A laboratory simulation of Arabidopsis seed dormancy cycling provides new insight into its regulation by clock genes and the dormancy-related genes DOG1, MFT, CIPK23 and PHYA

Abstract

Environmental signals drive seed dormancy cycling in the soil to synchronize germination with the optimal time of year, a process essential for species' fitness and survival. Previous correlation of transcription profiles in exhumed seeds with annual environmental signals revealed the coordination of dormancy-regulating mechanisms with the soil environment. Here, we developed a rapid and robust laboratory dormancy cycling simulation. The utility of this simulation was tested in two ways: firstly, using mutants in known dormancy-related genes [DELAY OF GERMINATION 1 (DOG1), MOTHER OF FLOWERING TIME (MFT), CBL-INTERACTING PROTEIN KINASE 23 (CIPK23) and PHYTOCHROME A (PHYA)] and secondly, using further mutants, we test the hypothesis that components of the circadian clock are involved in coordination of the annual seed dormancy cycle. The rate of dormancy induction and relief differed in all lines tested. In the mutants, dog1-2 and mft2, dormancy induction was reduced but not absent. DOG1 is not absolutely required for dormancy. In cipk23 and phyA dormancy, induction was accelerated. Involvement of the clock in dormancy cycling was clear when mutants in the morning and evening loops of the clock were compared. Dormancy induction was faster when the morning loop was compromised and delayed when the evening loop was compromised.

Keywords: circadian clock; circannual rhythm; germination; thermal time.

Figure 1

Induction of secondary dormancy in Cape Verde Island (Cvi) in response to cold stratification and decreasing water potential. Primary dormant Cvi seeds were incubated at 5 °C/dark on water or a range of water potentials from −0.4 to −1.2 MPa. At increasing periods of time, dormancy status was determined by measuring germination following transfer of seeds to (a) water or 10 mm KNO₃ or (b) a buffer control or 250 μ m GA_4 + 7 buffered at pH 5.0 at 20 °C/light for 28 d. Data are mean ± SE (n = 3). Absence of error bars indicates that SE is smaller than the symbol. GA, gibberellins.

Figure 2

Simulated dormancy cycling in Cape Verde Island. Seeds were incubated at 5 °C/dark at −1.2 MPa for up to 21 d before being transferred to water at 25 °C/dark. At increasing periods of time, dormancy status was determined by measuring germination following transfer of seeds to a buffer control, 50 or 250 μ m GA_4 + 7 buffered at pH 5.0 at 20 °C/light for 28 d. Data are mean ± SE (n = 3). Absence of error bars indicates that SE is smaller than the symbol. GA, gibberellins.

Figure 3

Simulated dormancy cycling in Col‐0, Ler and mutants in dormancy‐related and clock genes. Following 5 °C/dark at −1.0 MPa for 28 d, seeds were transferred to water and incubated in the dark at 25 °C for 14 d before transferring to 5 °C/dark. At increasing intervals, dormancy status was determined by measuring germination on water at 25 °C/light for 14 d. (a) Dormancy‐related mutants, (b) circadian clock mutants and (c) CCA1 and LHY overexpressing lines. Data are mean ± SE (n = 3). Absence of error bars indicates that SE is smaller than the symbol.

Figure 4

Seasonal coordination of clock gene transcription in winter (Cvi) and summer annual (Bur) ecotypes. Depth of dormancy in (a) Cvi [time to 50% after‐ripening (AR50)] and (e) Bur [base water potential (Ψb)] with soil temperature at seed depth (data from Footitt et al. 2011, 2013). Transcription profiles of the morning genes CCA1 and LHY and the evening gene TOC1 in (b) Cvi and (f) Bur. Transcription profiles of GI and PRR7 in (c) Cvi and (g) Bur. Transcription profiles of evening complex genes ELF3 and LUX in (d) Cvi and (h) Bur. Cvi, Cape Verde Island; Bur, Burren.

Figure 5

Thermal time analysis of dormancy induction at high temperature following low‐temperature conditioning. (a) Wild type (Col‐0) and the dormancy mutant, dog1‐2, and (b) wild type (Col‐0) and the circadian clock mutants, lhy20 cca1‐1 toc1‐2 and prr5‐11 prr7‐11 prr9‐10. Data from Figs 3, S1 and S2 are replotted against thermal time (sum of temperature above 0 °C) for secondary dormancy induction at 20, 25 and 30 °C. The response to thermal time fits the following relationships: exponential decay (three parameters) regressions describe Col‐0 (R ² = 0.972), lhy20 cca1‐1 toc1‐2 (R ² = 0.897) and prr5‐11 prr7‐11 prr9‐10 (R ² = 0.860), while a linear regression describes dog1‐2 (R = 0.928). The same data for Col‐0 appear in (a) and (b).

Figure 6

ABA sensitivity of dormancy and clock mutants. Following 3 d at 5 °C/dark on water seeds were transferred to ABA (10–250 nm) in buffer at pH 5.0, and cumulative germination was recorded during incubation at 25 °C/light over 14 d. Final germination at each concentration after 14 d (a) and (d). The time to 50% germination (b) and (e) in hours (h) of data in (a) and (d), respectively. Cumulative germination of dormancy mutants in the presence of 100 nm ABA (c). Cumulative germination of clock mutants in the presence of 50 nm ABA (f). Data are mean ± SE (n = 3). Absence of error bars indicates that SE is smaller than the symbol. ABA, abscisic acid.