tinker and Mather, 1995) were used for QTL detection. Using the MQTL software, each data set was analyzed with 1,000 permutations, a 5 cM walk speed, and a Type I error rate of 5%. Fifteen background markers with approximately even spacing were specified, with a maximum of three background markers per linkage group. Multiple environment data sets were first analyzed for QTL x environment (QTL xE) interaction. In the absence of interaction, analyses were conducted on means. Reactions to leaf scald and BYDV, while scored quantitatively, showed frequency distributions indicative of monogenic inheritance (Figure 2). Phenotypes for these traits were, therefore, scored as segregating alleles at a locus and appear in the maps as Rphx, Rrsx, and Ryd2.

Stripe rust: Stripe rust was assessed in three field tests at Toluca, Mexico under epidemic conditions created with local inoculum. Disease was rated by the modified Cobb Scale (Melchers and Parker, 1922) at anthesis on a whole-plot basis. Additional details are available in Chen et al. (1994). Seedling reactions to an isolate of P. striiformis collected in Montana in 1993 were measured under controlled environment conditions. The reaction of a set of differentials to this isolate was comparable to that seen for the European isolate of race 24. A three-replicate RCB design was used. Genotypes were grown in growth chambers with a 12 h daily photoperiod and a day: night temperature regime of 16:24° C. Seedlings were inoculated at the time of second leaf appearance. Seedlings were misted with distilled water and dusted with a talcum/urediospore mixture. A 24 h dew period at 7° C followed the inoculation. Fourteen days after inoculation, plants were rated using a 00 - 4 scale (00, 01 = resistant, 2 = intermediate, 3,4 = susceptible).

Leaf rust: Reaction to an epidemic created with inoculum reflecting local virulence types (principally races 8, 19, and 30) was scored as to the presence or absence of disease in single replicate head rows at Ciudad Obregon, Mexico in 1992. Reaction to a single isolate (ND8702) of P. hordei was assessed as described by Steffenson et al. (1995). This isolate is representative of the leaf rust virulence types that have predominated for over a decade in the Midwest US.

Scald: Reaction to scald was rated, using a 1 -10 scale, under an epidemic created with local inoculum at Toluca, Mexico. Each genotype was represented by a single head row.

BYDV: The population and parents were infested at tillering with 10 - 20 aphids per plant using PAV strain "Cloutier" and rated as described by Comeau (1992).

Agronomic and malting quality traits: The population and parents were grown in three Oregon environments: Corvallis, 1992; Klamath Falls, 1993, and Corvallis, 1993. The Corvallis, 1992 experiment consisted of a single head row of each genotype. In the second two tests, two-replicate randomized complete blocks were used. Plot sizes were 7.4 m2 at Corvallis and 6.3 m2 at Klamath Falls. Agronomic practices, including seeding rate, fertilizer formulation and rate, weed control, and irrigation scheduling were in accordance with recommended practices for each location. Plots were combine-harvested. Low level scald and leaf rust epidemics were observed in the Corvallis, 1992 experiment. Only malting quality data were obtained from this environment. Yield and malting quality data were obtained from the remaining tests. No appreciable symptoms of any disease developed in either of these tests. Malt analyses were performed at the USDA/ARS Cereal Crops Research Unit, Madison WI, USA as described by Tragoonrung et al. (1990). Phenotype data are provided in files gcphe1.txt, gcphe2.txt, and gcphe3.txt. Stripe rust resistance QTL were initially identified with the MAPMAKER/QTL software (Lincoln, 1992) and were reported by Chen et al. (1994). The same stripe rust phenotype data were used for QTL detection in this report. However, QTL were detected using MQTL (Tinker and Mather, 1995) and a modified genotype data set. This is the first report of QTL effects for the other traits described in this report.

Results and Discussion

Inheritance of the traits scored in this population ranged from qualitative to quantitative. Of the qualitative traits, the ratios of resistant to susceptible genotypes for leaf rust, scald, and BYDV were 55:55, 54:56, and 58:52, respectively, not significantly different from the 1:1 ratios expected for monogenic inheritance. The leaf rust resistance locus (Rphx) mapped to chromosome 1. Jin et al. (1993) reported that Rph3 is located on chromosome 1. Allelism tests will be required to determine the relationship of Rphx and Rph3. BYDV resistance mapped to chromosome 3, where we designate the locus as Ryd2. This designation is based on the assumption that backcrossing the Ryd2 resistance allele into Bowman was successful and on the reaction of the population to the PAV isolate. This chromosome location corresponds to published reports (Collins et al., 1996). Scald resistance also mapped to chromosome 3, 7.3 cM from the Ryd2 locus. A number of reports position scald resistance loci near the centromere on chromosome 3 (Barua et al., 1993; Takahashi, 1983). In this population, BYDV and scald resistance are linked in repulsion.

When kernel weight was mapped as a QTL, BOWBC (the 2-row parent) contributed the favorable allele and the mv locus accounted for 77% of the variation in phenotypic trait expression. As this population was derived from a 6-row x 2-row cross, the phenotypic distribution for kernel weight can be ascribed to effects of the mv locus on chromosome 2, which determines fertility of lateral florets. Ever since the domestication of barley some 10,000 years ago, it has been recognized that, in general, 2-row barley has higher kernel weight than 6-row barley. Spike morphology is not trivial. It accounts for very different perceptions (perhaps with little supporting genetic or phenotypic evidence) of what genotypes are suitable for malting and the manufacture of beer. Many maltsters and brewers in the US insist on 6-row barley. In much of the rest of the world, malting barleys are, by definition, 2-rowed. Kjaer et al. (1995) provided a detailed report on the mv locus and its effects on agronomic traits.

Phenotypic frequency distributions and sCIM QTL genome scans are juxtaposed in Figure 2. The horizontal bar indicates the test statistic significance threshold. The only case of significant QTL x E interaction was for field stripe rust reaction, where the QTL x E interaction test statistic (maximum 2.9) exceeded the simple interaction threshold (2.7) on chromosome 4 in the AGB472 - WG114 interval. Tinker and Mather (1995) described the difficulties of establishing thresholds for sCIM with multiple environment data. Given the absence of other significant interaction, analyses of phenotypes scored in more than one environment were based on means.

For field stripe rust reaction, effects exceeding the test statistic threshold were detected on the full length of the linkage group on chromosome 7, with a large peak in the Ale - CDO57 interval and a smaller proximal peak that coincides with ABG387c. BOWBC contributed the larger value (susceptible) allele at both intervals. The additive effects of alleles at the two QTL peaks were 32.5 and 22.9%, respectively. Together, these loci accounted for 48% of the phenotypic variation in trait expression. These results agree with our earlier analysis, based on the interval mapping procedures of MAPMAKER/QTL. These map locations do not correspond to any reported stripe rust resistance genes in barley or to stripe rust resistance loci on the homoeologous chromosomes of wheat. (Chen et al., 1994).

With regard to stripe rust, the situation on chromosome 4 is complex. Based on interval mapping, Chen et al. (1994) reported a stripe rust resistance QTL on chromosome 4 in the ABG397 - Bmy1 interval (LOD 2.2). With the sCIM analysis, the maximum test statistic (7.2) was reached at ABG472 - WG114, which was below the significance threshold (11.1.) As previously noted, significant QTL x E interaction was found in the ABG472 - WG114 interval. In each of the individual field environment data sets, BOWBC contributed the susceptible allele throughout the chromosome 4 region, but the effect was significant only in SIM analysis and only in one dataset where it reached a maximum in the ABG397 - Bmy1 interval. Thus, the interaction was due to change in magnitude of response but QTL position and significance remain ambiguous.

In the analysis of the controlled environment stripe rust data, both the SIM and sCIM analyses detected a significant effect on chromosome 4. The plots were coincident, with a maximum sCIM test statistic of 26.2, coincident with WG114. BOWBC contributed the larger value (susceptible) allele. No significant effects were detected on chromosome 7. A QTL not apparent in the field data was detected in the terminal interval of the chromosome 6 linkage group (ABG378b - ABG458), with BOWBC contributing the susceptible allele. Together, these loci accounted for 30% of the phenotypic variation in trait expression.

Phenotyping based on controlled environment assessment of seedling stripe rust resistance with this isolate would not effectively select for resistance to the virulence encountered under field conditions in Mexico. The difference between the two measures of resistance could be due to a number of factors, including seedling vs. adult plant resistance genes, different pathogenicity of the isolate vs. the field inoculum encountered in Mexico, or the effects of genes (such as temperature sensitive genes) that influenced symptom development under growth chamber conditions. The sCIM data support the argument of seedling vs. adult plant resistance, as a QTL peak under field conditions, while not significant, does coincide with the maximum seen under controlled environment conditions (i.e. a maximum at WG114). As the field data were obtained later in plant development (i.e. anthesis) the effects of resistance genes operative only in the seedling or juvenile phase could have been obscured.

Phenotyping for non-target traits provides evidence for an alternative explanation. A significant QTL for heading date was detected in the ABG397-Bmy1 interval, with LUGC contributing the later maturity allele (an additive effect of 3.9 days). The field resistance QTL detected with SIM in the ABG397-Bmy1 interval may be due to an escape mechanism attributable to maturity. Maturity is not a trivial issue, as extreme earliness or lateness is unacceptable in most production environments. Both BOWBC and LUGC headed within the range of adapted barley varieties, so later maturity would not, in this germplasm, represent a repulsion linkage. Hayes et al. (1993) also reported heading date effects in the vicinity of Bmy1, and these may reflect allelic variation at the sh locus (Takahashi and Yasuda, 1970).

The mv locus had a profound effect on grain sizing and quality, but not yield. Interestingly, high kernel weight and high protein were both associated with the two-row parent at this locus. Of the genome regions scanned in this population, this was the only region with significant determinants of expression for these two characters. LUGC, the six-row parent, contributed favorable QTL alleles for alpha-amylase and malt extract at the linkage group on chromosome 2, and these may be pleiotropic effects of the mv locus. LUGC also contributed favorable alleles for alpha amylase on chromosomes 1, 4, and 7 and for malt extract on chromosome 6. The diastatic power QTL on chromosome 4 is likely effects of the Bmy1 ( Beta-amylase) locus (Kreis et al. 1987).

The resistant parent, LUGC, contributed all favorable alleles for alpha-amylase, diastatic power, and malt extract. Alpha amylase and diastatic power QTL map near the putative stripe rust resistance locus on chromosome 4 (a coupling effect) and distal to the chromosome 7 resistance locus (also a coupling effect). LUGC was also the contributor of favorable alleles for grain yield at QTL on chromosomes 5, 6, and 7. The chromosome 7 QTL coincides with the stripe rust resistance QTL. An expanded comparative summary of QTL effects in this population and other barley mapping populations is available in Hayes et al. (1996).

Linkage and/or pleiotropy of target and non-target traits are important in this population. QTL for stripe rust resistance were found in the same region of the genome as QTL for quality and agronomic performance. In this cross combination, associations were in coupling. Coupling linkages of QTL for alpha amylase and malt extract were found distal to the Rphx locus. No agronomic or malting quality QTL were found on chromosome 3 in the vicinity of the Ryd2 and Rrsx loci. Comeau and Jedlinski (1990) reported difficulties in combining good agronomic performance with the Ryd2 locus. The differences may be due to the germplasm employed, recombination breaking repulsion linkages in the course of backcrossing the Ryd2 locus into Bowman, or differences in Ryd2 alleles. Based on these QTL data, intra-population selection for multiple resistance, agronomic performance, and quality should be feasible. It remains to be seen if the non-target QTL are significant sources of variation as resistance loci from LUGC and BOWBC are introgressed into unrelated germplasm. It is, however, reassuring to know that the associations of target and non-target traits in LUGC are positive. This genotype is testimony to the effectiveness of the ICARDA/CIMMYT effort to select for multiple resistance and agronomic performance.

Figure 1: Chromosome regions mapped in the LBIran/UNA8271//Gloria/ Come x Bowman (BC) population. Distances are in CM.

Figure 2: Phenotypic frequency distributions and QTL scans for quantitative traits measured in the LBIran/UNA8271//Gloria/ Come x Bowman (BC) population. In the QTL scans, the parent contributing the larger value allele is identified as "A" for LBIran/UNA8271//Gloria/ Come and "B" for Bowman (BC). The horizontal bar indicates the 5% significance threshold.


Barua, U.M., K.J. Chalmers, C.A. Hackett, W.T.B. Thomas, W. Powell, and R. Waugh. 1993. Identification of RAPD markers linked to a Rhynchosporium secalis resistance locus in barley using near-isogenic lines and bulked segregant analysis. Heredity 71: 177-184.

Chen, F., D. Prehn, P.M. Hayes, D. Mulrooney, A. Corey, and H. Vivar. 1994. Mapping genes for resistance to barley stripe rust (Puccinia striiformis f. sp. hordei). Theor. Appl. Genet. 88: 215-219.

Collins, N.G., C.M. Paltridge, C.M. Ford, and R.H. Symons. 1996. The Yd2 gene for barley yellow dwarf virus resistance maps close to the centromere on the long arm of barley chromosome 3. Theor. Appl. Genet. (in press).

Comeau, A., and H. Jedlinski. 1990. Successful breeding for barley yellow dwarf virus resistance or tolerance: a systematic approach related to other agronomic characteristics. p 441-451. In: Burnett, P.A. (ed.) World perspectives on barley yellow dwarf. CIMMYT, Mexico.

Comeau, A. 1992. The use of artificial inoculation with viruliferous aphids in barley yellow dwarf virus research. p 177-182. In. Comeau, A., Makkouk, K. (eds). BYDV in WANA. ICARDA, Aleppo, Syria.

Dubin, H.J., and R.W. Stubbs. 1986. Epidemic spread of barley stripe rust in South America. Plant Dis. 70:141-144.

Hayes, P.M., B.H. Liu, S.J. Knapp, F. Chen, B. Jones, T. Blake, J. Franckowiak, D. Rasmusson, M. Sorrells, S.E. Ullrich, D. Wesenberg, and A. Kleinhofs. 1993. Quantitative trait locus effects and environmental interaction in a sample of North American barley germplasm. Theor. Appl. Genet. 87:392-401.

Hayes, P.M., F.Q. Chen, A. Kleinhofs, A. Kilian, and D. Mather. 1996. Barley genome mapping and its applications. p 229 - 249. In: P.P. Jauhar (ed.). Methods of Genome Analysis in Plants. CRC Press, Boca Raton, USA.

Holloway, J. and Knapp. 1994. GMendel 3.0.Users Guide.

Jin Y., G.D. Statler, J.D. Franckowiak, and B.J. Steffenson. 1993. Linkage between leaf rust resistance genes and morphological markers in barley. Phytopath. 83:220-223.

Jin Y., and B.J. Steffenson. 1994. Inheritance of resistance to Puccinia hordei in cultivated and wild barley. J. Heredity. 85:451-454.

Kjaer, B., J. Jensen, and H. Giese. 1995. Quantitative trait loci for heading date and straw characters in barley. Genome 38: 1098-1104.

Kreis, M., M. Williamson, B. Buxton , J. Pywell, J. Hejgaard, and I. Svendsen . 1987. Primary structure and differential expression of ß-amylase in normal and mutant barleys Eur J Biochem 169:517-525.

Lincoln, S., M. Daly and E. Lander. 1992. Mapping genes controlling quantitative traits with MAPMAKER/QTL/QTL 1.1. Whitehead Inst. Tech. Rep. 2nd ed.

Marshall, D. and R.L. Sutton. 1995. Epidemiology of stripe rust, virulence of Puccinia striiformis f.sp. hordei, and yield loss in barley. Plant Dis. 79:732-737.

Melchers, L.E., and J.H. Parker. 1922. Rust resistance in winter wheat varieties. USDA Agr. Bull. 1046.

Roelfs, A.P. and J. Huerta-Espino. 1994. Seedling resistance in Hordeum to barley stripe rust from Texas. Plant Dis. 78:1046-1049.

Steffenson, B., Y. Jin, B.G. Rossnagel, J.B. Rasmussen, and K. Kao. 1995. Genetics of multiple disease resistance in a doubled haploid population of barley. Plant Breeding 114:50-54.

Takahashi R. and S. Yasuda. 1970. Genetics of earliness and growth habit in barley. p. 388-408. In: Nilan, RA (ed) Barley Genetics II Proc 2nd Int Barley Genet Symp, Pullman, WA. 1970 Washington State University Press.

Takahashi, R. 1983. Coordinator's report: chromosome 3. Barley Genet. Newsl. 13:89-93.

Tanksley, S.D. and J.C. Nelson. 1995. Advanced backcross QTL analysis: a method for the simultaneous discovery and transfer of valuable QTLs from unadapted germplasm into elite breeding lines. Theor. Appl. Genet (in press)

Tinker, N. and D. Mather. 1995. Methods for QTL analysis in progeny replicated in multiple environments. JQTL.

Tragoonrung, S., P.M. Hayes, and S. Broich. 1990. Near-infrared reflectance estimates of grain protein and malt extract in hill and row plot evaluations of spring malting barley. Can. J. Plant Sci. 70:71-78.