Genetics of resistance to ear diseases and mycotoxin accumulation in the pathosystems maize/Fusarium and wheat/Fusarium

摘要

Infection of ears of maize with Fusarium graminearum (FG) reduces yield and, more important, contaminate the harvest with mycotoxins. F. verticillioides (FV) is an economically important cause of ear rot. Among other mycotoxins, FV produces the fumonisins (FUM) and FG produces deoxynivalenol (DON) and zearalenone (ZEA). All three mycotoxins are harmful to humans and animals. Therefore, the European Union released legally enforceable limits. One alternative to reduce ear rot severity and mycotoxin concentrations is breeding and growing varieties resistant to Fusarium infections. However, few is known about breeding parameters for resistance to Fusarium infections and mycotoxin accumulation in European maize breeding material. The main objective of this thesis was to draw conclusions for breeding of resistance to ear rot and mycotoxin accumulation with special attention on three European maize maturity groups. We investigated methodical aspects like (1) the comparison of natural and artificial inoculation to evaluate ear rot resistance and (2) the necessity of separate testing of FV and FG. Furthermore, quantitative-genetic parameters like heritabilities and correlations were estimated to draw conclusions about (3a) genetic variation in line and testcross performance and the relationships (3b) between ear rot severity and mycotoxin concentrations in lines and testcrosses and (3c) between line and testcross performance. Three maturity groups (early, mid-late, late) each comprising about 150 maize inbred lines were evaluated for ear rot resistance to FV. The same genotypes of the early maturity group were additionally evaluated for resistance to FG in separate, but adjacent trials. Field evaluation was conducted in two to six environments with silk channel inoculation and natural infection, respectively. In the late maturity group kernel inoculation was conducted additionally. Out of the 150 lines, 50 to 60 lines per maturity group were crossed with two unrelated testers of the opposite heterotic group. The concentrations of toxins FUM, DON and ZEA of the chosen lines and their testcrosses were analyzed by immunotests. Despite significant genotypic differences among the inbred lines after inoculation or natural infections, inoculation was found to be superior due to easier visual differentiation and increased accuracy. Therefore, inoculation should be conducted. In the late maturity group silk channel inoculation (simulating infection over the silks) and kernel inoculation (simulating secondary infection after wounding) were appropriate since both caused similar ear rot severity. However, both inoculation methods should be tested separately due to only moderate correlations between them. In the early maturity group resistance to FG or FV should be tested separately due to moderate correlations. Significant genotypic variances in large sets and subsets of lines and also in testcrosses revealed that there is genetic variation in all maturity groups and also within heterotic groups. In the flint group less lines were resistant to FV and FG than in dents indicating that resistance needs improvement, i.e. by introgression of resistance alleles followed by recurrent selection. Significant genotype x environment interactions may complicate selection and, therefore, multi-environmental trials are required for an accurate selection. High genotypic correlations between ear rot rating and mycotoxin concentrations were found among lines and testcrosses. The cost efficient indirect selection for mycotoxin concentrations based on ear rot rating could increase response to selection by testing more genotypes and/or in more test environments assuming a fixed budget. This should increase selection intensity and/or heritability. Moderate genotypic correlations between line and testcross performance were. One moderately to highly susceptible tester is sufficient due to high genotypic correlations between testcrosses of different testers. Both indicates a mainly additive gene action, but also non-additive gene action may play a role in some crosses. Selection for testcross performance based on line performance was less effective when calculating relative efficiencies. Different scenarios have been identified: (1) In Central Europe mainly resistance to ear rot in lines needs to be tested to ensure high seed quality, whereas resistance in testcrosses is not important due to low natural infection. (2) In Southern Europe, where high natural infections occur regularly, parallel selection for resistance to ear rot in lines and testcrosses is important. One susceptible tester should be used for creation of testcrosses. For selection in lines all parental lines should be inoculated but only lines selected out of testcrosses for agronomic traits would be rated afterwards saving resources. This is feasible due to later harvest date of lines than of testcrosses.