BREEDING FOR SPRING RADIATION FROST TOLERANCE IN BARLEY
(HORDEUM VULGARE L.)
J.L. Reinheimer, A.R. Barr, G.K. McDonald and J.K. Eglinton
School of Agriculture and Wine, University of Adelaide, Waite Institute, PMB 1, Glen
Osmond, SA 5064, Australia
While there has been much research carried out into freezing tolerance during vegetative
growth of barley, especially in the northern hemisphere, few comprehensive studies have been
undertaken to determine whether there is genetic variation for spring radiation frost tolerance.
This is the most common form of frost damage in barley in Australia and can damage crops
from head emergence through to late grain fill.
Plant breeders have made little progress in reproductive frost tolerance due to spatial variation
of field trials and overriding maturity effects which favours frost avoidance rather than true
tolerance. Since factors such as maturity can confound comparisons between genotypes, tools
such as frost simulation chambers and QTL analysis can be used to separate effects of
maturity and plant type from true frost tolerance. Using markers closely linked to QTL,
marker-assisted selection may improve the efficiency of frost tolerance screening by reducing
these confounding effects.
MATERIALS AND METHODS
Genetic variation for spring radiation frost tolerance
A survey of diverse germplasm collected from throughout the world was undertaken to
identify genetic variation for reproductive frost tolerance in barley. The genotypes collected
included landraces, wild barley introgressions, winter types, Australian varieties and mapping
population parents. Two sites prone to radiation frost at Loxton and Black Rock in South
Australia, were selected to locate reproductive frost tolerance screening nurseries. Early
seeding times in autumn were used to encourage flowering during the peak frost risk period.
Irrigation was used to promote early growth and to supplement rainfall events during the
growing season. Four seeding times were required to allow for maturity differences between
genotypes and to collect data from multiple frost events during the season. Each entry was
sown in a single 1.5m row/seeding time. After each frost event, ten tillers/entry within each
Zadoks maturity class in the range of Z59-Z77 were tagged (Zadoks et al., 1974). Ten to
twenty days after a frost event, tagged heads were assessed on percentage sterility and grain
damage. A discriminating frost event occurred on 30 June 2001, where temperatures reached
–4.6oC at crop canopy height causing extensive damage to genotypes in the Loxton nursery.
ANOVA was used to analyses genotype effects on sterility, grain damage and total damage
percentage (%sterile + % damaged grain).
The Arapiles × Franklin mapping population was grown in a randomised complete block
design with two replicates at Horsham, Victoria by Mr David Moody, DPI Victoria. Each
entry was sown in a 6 meter × 6 row (150mm spacing) plot on 29 June 2001. A frost event on
11 October 2001, where the minimum temperature was –1.0oC at Stevenson screen, caused
extensive sterility and grain damage. Ten heads were randomly sampled from field plots of
both replicates of the Arapiles × Franklin mapping population after all genotypes had reached
physiological maturity. This material was then assessed for sterility (florets that did not
produce grain) and grain damage (florets that produced grain but did not complete grain fill).
Percent sterility and percent grain damage was calculated based on the number of potential
florets that could produce grain. ANOVA was used to analyse genotype effects on frost
damage and means used for QTL analysis.
The Galleon × Haruna Nijo and Amagi Nijo × WI2585 mapping populations were sown at
Loxton S.A. in April 2002 at multiple seeding times. On 2 July the minimum canopy
temperature of –4.5oC was recorded. On 21 July a minimum canopy temperature of –2.8oC
was recorded. Both events caused damage to reproductive tissue of genotypes post spike
emergence. Data was analysed using ANOVA and means analysed using Qgene version 3.04
with a LOD significance threshold of 3.
Frost Simulation Chamber
Seven barley lines were used for frost tolerance screening based on field frost tolerance
screening data collected in 2001 ranging from tolerant to susceptible.
Five seeding times of six plants of each line were planted from 08/07/02 to 05/08/02 and were
grown in a glass house until heading. All plants with tillers that were in the maturity range of
Z49 to Z83 (Zadoks et al., 1974) were selected for frost tolerance screening. The specific
developmental stage of each tiller was assessed and tagged. Plants from each line were
randomly selected for exposure to one of three treatments varying in frost duration and
intensity. Analysis of frost chamber data was conducted using residual maximum likelihood
(REML) analysis in Genstat 5.2 (Genstat 5 Committee, 1993). REML was used to analyse
variance components in a linear mixed model, which was then used to produce means for
varieties, treatments, and analysing their interaction using the unbalanced data.
RESULTS AND DISCUSSION
Genetic variation for frost tolerance
A significant genotype effect (P<0.001) on frost-induced sterility was observed at the Loxton
frost tolerance screening nursery during the 2001 season. The Japanese varieties Amagi Nijo
and Haruna Nijo had significantly lower frost-induced sterility than Australian commercial
varieties (Figure 1). Mapping populations had been developed from the crosses Amagi Nijo ×
WI2585 and Haruna Nijo × Galleon with parents differing significantly in frost-induced
sterility. This result suggested that genetic variation for frost tolerance existed and that the
trait may be segregating in at least two of the Australian developed mapping populations.
Frost induced sterility - Loxton 2001
L.S.D. = 14.3
Figure 1. Frost induced sterility of a subset of lines screened in the 2001 Loxton frost
QTL mapping frost tolerance
Due to maturity effects in the whole Arapiles × Franklin mapping data, a sub set of the
population was selected for QTL analysis. Based on maturity data collected from the
population, a reduced maturity effect was observed when mapping frost tolerance traits using
the sub population. A significant (P<0.001) genotype effect on grain damage and total frost
damage was observed. Frost induced total damage (grain damage and sterility) mapped to
chromosome 5HL with the Arapiles allele contributing to the higher damage score (Table 1).
A significant genotype effect (P<0.001) on frost-induced sterility was observed for both the
Galleon × Haruna Nijo and Amagi Nijo × WI2585 mapping populations. In both populations
frost induced sterility mapped to 5HL and 2HL (Table 1).
Table 1. QTL mapped and associated markers in the Arapiles × Franklin (A × F), Galleon ×
Haruna Nijo (G × HN) and Amagi Nijo × WI2585 (AN × WI) mapping populations.
Trait Population Chromosome Marker LOD R2 Parent
Total Damage A×F 5H AA_CCG282 4.01 0.17 Arapiles
Sterility G × HN 2H HVM54 4.99 0.42 Galleon
Sterility G × HN 5H AWBMA13b 3.54 0.44 Galleon
Sterility (2 July) AN × WI 5H MWH514 5.86 0.32 WI2585
Sterility (21 July) AN × WI 2H HVM54 3.40 0.18 WI2585
Sterility (21 July) AN × WI 5H BCD265a 4.47 0.26 WI2585
Using consensus markers to identify the relative chromosomal position of the frost tolerance
QTL in each population it was identified that the 5H QTL in the three different populations
mapped to the same relative location (Figure 2A). This region has been identified as the
location of the Vrn-H1 locus in winter × spring mapping populations (Laurie et al., 1995) as
well as the location of winter hardiness traits in the Dicktoo × Morex mapping population
(Hayes et al., 1993). The results suggest that this region on chromosome 5HL conditions
frost tolerance of both reproductive and vegetative plant tissue challenging the traditional
view that these traits are independent. The QTL located on chromosome 2HL was strongly
associated with SSR marker HVM54 in both the Galleon × Haruna Nijo and Amagi Nijo ×
WI2585 mapping populations (Figure 2B). No known stress tolerance or developmental
response has been mapped to this location suggesting that the identified locus may condition a
stress response not previously recognized.
5HL A 2HL B
A×F G × HN AN × WI G × HN AN × WI
Figure 2. Chromosome locations of QTL mapped on chromosome 5HL(A) and 2HL(B) in the
Arapiles × Franklin (A × F), Galleon × Haruna Nijo (G × HN) and Amagi Nijo × WI2585
(AN × WI) mapping populations.
Using data collected from the three frost treatments, a significant genotype effect on both
sterility and grain damage (p=<0.001) was observed. There was no significant simulated frost
treatment effect. There was however a significant genotype*treatment interaction
Table 2 – Predicted means of sterility Table 3 – Predicted means of grain
over three frost treatments. damage over three frost treatments.
Genotype Sterility (%) Genotype Grain damage (%)
Beka 25.7 Beka 0.5
Franklin sib 36.4 Haruna Nijo 1.8
Haruna Nijo 37.3 Franklin sib 2.9
ICARDA#66 51.1 ICARDA#66 4.8
Schooner 55.3 Schooner 5.1
ICARDA#81 62.1 ICARDA#81 5.8
ICARDA#70 67.6 ICARDA#70 8.3
s.e.d. 8.3 s.e.d. 3.4
Rankings of field based frost sterility, simulated frost sterility and grain damage were similar
(Figure 1, Tables 2 and 3). However, rankings within each frost treatment changed slightly as
identified by a significant genotype*treatment interaction (data not shown). Confounding
effects such as very high powdery mildew levels may have also contributed to the observed
variation, especially on the very susceptible line ICARDA#81.
From the preliminary results, the frost simulation chamber has potential to be used for routine
frost tolerance screening in barley (and presumably other cereals). However, further study
will be required to determine the level of reproducibility and experimental error within the
Marker assisted selection may provide a more efficient screening method for spring radiation
frost tolerance, enabling higher throughput in early generation screening of large populations
in a breeding program. Frost chamber screening will enable more detailed physiological
measurements to be made on tolerant germplasm to reduce confounding effects such as
maturity, on selection as well as identifying specific mechanisms of frost tolerance.
The identification of genetic variation for frost tolerance in barley and its genetic location
gives potential for targeted approaches, such as functional genomics analysis, to identify exact
gene locations and function. The homoeology of the closely linked vernalisation (Vrn-H1),
frost tolerance loci on chromosome 5H in barley with the vernalisation (Vrn), frost tolerance
(Fr) genes on the group 5 chromosomes in wheat (Galiba et al., 1995) indicates that they may
have a similar function, suggesting genetic variation for spring radiation frost tolerance in
wheat may also exist.
The authors acknowledge the contributions of Mr David Moody and the VDPI barley
breeding field team for trial design and management; Mr Richard Saunders from SARDI
Loxton, for aiding in trial management; The South Australian Barley Improvement Program
field team for their support in field operations; and the GRDC for their financial support
through project UA511.
Zadoks J.C., Chang T.T., Konzak C.F. (1974) A decimal code for the growth stages of
cereals. Weed Res. 14: 415-421.
Hayes P.M., Blake T., Chen T.H.H., Trangoonrung S., Chen F., Pan A., Liu B. (1993)
Quantitative trait loci on barley (Hordeum vulgare L.) chromosome 7 associated with
components of winter hardiness. Genome 36: 66-71.
Laurie D.A., Pratchett N., Bezant J.H., Snape J.W. (1995) RFLP mapping of five major genes
and eight quantitative trait loci controlling flowering time in a winter × spring barley
(Hordeum vulgare L.) cross. Genome 38: 575-585.
Galiba G., Quarrie S.A., Sutka J., Morgounov A., Snape W.J. (1995) RFLP mapping of the
vernalization (Vrn1) and frost resistance (Fr1) genes on chromosome 5A of wheat. Theor.
Appl. Genet. 90: 1174-1179.