Investigating disease tolerance to Zymoseptoria tritici in wheat

Disease tolerance is defined as the ability to maintain grain yield in the presence of disease and could be a potential defence mechanism to be incorporated into breeding programmes. It is an attractive goal, as disease tolerance has the potential to be a broad-spectrum, durable defence mechanism while exerting little selection pressure on pathogen populations. Relatively little is known about how disease tolerance is conferred, but most of the hypotheses suggest resource capture and resource-use traits such as large green canopy area, increased light extinction coefficient and a high source to sink balance. Disease tolerance in current wheat genotypes is generally associated with low yield potential, and for disease tolerance to be incorporated into commercial breeding it is important to determine whether this link can be disassociated. In this study, an attempt was made to identify physiological traits conferring disease tolerance to Septoria tritici blotch (STB) in winter wheat. Wheat genotypes contrasting in disease tolerance were selected for in-depth phenotyping of selected physiological traits to determine their association with disease tolerance. A number of publications have attempted to link disease tolerance to physiological traits in wheat, based on their yield loss to disease symptom relationship. However, in this study it was proposed that variation in non-symptomatic disease could influence the appearance of disease tolerance which has not previously been investigated. The ratio of in-leaf pathogen biomass to visual disease symptoms was studied in both controlled-environment experiments and in field experiments to determine whether a high in-leaf pathogen biomass was associated with disease tolerance. Two field experiments were conducted during the field seasons 2011/12 and 2013/14 at Teagasc Oak Park, Carlow, Ireland and ADAS Rosemaund, Herefordshire, UK, respectively. A field experiment was also conducted in 2012/13 at Teagasc Oak Park, but due to dry conditions and little disease presence this field experiment was excluded from nearly all experimental analyses. In each experiment, there were two fungicide treatments, non-target disease control and full disease control. In order to increase genetic variability, 38 selected lines from a L14 x Rialto doubled-haploid (DH) mapping population developed by the International Maize and Wheat Improvement Centre (CIMMYT) were screened alongside 10 UK-adapted reference genotypes for contrasting disease tolerance in 2012. Tolerance was quantified as yield loss per unit of green lamina area index (GLAI) loss to disease. L14 is a CIMMYT spring wheat large-ear phenotype advanced line and Rialto is a UK winter wheat which has high radiation-use efficiency and stem soluble carbohydrate. The DH lines displayed an increased range of disease tolerance compared to the UK-adapted reference genotypes. Selected genotypes were subjected to in-depth phenotyping for an extended range of physiological traits in 2014 to identify traits associated with increased disease tolerance. The traits measured included pre- and post- anthesis radiation interception, light extinction coefficient at anthesis, pre- and post anthesis radiation-use efficiency and stem water soluble carbohydrate accumulation at ear emergence + 7 days. In general, there was a wide range of physiological traits displaying weak associations with disease tolerance. The main traits associated with disease tolerance were related to large and/or maintained source capacity in the presence of disease, such as increased GLAI at anthesis and increased post-anthesis light interception. There was also a general association with low grain yield in the absence of disease and decreased harvest index. Increased disease tolerance was associated with high source capacity and low sink capacity, and there was an association between a high source to sink balance, measured as increased Healthy Area Duration (HAD) per grain, and disease tolerance. The impact of genotype variation on the amount of non-symptomatic disease to visual disease expression was investigated in controlled-environment (CE) experiments. In-leaf Zymoseptoria tritici fungal biomass (pathogen load) was quantified by a Real Time qPCR assay targeting the β-tubulin gene (Accession no. AY547264) and compared to visual disease expression. In the first CE experiment, two wheat genotypes were exposed to increasing concentrations of Z. tritici inoculum. There were differences in rates of pathogen development and pathogen presence between inoculum concentrations in both visual disease symptoms and pathogen loads. In the following CE experiment, a wider range of genotypes exposed to a high inoculum level were shown to differ significantly in the relationship between visual disease symptoms and pathogen loads. In order to determine the impact of genotype variation on the visual disease symptoms to pathogen load ratio, flag leaves of genotypes screened for in-field disease tolerance in 2012 and 2014 were analysed. Large variations in the disease symptoms to pathogen load ratio were identified, which has not previously been shown in wheat experiments. An attempt was made to relate the visual symptoms – pathogen load ratio to non-lesion green area loss as a measure of a potential metabolic cost of increased pathogen pressure, but no such relationship was found. An increased pathogen load per unit visual symptoms did not account for larger yield losses than predicted for a given disease level and there was no direct relationship between symptom expression - pathogen load ratios and disease tolerance. The consistency of high/low displays of disease tolerance calculated by different disease measures was investigated using three different ways of measuring disease; HAD, area under disease progress curve (AUDPC) and pathogen DNA quantified by qPCR. In general, the two measures of pathogen presence (AUDPC and pathogen load) tended to quantify disease tolerance similarly, while the HAD-based tolerance contrasted. There were also differences in which traits were associated with disease tolerance for the different methods of calculating tolerance; the calculations based on AUDPC and pathogen DNA tended to associate a decreased source capacity to disease tolerance while the HAD-based tolerance indicated an association with increased source capacity. All methods, however, indicated that a low yield potential was associated with disease tolerance. In conclusion, there was a large range of disease tolerance found in the field experiments compared to previous investigations. The HAD-based disease tolerance seems to be mainly related to a large source capacity and a low sink capacity. However, the genotype ratings of high/low disease tolerance and associated physiological traits seem to vary according to the method of calculating tolerance. There were large differences in the ratio of visual symptoms-pathogen load between genotypes; even though this did not have a direct impact on disease tolerance or yield loss it could potentially be associated with increased metabolic costs