Greg J. Rebetzke

Breeding for improved water productivity in temperate cereals: phenotyping, quantitative trait loci, markers and the selection environment.

Consistent gains in grain yield in dry environments have been made by empirical breeding although there is disturbing evidence that these gains may have slowed. There are few examples where an understanding of the physiology and the genetics of putative important drought-related traits has led to improved yields. Success will first depend on identifying the most important traits in the target regions. It will then depend on accurate and fast phenotyping, which, in turn, will lead to: (1) trait-based selection being immediately transferable into breeding operations and (2) being able to identify the underlying genes or the important genomic regions (quantitative trait loci), perhaps leading to efficient marker-based selection (MBS). Genetic complexity, extent of genotypeenvironment (GE) interaction and sampling cost per line will determine value of phenotyping over MBS methods. Here, we review traits of importance in dry environments and review whether molecular or phenotypic selection methods are likely to be the most effective in crop improvement programs and where the main bottlenecks to selection are. We also consider whether selection for these traits should be made in dry environments or environments where there is no soil water limitation. The development of lines/ populations for trait validation studies and for varietal development is also described. We firstly conclude that despite the spectacular improvements in molecular technologies, fast and accurate phenotyping remains the major bottleneck to enhancing yield gains in water-limited environments. Secondly, for most traits of importance in dry environments, selection is generally conducted most effectively in favourable moisture environments.

Genotypic variation in water-soluble carbohydrate accumulation in wheat

The water-soluble carbohydrate (WSC) that accumulates in the stems of wheat during growth can be an important contributor to grain filling, particularly under conditions when assimilation is limited, such as during end-of-season drought. WSC concentration was measured at anthesis across a diverse set of wheat genotypes over multiple environments. Environmental differences in WSC concentration were large (means for the set ranging between 108 and 203 mg g-1 dry weight), and there were significant and repeatable differences in WSC accumulation among genotypes (means ranging from 112 to 213 mg g-1 dry weight averaged across environments), associated with large broad-sense heritability (H = 0.90 ± 0.12). These results suggest that breeding for high WSC should be possible in wheat. The composition of the WSC, examined in selected genotypes, indicated that the variation in total WSC was attributed mainly to variation in the fructan component, with the other major soluble carbohydrates, sucrose and hexose, varying less. The degree of polymerisation (DP) of fructo-oligosaccharides was up to ~13 in samples where higher levels of WSC were accumulated, owing either to genotype or environment, but the higher DP components (DP > 6) were decreased in samples of lower total WSC. The results are consistent with fructan biosynthesis occurring via a sequential mechanism that is dependent on the availability of sucrose, and differences in WSC contents of genotypes are unlikely to be due to major mechanistic differences.

Implementation of markers in Australian wheat breeding

Genetic associations of morphological, biochemical, and DNA markers with economically important traits can be used for indirect selection of the traits. Chromosomal linkage between pseudo-black chaff and the stem-rust resistance gene Sr2, and between the red glume gene (Rg1) and the stripe rust resistance gene Yr10, have been used in this way for many years. Similarly, linkages between disease resistance genes, such as Sr38, Lr37, and Yr17, have been used to achieve resistance to multiple diseases while selection is performed for resistance to one disease. Alleles at the Glu loci, assessed as protein differences, have been used as predictors of dough strength. More recently, DNA markers have been developed and used, especially to select for resistance to cereal cyst nematode, a trait which is difficult and expensive to assess with conventional bioassays. We found that the major use of DNA markers was for selection for traits of substantial economic importance, which were primarily determined by a single gene, and where the non-marker assay was expensive and unreliable. The other uses of markers were for pyramiding several genes influencing one trait, or for rapid backcrossing.

The effect of different height reducing genes on the early growth of wheat

Genes that reduce height without compromising seedling vigour or coleoptile length have great potential for wheat improvement. We therefore investigated the effects of various reduced height (Rht) genes on the early stages of plant development, using a combination of near isogenic, recombinant, mutant and wild type comparisons. Gibberellin (GA) insensitivity caused by Rht-B1b or Rht-D1b was associated with reduced leaf elongation rate and coleoptile length. Similar results were found for two other sources of dwarfing, Rht11 and Rht17. We found one class of Rht genes (e.g. Rht8) which had no effect on coleoptile length, leaf elongation rate or responsiveness to GA, indicating that these dwarfing genes may act later in wheat development to reduce height and increase harvest index, without affecting early growth. A third class of Rht genes was found in three durum backgrounds. These had reduced coleoptile lengths and leaf elongation rates, but had a greater response to GA than the corresponding tall varieties. We discuss these results in relation to the possible mechanisms underlying the reduction in height and the suitability of the different Rht genes for wheat improvement.

Large root systems: are they useful in adapting wheat to dry environments?

There is little consensus on whether having a large root system is the best strategy in adapting wheat (Triticum aestivum L.) to water-limited environments. We explore the reasons for the lack of consensus and aim to answer the question of whether a large root system is useful in adapting wheat to dry environments. We used unpublished data from glasshouse and field experiments examining the relationship between root system size and their functional implication for water capture. Individual root traits for water uptake do not describe a root system as being large or small. However, the recent invigoration of the root system in wheat by indirect selection for increased leaf vigour has enlarged the root system through increases in root biomass and length and root length density. This large root system contributes to increasing the capture of water and nitrogen early in the season, and facilitates the capture of additional water for grain filling. The usefulness of a vigorous root system in increasing wheat yields under water-limited conditions maybe greater in environments where crops rely largely on seasonal rainfall, such as the Mediterranean-type environments. In environments where crops are reliant on stored soil water, a vigorous root system increases the risk of depleting soil water before completion of grain filling.

Genomic regions for canopy temperature and their genetic association with stomatal conductance and grain yield in wheat

Stomata are the site of CO2 exchange for water in a leaf. Variation in stomatal control offers promise in genetic improvement of transpiration and photosynthetic rates to improve wheat performance. However, techniques for estimating stomatal conductance (SC) are slow, limiting potential for efficient measurement and genetic modification of this trait. Genotypic variation in canopy temperature (CT) and leaf porosity (LP), as surrogates for SC, were assessed in three wheat mapping populations grown under well-watered conditions. The range and resulting genetic variance were large but not always repeatable across days and years for CT and LP alike. Leaf-to-leaf variation was large for LP, reducing heritability to near zero on a single-leaf basis. Replication across dates and years increased line-mean heritability to ~75% for both CT and LP. Across sampling dates and populations, CT showed a large, additive genetic correlation with LP (rg=-0.67 to -0.83) as expected. Genetic increases in pre-flowering CT were associated with reduced final plant height and both increased harvest index and grain yield but were uncorrelated with aerial biomass. In contrast, post-flowering, cooler canopies were associated with greater aerial biomass and increased grain number and yield. A multi-environment QTL analysis identified up to 16 and 15 genomic regions for CT and LP, respectively, across all three populations. Several of the LP and CT QTL co-located with known QTL for plant height and phenological development and intervals for many of the CT and LP quantitative trait loci (QTL) overlapped, supporting a common genetic basis for the two traits. Notably, both Rht-B1b and Rht-D1b dwarfing alleles were paradoxically positive for LP and CT (i.e. semi-dwarfs had higher stomatal conductance but warmer canopies) highlighting the issue of translation from leaf to canopy in screening for greater transpiration. The strong requirement for repeated assessment of SC suggests the more rapid CT assessment may be of greater value for indirect screening of high or low SC among large numbers of early-generation breeding lines. However, account must be taken of variation in development and canopy architecture when interpreting performance and selecting breeding lines on the basis of CT.

A multisite managed environment facility for targeted trait and germplasm phenotyping

Field evaluation of germplasm for performance under water and heat stress is challenging. Field environments are variable and unpredictable, and genotype×environment interactions are difficult to interpret if environments are not well characterised. Numerous traits, genes and quantitative trait loci have been proposed for improving performance but few have been used in variety development. This reflects the limited capacity of commercial breeding companies to screen for these traits and the absence of validation in field environments relevant to breeding companies, and because little is known about the economic benefit of selecting one particular trait over another. The value of the proposed traits or genes is commonly not demonstrated in genetic backgrounds of value to breeding companies. To overcome this disconnection between physiological trait breeding and uptake by breeding companies, three field sites representing the main environment types encountered across the Australian wheatbelt were selected to form a set of managed environment facilities (MEFs). Each MEF manages soil moisture stress through irrigation, and the effects of heat stress through variable sowing dates. Field trials are monitored continuously for weather variables and changes in soil water and canopy temperature in selected probe genotypes, which aids in decisions guiding irrigation scheduling and sampling times. Protocols have been standardised for an essential core set of measurements so that phenotyping yield and other traits are consistent across sites and seasons. MEFs enable assessment of a large number of traits across multiple genetic backgrounds in relevant environments, determine relative trait value, and facilitate delivery of promising germplasm and high value traits into commercial breeding programs.

Longer coleoptiles improve emergence through crop residues to increase seedling number and biomass in wheat (Triticum aestivum L.)

Crop residues protect soils from erosion, reduce soil water evaporation and increase soil organic matter. Yet management of stubbles for cropping can be difficult. Surface-retained residue can act as a mechanical barrier to slow emergence and reduce seedling biomass. Longer coleoptiles improve seedling emergence with deep sowing and may assist where stubble load is large. In a glasshouse study, six wheat and barley genotypes were sown at 30 and 50 mm depth into pots containing pasteurised soil. Unweathered sorghum, canola and wheat stubble were added at 0, 3 and 6 t/ha equivalents to the soil surface and pots watered above or below the stubble. Stubble species and watering regime had little effect on seedling growth. However, deeper sowing and increased stubble mass adversely affected most seedling characteristics particularly slowing seedling emergence and reducing tiller number to decrease plant biomass (environmental correlations (re) of −0.98** and 0.88**, respectively). Shorter coleoptile Rht-B1b wheats ‘Banks’ and ‘Janz’, and barley ‘Beecher’ emerged slower and abnormally with thicker stubble, and had more sterile tillers to reduce total tiller number and biomass. Deeper crowns for these genotypes also resulted in proportionally less biomass located above the stubble. The converse was true of long coleoptile ‘Vigour 18’, ‘Halberd’, and its Rht8 progeny, ‘HM14bS’ which were less affected by stubble mass and sowing depth. In a corresponding field study, increasing wheat stubble mass from 0 to 3 and 6 t/ha delayed seedling emergence and decreased plant number to reduce biomass. Short coleoptile wheat genotypes ‘Hartog’ and ‘Janz’ emerged slower and produced less biomass at 3 and 6 t/ha of stubble than long coleoptile wheat genotypes ‘Halberd’ and ‘HM14bS’. Emergence of seedlings sown at 50 mm depth with 6 t/ha overlying stubble was similar to that sown at 120 mm with no stubble, reflecting the similar impact of retained residues to deep sowing. Genetic variation for coleoptile length and availability of gibberellin-responsive dwarfing genes such as Rht8 will allow development of long coleoptile wheats for deep sowing or where stubble retention is practiced.

Guiding deployment of resistance in cereals using evolutionary principles.

Genetically controlled resistance provides plant breeders with an efficient means of controlling plant disease, but this approach has been constrained by practical difficulties associated with combining many resistance genes together and strong evolutionary responses from pathogen populations leading to subsequent resistance breakdown. However, continuing advances in molecular marker technologies are revolutionizing the ability to rapidly and reliably manipulate resistances of all types – major gene, adult plant and quantitative resistance loci singly or multiply into individual host lines. Here, we argue that these advances provide major opportunities to deliberately design deployment strategies in cereals that can take advantage of the evolutionary pressures faced by target pathogens. Different combinations of genes deployed either within single host individuals or between different individuals within or among crops, can be used to reduce the size of pathogen populations and generate patterns of disruptive selection. This will simultaneously limit immediate epidemic development and reduce the probability of subsequent evolutionary change in the pathogen for broader infectivity or increased aggressiveness. The same general principles are relevant to the control of noncereal diseases, but the most efficacious controls will vary reflecting the range of genetic options available and their fit with specific ecology and life‐history combinations.

Genetic control of duration of pre-anthesis phases in wheat (Triticum aestivum L.) and relationships to leaf appearance, tillering, and dry matter accumulation

The duration of pre-anthesis developmental phases is of interest in breeding for improved adaptation and yield potential in temperate cereals. Yet despite numerous studies on the genetic control of anthesis (flowering) time and floral initiation, little is known about the genetic control of other pre-anthesis phases. Furthermore, little is known about the effect that changes in the duration of pre-anthesis phases could have on traits related to leaf appearance and tillering, or dry matter accumulation before terminal spikelet initiation (TS). The genetic control of the leaf and spikelet initiation phase (LS; from sowing to TS), the stem elongation phase (SE; from TS to anthesis), and, within the latter, from TS to flag leaf appearance and from then to anthesis, was studied in two doubled-haploid, mapping bread wheat populations, Cranbrook×Halberd and CD87×Katepwa, in two field experiments (ACT and NSW, Australia). The lengths of phases were estimated from measurements of both TS and the onset of stem elongation. Dry weight per plant before TS, rate of leaf appearance, tillering rate, maximum number of tillers and number of leaves, and dry weight per plant at TS were also estimated in the Cranbrook×Halberd population. More genomic regions were identified for the length of the different pre-anthesis phases than for total time to anthesis. Although overall genetic correlations between LS and SE were significant and positive, independent genetic variability between LS and SE, and several quantitative trait loci (QTLs) with different effects on both phases were found in the two populations. Several of these QTLs (which did not seem to coincide with reported major genes) could be of interest for breeding purposes since they were only significant for either LS or SE. There was no relationship between LS and the rate of leaf appearance. LS was strongly and positively correlated with dry weight at TS but only slightly negatively correlated with early vigour (dry weight before TS). Despite significant genetic correlations between LS and some tillering traits, shortening LS so as to lengthen SE without modifying total time to anthesis would not necessarily reduce tillering capacity, as QTLs for tillering traits did not coincide with those QTLs significant only for LS or SE. Therefore, the study of different pre-anthesis phases is relevant for a better understanding of genetic factors regulating developmental time and may offer new tools for fine-tuning it in breeding for both adaptability and yield potential.