Vegetative and reproductive phase change and phenology are economically and ecologically important traits. Trees typically require several years of growth before flowering and once mature, seasonal control of the transition to flowering and flower development is necessary to maintain vegetative meristems and for reproductive success. Members of two related gene subfamilies, FLOWERING LOCUST (FT) and TERMINAL FLOWER1 (TFL1)/CENTRORADIALIS (CEN)/BROTHER OF FT AND TFL1 (BFT), have antagonistic roles in flowering in diverse species and roles in vegetative phenology in trees, but many details of their functions in trees have yet to be resolved. Here, we used CRISPR/Cas9 to generate single and double mutants involving the five Populus FT and TFL1/CEN/BFT genes. ft1 mutants exhibited wild-type-like phenotypes in long days and short days, but after chilling to release dormancy showed delayed bud flush and GA3 could compensate for the ft1 mutation. After rooting and generating some phytomers in tissue culture, both cen1 and cen1ft1 mutants produced terminal as well as axillary flowers, indicating that the cen1 flowering phenotype is independent of FT1. CEN1 showed distinct circannual expression patterns in vegetative and reproductive tissues and comparison with the expression patterns of FT1 and FT2 suggest that the relative levels of CEN1 compared to FT1 and FT2 regulate multiple phases of vegetative and reproductive seasonal development.

Vegetative and reproductive phase change and phenology are economically and ecologically important traits. Trees typically require several years of growth before flowering and once mature, seasonal control of the transition to flowering and flower development is necessary to maintain vegetative meristems and for reproductive success. Members of two related gene subfamilies, FLOWERING LOCUST (FT) and TERMINAL FLOWER1 (TFL1)/CENTRORADIALIS (CEN)/BROTHER OF FT AND TFL1 (BFT), have antagonistic roles in flowering in diverse species and roles in vegetative phenology in trees, but many details of their functions in trees have yet to be resolved. Here, we used CRISPR/Cas9 to generate single and double mutants involving the five Populus FT and TFL1/CEN/BFT genes. ft1 mutants exhibited wild-type-like phenotypes in long days and short days, but after chilling to release dormancy showed delayed bud flush and GA3 could compensate for the ft1 mutation. After rooting and generating some phytomers in tissue culture, both cen1 and cen1ft1 mutants produced terminal as well as axillary flowers, indicating that the cen1 flowering phenotype is independent of FT1. Some axillary meristems initially generated phytomers and in potted plants, the timing of flowering in these shoots correlated with upregulation of FT2 in maturing leaves, suggesting that, in long days, CEN1 antagonizes FT2 promotion of flowering but enables FT2 promotion of shoot growth by maintaining indeterminacy of the shoot apical meristem. CEN1 showed distinct circannual expression patterns in vegetative and reproductive tissues and comparison with the expression patterns of FT1 and FT2 suggest that the relative levels of CEN1 compared to FT1 and FT2 regulate multiple phases of vegetative and reproductive seasonal development.