Natural Variation and Timing of Stress Responses Promote and Predict Heterosis

Hybrid plants and animals grow larger and more vigorously than the parents, a common phenomenon known as hybrid vigor or heterosis. Heterosis often correlates with the genetic distance between hybridizing parents, but the mechanism for this is largely unknown. We found that genetic distance was correlated with natural variation of stress responses under the control of the circadian clock. CIRCADIAN-CLOCK-ASSOCIATED1 (CCA1) and LATE-ELONGATED-HYPOCOTYL (LHY) or CCA1 alone mediate expression amplitudes and periods of these stress-responsive genes in stress and non-stress conditions. In Arabidopsis thaliana intraspecific hybrids, genome-wide expression of many biotic and abiotic stress-responsive genes was diurnally repressed to promote biomass heterosis, which is associated with several biomass quantitative trait loci (QTLs). Expression differences in four selected stress-responsive genes among ten ecotypes could be used to predict heterosis in their hybrids. Parent plants with larger expression differences between stress-responsive genes produced higher-vigor hybrids, while those with smaller differences produced lower-vigor hybrids. Stress-responsive genes were epigenetically repressed in the hybrids under normal conditions but induced during times of stress at certain times of the day, balancing the tradeoff between stress responses and growth. Consistently, repressing the stress genes in diploids increased growth vigor. We demonstrate how hybrids manipulate diurnal stress-responsive gene expression to enhance growth vigor. Both circadian and epigenetic regulation play key roles in the altered expression of stress-responsive genes in hybrids. Our findings provide a conceptual advance and mechanistic understanding of heterosis, as well as selection criteria for parents to be effectively used for producing high-yield hybrids. Overall design: Examination of gene expression in Arabidopsis thaliana F1 hybrids between Col and C24 and 3 time points using mRNA-seq