Phytochrome B and REVEILLE1/2-mediated signalling controls seed dormancy and germination in Arabidopsis

Introduction

Seeds mediate the alternation of generations in flowering plants and have been a staple food throughout human civilization. Primary seed dormancy is acquired during seed maturation and reaches a high level in freshly harvested seeds and maintained for a certain period that allows seeds to survive under unfavourable conditions and prevents pre-harvest sprouting, and is thus an important aspect of plant fitness1,2,3. Under optimal conditions, the release of dormancy by after-ripening and the successful germination and establishment of a robust seedling are critical for the propagation of the plant species4,5. Dormancy and germination are two distinct but closely connected physiological processes. The dormancy-to-germination transition is a critical developmental step in the life cycle of plants that is determined by both genetic factors and environmental influences2,3,4. The phytohormones gibberellin (GA) and abscisic acid (ABA) primarily and antagonistically regulate the seed status; GA represses dormancy and promotes germination, whereas ABA has the opposite effects. The signalling pathways that GA and ABA control seed germination have been extensively studied2,6,7,8,9.

Previous genetic analyses have revealed many regulators that affect the induction, maintenance and release of seed dormancy2,3. Studies of natural variation have identified multiple quantitative trait loci (QTL) that contribute to dormancy in wild populations of Arabidopsis thalianaand some crops10,11,12,13,14,15. Among them, DELAY OF GERMINATION 1(DOG1) is a major QTL in a recombinant inbred line population ofArabidopsis10,16. DOG1 protein levels in freshly harvested dry seeds strongly correlate with the time required for after-ripening17. A recent study showed that DOG1 regulates primary seed dormancy through a microRNA pathway18. However, the molecular function and regulation ofDOG1 remain elusive.

Light is a major environmental signal that oppositely modulates the levels of GA and ABA, and affects seed germination8,19,20. Among plant photoreceptors, the red and far-red-light-absorbing phytochromes are essential for light promotion of germination21. In the model speciesArabidopsis, five genes (PHYA to PHYE) encode phytochrome apoproteins22,23. Phytochrome B (phyB) predominantly triggers red/far-red-light-reversible seed germination, whereas phyA mediates distinct, very low fluence responses in red and far-red light24,25,26,27,28,29,30,31. phyA- and phyB-dependent induction of germination are spatially separated in the endosperm and embryo32. phyE is required for germination in continuous far-red light33. A recent study shows that phyE and phyD stimulate germination at very low red/far-red ratios and, surprisingly, phyC antagonizes the promotion of germination by light34. At the molecular level, far-red light converts phyochromes into the inactive Pr form, which inhibits seed germination, whereas a subsequent red-light pulse reverts it to active Pfr and induces germination22,35. Light-activated phyB interacts with and promotes the degradation of a negative regulator, PHYTOCHROME-INTERACTING FACTOR 1 (PIF1, also known as PIL5)36,37. PIF1 directly regulates the expression of several downstream genes, including GA-INSENSITIVE,REPRESSOR OF GA1-3 and SOMNUS, which modulates GA responsiveness, GA and ABA biosynthesis and subsequent seed germination38,39,40,41. PIF6 was previously shown to regulate the primary seed dormancy42. Although phytochromes are involved in regulating seed dormancy21,28,43, the underlying molecular mechanism was hitherto unknown.

REVEILLE1 (RVE1) belongs to a subfamily of Myb-like transcription factors that includes CIRCADIAN CLOCK-ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL clock components44,45. RVE1 regulates hypocotyl growth by integrating the circadian clock and auxin pathways44. We previously revealed that RVE1 modulates chlorophyll biosynthesis and seedling de-etiolation in Arabidopsis46. In this study, we identified RVE1 and RVE2 as activators of seed dormancy and further provided molecular and genetic evidence to show that RVE1 and RVE2 promote primary seed dormancy and repress red/far-red-light-mediated germination downstream of phyB in A. thaliana. We found that the transcription of RVE1, RVE2 and DOG1 is reduced by phyB-Pfr, the active form of phyB. We also demonstrate that RVE1 directly inhibitsGIBBERELLIN 3-OXIDASE 2 (GA3ox2) transcription and subsequently suppresses bioactive GA biosynthesis. Therefore, we reveal a genetic pathway that links light input with internal factors to control seed dormancy and germination that can potentially optimize seed adaptability to changing environments.

Results

RVE1 and RVE2 regulate seed dormancy and germination

We first examined the role of RVE1 in light-induced seed germination using 2- to 5-month post-harvest seeds. The after-ripening seeds were exposed to different light conditions without cold stratification (Fig. 1a). The phyB photoreceptor positively controls red/far-red reversible seed germination25,26,27. As shown in Fig. 1b,c, under phyB-off conditions (grown in darkness interrupted by a 5-min pulse of far-red light to inactivate phyB), Columbia (Col) wild-type seeds did not germinate, whereas the germination frequency of rve1-null mutant (Supplementary Fig. 1a,b) was close to 40%. Under darkness or phyB-on conditions (grown in darkness interrupted by a 5-min pulse of far-red light followed by 5-min pulse of red light to activate phyB), 100% of the rve1seeds germinated, similar to Col. Remarkably, RVE1 overexpression (RVE1-OX) transgenic seeds failed to germinate under all conditions (Fig. 1b,c). It was further supported by a time-course germination assay (Supplementary Fig. 1c). The pif1 seeds, which were used as a control, germinated under all conditions tested, as reported previously (Fig. 1b,c)38. These observations suggest that RVE1 negatively regulates phyB-mediated seed germination.

Figure 1: RVE1 regulates both seed dormancy and germination.

(a) Light irradiation treatment in the experiments. Post-harvest seeds were irradiated with white light (WL) for 1 h (starting from seed sterilization) and were then exposed to far-red (FR) light for 5 min (phyB-off) or followed by 5 min of red (R) light (phyB-on). Seeds were then kept in darkness and germination frequencies were recorded after 4 days. (b) Quantification of the germination frequencies of seeds under different conditions as shown in a. (c) Representative images of seed germination assays of Col, rve1 and RVE1-OX seeds under the light conditions shown in a. (d–g) Percentage of seed germination. Freshly harvested seeds were kept in darkness (d) or under white light (e) for 3 days, or seeds were stratified at 4 °C for 1 day (f) or 3 days (g) in darkness before being exposed to 3 days of white light treatment at 22 °C. (h) Germination percentage of post-harvest seeds of rve1,rve2 and rve1/rve2 grown under the phyB-off condition. (i) Dormancy phenotype of freshly harvested seeds of rve1, rve2 and rve1/rve2 grown in darkness for 4 days. For b and d–i, mean±s.d., n=3.