IntroductionFrost events lead to A$360 million of yield losses annually to the Australian wheat industry, making improvement of chilling and frost tolerance an important trait for breeding.ObjectivesThis study aimed to use metabolomics and lipidomics to explore genetic variation in acclimation potential to chilling and to identify metabolite markers for chilling tolerance in wheat.MethodsWe established a controlled environment screening assay that is able to reproduce field rankings of wheat germplasm for chilling and frost tolerance. This assay, together with targeted metabolomics and lipidomics approaches, were used to compare metabolite and lipid levels in flag leaves of two wheat varieties with contrasting chilling tolerance.ResultsThe sensitive variety Wyalkatchem showed a strong reduction in amino acids after the first cold night, followed by accumulation of osmolytes such as fructose, glucose, putrescine and shikimate over a 4-day period. Accumulation of osmolytes is indicative of acclimation to water stress in Wyalkatchem. This response was not observed for tolerant variety Young. The two varieties also displayed significant differences in lipid accumulation. Variation in two lipid clusters, resulted in a higher unsaturated to saturated lipid ratio in Young after 4 days cold treatment and the lipids PC(34:0), PC(34:1), PC(35:1), PC(38:3), and PI(36:4) were the main contributors to the unsaturated to saturated ratio change. This indicates that Young may have superior ability to maintain membrane fluidity following cold exposure, thereby avoiding membrane damage and water stress observed for Wyalkatchem.ConclusionOur study suggests that metabolomics and lipidomics markers could be used as an alternative phenotyping method to discriminate wheat varieties with differences in cold acclimation.Electronic supplementary materialThe online version of this article (10.1007/s11306-019-1606-2) contains supplementary material, which is available to authorized users.

The number of rachis nodes (spikelets) on a wheat spike is a component of grain yield that correlates with flowering time. The genetic basis regulating flowering in cereals is well understood, but there are reports that flowering time can be modified at a high frequency by selective breeding, suggesting that it may be regulated by both epigenetic and genetic mechanisms. We investigated the role of DNA methylation in regulating spikelet number and flowering time by treating a semi-spring wheat with the demethylating agent, Zebularine. Three lines with a heritable increase in spikelet number were identified. The molecular basis for increased spikelet number was not determined in 2 lines, but the phenotype showed non-Mendelian inheritance, suggesting that it could have an epigenetic basis. In the remaining line, the increased spikelet phenotype behaved as a Mendelian recessive trait and late flowering was associated with a deletion encompassing the floral promoter, FT-B1. Deletion of FT-B1 delayed the transition to reproductive growth, extended the duration of spike development, and increased spikelet number under different temperature regimes and photoperiod. Transiently disrupting DNA methylation can generate novel flowering behaviour in wheat, but these changes may not be sufficiently stable for use in breeding programs.

Water stress during reproductive growth is a major yield constraint for wheat ( Triticum aestivum L). We previously established a controlled environment drought tolerance phenotyping method targeting the young microspore stage of pollen development. This method eliminates stress avoidance based on flowering time. We substituted soil drought treatments by a reproducible osmotic stress treatment using hydroponics and NaCl as osmolyte. Salt exclusion in hexaploid wheat avoids salt toxicity, causing osmotic stress. A Cranbrook x Halberd doubled haploid (DH) population was phenotyped by scoring spike grain numbers of unstressed (SGNCon) and osmotically stressed (SGNTrt) plants. Grain number data were analyzed using a linear mixed model (LMM) that included genetic correlations between the SGNCon and SGNTrt traits. Viewing this as a genetic regression of SGNTrt on SGNCon allowed derivation of a stress tolerance trait (SGNTol). Importantly, and by definition of the trait, the genetic effects for SGNTol are statistically independent of those for SGNCon. Thus they represent non-pleiotropic effects associated with the stress treatment that are independent of the control treatment. QTL mapping was conducted using a whole genome approach in which the LMM included all traits and all markers simultaneously. The marker effects within chromosomes were assumed to follow a spatial correlation model. This resulted in smooth marker profiles that could be used to identify positions of putative QTL. The most influential QTL were located on chromosome 5A for SGNTol (126cM; contributed by Halberd), 5A for SGNCon (141cM; Cranbrook) and 2A for SGNTrt (116cM; Cranbrook). Sensitive and tolerant population tail lines all showed matching soil drought tolerance phenotypes, confirming that osmotic stress is a valid surrogate screening method.