DOG1 controls dormancy independently of ABA core signaling kinases regulation by preventing AFP dephosphorylation through AHG1

Abstract

Seed dormancy determines germination timing, influencing seed plant adaptation and overall fitness. DELAY OF GERMINATION 1 (DOG1) is a conserved central regulator of dormancy cooperating with the phytohormone abscisic acid (ABA) through negative regulation of ABA HYPERSENSITIVE GERMINATION (AHG) 1 and AHG3 phosphatases. The current molecular mechanism of DOG1 signaling proposes it regulates the activation of central ABA-related SnRK2 kinases. Here, we unveil DOG1's functional autonomy from the regulation of ABA core signaling components and unravel its pivotal control over the activation of ABSCISIC ACID INSENSITIVE FIVE BINDING PROTEINs (AFPs). Our data revealed a molecular relay in which AFPs' genuine activation by AHG1 is contained by DOG1 to prevent the breakdown of maturation-imposed ABA responses independently of ABA-related kinase activation status. This work offers a molecular understanding of how plants fine-tune germination timing, while preserving seed responsiveness to adverse environmental cues, and thus represents a milestone in the realm of conservation and breeding programs.

Fig. 1.. Proteome and phosphoproteome analyses of freshly harvested NIL-DOG1 and dog1-1 seeds in the dry and 6-hour water imbibed state.

(A) Schematic workflow of the proteome analysis. Peptides for proteome and phosphoproteome analyses were prepared from the same seed extract and analyzed in parallel. (B) Germination capacity of NIL-DOG1 and dog1-1 seeds at the sampling time of the proteomic experiments (means ± SD, n = 3 biological replicates, unpaired t test P value is indicated). Scale bars, 2 mm. (C and D) Volcano plots showing the regulation of (C) protein group abundance or (D) phosphopeptide abundance in dry (left) and 6-hour imbibed (right panel) seeds of dog1-1 compared to NIL-DOG1. (E) Number of protein groups and (F) p-peptides with significant changes (P value < 0.05) in abundance between genotypes. In (C) to (F), down- and up-regulation are shown in blue and red, respectively. (G) GO enrichment for biological processes in dry or imbibed (imb.) seeds of proteins significantly regulated between genotypes. (H) GO biological process (top) and cellular component enrichment (bottom) of all (up- and down-regulated) phosphoproteins with significantly regulated p-peptide abundances either in dry, imbibed, or in both conditions. h, hours.

Fig. 2.. ABA core signaling is primed in dog1-1 seeds.

(A) Schematic representation of SnRK2III protein (using SnRK2.6 as an example) and of amino acid sequence alignment of Arabidopsis SnRK2s’ activation loops. The tryptic p-peptide more abundant in dog1-1 compared to NIL-DOG1 is highlighted in red. Phosphorylation of SnRK2III activation loop is mediated by RAF kinases. (B) MS² spectrum of the tryptic peptide phosphorylated at S171 (in SnRK2.6). m/z, mass/charge ratio. (C) MS¹ intensities of the SnRK2III activation loop p-peptide in dry NIL-DOG1 and dog1-1 seeds (means ± SD, n = 3 biological replicates). (D) Western blot quantification of SnRK2IIIs in dry NIL-DOG1 and dog1-1 seeds. The bar chart shows protein SnRK2III accumulation normalized to the histone 3 (H3) loading control (means ± SD, n = 4 biological replicates). (E) Phosphorylation of SnRK2III activation loop between dry NIL-DOG1 and dog1-1 seeds after normalization of peptide intensity to the protein abundance change (means ± SD, n = 3 biological replicates). (F) Kinase activity of YFP-SnRK2.6 immunoprecipitated from dry seed of NIL-DOG1 and dog1-1 background. YFP-SnR2.6–bound beads were used to phosphorylate MBP in vitro in the presence of ³²P γ-ATP. The phosphorylation of MBP with radiolabeled phosphate was assessed using autoradiography (AR), loading of MBP was evaluated by Coomassie Brilliant Blue (CBB) staining, and the amount of YFP:SnRK2.6 was investigated with anti-GFP. Signal intensities of phosphorylated MBP (autoradiography), total MBP (CBB), and YFP-SnRK2.6 (@GFP) amounts (bands marked with arrows) were determined by densitometric analyses. The bar chart shows the normalized (AR/CBB/@GFP) kinase activity (means ± SD, n = 4 biological replicates). IP from nontransgenic NIL-DOG1 protein extract served as the negative control. (G) Expression of the 14 Arabidopsis ABA receptors from the RCAR family in dry seeds of NIL-DOG1(means ± SD, n = 3 biological replicates). The data presented in (C) to (G) show levels relative to NIL-DOG1 (arbitrary mean value of 1), and unpaired t test P values are indicated. In (D) and (F), “#” indicates independent biological replicates.

Fig. 3.. AFPs are substrates of AHG1, and AFP2 functionally requires AHG1 to control germination in vivo.

(A) MS intensity of AFP1 p-peptides containing {pS}115 (means ± SD, n = 3 biological replicates). (B) Western blot quantification of AFP1 in dry NIL-DOG1 and dog1-1 seeds. The bar chart shows protein accumulation of AFP1 normalized to the actin loading control (means ± SD, n = 4 biological replicates). (C) Phosphorylation of AFP1 {pS}115 between dry NIL-DOG1 and dog1-1 seeds after normalization of peptide intensity to protein abundance change (means ± SD, n = 3 biological replicates). (D) Family-wide yeast two-hybrid screen between Arabidopsis clade A PP2Cs and AFPs. Control medium lacking leucine and tryptophan, −LW; interaction medium additionally lacking histidine, −LWH; 3-AT, 3-aminotriazole. Interaction of ABI2 with SnRK2.6 (+) served as positive control. Negative control assays for BD-AFP self-activation are shown in fig. S10. (E) In vitro phosphatase assays using HIS:MBP:AHG1 and synthetic phosphorylated peptides of AFP1{pS}115, AFP2{pS}112, and SnRK2III activation loop (i.e., SnRK2.6{pS}171). Assays in which the substrate or the enzyme was omitted served as negative controls (means ± SD, n = 3 independent assays). The top shows a representative picture of the colorimetric assay result. A positive control for SnRK2.6{pS}171 dephosphorylation is presented in fig. S12. (F) Germination capacity of seed progenies from 35S::YFP:AFP2(+/−) in Col-0, ahg1-5, or ahg3-2 mutant backgrounds compared to their parental lines at harvest and after 6 weeks after harvest (WAH) (means ± SD, n = 6 biological replicates). The data presented in (A) to (C) show levels relative to NIL-DOG1 (arbitrary mean value of 1), and unpaired t test P values are indicated. In (B), “#” indicates independent biological replicates. In (E) and (F), letters on the top of the bars indicate significantly different groups using a one-way analysis of variance (ANOVA) with a Tukey post hoc test (α = 0.05).

Fig. 4.. AFPs are required by the DOG1-PP2C module and operate specifically downstream of AHG1 in the control of seed dormancy.

(A) Germination capacity of single and multiple afps mutants in a dog1-2 or (B) an ahg3-2 dog1-2 genetic background at harvest, 6 and 12 weeks after harvest (WAH) (means ± SD, n = 6 biological replicates). Letters on the top of the bars indicate significantly different groups using a one-way ANOVA with a Tukey post hoc test (α = 0.05). (C) Representative pictures showing the germination capacity of key combinatorial mutants after 32 WAH resuming the importance of AFP1 and AFP2 for the control of germination by the AHG1-related branch of the DOG1-PP2C module. Scale bars represent 2 mm.

Fig. 5.. Proposed model for the accommodation of both dormancy and stress reactiveness during seed imbibition by a bimodular control of ABA responses.

Schematic representation of ABA dynamics (top) and the DOG1-PP2C module’s operation alongside ABA core signaling in dormant (bottom left), nondormant unstressed (bottom middle), and nondormant stressed seeds (bottom right) during imbibition. ABA levels decrease upon imbibition regardless of the seed’s stress or dormancy status. However, in dormant and stressed seeds, ABA is de novo synthesized at later imbibition stages to prevent sprouting after prolonged hydration. In dormant seeds imbibed under optimal conditions, DOG1 is active, inhibiting AFP activation by AHG1 and maintaining maturation-imposed ABA responses, independent of the early ABA drop. In later stages, active seed maturation responses induce ABA de novo synthesis, which, alongside the priming of SnRK2III activation, sustain ABA responses post-imbibition and prevent germination. In nondormant seeds (e.g., after-ripened or dog1 mutants) imbibed under optimal conditions, DOG1 is inactive (dashed border; shaded), restoring AHG1 and AHG3 functions. AHG1 dephosphorylates AFP1 and AFP2, switching off ABA responses from maturation and leading to AFP disposal in the absence of stress during early imbibition. This disruption of ABA responses, combined with insufficient ABA de novo synthesis in a stress-free environment, promotes germination and growth. In nondormant seeds imbibed under stress, DOG1 remains inactive, activating AFP and destabilizing ABA responses from maturation. While SnRK2III activation remains primed (in dog1 mutants, the activating phosphorylation mark is preserved), high ABA levels from stress-induced de novo synthesis at later imbibition stages prevent PP2CA-mediated SnRKIII repression, rapidly reactivating kinases and triggering ABA responses that halt germination. The question mark indicates unknown, AFP-independent AHG3 functions in germination control. Molecular events during early or late imbibition are shown as brown or blue connectors, respectively, with the DOG1-PP2C module (brown) and ABA core signaling pathway (blue) boxed.