Transcriptomic response to divergent selection for flowering time in maize reveals convergence and key players of the underlying gene regulatory network

Artificial selection experiments are designed to investigate phenotypic evolution of complex traits and its genetic bases. Here we focused on flowering time, a trait of key importance for plant adaptation and life-cycle shifts. We undertook divergent selection experiments from two maize inbred lines. After 13 generations of selection, we obtained a time-lag of roughly two weeks between Early- and Late- populations. We used this material to characterize the genome-wide transcriptomic response to selection in the shoot apical meristem before, during and after floral transition in field conditions during two consecutive years. We validated the reliability of performing RNA-sequencing in uncontrolled conditions. We found that roughly half of maize genes were expressed in the shoot apical meristem, 59.3% of which were differentially expressed. We detected a majority of genes with differential expression between inbreds and across meristem status, and retrieved a subset of 2,451 genes involved in the response to selection. Among these, we found a significant enrichment for genes with known function in maize flowering time. Furthermore, they were more often shared between inbreds than expected by chance, suggesting convergence of gene expression. We discuss new insights into the expression pattern of key players of the underlying gene regulatory network including the Zea mays genes CENTRORADIALIS (ZCN8), RELATED TO AP2.7 (RAP2.7), MADS4 (ZMM4), KNOTTED1 (KN1), GIBBERELLIN2-OXIDASE1(GA2ox1), as well as alternative scenarios for genetic convergence.