Temperature cues and molecular control of masting.

Mast flowering (masting) is synchronised highly variable flowering by a population of perennial plants over a wide geographical area. Masting causes widespread community effects, including in New Zealand negative impacts on threatened birds, therefore prediction of masting events is important. The ΔT model says that flowering effort is well predicted by a positive temperature difference in successive summers, but the molecular mechanism which could measure a temperature difference between years is unknown. The aim of this study was to identify flowering-time regulatory genes that might respond to the ∆T cue and initiate flowering in two masting species, Celmisia lyallii (Asteraceae) and Chionochloa pallens (Poaceae). Translocation experiments were performed moving C. lyallii and C. pallens plants to higher or lower altitudes, thus manipulating ∆T cues in order to either induce or repress flowering. Leaf samples were collected over time and retrospectively classified by whether the plants subsequently flowered or remained vegetative. Genome-wide transcriptomic and targeted gene expression analyses were applied to identify changes in the expression of the flowering- time control genes associated with the floral induction. RNA-seq analysis identified ~ 45% of the orthologous flowering-time genes known from model plants in C. lyallii and C. pallens, suggesting the conservation of the flowering network in these wild alpine plants. Translocated plants exposed to a large positive ΔT flowered heavily, consistent with the ΔT model. Structural and genetic analysis identified a novel homologue of TFL1 (CpATFL1, an “anti-TFL1”) that promotes flowering in arabidopsis, in contrast to its repressive activity in model plant species. Gene expression analysis suggested that ClSOC1 and CpATFL1 promoted flowering in C. lyallii and C. pallens¸ respectively, in response to the ∆T cue. Additionally, potential floral promoters were identified that responded to the temperature signal and which may then upregulate ClSOC1 and CpATFL1 including PIF-family genes, miR172 and several autonomous pathway genes including FLK, FY and LD. Induction of flowering was also associated with the suppression of known floral repressors in both C. lyallii and C. pallens, including SVP, AP2-family genes, VRN2 and FRI. Increased activity of epigenetic genes including REF6, FLD, HDA6 and MSI1 in response to ∆T suggests a potential mechanism to remember temperature signal for two years. These could then either maintain the expression of floral promoters or suppress floral repressors over the period to induce masting. Further, the differential summer temperature led to differential regulation in the sugar and hormonal signaling pathways which may lead to the suppression of floral repressors leading to the initiation of flowering. This research has provided a mechanistic understanding of floral induction in masting plants and evidence of their ability to imprint various environmental cues in the form of a ‘plastic memory’. Additionally, molecular field studies suggested the ∆T model could be improved by including temperatures three years before flowering (the “∆T3” model) in C. lyallii and C. pallens. This modified ∆T model was supported by analysis of observational data from the field over the past 25 years. These results support and expand the ΔT model of mast flowering in two species and should allow better prediction of masting events, thereby assisting conservation programs for the protection of endangered fauna