Control of seed dormancy and germination by DOG1-AHG1 PP2C phosphatase complex via binding to heme

Control of seed dormancy and germination by DOG1-AHG1 PP2C phosphatase complex via binding to heme ARTICLE DOI: 10.1038/s41467-018-04437-9 OPEN Control of seed dormancy and germination by DOG1-AHG1 PP2C phosphatase complex via binding to heme 1,8 2 3 4 1 1 Noriyuki Nishimura , Wataru Tsuchiya , James J. Moresco , Yuki Hayashi , Kouji Satoh , Nahomi Kaiwa , 1 4,5 6 3 7 Tomoko Irisa , Toshinori Kinoshita , Julian I. Schroeder , John R. YatesIII , Takashi Hirayama & Toshimasa Yamazaki Abscisic acid (ABA) regulates abiotic stress and developmental responses including reg- ulation of seed dormancy to prevent seeds from germinating under unfavorable environ- mental conditions. ABA HYPERSENSITIVE GERMINATION1 (AHG1) encoding a type 2C protein phosphatase (PP2C) is a central negative regulator of ABA response in germination; however, the molecular function and regulation of AHG1 remain elusive. Here we report that AHG1 interacts with DELAY OF GERMINATION1 (DOG1), which is a pivotal positive regulator in seed dormancy. DOG1 acts upstream of AHG1 and impairs the PP2C activity of AHG1 in vitro. Furthermore, DOG1 has the ability to bind heme. Binding of DOG1 to AHG1 and heme are independent processes, but both are essential for DOG1 function in vivo. Our study demonstrates that AHG1 and DOG1 constitute an important regulatory system for seed dormancy and germination by integrating multiple environmental signals, in parallel with the PYL/RCAR ABA receptor-mediated regulatory system. Radiation Breeding Division, Institute of Crop Science, National Agriculture and Food Research Organization, 2425 Kamimurata, Hitachiohmiya, Ibaraki 319- 2293, Japan. Structural Biology Team, Advanced Analysis Center, National Agriculture and Food Research Organization, Tsukuba, Ibaraki 305-8602, Japan. 3 4 Department of Molecular Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA. Division of Biological Science, Graduate School of Science, Nagoya University, Nagoya 464-8602, Japan. Institute of Transformative Bio-Molecules (WPI-ITbM), Nagoya University, Chikusa, Nagoya 464-8602, Japan. Division of Biological Sciences, Cell and Developmental Biology Section, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0116, USA. Institute of Plant Science and Resources, Okayama University, 2-20-1 Chuo, Kurashiki, Okayama 710- 0046, Japan. Present address: Division of Basic Research, Institute of Crop Science, National Agriculture and Food Research Organization, 2-1-2 Kannondai, Tsukuba, Ibaraki 305-8518, Japan. These authors contributed equally: Noriyuki Nishimura, Wataru Tsuchiya. Correspondence and requests for materials should be addressed to N.N. (email: nonishi@affrc.go.jp) NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 1 | | | 1234567890():,; ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 s sessile organisms, plants have evolved a number of PHOSPHATASE 2CA (PP2CA), HIGHLY ABA-INDUCED unique mechanisms to adapt to environmental changes. PP2C GENE1 (HAI1), HAI2/AKT1 INTERACTING PP2C ASeed dormancy, which is increased during seed matura- (AIP1), and HAI3. AHG1 and AHG3 have been identified as tion, is one strategy in plants to prevent the seeds from germi- genetic loci that are involved in the ABA response in seed ger- 1–4 33 nating under unfavorable environmental conditions . The mination . ahg1 and ahg3 mutants show a strong ABA hyper- ability to regulate seed dormancy is considered an important trait sensitive phenotype in germination compared to other PP2C 34–36 for the domestication of crops, because reducing seed dormancy single mutants . Since the expression levels of AHG1 and leads to pre-harvest sprouting during crop production, while an AHG3 are higher in dry seeds and increased during the seed increase in seed dormancy makes it difficult for the plants to maturation stage or in the presence of ABA, we proposed that the germinate in the field. Thus, seed dormancy and germination are expression levels of PP2C genes in seeds are likely to be a major strictly connected by developmental and environmental condi- factor contributing to ABA response in seed germination. A tions. Under favorable conditions, the phase transition from seed detailed analysis indicated that AHG1 and AHG3 have both dormancy to germination is controlled by the plant hormones overlapping and distinct functions . Consistent with this abscisic acid (ABA) and gibberellin (GA). In seeds, GA is known observation, PYL/RCARs terminated the PP2C activity of AHG3, to induce germination and inhibit seed dormancy, while ABA but not that of AHG1 in the presence of ABA , suggesting that 1,4,5 antagonizes GA signaling . ABA regulates abiotic stress and AHG1 functions in a unique regulatory system independent of developmental responses including seed maturation, regulation of PYL/RCAR ABA receptors. seed dormancy and germination, growth regulation, and stomatal In order to understand the molecular function and regulation 6,7 closure , and has recently been shown to transiently elevate of AHG1 in the ABA signaling pathway, we have conducted 8,9 heme levels . Heme, which is an iron-binding protoporphrin IX, experiments to co-purify AHG1-interacting proteins in Arabi- is a key molecule that regulates diverse biological activities dopsis using affinity column-based purification. Interestingly, including light respiration, secondary metabolism, and signal DOG1 has been identified as an in vivo interactor of AHG1. Our 10,11 transduction , however, its role in ABA signaling is mostly epistatic analysis demonstrates that DOG1 acts upstream of unknown. AHG1 and reduces the PP2C activity of AHG1 in vitro. Fur- Genetic analyses have identified many loci involved in seed thermore, we find that DOG1 is an α-helical protein that has the 1–3,7 dormancy and germination . DELAY OF GERMINATION1 ability to bind both AHG1 and heme. Binding of DOG1 to AHG1 (DOG1) encodes a protein of unknown biochemical function and and heme are independent processes, but both are essential for was first identified in Arabidopsis as a major quantitative trait DOG1's function in vivo. Our study unveils a novel regulatory locus (QTL) for an increase in seed dormancy . The degree of system of seed dormancy and germination regulated by ABA seed dormancy was determined by the abundance of DOG1 signaling through a DOG1–AHG1 interaction, in parallel with protein in freshly harvested seeds, therefore, it was proposed that PYL/RCAR ABA receptor-dependent regulation. DOG1 is a timer for release from dormancy . While DOG1 has been proposed to function independent of ABA signaling ,a recent study has shown that both DOG1 and ABA signaling Results function in both seed dormancy and seed maturation . Thus Physical interaction between AHG1 and PYR1. We previously DOG1 is a pivotal regulator of seed dormancy. identified PYL/RCAR ABA receptors as in vivo ABI1-interacting The core ABA signaling mechanism is widely thought to be proteins by a combination affinity column purification, using composed of three major components: Pyrabactin Resistance 1 YFP-ABI1 overexpressing plants (YFP-ABI1ox) with LC-MS/ (PYR1)/PYR1-Like (PYL)/Regulatory Components of ABA MS . To assess whether PYR1 was able to interact with all nine Receptor (RCAR) ABA receptors, group A type 2C protein group A PP2Cs including AHG1, which we previously identified phosphatases (PP2Cs), and subclass III sucrose nonfermenting-1- as a central negative regulator of ABA signaling in seeds (Sup- 6,7,15–18 36 related protein kinase2s (SnRK2s) . In the presence of plementary Fig. 1a) , we performed yeast two-hybrid assays ABA, ABA-bound PYR/PYL/RCARs activates SnRK2s through (Fig. 1a). A previous study found that ABI1 subfamily members the inhibition of phosphatase activity of PP2Cs. Most group A interacted with PYR1 in an ABA-dependent manner . Interest- PP2Cs are negatively regulated by PYL/RCAR ABA receptors and ingly, in our investigation, AHG1 subfamily members, except for control the activation of target proteins such as subclass III AHG3, did not interact with PYR1, even in the presence of ABA SnRK2 members including SnRK2.2, SnRK2.3 and SnRK2.6/ (Fig. 1a). When HA-PYR1 was co-expressed with either YFP- OST1 in an ABA-dependent manner to evoke physiological AHG1 or YFP-AHG3 in Nicotiana benthamiana, HA-PYR1 co- 15,16,19,20 responses . The snrk2.2snrk2.3snrk2.6 triple loss-of- immunoprecipitated with YFP-AHG3 in an ABA-dependent function mutant was shown to exhibit vivipary and is almost manner, but not detectably with YFP-AHG1 (Fig. 1b). Some PYL/ completely unresponsive to ABA, supporting the evidence that RCAR-GFP fusion proteins, which interact with ABI1 subfamily these kinases are important for the regulation of seed dormancy members in vivo, have been reported to localize in both the 21–24 15,16,38 and germination regulated by ABA signaling . Activated cytoplasm and the nucleus . With the exception of HAI2/ SnRK2s phosphorylate downstream targets including b-ZIP type AIP1, YFP fused to AHG1 subfamily members predominantly transcriptional factors ABA INSENSITIVE5 (ABI5) and abscisic localized in the nucleus when transiently expressed in N. ben- 7,21,22,25 acid responsive elements (AREB)-binding factors (ABFs) , thamiana protoplasts, whereas YFP fused to all the and the ion channel SLOW ANION CHANNEL-ASSOCIATED1 ABI1 subfamily members and HAI2/AIP1 were observed in the 26–28 29,30 (SLAC1) that is involved in stomatal response . cytoplasm and the nucleus, consistent with the results from the The Arabidopsis genome, which includes more than 76 PP2C previous report (Supplementary Fig. 1b) . These data suggest 31,32 genes , has nine group A PP2Cs that function as central that the functional characteristics of AHG1 subfamily members negative regulators of ABA signaling . Based on the sequence are distinct from those of ABI1 subfamily members in ABA alignment, the group A PP2Cs can be classified into two sub- response. families named ABI1 and ABA HYPERSENSITIVE GERMINA- We previously demonstrated that the ahg1-1ahg3-1 double TION1 (AHG1). The ABI1 subfamily is formed by ABI1, ABI2, mutant exhibited strong ABA-hypersensitive phenotypes in seed HYPERSENSITIVE TO ABA1 (HAB1) and HAB2, while germination . To examine the genetic and physiological AHG1 subfamily is formed by AHG1, AHG3/PROTEIN relationship among AHG1 subfamily proteins in ABA response, 2 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE a b AD BD 1 1/10 1/100 1 1/10 1/100 1 1/10 1/100 AHG1 PYR1 AHG3 PYR1 YFP- AHG1 AHG3 HA- PYR1 PYR1 HAI1 PYR1 ABA –+ –+ HAI2 PYR1 Anti-GFP 75 IP with HAI3 PYR1 anti-HA 25 Anti-HA ABI1 PYR1 ABI2 PYR1 Anti-GFP HAB1 PYR1 Input Anti-HA HAB2 PYR1 Empty PYR1 QDO/A/ABA QDO/A DDO Fig. 1 Physical interaction of group A PP2Cs with PYR1. a Yeast two-hybrid analysis of group A PP2Cs with PYR1. Y2H gold cells transformed with GAL4BD-PYR1 and GAL4AD-PP2Cs, as indicated. A series of tenfold serial dilutions were spotted onto DDO (Double dropout medium: SD/-Leu/-Trp), QDO/A (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aureobasidin A) and QDO/A/ABA (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aureobasidin A and ABA) medium agar plates for 7 days after inoculation. b Interaction of PYR1 with AHG1 and AHG3. HA-PYR1 co-immunoprecipitates with YFP-AHG3, but not YFP-AHG1, in an ABA-dependent manner. Total protein extracts from transformed N. benthamiana leaves were harvested 4 days after inoculation and were treated with or without 100 μM ABA for 24 h before harvesting. After co- immunoprecipitation using anti-HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies a b ahg1-1ahg3-1 Col Col ahg1-1ahg3-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1hai3-1 ahg1-1ahg3-1hai3-1 100 100 60 60 40 40 0 0 0 0.01 0.05 0.1 0.3 0 0.01 0.05 0.1 0.3 ABA (μM) ABA (μM) c d 100 100 80 80 Col ahg1-1ahg3-1 60 60 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1hai3-1 40 40 Col ahg1-1ahg3-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 20 20 ahg1-1ahg3-1hai3 -1 01234 567 0123456 7 Days after stratification Days after stratification Fig. 2 The triple mutants of PP2C show ABA hypersensitivity. a, b Germination efficiencies (a) and post-germination growth efficiencies (b)of pp2c double and triple mutant lines (ahg1-1ahg3-1, ahg1-1ahg3-1hai1-2, ahg1-1ahg3-1aip1-1, and ahg1-1ahg3-1hai3-1), and wild-type were examined in the presence of various concentrations of ABA at 3 days (a) and 7 days (b) after stratification. c, d Germination efficiencies of pp2c double and triple mutant lines and wild- type were examined with stratification for 0 days (c) or 4 days (d). Error bars show s.d. of three independent experiments using the same seed batch (a–d) we obtained ahg1-1ahg3-1hai1-2, ahg1-1ahg3-1aip1-1, and ahg1- mutations affect the response to exogenously applied ABA or seed 1ahg3-1hai3-1 triple mutants. The germination (radicle emer- dormancy, we tested the effect of stratification length on gence) and post-germination growth (seedling with expanded germination and post-germination growth efficiencies of these green cotyledons) efficiencies of ahg1-1ahg3-1hai3-1 triple lines in the absence of exogenous ABA. Without stratification, the mutant were remarkably reduced in the presence of as little as germination and post-germination growth efficiencies of the 0.05 μM ABA (Fig. 2a, b). To determine whether multiple ahg1-1ahg3-1hai1-2 and ahg1-1ahg3-1hai3-1 triple mutants were NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 3 | | | Germination (%) Germination (%) Post-germination growth (%) Germination (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 dramatically reduced (Fig. 2c and Supplementary Fig. 2a). root growth response as expected, whereas they showed normal Interestingly, the germination and post-germination growth plant sizes and similar ABA sensitivities in stomatal responses, efficiencies of ahg1-1ahg3-1hai3-1 were still reduced after compared to control plants (Fig. 3b–d and Supplementary stratification for 4 days (Fig. 2d and Supplementary Fig. 2b). Fig. 3a–c). These observations indicate that these triple mutant seeds have a Using the YFP-AHG1ox plants, AHG1-interacting protein deeper dormancy, and further suggested that at least AHG1, candidates were co-purified with YFP-AHG1. A GFP affinity AHG3, and HAI3 of the AHG1 subfamily members are involved column was loaded with whole-protein extracts from 3-week-old in the regulation of seed dormancy. YFP-AHG1ox or YFPox plants treated with or without ABA. Western blot analysis with an anti-GFP antibody confirmed that YFP-AHG1 or YFP were purified properly (Supplementary Identification of AHG1-interacting proteins. To further address Fig. 4a). Upon SDS-PAGE gel staining with Oriole, some visible the molecular function and regulation of AHG1, we generated bands specific to YFP-AHG1ox samples were detected (Supple- transgenic Arabidopsis plants overexpressing YFP-AHG1 (YFP- mentary Fig. 4b). Mass spectrometric analyses of three indepen- AHG1ox), HA-AHG1 (HA-AHG1ox), and YFP (YFPox) under dent samples with or without ABA treatment identified proteins the CaMV 35S promoter. Fluorescence microscopic analysis co-purified with the YFP-AHG1 (Supplementary Data 1). The showed that the YFP-AHG1 proteins localized in the nucleus specificity of the proteins purified by YFP affinity purification was (Fig. 3a) in the YFP-AHG1ox lines was consistent with the data of confirmed in parallel with experiments using the YFPox plants our transient expression analysis (Supplementary Fig. 1b). We (Supplementary Data 2). We selected the AHG1-interacting previously reported that compared to the YFPox control plants, proteins that were detected in at least three YFP-AHG1ox the YFP-ABI1ox plants were small and exhibited strong ABA- samples, but not in all of the YFPox control samples and insensitive phenotypes in seed germination, root growth, and validated their affinity to AHG1 by yeast two-hybrid assays. As a stomatal responses . The YFP-AHG1ox and HA-AHG1ox plants consequence, four AHG1-interacting proteins were identified showed ABA-insensitive phenotypes in seed germination and in a b YFP YFP-AHG1 YFP YFP-AHG1 ABA (μM) c d YFP YFP-AHG1 YFP YFP-AHG1 100 100 80 80 60 60 40 40 20 20 0 0 010 020 ABA (μM) ABA (μM) Fig. 3 Overexpression of YFP-AHG1 causes ABA insensitivity. a Morphology and subcellular localization of Arabidopsis plants overexpressing YFP (left) and YFP-AHG1 (right) at the rosette plant stage. Plants were grown for 6 weeks in soil. Scale bars, 20 μm. b Post-germination growth efficiencies of overexpressing YFP and YFP-AHG1 lines in the presence of various concentrations of ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch. c ABA-dependent root growth responses of overexpressing YFP and YFP-AHG1 lines. Seedlings were germinated and grown on hormone-free MS plates for 5 days and then transferred to MS plates with or without 10 µM ABA. Root length was measured 4 days after the transfer. Error bars show s.e.m. of three independent experiments. An asterisk indicates significant difference between the corresponding values (*P < 0.05; Tukey–Kramer test). d ABA-induced stomatal closure in overexpressing YFP and YFP-AHG1 lines. The epidermal tissues isolated from dark-adapted 4-to 6-week-old plants were incubated in basal buffer (5 mM MES-BTP pH 6.5, 50 mM KCl, and 0.1 mM CaCl ). Pre-illuminated epidermal −2 −1 −2 −1 tissues were incubated under light (red light at 50 µmol m s and blue light at 10 µmol m s ) for 2.5 h with or without 20 µM ABA. Error bars show s. e.m. of three independent experiments (35 stomata per experiment and condition). No significant difference between the corresponding values (P > 0.05; Tukey–Kramer test) 4 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | Root elongation (%) Stomatal aperture (%) Post-germination growth (%) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE (Supplementary Fig. 4c and Supplementary Tables 1,2). The (Fig. 4b), confirming the specific interaction between DOG1 and identified AHG1-interacting proteins include a known ABA- AHG1 subfamily members. To evaluate the biological relevance signaling component, ABI FIVE BINDING PROTEIN 2 (AFP2), of the interaction between DOG1 and AHG1 in plant, we gen- 39,40 which is characterized as an ABI5-interacting protein . The erated transgenic Arabidopsis plants overexpressing YFP-DOG1 AFP family proteins have been reported to form a transcriptional (YFP-DOG1ox) that was detected in the cytoplasm and nucleus co-repressor complex with TOPLESS (TPL) , which interestingly (Supplementary Fig. 5a, b). The germination and post- was identified as AHG1-interacting protein in this study. FVE/ germination growth efficiencies of the YFP-DOG1ox lines were MSI4 was also identified as AHG1-interacting protein, which apparently reduced, compared to those of the YFPox control line previously is reported to be involved in the flowering time in the presence of ABA (Fig. 4c and Supplementary Fig. 5c), regulation via epigenetic modifications . Intriguingly, we identi- indicating that DOG1 inhibits germination in an ABA-dependent fied DOG1 as AHG1-interacting protein, which has been shown manner. In contrast, the inhibitory effect of ABA on root growth to be a pivotal regulator of seed dormancy and confirmed that and stomatal responses in the YFP-DOG1ox lines were similar to DOG1 mRNA was expressed in both YFP-AHG1ox and YFPox those in the control lines (Supplementary Fig. 5d, e). In contrast, plants (Supplementary Fig. 4d). We decided to focus on DOG1 the AHG1ox lines showed ABA-insensitive phenotypes in root because we suspected that AHG1 has a specific function in seed growth (Fig. 3c and Supplementary Fig. 3b), suggesting that germination . AHG1 and DOG1 have overlapping, but distinct physiological functions. Based on the sequence alignment, DOG1 has five additional Physical interaction between AHG1 and DOG1. To clarify members in the Arabidopsis genome named DOG1-Like 1 to whether the interaction with DOG1 is specific to AHG1, we first DOG1-Like 5 (DOGL1 to DOGL5) (Supplementary Fig. 6). We tested the direct physical interaction between DOG1 and all nine examined whether AHG1 is able to interact with all DOGLs by group A PP2Cs in yeast two-hybrid assays. DOG1 could interact yeast two-hybrid assay. For DOGL5, an alternative splicing form, with all of AHG1 subfamily members, but not ABI1 subfamily DOGL5.2, with higher similarity to DOG1, was used in this study. members (Fig. 4a). To assess the yeast two-hybrid data, we per- The results showed that AHG1 interacts with DOGL3 and formed co-immunoprecipitation experiments. HA-DOG1 co- DOGL5.2 (Supplementary Fig. 7a). DOGL3 and DOGL5.2 were immunoprecipitated with YFP-AHG1 and YFP-AHG3 in an also able to interact with most of the AHG1 subfamily members, ABA-independent manner, but not detectably with YFP-ABI1 a b AD BD 1 1/10 1/100 1 1/10 1/100 AHG1 DOG1 YFP- AHG1 AHG3 ABI1 HA- DOG1 DOG1 DOG1 AHG3 DOG1 ABA –+ –+ –+ HAI1 DOG1 Anti-GFP HAI2 DOG1 IP with anti-HA HAI3 DOG1 Anti-HA ABI1 DOG1 ABI2 DOG1 Anti-GFP HAB1 DOG1 Input HAB2 DOG1 Anti-HA Empty DOG1 QDO/A DDO YFP YFP-DOG1 #1 YFP-DOG1 #2 YFP-DOG1 #3 0 0.3 ABA (μM) Fig. 4 Physical interaction of AHG1 with DOG1. a Yeast two-hybrid analysis of group A PP2Cs with DOG1. Y2H gold cells transformed with GAL4BD-DOG1 and GAL4AD-PP2Cs, as indicated, and were spotted onto DDO (Double dropout medium: SD/-Leu/-Trp) and QDO/A (Quadruple dropout medium: SD/- Ade/-His/-Leu/-Trp supplemented with Aureobasidin A) medium agar plates for 7 days after inoculation. b Interaction of DOG1 with AHG1, AHG3, and ABI1. HA-DOG1 co-immunoprecipitates with YFP-AHG1 and YFP-AHG3, but not YFP-ABI1. Total protein extracts from transformed N. benthamiana leaves were harvested 4 days after inoculation and were treated with or without 100 μM ABA for 24 h before harvesting. After co-immunoprecipitation using anti- HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies. c Post-germination growth efficiencies of overexpressing YFP-DOG1 and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 5 | | | Post-germination growth (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 but not with ABI1 subfamily members (Supplementary Fig. 7b, c). DOG1 functions upstream of AHG1 in the ABA signaling In the transgenic Arabidopsis plants overexpressing YFP-DOGL3 pathway, and led us to the idea that DOG1 directly regulates the (YFP-DOGL3ox) and YFP-DOGL5.2 (YFP-DOGL5.2ox), YFP PP2C activity of AHG1 in an ABA-dependent manner. To test fluorescence was detected in the cytoplasm and the nucleus this hypothesis, we examined the phosphatase activity of (Supplementary Fig. 8a, b). The YFP-DOGL3ox lines exhibited recombinant truncated AHG1 in the presence or absence of lower germination and post-germination growth efficiencies recombinant DOG1 or ABA using a synthetic phosphopeptide , similar to the YFP-DOG1ox line, when compared to the YFPox corresponding to the regulatory phosphorylation site of SnRK2s control line in the presence of ABA (Supplementary Fig. 8c, d). (HSQPK(pS)TVGTP) and an artificial substrate for phospha- Surprisingly, the YFP-DOGL5.2ox lines did not show an ABA- tase 2A, 2B, and 2C (RRA(pT)VA) in vitro. Interestingly, DOG1 hypersensitive phenotype in seed germination (Supplementary impaired the PP2C activity of truncated AHG1 for synthetic Fig. 8c, d), suggesting that DOGL5.2 may not function like SnRK2s phosphopeptide, but not for the artificial substrate, DOG1, even though it has the ability to interact with AHG1. regardless of ABA in our in vitro assay conditions (Fig. 5c, Supplementary Fig. 10). These findings imply that DOG1 reg- ulates the activation state of SnRK2s through the inhibition of the Genetic and functional interactions between AHG1 and PP2C activity of AHG1. DOG1. To investigate the genetic relationship between AHG1 and DOG1, we constructed ahg1-1dog1-2 and ahg1-1dog1-3 double mutants, and transgenic plants overexpressing both HA- The N-terminal portion of DOG1 interacts with AHG1.To AHG1 and YFP-DOG1 fusion proteins (Supplementary Fig. 9a). determine the regions of DOG1 required for AHG1 interaction, The germination and post-germination growth efficiencies of various deleted forms of DOG1 fused to YFP were constructed ahg1-1dog1-2 and ahg1-1dog1-3 double mutants were apparently and their ability to interact with AHG1 was examined (Supple- reduced in the presence of ABA, similar to those of ahg1-1 mentary Fig. 11a, b). Since some deleted forms of DOG1 that (Fig. 5a and Supplementary Fig. 9b). Correspondingly, the HA- excluded a conserved region were difficult to express (Supple- AHG1ox/YFP-DOG1ox double-expression line showed a strong mentary Fig. 11b; Lane 3,4,5), we could not evaluate the results in ABA-insensitive phenotype, similar to the HA-AHG1ox line those samples. HA-AHG1 co-immunoprecipitated with YFP- (Fig. 5b and Supplementary Fig. 9c). These results suggest that DOG1 , but not detectably or very weakly with YFP- Δ257-291 a b Col ahg1-1 dog1-2 dog1-3 ahg1-1dog1-2 ahg1-1dog1-3 YFP HA-AHG1 YFP-DOG1 HA-AHG1/YFP-DOG1 0 0.3 0.5 0.8 0 0.3 1 ABA (μM) ABA (μM) AHG1 AHG1 +DOG1 ΔN ΔN ABA (μM) Fig. 5 Genetic and functional interactions between AHG1 and DOG1. a Post-germination growth efficiencies of the single and double mutant lines (ahg1-1, dog1-2, dog1-3, ahg1-1dog1-2, and ahg1-1dog1-3) and wild-type were examined in the presence of various concentrations of ABA at 7 days after stratification. b Post-germination growth efficiencies of the double overexpressing (HA-AHG1ox/YFP-DOG1ox), parental (HA-AHG1ox and YFP-DOG1ox), and control YFP lines were examined in the presence of various concentrations of ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch (a, b). c The PP2C activities of the truncated AHG1 were measured with or without DOG1 or ABA using the ΔN synthetic phosphopeptide, corresponding to the regulatory phosphorylation site of SnRK2s (HSQPK(pS)TVGTP) as a substrate. Error bars show s.d. of three independent experiments 6 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | % of PP2C activity Post-germination growth (%) Post-germination growth (%) YFP-DOG1 #2 YFP YFP-DOG1 #1 YFP-DOG1 #2 Δ1–18 Δ1–18 0 0.3 NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE a b KDa Anti-GFP IP with anti-HA Anti-HA Anti-GFP Input Anti-HA ABA (μm) Fig. 6 Physical interaction of DOG1 with AHG1 affects DOG1 function. a Post-germination growth efficiencies of overexpressing YFP-DOG1 , YFP-DOG1, Δ1–18 and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch. b Interaction of the N-terminal portion of DOG1 with AHG1. HA-AHG1 co-immunoprecipitates with the deleted forms of YFP-DOG1. Total protein extracts from transformed N. benthamiana leaves were harvested 5 days after inoculation. After co-immunoprecipitation using anti-HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies DOG1 (Supplementary Fig. 11b). In the transgenic Arabi- 84 nM for the N-terminal His -tagged DOG1 (Supplementary Δ1-18 6 dopsis plants overexpressing YFP-DOG1 (YFP-DOG1 Fig. 13f and Supplementary Table 3). Δ1-18 Δ1- ox), YFP fluorescence was detected in the cytoplasm and the To understand the secondary structure of DOG1, we nucleus (Supplementary Fig. 12a, b). The germination and post- performed Far-UV circular dichroism (CD) analysis. The apo- germination growth efficiencies of these lines were not apparently DOG1 showed a spectrum with negative peaks at 222 and 208 nm reduced, compared to those of the YFP-DOG1 line in the pre- and a positive peak at 193 nm (Supplementary Fig. 14a), sence of ABA, and were similar to those of the YFPox control line indicating that DOG1 is a typical α-helical protein. This result (Fig. 6a and Supplementary Fig. 12c), suggesting that the N- is consistent with the secondary structures predicted by Jpred4 45,46 terminal 18 residues are necessary for the DOG1 function, and and Phyre2 (Supplementary Fig. 6). The CD spectrum of the further implying that the interaction with AHG1 is important for heme-bound DOG1 is nearly superimposable with that of the DOG1 function. To further narrow the region required for apo-DOG1, suggesting that heme coordination does not affect the interaction with AHG1, additional deleted forms of YFP-DOG1 secondary structure of DOG1, but may induce tertiary structural were constructed and their abilities to interact with AHG1 were changes. examined. HA-AHG1 co-immunoprecipitated with YFP- DOG1 , YFP-DOG1 , and YFP-DOG1 , but not Δ1–6 Δ1–12 Δ7–12 Heme-binding site is essential for DOG1 function. According detectably or very weakly with YFP-DOG1 and YFP- Δ7–18 to Li et al. , five different amino acids, His, Met, Cys, Tyr, and DOG1 , like YFP-DOG1 (Fig. 6b). Thus, we con- Δ13–18 Δ1–18 Lys can preferentially function as axial ligands to heme, and cluded that the six-residue sequence of DOG1 spanning histidine is the dominant residue (ca. 80%). To confirm the position 13–18, DSYLEW, is essential for interacting with AHG1 possible involvement of histidine residues in the binding of heme, (Fig. 6b). we made mutant DOG1 proteins (H39A, H71A, H153A, H245A, and H249A) in which each of the five histidine residues were DOG1 is an α-helical heme-binding protein. To further address substituted with alanine and measured their electronic absorption the function of DOG1, we expressed the recombinant DOG1 in spectra. The wild-type and mutant DOG1 proteins in apo forms Escherichia coli under two different conditions, short-time (16 h) were incubated with an excess of hemin and passed through a gel expression in LB medium (condition I) and long-time (50–60 h) filtration column to remove the unbound hemin. This treatment expression in an enriched TB medium (condition II). Interest- produced the fully heme-bound, reddish-brown colored wild-type ingly, reddish-brown colored DOG1 was obtained from cells and mutant DOG1 proteins without free hemin. All the DOG1 under condition II, while colorless DOG1 was obtained from cells mutants, except for H245A mutant, exhibited the similar spectral under condition I, suggesting that the colored DOG1 may have characteristics as the wild-type DOG1 (Fig. 8a, b). Interestingly, H245A the potential to bind to a small chromophore molecule such as DOG1 showed drastic spectral changes with the appearance heme (Fig. 7a). The absorption spectrum of the colored DOG1 of new peaks around 390 and 645 nm (Fig. 8a). These peaks are a 3+ exhibited characteristics of heme protein complex peaks, δ peak characteristic of the pentacoordinate high-spin heme-iron Fe at 360 nm, γ (Soret) peak at 425 nm, β peak at 543 nm, and α peak form, indicating that His245 could function as an axial ligand for around 575 nm appeared as a shoulder of β peak (Fig. 7b). These bound heme in wild-type DOG1 (Fig. 8a). However, the char- peaks are consistent with a typical hexacoordinate low spin Fe acteristic peaks for the hexacoordinate low-spin heme at 360 and H245A (III) heme. Analyzing the spectrum as described in the Methods 420 nm were still observed for DOG1 , albeit at lower revealed that about 70% of the colored DOG1 expressed in E. coli intensities than the wild-type DOG1. under condition II was bound to heme. The heme binding ability According to the secondary structure predicted by the Jpred4 of DOG1 was further demonstrated by titration of hemin to the and Phyre2 programs, His249 and His245 are located close to colorless apo-DOG1 expressed in E. coli under condition I each other in the same α-helix (Supplementary Fig. 6). To test (Fig. 7c and Supplementary Fig. 13a, b). For nonlinear curve whether His249 is an axial ligand for the hexacoordinated form in H245A fitting of the data, an increase in the Soret peak against hemin the DOG1 , we measured the spectral properties of the concentration was fit to a 1:1 stoichiometric heme binding, and double amino acid substitution (H245A and H249A) DOG1 the K values were 59 nM for the untagged DOG1 (Fig. 7d) and mutant protein. The fully heme-bound, reddish-brown colored NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 7 | | | YFP-DOG1 - HA-AHG1 YFP-DOG1 - HA-AHG1 Δ1–18 YFP-DOG1 - HA-AHG1 Δ1–6 YFP-DOG1 - HA-AHG1 Δ1–12 YFP-DOG1 - HA-AHG1 Δ7–18 YFP-DOG1 - HA-AHG1 Δ7–12 YFP-DOG1 - HA-AHG1 Δ13–18 Post-germination growth (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 0.8 DOG1 DOGL3 a DOGL5.2 0.6 0.4 β × 2.5 0.2 0.0 300 400 500 600 700 800 Wavelength (nm) c d 1.4 1.4 1.2 1.2 γ _425 nm 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 α 0.2 0.0 0.0 300 400 500 600 700 800 05 10 15 20 Wavelength (nm) Hemin (μM) Fig. 7 Heme binding properties of DOG1. a E. coli expressed N-terminal His -tagged DOG1 (29 μM), DOGL3 (40 μM), and DOGL5.2 (40 μM) for 50-60 h in TB medium (condition II). b Electronic absorption spectra of the DOG1 (11 μM), DOGL3 (15 μM), and DOGL5.2 (15 μM) shown in a. c Electronic absorption spectra of the untagged DOG1 (9.4 μM) after the addition of hemin at the amount of up to ca. 2 mol equivalents of protein. d The γ (Soret) peak absorbance at 425 nm plotted as a function of hemin concentration. Nonlinear curve fitting of experimental data as described in the Methods produced K −1 −1 −1 −1 = 59 nM, ε = 79.4 mM cm , ε = 27.4 mM cm , and x = 0.81. ε , and ε are the extinction coefficients of the DOG1-bound heme and the free DH H DH H hemin at 425 nm, respectively, while “x” represents a fraction of the active hemin, which can be incorporated into DOG1 H245AH249A DOG1 showed a dramatically perturbed electronic Fig. 14e). These data suggest that the interaction with AHG1 of absorption spectrum, indicating that a pentacoordinate high-spin DOG1 is independent of its heme coordination. H245AH249A heme is dominant in DOG1 (Fig. 8c). This result YFP-DOG1ox and YFP-DOGL3ox lines showed a strong ABA- strongly suggests that the new axial ligand is His249 in the hypersensitive phenotype in seed germination, while YFP- H245A hexacoodinate low-spin heme of DOG1 (Supplementary DOGL5.2ox lines did not (Fig. 4c and Supplementary Fig. 8d), Fig. 14b). These data suggest that two histidine residues, His245 even though all three proteins had the ability to interact with and His249, would be located in close proximity to each other AHG1. According to the sequence alignment, histidine residues and would act as an alternative axial ligand in wild-type DOG1. His245 and His249, which bind to heme in DOG1, are conserved H245A In support of this idea, DOG1 expressed under condition II in DOGL3, but not in DOGL5.2 (Supplementary Fig. 6). H245AH249A showed a faint red color, while DOG1 was almost Recombinant DOGL3 expressed under condition II was faint colorless (Supplementary Fig. 14c). As shown in Supplementary red in color (Fig. 7a). Analysis of the electronic absorption Fig. 14d, the absorption spectra of these samples depicted in spectrum revealed that heme content in the colored DOGL3 was Supplementary Fig. 14c are largely different from those for the ca. 8% (Fig. 7b). In contrast, recombinant DOGL5.2 expressed fully heme-bound, reddish-brown colored forms of the corre- under the same condition was colorless (Fig. 7a, b) and heme sponding mutants shown in Fig. 8c. Hemin titration experiments binding was not observed, even when it was treated with excess H245A provided K values of 129 nM for DOG1 and 918 nM for hemin. These results strongly suggest that the heme-binding H245AH249A DOG1 (Supplementary Fig. 13 and Supplementary abilities of DOG1, DOGL3, and DOGL5.2 are correlated well to Table 3), further confirming their lower heme-binding affinities their abilities to ABA-hypersensitive phenotype in seed germina- than wild-type DOG1. It is noteworthy that DOG1 lacking the N- tion (Fig. 4c and Supplementary Fig. 8d). terminal 18 residues, which is required for the interaction with To see whether the heme binding via His245 and His249 is AHG1 (Fig. 6a, b), showed a reddish-brown color and spectral essential for DOG1 function in plants, we generated transgenic H245A characteristics similar to wild-type DOG1 (Supplementary Arabidopsis plants overexpressing YFP-DOG1 (YFP- H245A H245AH249A H245A- Fig. 14c, d). When HA-AHG1 was co-expressed with either DOG1 ox) and YFP-DOG1 (YFP-DOG1 H245A H245AH249A H249A YFP-DOG1, YFP-DOG1 or YFP-DOG1 protein ox) that were detected in the cytoplasm and the nucleus H245A in N. benthamiana, HA-AHG1 co-immunoprecipitated with (Supplementary Figs 15a, b and 16a, b). The YFP-DOG1 ox H245A H245AH249A YFP-DOG1, YFP-DOG1 , and YFP-DOG1 at lines exhibited lower germination and post-germination growth very similar efficiencies, regardless of hemin (Supplementary efficiencies similar to YFP-DOG1ox line, compared to the YFPox 8 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | DOG1 DOGL3 DOGL5.2 Absorbance Absorbance Absorbance (425 nm) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE ab 0.8 0.8 425 DOG1 DOG1 H245A H39A DOG1 DOG1 H71A DOG1 0.6 0.6 H153A DOG1 H249A DOG1 0.4 0.4 × 2.5 × 2.5 0.2 0.2 0.0 0.0 300 400 500 600 700 800 300 400 500 600 700 800 Wavelength (nm) Wavelength (nm) cd YFP YFP-DOG1 #2 0.8 H245AH249A YFP-DOG1 #1 H245AH249A YFP-DOG1 #2 H245A H245AH249A DOG1 YFP-DOG1 #3 H245AH249A DOG1 Hemin 0.6 0.4 × 2.5 0.2 20 0 0.3 0.0 ABA (μM) 300 400 500 600 700 800 Wavelength (nm) Fig. 8 An alternative axial ligand to heme in the wild-type DOG1. a Comparison of electronic absorption spectra of the fully heme-bound wild-type DOG1 and its H245A single mutant. b Overlay of electronic absorption spectra of the fully heme-bound wild-type DOG1 and its His-to-Ala single mutants (H39A, H71A, H153A, and H249A). c Comparison of electronic absorption spectra of the fully heme-bound H245A single mutant DOG1 and H245AH249A double mutant DOG1, as well as free hemin. The fully heme-bound wild-type and mutant DOG1 proteins were prepared by incubating their apo forms with an excess of hemin, followed by a gel filtration column to remove the unbound hemin. All the spectra shown in a–c were measured for the N-terminal His - H245AH249A tagged proteins at concentration of ca. 8 μM. d Post-germination growth efficiencies of overexpressing YFP-DOG1 , YFP-DOG1, and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch control line in the presence of ABA (Supplementary Fig. 15c,d). To examine the AHG1-specific regulatory system, we suc- H245AH249A In contrast, the YFP-DOG1 ox lines showed similar cessfully identified four AHG1-interacting proteins (DOG1, germination and post-germination growth efficiencies to the AFP2, TPL, and FVE/MSI4) (Supplementary Fig. 4c), and char- YFPox control line in the presence of ABA (Fig. 8d and acterized the unique properties in the interactions with AHG1, Supplementary Fig. 16c). These in vivo results correlate well with AHG3, and ABI1. Among these AHG1-interacting proteins, the in vitro spectroscopic results and support the biological DOG1 was particularly enticing because AHG1 functions pri- relevance of heme binding for DOG1 function in ABA signaling. marily in the seeds . DOG1 is one of the pivotal regulators of seed dormancy and germination, although its molecular function 2–4 is largely unknown . A recent study has shown that DOG1 also regulates flowering time through microRNA, suggesting that Discussion The solved crystallographic structures of ABA-PYL/RCAR-PP2Cs DOG1 is expressed and functions in mature plants (Supple- mentary Fig. 4d) . Interestingly, AHG1 subfamily members complexes revealed that a conserved tryptophan residue in PP2Cs, excluding AHG1, inserted in the entrance of the internal interact with DOG1, while no other ABI1 subfamily members do (Fig. 4a, b). We also demonstrate the clear genetic and functional cavity of PYL/RCARs formed a water-mediated hydrogen-bond 47–49 interactions between AHG1 and DOG1 loci (Fig. 5). In addition, network with the ABA-bound PYL/RCARs complex . AHG1 removing the N-terminal portion of DOG1 required for the lacks this conserved tryptophan residue and is therefore predicted 20,50 AHG1-interaction impaired the ability of DOG1 to confer an to be unable to interact with ABA-bound PYL/RCARs . This ABA-hypersensitive phenotype in seed germination, indicating idea is supported by the fact that none of the PYL/RCAR ABA that the interaction with AHG1 is indispensable for DOG1 receptors were identified as AHG1-interacting proteins in this study (Supplementary Data 1 and Supplementary Tables 1, 2). function (Fig. 6). Therefore, it is likely that AHG1 and DOG1 constitute an alternative regulatory system, distinct from the PYL/ These observations led us to believe that AHG1 functions in a unique regulatory system that is independent of PYL/RCARs. RCAR-PP2C regulatory system, to control seed dormancy and NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 9 | | | Absorbance Absorbance Post-germination growth (%) Absorbance ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 germination. Since AHG1 and DOG1 homologs are conserved hypersensitive phenotype in germination (Fig. 8), strongly sup- 32,52,53 among higher plants , the DOG1-AHG1 regulatory system porting the importance of heme coordination for DOG1. Heme is may be common. Recently, the same region of DOG1 responsible a well-known key molecule that regulates diverse biological 10,11 for AHG1 interaction had been reported to be a self-dimerization activities . Some heme-binding proteins are reported to 54 11 site . While it remains unclear how self-dimerization is involved function as sensors for oxygen and nitric oxide (NO) , including 58,59 in DOG1 function, DOG1 might be a core protein in a large the widely studied mammalian NO sensor, Guanylate cyclase , protein complex, regulating components including itself and and FixL, a nitrogen-fixing bacteria which is a well-known oxygen AHG1. It is noteworthy that the HA-AHG1 levels seemed to be sensor . Reactive oxygen species (ROS) and NO counteract ABA less accumulated when co-expressed with the deleted forms of to change the phase transition from seed dormancy to germina- 61,62 YFP-DOG1 that can interact with HA-AHG1, in comparison tion . Similarly, heme-binding DOG1 might monitor the ROS with deleted forms of YFP-DOG1 that cannot interact with HA- and the NO levels affected by physiological or developmental AHG1 (Fig. 6b). Although we could not exclude the possibility stimuli in seeds. Indeed, DOG1 has been shown to undergo post- that this is caused by the experimental conditions, DOG1- translational modifications (PTMs) during after-ripening , interaction might affect the AHG1 level. implying that DOG1 is regulated by PTMs related to ROS and 2,62 Epistatic analysis suggested that DOG1 functions upstream of NO, such as cysteine oxidation and S-nitrosylation , and might AHG1 in ABA signaling in seed germination (Fig. 5). AHG1 was be the hub regulator integrating environmental signals. In addi- reported to interact with SnRK2.3 in vivo and regulate the acti- tion, there are reports showing a link between heme and ABA 19,20 vation state of OST1/SnRK2.6 . In our investigation, DOG1 response. The heme scavenger tryptohan-rich sensory protein could reduce the PP2C activity of AHG1, regardless of ABA (TSPO) is principally detected in dry seeds, and is accumulated in vitro with a synthetic SnRK2s phosphorylation site peptide, but under ABA treatments in Arabidopsis, and TSPO overexpression not with a conventional PP2C substrate (Fig. 5c, Supplementary lines show a weak ABA-hypersensitive phenotype in seed 8,63 Fig. 10). We thought that if DOG1 negatively regulates the AHG1 germination . PP2C activity, the HA-AHG1ox/YFP-DOG1ox line would show a The results presented here, along with the previous studies, similar phenotype to the YFP-DOG1ox line. Indeed, the PYL5/ lead us to propose a model as a working hypothesis for the reg- HAB1 double expression lines exhibited a similar phenotype to ulation of seed dormancy and germination (Fig. 9). As previously the PYL5ox expression line, indicating that the PYL5 ABA reported, ABA directly regulates PYL/RCAR-PP2C and down- receptor prohibited HAB1 function . However, the HA- stream components such as SnRK2s and ABI5, which in turn AHG1ox/YFP-DOG1ox double expression line showed an ABA- control seed dormancy and germination. The present research insensitive phenotype similar to the HA-AHG1ox line (Fig. 5b). points to the hypothesis that DOG1 and AHG1 constitute Presumably, the over-production of HA-AHG1 overcame the another parallel regulatory pathway, in which DOG1 integrates inhibitory effect of the over-produced YFP-DOG1. It is possible environmental signals or physiological conditions other than that DOG1 needs additional modifications or interactions with ABA, and the DOG1-AHG1 complex regulates downstream other components for its function. Such a modification or con- components including SnRK2s and ABI5. Presumably, the structing a large complex with multicomponents might be downstream components of these two regulatory pathways required for the proper DOG1 regulation of AHG1. overlap. In addition, there are several PP2Cs that appear to The DOG1-AHG1 complex may also regulate other interactors function in both the pathways, suggesting that both the intense of AHG1. One good candidate is ABI5, which is one of the major cross-talk and the signal integration between the two pathways 55,56 . We identified determinants of seed germination regulation are essential for the regulation of seed dormancy and germination AFP2 and TPL as AHG1-interacting proteins (Supplementary to maximize the environmental adaptability of the plant life cycle. Fig. 3c). AFP negatively regulates ABI5 controlling its protein levels through the ubiquitin-mediated system , and is also reported to interact with TPL to form a transcriptional co- Environmental and repressor complex . Recent studies indicated that DOG1 affects developmental cues the expression level of ABI5, AFPs, and some group A PP2C genes . ABI5 is one of the well-known major substrates of 22,25 subgroup III SnRK2s , and both of these components are also 19,57 reported to interact with AHG1 . Taken together, these Heme ABA observations suggest that the DOG1-AHG1 complex is likely to regulate ABI5 function through SnRK2s and AFPs-TPL repressor complex. ABA Heme Although AHG3 and other AHG1 subfamily members have PYL/RCARs DOG1 been reported to interact with some of PYL/RCAR ABA recep- tors , they all are able to interact with DOG1 (Fig. 4a). Thus these members function like AHG1 in the DOG1-dependent system. This idea is supported by the data that the triple loss-of-function AHG1/ ABI1/ mutants of these PP2Cs showed strong ABA hypersensitivity and AHG3 AHG3 increased the seed dormancy phenotypes in seeds (Fig. 2). Their ability to interact with both DOG1 and PYL/RCAR ABA recep- tors indicates that these PP2Cs can be important hub regulators connecting the two distinct regulatory pathways to integrate ABA responses multiple signals and fine-tune the regulation of seed dormancy and germination. In this study, we demonstrated that DOG1 binds to heme Seed dormancy (Fig. 7), and inferred that histidine residues function as an axial ligand for the bound heme. Conversion of histidine residues to Fig. 9 A hypothesized regulatory system of seed dormancy. See the text for alanines abolished the DOG1 activity to confer the ABA- detail 10 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE We note that, while a prior version of this manuscript was strain GV3101 carrying the gene of interest was used to transform wild-type Col-0 with plants by the floral dip method . Transgenic plants were screened for under revision, the physical and genetic interaction of DOG1 with hygromycin or bialaphos resistance, and the homozygous T3 lines were isolated. PP2Cs was reported . The conclusion of that study, that PP2Cs Subcellular localizations of YFP-AHG1, YFP-DOG1, YFP-DOGL3, and YFP- interact with and are epistatic to DOG1, is consistent with the DOGL5.2 in the root were analyzed by fluorescence microscope (Nikon Eclipse current work. The reduction of PP2C activity of AHG1 Ni). Transgenic plants were verified by immunoblotting using a 1:10,000 diluted anti-GFP antibody (in this study), followed by a 1:50,000 diluted anti-Rabbit HRP- and heme-binding data here provide further biochemical details conjugate (Pierce #31460). Detection was performed using SuperSignal West Dura regarding the function of DOG1. extended Duration Substrate (Pierce), according to the manufacturer’s instructions. The uncropped full-scan data of immunoblots and gel are shown in Supplementary Fig. 17. Methods Plant materials and growth conditions. Arabidopsis thaliana wild-type Columbia 35 33 (Col), ahg1-1 (see ref. ), ahg3-1 (see ref. ), hai1-2 (SALK_1082282), aip1-1 Yeast two-hybrid assay. The Matchmaker Gold Yeast two-hybrid system (SALK_090738), hai3-1 (SALK_033011), dog1-2 (see ref. ), and dog1-3 (Clontech) was used according to the manufacturer’s instruction. Full-length PYR1, (SALK_000867) were used in this study. Plants were grown on MS plates con- group A PP2Cs, DOG1/DOGLs, and AHG1-interacting proteins cDNAs were cloned taining 1× Murashige and Skoog salt mix, 1% sucrose, 2.5 mM MES (pH5.8), and into pACTGW-attR and pASGW-attR. Each vector was transformed into yeast 0.8% agar or on soil at 22 °C under 16 h light/8 h dark cycles. Seeds were first Y2H gold strain (Clontech). The transformed yeast cells were sprayed onto DDO stratified at 4 °C for 4 days and then transferred to a growth chamber, unless (Double dropout medium: SD/-Leu/-Trp) medium agar plates and incubated at 30 otherwise indicated. °C for 5 days. A series of tenfold serial dilutions were spotted onto DDO, QDO/A (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aur- Germination and growth assays. For germination assays, approximately 50 seeds eobasidin A), and QDO/A/ABA (Quadruple dropout medium: SD/-Ade/-His/- were sown on MS plates containing various concentration of ABA (Sigma). Ger- Leu/-Trp supplemented with Aureobasidin A and 10 μM ABA) medium agar plates mination (radicle emergence) and post-germination growth (seedling with for 7 days after inoculation. expanded green cotyledons) were scored daily for 3 days and 7 days. For seed dormancy assays, approximately 50 seeds were sown on hormone-free MS plates. Co-immunoprecipitation experiments in N. benthamiana. Full-length AHG1, Germination (radicle emergence) and post-germination growth (seedling with AHG3, ABI1, and DOG1 (WT and mutants) cDNAs were cloned into pH35YG, expanded green cotyledons) were scored daily for 7 days. For root growth inhi- and the full-length PYR1, AHG1, and DOG1 cDNAs were cloned into pEarleyGate bition assays, 12–15 seedlings were germinated and grown on hormone-free MS 201. A. tumefaciens strain GV3101 carrying the gene of interest was used and plates for 5 days and then transferred to MS plates containing various con- infiltrated at an OD of 0.5, together with p19 strain in N. benthamiana. After centration of ABA (Sigma) for 4 days. For ABA-induced stomatal closure assays, 65 4 days of infiltration, the infiltrated leaves were sprayed with water containing stomatal apertures in the abaxial epidermis were measured microscopically . The 0.01% Silwet L-77 (BMS) and either 100 μM ABA or 50 μM hemin for 24 h prior to epidermal tissues isolated from dark-adapted 4- to 6-week-old plants were incu- leaf excision. For protein extraction, N. benthamiana leaves (0.75 g) were harvested bated in basal buffer (5 mM MES-BTP pH 6.5, 50 mM KCl, and 0.1 mM CaCl ). and ground to a powder in liquid nitrogen. Ground tissues were resuspended into The epidermal tissues were pre-illuminated under light [blue light (Stick-B-32; −2 −1 1.5 mL of extraction buffer (50 mM Na-phosphate pH 7.4, 150 mM NaCl, 0.1% EYELA) at 10 μmol m s superimposed on background red light (LED-R; NP-40, 1 mM DTT and 1× protease inhibitor cocktail (Sigma)). Crude extracts −2 −1 EYELA) at 50 μmol m s ] for 2.5 h, then incubated with or without 20 µM ABA were then centrifuge at 20,000×g for 30 min at 4 °C. The supernatant was passed for 2.5 h under light (same as above). through a Miracloth (Calbiochem) and used for each immunoprecipitation as an input. The input was incubated with 40 μL anti-HA matrix (Roche #11815016001) Vector constructions. Full-length ABI1, ABI2, HAB1, and PYR1 cDNAs were for 3 h at 4 °C. Immunocomplexes were washed four times with the extraction previously cloned into entry vector (Invitrogen). Full-length AHG1, AHG3, buffer without the protease inhibitor cocktail. After washing, the matrix was HAI1, HAI2/AIP1, HAI3, DOG1, DOGL1, DOGL2, DOGL3, DOGL4, and AFP2 resupended into 50 μL 2× SDS sample buffer for 5 min at 95 °C. The protein cDNAs were cloned into pENTR vector (Invitrogen) and sequenced. The entry samples were separated by 12.5% SDS-PAGE gel (ATTO) and blotted onto an clones, which are inserted into full-length HAB2 (CIW00104), TPL (G09916), Immobilon-P membrane (Millipore), and immunodetected using 1:10,000 diluted DOGL5.2 (PENTR221-AT3G14880), FVE/MSI4 (U15529) cDNAs, were ordered anti-GFP (in this study) or anti-HA (Roche #11583816001) antibodies, followed by from ABRC and sequenced. Site-directed mutagenesis was performed with the a 1:50,000 diluted anti-Rabbit HRP-conjugate (Pierce #31460) or anti-Mouse HRP- KOD-Plus-mutagenesis kit (TOYOBO) using entry clones as templates. The entry conjugate (Pierce #31430). The uncropped full-scan data of immunoblots and gel clones were transferred to the destination vectors pACTGW-attR, pASGW-attR, are shown in Supplementary Fig. 17. pH35YG, and pEarleyGate 201 by Gateway LR clonase II enzyme mix recombi- nation reaction. Affinity column purification of YFP-AHG1 complexes. The three-week-old seedlings (10–20 gFW) grown on MS plates were incubated for 2 h in water (-ABA YFP fusion protein expression analyses. Full-length group A PP2Cs cDNAs samples) before being treated with 100 μM ABA for 24 h (+ABA samples). Sam- were cloned into pH35YG. Agrobacterium tumefaciens strain GV3101 carrying the ples were ground to a powder in liquid nitrogen and resuspended into 2× times gene of interest was used and infiltrated at an OD of 0.5 together with p19 strain extraction buffer (50 mM Na-phosphate pH 7.4, 150 mM NaCl, 0.1% NP-40, 1 mM in N. benthamiana. Mesophyll protoplasts were isolated from leaves after 5 days of DTT and 1× protease inhibitor cocktail (Sigma)). Crude extracts were then cen- infiltration, according to instructions . Fluorescence imaging was analyzed by trifuged at 20,000×g for 30 min at 4 °C. The complete supernatant was passed confocal microscopy (Nikon Eclipse TE2000-U) using 488 nm excitation and through a Miracloth (Calbiochem) and 0.45 μm syringe filter (Starlab Scientific), 500–550 nm emission filters for YFP. and loaded onto an anti-GFP (in this study) conjugated 1 mL HiTrap NHS- activated HP column (GE Healthcare). Anti-GFP affinity columns were generated according to the manufacturer’s instructions (GE Healthcare). After a 50 to 100 mL GFP expression and antiserum preparation. sGFP gene was cloned into the wash with the extraction buffer without the protease inhibitor cocktail, YFP control NdeI/XhoI sites of the expression vector pET28a (+) (Novagen) and transformed and complexed AHG1-interacting proteins were eluted into 10 mL elution buffer into E. coli. strain BL21 Rossetta2 (DE3) (Merck). The E.coli. cells carrying (0.3 M Glycine-HCl pH 3.0, 1 mM DTT and 1× protease inhibitor cocktail expression plasmid were grown at 37 °C, and protein expression was induced by (Sigma)), and were fractionated each (0.5 mL per fraction), and were TCA pre- the addition of IPTG to 1 mM at OD 0.6–0.7 in LB medium. After a 3 h cipitated. Western blot was performed using a 1:10,000 diluted anti-GFP antibody incubation, the cells were harvested by 4000×g centrifugation for 10 min at 4 °C (in this study), followed by a 1:50,000 diluted anti-Rabbit HRP-conjugate (Pierce and the pellet was resuspended in the sonic buffer (50 mM sodium phosphate #31460). Gel staining was performed using Oriole fluorescent gel stain (Bio-Rad). buffer pH 7.0, 300 mM NaCl, 10 mM imidazole, EDTA-free protease inhibitor The uncropped full-scan data of immunoblots and gel are shown in Supplementary cocktail (Roche)). Cells were sonicated on ice seven times for 30 s with regular Fig. 17. resting intervals. A supernatant was obtained after centrifugation at 20,000×g for 10 min at 4 °C and incubated with TALON metal resin (Clontech) for 30 min at 4 ° C. After being washed three times in the sonic buffer without the protease inhibitor Mass spectrometry and data analysis. Purified proteins were reduced with 5 mM cocktail, the matrix was resuspended with an elution buffer (50 mM sodium Tris (2-carboxyethyl) phosphine hydrochloride (Sigma-Aldrich) and alkylated. phosphate buffer pH 7.0, 300 mM NaCl, 250 mM imidazole) and incubated for 30 Proteins were digested for 18 h at 37 °C in 2 M urea, 100 mM Tris pH 8.5, 1 mM min. The protein was further purified with Sephacryl S-100 (GE Healthcare) using CaCl with 2 μg trypsin (Promega). Analysis was performed using an Eksigent a solution of PBS buffer. Antisera against GFP were generated by Scrum Inc. and nanopump and a Thermo LTQ Orbitrap using an in-house built electrospray were evaluated by immunoblot for further use in this study as anti-GFP antibody. stage . Protein and peptide identification and protein quantitation were done with The construction of transgenic Arabidopsis lines. Full-length AHG1, DOG1 (WT Integrated Proteomics Pipeline—IP2 (Integrated Proteomics Applications, Inc., San and mutants), DOGL3, and DOGL5.2 cDNAs were cloned into pH35YG and full- Diego, CA. http://www.integratedproteomics.com/). Tandem mass spectra were length AHG1 cDNA was cloned into pEarleyGate 201. Agrobacterium tumefaciens extracted from raw files using RawConverter and were searched against the NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 11 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 TAIR10_pep_20101214 database (https://www.arabidopsis.org/index.jsp) with spectrophotometer (VARIAN) between 250 and 800 nm at room temperature in 70,71 YFP-AHG1, YFP, and reversed sequences added using ProLuCID . The search buffer B. For the titration experiment, hemin stock solution was prepared by space included all full-tryptic and half-tryptic peptide candidates for the tryptic dissolving 4 mg hemin (Wako) in 500 μL of 0.1 M NaOH and then centrifuged at digest with static modification of 57.02146 on cysteine. Peptide candidates were 20,000×g for 15 min. Concentration of the hemin stock solution was determined filtered using DTASelect , with these parameters: -p 2 -y 1 --trypstat --pfp 0.01 −1 −1 73 using an extinction coefficient value of 58,440 M cm at 385 nm . The solution 69,72 --extra --pI -DM 10 --DB --dm -in -t 1 --brief –quiet . was diluted with 20 mM Tris-HCl buffer pH 8.5, containing 5 mM DTT just before use. The hemin titration experiments were carried out on the untagged and His - tagged DOG1, and its His -tagged H245A and H245AH249A mutants by stepwise Double and triple mutants and double-expression lines. The ahg1-1ahg3-1 addition of a 1 μL aliquot of 1 mM hemin solution to 1 mL of 10 μM apo-proteins double mutant was crossed with hai1-2, aip1-1, and hai3-1. Triple mutants were in 20 mM Tris-HCl buffer pH 8.5, containing 5 mM DTT. Samples were incubated obtained from F2 or F3 progeny using mutant-specific CAPS markers and primer for 5 min at room temperature before measuring the absorption spectra. The sets (Supplementary Table 4). The ahg1-1 mutant was crossed with dog1-2 and stoichiometry and affinity of hemin binding to apo-DOG1 were determined by dog1-3. Double mutants were obtained from F2 or F3 progeny using mutant- nonlinear curve fitting of an increase in the absorbance at 425 nm for the untagged specific CAPS markers and primer sets (Supplementary Table 4). HA-AHG1ox line H245A and His -tagged DOG1 at 418 nm for the His -tagged DOG1 and at 416 nm 6 6 was crossed with YFP-DOG1ox #2 line. The double-expression homo line was H245AH249A for the His -tagged DOG1 , as described in the Methods. To obtain the screened for hygromycin and bialaphos resistance, and obtained from homozygous absorption spectra of the fully heme-bound wild-type DOG1 and its mutants, the T3 lines. corresponding proteins in their apo forms were supplemented with 50 μM hemin dissolved in DMSO and incubated for 20 min at room temperature before gel- RNA isolation and RT-PCR experiments. Total RNA was isolated with a RNeasy filtration chromatography, which removed the unbound hemin. It is worth men- Plant mini kit (Qiagen). After treatment with RNase-free DNase I (Qiagen), first- tioning that the fully heme-bound proteins prepared hemin in NaOH and in strand cDNA was synthesized from 1 μg total RNA using Superscript III first- DMSO displayed essentially the same absorption spectra including peak positions strand synthesis super mix for qRT-PCR (Invitrogen), according to the manu- and heights. Protein concentration was estimated, as described in the Methods. facturer’s instructions. DNA fragments for AHG1, DOG1 and 18s rRNA were amplified for 20 or 30 PCR cycles by using gene specific primers (Supplementary Dissociation constant analysis. The dissociation constants, K , of heme to DOG1 Table 4). The uncropped full-scan data of gel are shown in Supplementary Fig. 18. and its H245A and H245AH249A mutants were determined by fitting the increase of absorption at 425 nm for DOG1, 418 nm for the H245A mutant, and 416 nm for the H245AH249A mutant using following equations. Cloning for protein expression. Full-length DOG1 was cloned into the NdeI/ BamHI sites of the expression vector pCold I (TaKaRa) and pET30b (+) (Nova- A ¼ ε ½DHþ ε ½H ; obs DH H gen) to obtain the N-terminal His -tagged and untagged DOG1, respectively. Full- length DOGL3 and DOGL5.2 cDNAs were inserted between NdeI/XhoI and NdeI/ BamHI sites of the pET28b (+) (Novagen), respectively, containing an N-terminal ½DH¼ ð½D þ x½H þ K  fð½D þ x½H þ K Þ d d total total total total His -tag. The catalytic PP2C domain of AHG1 (residues 104-416) cDNA was 1=2 cloned in the pET28b (+) using NdeI/BamHI sites with In-Fusion system (Clon- 4x½H ½D g Þ=2; total total tech). Site-directed mutagenesis was performed with the QuickChange mutagenesis protocol (Stratagene). All plasmid constructs were confirmed by DNA sequencing and were introduced in the Escherichia coli BL21 (DE3) strain (Merck). ½H ¼½H ½DH and ½D ¼½DHþ½D ; f total total f where [DH] represents the concentration of the heme-bound protein while [D] Protein expression and purification. To obtain apo- or heme-bound DOG1 and [H] are concentrations of the free protein and free hemin, respectively. The expression, E. coli cells carrying expression plasmid were grown at 37 °C to reach extinction coefficients of the protein-bound heme, ε , and the free hemin, ε , DH H an OD of 0.5–0.6 in LB medium or TB medium, respectively, each containing were treated as variable parameters. The “x” represents a fraction of the active antibiotics. Apo-DOG1 was induced by addition of IPTG to 0.2 mM and cell hemin, which can be incorporated into DOG1 and its relatives more easily than the growth was continued for 16 h at 16 °C in LB medium (condition I). Heme-bound aggregated hemin. We treated this “x” as a variable parameter because it is difficult DOG1 was induced with 0.1 mM IPTG at 16°C for 50–60 h in TB medium to estimate the exact value due to co-existence of the monomeric and aggregated (condition II). The harvested cells expressing His -tagged DOG1 were resuspended forms of hemin in the solution. Typically, the obtained “x” is larger than 0.8 in our in buffer A (20 mM Tris-HCl pH 8.5, 150 mM NaCl, 1 mM DTT, and 10% gly- experiments. cerol). The cells were lysed by sonication, and the debris was removed by cen- trifugation at 40,000×g for 30 min. The supernatant was purified with a 5 ml HisTrap HP column (GE Healthcare) and was further purified by HiLoad 26/60 Protein concentration. Since free heme molecules could be removed almost Superdex 75 prep grade (GE Healthcare) pre-equilibrated with buffer B (10 mM completely during the chromatographic purification processes, concentrations of Tris-HCl pH 8.5, 150 mM NaCl, and 5 mM DTT). DOGL3, DOGL5.2, and DOG1 chromatographically purified proteins, with the exception of the H245A and mutants were prepared in the same manner. The untagged DOG1-expressed cells H245AH249A mutants of DOG1, were estimated from the absorbance at 280 and were resuspended in buffer C (20 mM Tris-HCl pH 8.5, 5 mM DTT) and lysed by 425 nm using the following equations: sonication. Following centrifugation clarification, the supernatant was loaded to ½D ¼½DHþ½D ; total f HiLoad 16/10 Q sepharose HP column (GE Healthcare) equilibrated with buffer C, and eluted with linear gradient from 0 to 500 mM NaCl. Fractions containing the untagged DOG1 were applied to a HiLoad 26/60 Superdex 75 prep grade (GE ½DH¼ A =ε ð425Þ; 425 DH Healthcare) pre-equilibrated with buffer B. Further purification was performed by Mono Q 10/100 GL (GE Healthcare) equilibrated with buffer C. The untagged DOG1 was eluted in a linear gradient from 0 to 250 mM NaCl. The catalytic PP2C ½D ¼fA  ε ð280Þ½DHg=ε ð280Þ; f 280 DH D domain of AHG1 was expressed and purified by employing the same procedures used for the apo-DOG1. The purity of the expressed proteins was confirmed by where A and A are the observed absorbance at 280 and 425 nm. The values of 280 425 15% SDS-PAGE. A and A were used for the H245A and H245AH249A mutants of DOG1, 418 416 respectively. For the other proteins, the A values were used. The extinction PP2C enzyme assay. Phosphatase activity of AHG1 was measured in 50 μL coefficient of the apo-protein at 280 nm, ε (280), was evaluated from the amino reaction buffer containing 25 mM Tris-HCl (pH7.5), 10 mM MgCl , 1 mM DTT 2 acid sequence of the protein. The extinction coefficient of the heme-bound DOG1 with or without 50 μM ABA using Serine/Threonine phosphatase assay system at 425 nm, ε (425), was obtained from the K analysis of the hemin titration DH d (Promega). The phosphopeptide HSQPK(pS)TVGTP, corresponding to the reg- experiments, and then ε (280) was calculated dividing ε (425) by a factor of DH DH ulatory phosphorylation site of SnRK2s, was synthesized and purified by Scrum 1.43, the Reinheitszahl value (A /A ) determined for the fully heme-bound 425 280 Inc. Using synthetic SnRK2s phosphopeptide as substrate, the reaction buffer was DOG1. The ε (425) and ε (280) thus obtained were used for all the proteins. DH DH added, which is 100 μM SnRK2s phosphopeptide and 200 μM AHG1 with ΔΝ (1-103) Concentrations of the chromatographically purified H245A and H245AH249A or without 800 μM heme-bound DOG1. Using an artificial substrate for phos- mutants were estimated in the same manner using A and A , respectively, 418 416 phatase 2A, 2B, and 2C, reaction buffer was added, which is 25 μM RRA(pT)VA instead of A . peptide (Promega) and 50 μM AHG1 with or without 200 μM heme- ΔΝ (1-103) bound DOG1. After 30 min at 30 °C, each reaction was stopped with 50 μL Circular dichroism spectroscopy. Circular dichroism (CD) spectra of the purified molybdate dye solution. The absobance at 600 nm was measured with a plate recombinant DOG1 were recorded on a J-720W spectropolarimeter (Jasco) in the reader (GloMax-Multi + Detection system; Promega). far-UV range from 185 to 260 nm using a 0.2-mm thermostatted cell at 25 °C, using the following parameters: resolution, 0.2 nm; bandwidth, 1.0 nm; sensitivity, UV–visible absorption spectroscopy. Electronic absorption spectra of the pur- 50 mdeg; response, 1 s; speed, 100 nm/min; accumulation, 8. Solutions used to ified recombinant proteins were measured on a CARY 400 Bio ultraviolet-visible record CD spectra contained apo-DOG1 or heme-bound DOG1 dissolved in 10 12 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE mM sodium phosphate buffer pH 7.5, with 5 mM DTT at protein concentration of 25. Piskurewicz, U. et al. The gibberellic acid signaling repressor RGL2 inhibits 20 μM. Arabidopsis seed germination by stimulating abscisic acid synthesis and ABI5 activity. Plant Cell 20, 2729–2745 (2008). 26. Lee, S. C., Lan, W., Buchanan, B. B. & Luan, S. A protein kinase-phosphatase Data availability. The mass spectrometry proteomics data have been deposited to 74 pair interacts with an ion channel to regulate ABA signaling in plant guard the ProteomeXchange Consortium via the PRIDE partner repository with the cells. Proc. Natl Acad. Sci. USA 106, 21419–21424 (2009). dataset identifier PXD009392. All relevant data that support the findings of this 27. Geiger, D. et al. Activity of guard cell anion channel SLAC1 is controlled by study are available from the corresponding author upon reasonable request. drought-stress signaling kinase-phosphatase pair. Proc. Natl Acad. Sci. USA 106, 21425–21430 (2009). Received: 9 September 2016 Accepted: 1 May 2018 28. Vahisalu, T. et al. Ozone-triggered rapid stomatal response involves the production of reactive oxygen species, and is controlled by SLAC1 and OST1. Plant J. 62, 442–453 (2010). 29. Negi, J. et al. CO regulator SLAC1 and its homologues are essential for anion homeostasis in plant cells. Nature 452, 483–486 (2008). 30. Vahisalu, T. et al. SLAC1 is required for plant guard cell S-type anion channel References function in stomatal signalling. Nature 452, 487–491 (2008). 1. Bentsink, L. & Koornneef, M. Seed dormancy and germination. Arab. Book 6, 31. Schweighofer, A., Hirt, H. & Meskiene, I. Plant PP2C phosphatases: emerging e0119 (2008). functions in stress signaling. Trends Plant Sci. 9, 236–243 (2004). 2. Graeber, K., Nakabayashi, K., Miatton, E., Leubner-Metzger, G. & Soppe, W. J. 32. Xue, T. et al. Genome-wide and expression analysis of protein phosphatase 2C Molecular mechanisms of seed dormancy. Plant Cell Environ. 35, 1769–1786 in rice and Arabidopsis. BMC Genom. 9, 550 (2008). (2012). 33. Nishimura, N. et al. Isolation and characterization of novel mutants affecting 3. Nonogaki, H. Seed dormancy and germination-emerging mechanisms and the abscisic acid sensitivity of Arabidopsis germination and seedling growth. new hypotheses. Front. Plant Sci. 5, 233 (2014). Plant Cell Physiol. 45, 1485–1499 (2004). 4. Bassel, G. W. To grow or not to grow? Trends Plant Sci. 21, 498–505 (2016). 34. Yoshida, T. et al. ABA-hypersensitive germination3 encodes a protein 5. Shu, K., Liu, X. D., Xie, Q. & He, Z. H. Two faces of one seed: hormonal phosphatase 2C (AtPP2CA) that strongly regulates abscisic acid signaling regulation of dormancy and germination. Mol. Plant 9,34–45 (2016). during germination among Arabidopsis protein phosphatase 2Cs. Plant 6. Hubbard, K. E., Nishimura, N., Hitomi, K., Getzoff, E. D. & Schroeder, J. I. Physiol. 140, 115–126 (2006). Early abscisic acid signal transduction mechanisms: newly discovered 35. Kuhn, J. M., Boisson-Dernier, A., Dizon, M. B., Maktabi, M. H. & Schroeder, J. components and newly emerging questions. Genes Dev. 24, 1695–1708 (2010). I. The protein phosphatase AtPP2CA negatively regulates abscisic acid signal 7. Finkelstein, R. Abscisic acid synthesis and response. Arab. Book 11, e0166 transduction in Arabidopsis, and effects of abh1 on AtPP2CA mRNA. Plant (2013). Physiol. 140, 127–139 (2006). 8. Vanhee, C., Zapotoczny, G., Masquelier, D., Ghislain, M. & Batoko, H. The 36. Nishimura, N. et al. ABA-Hypersensitive Germination1 encodes a protein Arabidopsis multistress regulator TSPO is a heme binding membrane protein phosphatase 2C, an essential component of abscisic acid signaling in and a potential scavenger of porphyrins via an autophagy-dependent Arabidopsis seed. Plant J. 50, 935–949 (2007). degradation mechanism. Plant Cell 23, 785–805 (2011). 37. Nishimura, N. et al. PYR/PYL/RCAR family members are major in-vivo ABI1 9. Kobayashi, Y., Ando, H., Hanaoka, M. & Tanaka, K. Abscisic acid participates protein phosphatase 2C-interacting proteins in Arabidopsis. Plant J. 61, in the control of cell cycle initiation through heme homeostasis in the 290–299 (2010). unicellular red alga Cyanidioschyzon merolae. Plant Cell Physiol. 57, 953–960 38. Santiago, J. et al. Modulation of drought resistance by the abscisic acid (2016). receptor PYL5 through inhibition of clade A PP2Cs. Plant J. 60, 575–588 10. Mochizuki, N. et al. The cell biology of tetrapyrroles: a life and death struggle. (2009). Trends Plant. Sci. 15, 488–498 (2010). 39. Lopez-Molina, L., Mongrand, S., Kinoshita, N. & Chua, N. H. AFP is a novel 11. Li, T., Bonkovsky, H. L. & Guo, J. T. Structural analysis of heme proteins: negative regulator of ABA signaling that promotes ABI5 protein degradation. implications for design and prediction. BMC Struct. Biol. 11, 13 (2011). Genes Dev. 17, 410–418 (2003). 12. Bentsink, L., Jowett, J., Hanhart, C. J. & Koornneef, M. Cloning of DOG1,a 40. Garcia, M. E., Lynch, T., Peeters, J., Snowden, C. & Finkelstein, R. A small quantitative trait locus controlling seed dormancy in Arabidopsis. Proc. Natl plant-specific protein family of ABI five binding proteins (AFPs) regulates Acad. Sci. USA 103, 17042–17047 (2006). stress response in germinating Arabidopsis seeds and seedlings. Plant Mol. 13. Nakabayashi, K. et al. The time required for dormancy release in Arabidopsis Biol. 67, 643–658 (2008). is determined by DELAY OF GERMINATION1 protein levels in freshly 41. Pauwels, L. et al. NINJA connects the co-repressor TOPLESS to jasmonate harvested seeds. Plant Cell 24, 2826–2838 (2012). signalling. Nature 464, 788–791 (2010). 14. Dekkers, B. J. et al. The Arabidopsis DELAY OF GERMINATION 1 gene 42. Pazhouhandeh, M., Molinier, J., Berr, A. & Genschik, P. MSI4/FVE interacts affects ABSCISIC ACID INSENSITIVE 5 (ABI5) expression and genetically with CUL4-DDB1 and a PRC2-like complex to control epigenetic regulation interacts with ABI3 during Arabidopsis seed development. Plant J. 85, of flowering time in Arabidopsis. Proc. Natl Acad. Sci. USA 108, 3430–3435 451–465 (2016). (2011). 43. Fujii, H. et al. In vitro reconstitution of an abscisic acid signalling pathway. 15. Ma, Y. et al. Regulators of PP2C phosphatase activity function as abscisic acid Nature 462, 660–664 (2009). sensors. Science 324, 1064–1068 (2009). 16. Park, S. Y. et al. Abscisic acid inhibits type 2C protein phosphatases via the 44. Deana, A. D. et al. An investigation of the substrate specificity of protein PYR/PYL family of START proteins. Science 324, 1068–1071 (2009). phosphatase 2C using synthetic peptide substrates; comparison with protein phosphatase 2A. Biochim. Biophys. Acta 1051, 199–202 (1990). 17. Cutler, S. R., Rodriguez, P. L., Finkelstein, R. R. & Abrams, S. R. Abscisic acid: 45. Drozdetskiy, A., Cole, C., Procter, J. & Barton, G. J. JPred4: a protein emergence of a core signaling network. Annu. Rev. Plant Biol. 61, 651–679 (2010). secondary structure prediction server. Nucleic Acids Res. 43, W389–W394 (2015). 18. Raghavendra, A. S., Gonugunta, V. K., Christmann, A. & Grill, E. ABA 46. Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N. & Sternberg, M. J. The perception and signalling. Trends Plant. Sci. 15, 395–401 (2010). 19. Umezawa, T. et al. Type 2C protein phosphatases directly regulate abscisic Phyre2 web portal for protein modeling, prediction and analysis. Nat. Protoc. acid-activated protein kinases in Arabidopsis. Proc. Natl Acad. Sci. USA 106, 10, 845–858 (2015). 47. Melcher, K. et al. A gate-latch-lock mechanism for hormone signalling by 17588–17593 (2009). abscisic acid receptors. Nature 462, 602–608 (2009). 20. Antoni, R. et al. Selective inhibition of clade A phosphatases type 2C by PYR/ PYL/RCAR abscisic acid receptors. Plant Physiol. 158, 970–980 (2012). 48. Miyazono, K. et al. Structural basis of abscisic acid signalling. Nature 462, 609–614 (2009). 21. Fujii, H. & Zhu, J. K. Arabidopsis mutant deficient in 3 abscisic acid-activated 49. Yin, P. et al. Structural insights into the mechanism of abscisic acid signaling protein kinases reveals critical roles in growth, reproduction, and stress. Proc. Natl Acad. Sci. USA 106, 8380–8385 (2009). by PYL proteins. Nat. Struct. Mol. Biol. 16, 1230–1236 (2009). 22. Nakashima, K. et al. Three Arabidopsis SnRK2 protein kinases, SRK2D/ 50. Dupeux, F. et al. Modulation of abscisic acid signaling in vivo by an engineered receptor-insensitive protein phosphatase type 2C allele. Plant SnRK2.2, SRK2E/SnRK2.6/OST1 and SRK2I/SnRK2.3, involved in ABA Physiol. 156, 106–116 (2011). signaling are essential for the control of seed development and dormancy. Plant Cell Physiol. 50, 1345–1363 (2009). 51. Huo, H., Wei, S. & Bradford, K. J. DELAY OF GERMINATION1 (DOG1) regulates both seed dormancy and flowering time through microRNA 23. Fujita, Y. et al. Three SnRK2 protein kinases are the main positive regulators pathways. Proc. Natl Acad. Sci. USA 113, E2199–E2206 (2016). of abscisic acid signaling in response to water stress in Arabidopsis. Plant Cell Physiol. 50, 2123–2132 (2009). 52. Ashikawa, I., Abe, F. & Nakamura, S. Ectopic expression of wheat and barley 24. Brandt, B. et al. Calcium specificity signaling mechanisms in abscisic acid DOG1-like genes promotes seed dormancy in Arabidopsis. Plant Sci. 179, 536–542 (2010). signal transduction in Arabidopsis guard cells. eLife 4, e03599 (2015). NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 13 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 53. Graeber, K. et al. Cross-species approaches to seed dormancy and 72. Tabb, D. L., McDonald, W. H. & Yates, J. R. III DTASelect and contrast: tools germination: conservation and biodiversity of ABA-regulated mechanisms for assembling and comparing protein identifications from shotgun and the Brassicaceae DOG1 genes. Plant Mol. Biol. 73,67–87 (2010). proteomics. J. Proteome Res. 1,21–26 (2002). 54. Nakabayashi, K., Bartsch, M., Ding, J. & Soppe, W. J. Seed dormancy in 73. Dawson, R. M. C., Elliott, D. C., Elliott, W. H. & Jones, K. M. Data for Arabidopsis requires self-binding ability of DOG1 protein and the presence of biochemical research third edition, 230–231 (Oxford Science Publications, multiple isoforms generated by alternative splicing. PLoS Genet. 11, e1005737 1986). (2015). 74. Vizcaino, J. A. et al. 2016 Update of the PRIDE database and its related tools. 55. Finkelstein, R. R. & Lynch, T. J. The Arabidopsis abscisic acid response gene Nucleic Acids Res. 44(D1), D447–D456 (2016). ABI5 encodes a basic leucine zipper transcription factor. Plant Cell 12, 599–609 (2000). Acknowledgements 56. Lopez-Molina, L., Mongrand, S. & Chua, N. H. A postgermination We are grateful to Dr. Claire Delahunty and to Mr. Nathaniel Duncan for critical reading developmental arrest checkpoint is mediated by abscisic acid and requires the and comments on the manuscript. We thank Drs. Wim Soppe and Kazumi Nakabayashi ABI5 transcription factor in Arabidopsis. Proc. Natl. Acad. Sci. USA 98, for dog1-2 seeds, Dr. Paul Verslues for hai1-2, aip1-1 and hai3-1 seeds, Dr. Manabu 4782–4787 (2001). Nakayama for pACTGW-attR and pASGW-attR vectors, Dr. Taku Demura for pH35YG 57. Lynch, T., Erickson, B. J. & Finkelstein, R. R. Direct interactions of ABA- vector, Dr. Yasuo Niwa for sGFP vector, ABRC for providing plant materials or reagents, insensitive(ABI)-clade protein phosphatase(PP)2Cs with calcium-dependent Dr. Mitsuhiro Miyazawa for measuring CD spectra, and Ms. Izumi Odano for technical protein kinases and ABA response element-binding bZIPs may contribute to assistance. This research was supported by MEXT/JSPS KAKENHI (Grant Numbers turning off ABA response. Plant Mol. Biol. 80, 647–658 (2012). 23688044, 26712030, and 23119521) and MEXT as part of Joint Research Program 58. Arnold, W. P., Mittal, C. K., Katsuki, S. & Murad, F. Nitric oxide activates implemented at the Institute of Plant Science and Resources, Okayama University in guanylate cyclase and increases guanosine 3′:5′-cyclic monophosphate levels in Japan (2608, 2707, and 2805) to N.N., MEXT KAKENHI (15H05956) to T.K., National various tissue preparations. Proc. Natl Acad. Sci. USA 74, 3203–3207 (1977). Institutes of Health (GM060396) to J.I.S., the National Institute of General Medical 59. Gerzer, R., Radany, E. W. & Garbers, D. L. The separation of the heme and Sciences (8 P41 GM103533) to J.J.M and J.R.Y., and the National Institute of Agrobio- apoheme forms of soluble guanylate cyclase. Biochem. Biophys. Res. Commun. logical Sciences Strategic Fund to T.Y. 108, 678–686 (1982). 60. Gilles-Gonzalez, M. A., Ditta, G. S. & Helinski, D. R. A haemoprotein with kinase activity encoded by the oxygen sensor of Rhizobium meliloti. Nature Author contributions 350, 170–172 (1991). N.N., W.T., J.I.S., and T.Y. designed the research; N.N., W.T., J.J.M., Y.H., K.S., T.I., and 61. Liu, Y., Ye, N., Liu, R., Chen, M. & Zhang, J. H O mediates the regulation of 2 2 N.K. performed the experiments; N.N., W.T., J.J.M., Y.H., K.S., T.K., J.I.S., J.R.Y., T.H., ABA catabolism and GA biosynthesis in Arabidopsis seed dormancy and and T.Y. analyzed the data; N.N., W.T., J.J.M., J.I.S., T.H., and T.Y. wrote the manuscript. germination. J. Exp. Bot. 61, 2979–2990 (2010). 62. Arc, E., Galland, M., Godin, B., Cueff, G. & Rajjou, L. Nitric oxide implication in the control of seed dormancy and germination. Front. Plant Sci. 4, 346 Additional information (2013). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 63. Guillaumot, D. et al. The Arabidopsis TSPO-related protein is a stress and 018-04437-9. abscisic acid-regulated, endoplasmic reticulum-Golgi-localized membrane Competing interests: The authors declare no competing interests. protein. Plant J. 60, 242–256 (2009). 64. Née, G. et al. DELAY OF GERMINATION1 requires PP2C phosphatases of the ABA signalling pathway to control seed dormancy. Nat. Commun. 8,72 Reprints and permission information is available online at http://npg.nature.com/ (2017). reprintsandpermissions/ 65. Tsuzuki, T. et al. Mg-chelatase H subunit affects ABA signaling in stomatal guard cells, but is not an ABA receptor in Arabidopsis thaliana. J. Plant Res. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in 124, 527–538 (2011). published maps and institutional affiliations. 66. Yoo, S. D., Cho, Y. H. & Sheen, J. Arabidopsis mesophyll protoplasts: a versatile cell system for transient gene expression analysis. Nat. Protoc. 2, 1565–1572 (2007). 67. Clough, S. J. & Bent, A. F. Floral dip: a simplified method for Agrobacterium- Open Access This article is licensed under a Creative Commons mediated transformation of Arabidopsis thaliana. Plant J. 16, 735–743 (1998). Attribution 4.0 International License, which permits use, sharing, 68. Wolters, D. A., Washburn, M. P. & Yates, J. R. 3rd An automated adaptation, distribution and reproduction in any medium or format, as long as you give multidimensional protein identification technology for shotgun proteomics. appropriate credit to the original author(s) and the source, provide a link to the Creative Anal. Chem. 73, 5683–5690 (2001). Commons license, and indicate if changes were made. The images or other third party 69. He, L., Diedrich, J., Chu, Y. Y. & Yates, J. R. 3rd Extracting accurate precursor material in this article are included in the article’s Creative Commons license, unless information for tandem mass spectra by raw converter. Anal. Chem. 87, indicated otherwise in a credit line to the material. If material is not included in the 11361–11367 (2015). article’s Creative Commons license and your intended use is not permitted by statutory 70. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J. & Gygi, S. P. Evaluation of regulation or exceeds the permitted use, you will need to obtain permission directly from multidimensional chromatography coupled with tandem mass spectrometry the copyright holder. To view a copy of this license, visit http://creativecommons.org/ (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. licenses/by/4.0/. J. Proteome Res. 2,43–50 (2003). 71. Xu, T. et al. ProLuCID: an improved SEQUEST-like algorithm with enhanced sensitivity and specificity. J. Proteom. 129,16–24 (2015). © The Author(s) 2018 14 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals
Free
14 pages

Loading next page...
 
/lp/springer_journal/control-of-seed-dormancy-and-germination-by-dog1-ahg1-pp2c-phosphatase-YzSspI0Q4P
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
D.O.I.
10.1038/s41467-018-04437-9
Publisher site
See Article on Publisher Site

Abstract

ARTICLE DOI: 10.1038/s41467-018-04437-9 OPEN Control of seed dormancy and germination by DOG1-AHG1 PP2C phosphatase complex via binding to heme 1,8 2 3 4 1 1 Noriyuki Nishimura , Wataru Tsuchiya , James J. Moresco , Yuki Hayashi , Kouji Satoh , Nahomi Kaiwa , 1 4,5 6 3 7 Tomoko Irisa , Toshinori Kinoshita , Julian I. Schroeder , John R. YatesIII , Takashi Hirayama & Toshimasa Yamazaki Abscisic acid (ABA) regulates abiotic stress and developmental responses including reg- ulation of seed dormancy to prevent seeds from germinating under unfavorable environ- mental conditions. ABA HYPERSENSITIVE GERMINATION1 (AHG1) encoding a type 2C protein phosphatase (PP2C) is a central negative regulator of ABA response in germination; however, the molecular function and regulation of AHG1 remain elusive. Here we report that AHG1 interacts with DELAY OF GERMINATION1 (DOG1), which is a pivotal positive regulator in seed dormancy. DOG1 acts upstream of AHG1 and impairs the PP2C activity of AHG1 in vitro. Furthermore, DOG1 has the ability to bind heme. Binding of DOG1 to AHG1 and heme are independent processes, but both are essential for DOG1 function in vivo. Our study demonstrates that AHG1 and DOG1 constitute an important regulatory system for seed dormancy and germination by integrating multiple environmental signals, in parallel with the PYL/RCAR ABA receptor-mediated regulatory system. Radiation Breeding Division, Institute of Crop Science, National Agriculture and Food Research Organization, 2425 Kamimurata, Hitachiohmiya, Ibaraki 319- 2293, Japan. Structural Biology Team, Advanced Analysis Center, National Agriculture and Food Research Organization, Tsukuba, Ibaraki 305-8602, Japan. 3 4 Department of Molecular Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA. Division of Biological Science, Graduate School of Science, Nagoya University, Nagoya 464-8602, Japan. Institute of Transformative Bio-Molecules (WPI-ITbM), Nagoya University, Chikusa, Nagoya 464-8602, Japan. Division of Biological Sciences, Cell and Developmental Biology Section, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0116, USA. Institute of Plant Science and Resources, Okayama University, 2-20-1 Chuo, Kurashiki, Okayama 710- 0046, Japan. Present address: Division of Basic Research, Institute of Crop Science, National Agriculture and Food Research Organization, 2-1-2 Kannondai, Tsukuba, Ibaraki 305-8518, Japan. These authors contributed equally: Noriyuki Nishimura, Wataru Tsuchiya. Correspondence and requests for materials should be addressed to N.N. (email: nonishi@affrc.go.jp) NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 1 | | | 1234567890():,; ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 s sessile organisms, plants have evolved a number of PHOSPHATASE 2CA (PP2CA), HIGHLY ABA-INDUCED unique mechanisms to adapt to environmental changes. PP2C GENE1 (HAI1), HAI2/AKT1 INTERACTING PP2C ASeed dormancy, which is increased during seed matura- (AIP1), and HAI3. AHG1 and AHG3 have been identified as tion, is one strategy in plants to prevent the seeds from germi- genetic loci that are involved in the ABA response in seed ger- 1–4 33 nating under unfavorable environmental conditions . The mination . ahg1 and ahg3 mutants show a strong ABA hyper- ability to regulate seed dormancy is considered an important trait sensitive phenotype in germination compared to other PP2C 34–36 for the domestication of crops, because reducing seed dormancy single mutants . Since the expression levels of AHG1 and leads to pre-harvest sprouting during crop production, while an AHG3 are higher in dry seeds and increased during the seed increase in seed dormancy makes it difficult for the plants to maturation stage or in the presence of ABA, we proposed that the germinate in the field. Thus, seed dormancy and germination are expression levels of PP2C genes in seeds are likely to be a major strictly connected by developmental and environmental condi- factor contributing to ABA response in seed germination. A tions. Under favorable conditions, the phase transition from seed detailed analysis indicated that AHG1 and AHG3 have both dormancy to germination is controlled by the plant hormones overlapping and distinct functions . Consistent with this abscisic acid (ABA) and gibberellin (GA). In seeds, GA is known observation, PYL/RCARs terminated the PP2C activity of AHG3, to induce germination and inhibit seed dormancy, while ABA but not that of AHG1 in the presence of ABA , suggesting that 1,4,5 antagonizes GA signaling . ABA regulates abiotic stress and AHG1 functions in a unique regulatory system independent of developmental responses including seed maturation, regulation of PYL/RCAR ABA receptors. seed dormancy and germination, growth regulation, and stomatal In order to understand the molecular function and regulation 6,7 closure , and has recently been shown to transiently elevate of AHG1 in the ABA signaling pathway, we have conducted 8,9 heme levels . Heme, which is an iron-binding protoporphrin IX, experiments to co-purify AHG1-interacting proteins in Arabi- is a key molecule that regulates diverse biological activities dopsis using affinity column-based purification. Interestingly, including light respiration, secondary metabolism, and signal DOG1 has been identified as an in vivo interactor of AHG1. Our 10,11 transduction , however, its role in ABA signaling is mostly epistatic analysis demonstrates that DOG1 acts upstream of unknown. AHG1 and reduces the PP2C activity of AHG1 in vitro. Fur- Genetic analyses have identified many loci involved in seed thermore, we find that DOG1 is an α-helical protein that has the 1–3,7 dormancy and germination . DELAY OF GERMINATION1 ability to bind both AHG1 and heme. Binding of DOG1 to AHG1 (DOG1) encodes a protein of unknown biochemical function and and heme are independent processes, but both are essential for was first identified in Arabidopsis as a major quantitative trait DOG1's function in vivo. Our study unveils a novel regulatory locus (QTL) for an increase in seed dormancy . The degree of system of seed dormancy and germination regulated by ABA seed dormancy was determined by the abundance of DOG1 signaling through a DOG1–AHG1 interaction, in parallel with protein in freshly harvested seeds, therefore, it was proposed that PYL/RCAR ABA receptor-dependent regulation. DOG1 is a timer for release from dormancy . While DOG1 has been proposed to function independent of ABA signaling ,a recent study has shown that both DOG1 and ABA signaling Results function in both seed dormancy and seed maturation . Thus Physical interaction between AHG1 and PYR1. We previously DOG1 is a pivotal regulator of seed dormancy. identified PYL/RCAR ABA receptors as in vivo ABI1-interacting The core ABA signaling mechanism is widely thought to be proteins by a combination affinity column purification, using composed of three major components: Pyrabactin Resistance 1 YFP-ABI1 overexpressing plants (YFP-ABI1ox) with LC-MS/ (PYR1)/PYR1-Like (PYL)/Regulatory Components of ABA MS . To assess whether PYR1 was able to interact with all nine Receptor (RCAR) ABA receptors, group A type 2C protein group A PP2Cs including AHG1, which we previously identified phosphatases (PP2Cs), and subclass III sucrose nonfermenting-1- as a central negative regulator of ABA signaling in seeds (Sup- 6,7,15–18 36 related protein kinase2s (SnRK2s) . In the presence of plementary Fig. 1a) , we performed yeast two-hybrid assays ABA, ABA-bound PYR/PYL/RCARs activates SnRK2s through (Fig. 1a). A previous study found that ABI1 subfamily members the inhibition of phosphatase activity of PP2Cs. Most group A interacted with PYR1 in an ABA-dependent manner . Interest- PP2Cs are negatively regulated by PYL/RCAR ABA receptors and ingly, in our investigation, AHG1 subfamily members, except for control the activation of target proteins such as subclass III AHG3, did not interact with PYR1, even in the presence of ABA SnRK2 members including SnRK2.2, SnRK2.3 and SnRK2.6/ (Fig. 1a). When HA-PYR1 was co-expressed with either YFP- OST1 in an ABA-dependent manner to evoke physiological AHG1 or YFP-AHG3 in Nicotiana benthamiana, HA-PYR1 co- 15,16,19,20 responses . The snrk2.2snrk2.3snrk2.6 triple loss-of- immunoprecipitated with YFP-AHG3 in an ABA-dependent function mutant was shown to exhibit vivipary and is almost manner, but not detectably with YFP-AHG1 (Fig. 1b). Some PYL/ completely unresponsive to ABA, supporting the evidence that RCAR-GFP fusion proteins, which interact with ABI1 subfamily these kinases are important for the regulation of seed dormancy members in vivo, have been reported to localize in both the 21–24 15,16,38 and germination regulated by ABA signaling . Activated cytoplasm and the nucleus . With the exception of HAI2/ SnRK2s phosphorylate downstream targets including b-ZIP type AIP1, YFP fused to AHG1 subfamily members predominantly transcriptional factors ABA INSENSITIVE5 (ABI5) and abscisic localized in the nucleus when transiently expressed in N. ben- 7,21,22,25 acid responsive elements (AREB)-binding factors (ABFs) , thamiana protoplasts, whereas YFP fused to all the and the ion channel SLOW ANION CHANNEL-ASSOCIATED1 ABI1 subfamily members and HAI2/AIP1 were observed in the 26–28 29,30 (SLAC1) that is involved in stomatal response . cytoplasm and the nucleus, consistent with the results from the The Arabidopsis genome, which includes more than 76 PP2C previous report (Supplementary Fig. 1b) . These data suggest 31,32 genes , has nine group A PP2Cs that function as central that the functional characteristics of AHG1 subfamily members negative regulators of ABA signaling . Based on the sequence are distinct from those of ABI1 subfamily members in ABA alignment, the group A PP2Cs can be classified into two sub- response. families named ABI1 and ABA HYPERSENSITIVE GERMINA- We previously demonstrated that the ahg1-1ahg3-1 double TION1 (AHG1). The ABI1 subfamily is formed by ABI1, ABI2, mutant exhibited strong ABA-hypersensitive phenotypes in seed HYPERSENSITIVE TO ABA1 (HAB1) and HAB2, while germination . To examine the genetic and physiological AHG1 subfamily is formed by AHG1, AHG3/PROTEIN relationship among AHG1 subfamily proteins in ABA response, 2 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE a b AD BD 1 1/10 1/100 1 1/10 1/100 1 1/10 1/100 AHG1 PYR1 AHG3 PYR1 YFP- AHG1 AHG3 HA- PYR1 PYR1 HAI1 PYR1 ABA –+ –+ HAI2 PYR1 Anti-GFP 75 IP with HAI3 PYR1 anti-HA 25 Anti-HA ABI1 PYR1 ABI2 PYR1 Anti-GFP HAB1 PYR1 Input Anti-HA HAB2 PYR1 Empty PYR1 QDO/A/ABA QDO/A DDO Fig. 1 Physical interaction of group A PP2Cs with PYR1. a Yeast two-hybrid analysis of group A PP2Cs with PYR1. Y2H gold cells transformed with GAL4BD-PYR1 and GAL4AD-PP2Cs, as indicated. A series of tenfold serial dilutions were spotted onto DDO (Double dropout medium: SD/-Leu/-Trp), QDO/A (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aureobasidin A) and QDO/A/ABA (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aureobasidin A and ABA) medium agar plates for 7 days after inoculation. b Interaction of PYR1 with AHG1 and AHG3. HA-PYR1 co-immunoprecipitates with YFP-AHG3, but not YFP-AHG1, in an ABA-dependent manner. Total protein extracts from transformed N. benthamiana leaves were harvested 4 days after inoculation and were treated with or without 100 μM ABA for 24 h before harvesting. After co- immunoprecipitation using anti-HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies a b ahg1-1ahg3-1 Col Col ahg1-1ahg3-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1hai3-1 ahg1-1ahg3-1hai3-1 100 100 60 60 40 40 0 0 0 0.01 0.05 0.1 0.3 0 0.01 0.05 0.1 0.3 ABA (μM) ABA (μM) c d 100 100 80 80 Col ahg1-1ahg3-1 60 60 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 ahg1-1ahg3-1hai3-1 40 40 Col ahg1-1ahg3-1 ahg1-1ahg3-1hai1-2 ahg1-1ahg3-1aip1-1 20 20 ahg1-1ahg3-1hai3 -1 01234 567 0123456 7 Days after stratification Days after stratification Fig. 2 The triple mutants of PP2C show ABA hypersensitivity. a, b Germination efficiencies (a) and post-germination growth efficiencies (b)of pp2c double and triple mutant lines (ahg1-1ahg3-1, ahg1-1ahg3-1hai1-2, ahg1-1ahg3-1aip1-1, and ahg1-1ahg3-1hai3-1), and wild-type were examined in the presence of various concentrations of ABA at 3 days (a) and 7 days (b) after stratification. c, d Germination efficiencies of pp2c double and triple mutant lines and wild- type were examined with stratification for 0 days (c) or 4 days (d). Error bars show s.d. of three independent experiments using the same seed batch (a–d) we obtained ahg1-1ahg3-1hai1-2, ahg1-1ahg3-1aip1-1, and ahg1- mutations affect the response to exogenously applied ABA or seed 1ahg3-1hai3-1 triple mutants. The germination (radicle emer- dormancy, we tested the effect of stratification length on gence) and post-germination growth (seedling with expanded germination and post-germination growth efficiencies of these green cotyledons) efficiencies of ahg1-1ahg3-1hai3-1 triple lines in the absence of exogenous ABA. Without stratification, the mutant were remarkably reduced in the presence of as little as germination and post-germination growth efficiencies of the 0.05 μM ABA (Fig. 2a, b). To determine whether multiple ahg1-1ahg3-1hai1-2 and ahg1-1ahg3-1hai3-1 triple mutants were NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 3 | | | Germination (%) Germination (%) Post-germination growth (%) Germination (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 dramatically reduced (Fig. 2c and Supplementary Fig. 2a). root growth response as expected, whereas they showed normal Interestingly, the germination and post-germination growth plant sizes and similar ABA sensitivities in stomatal responses, efficiencies of ahg1-1ahg3-1hai3-1 were still reduced after compared to control plants (Fig. 3b–d and Supplementary stratification for 4 days (Fig. 2d and Supplementary Fig. 2b). Fig. 3a–c). These observations indicate that these triple mutant seeds have a Using the YFP-AHG1ox plants, AHG1-interacting protein deeper dormancy, and further suggested that at least AHG1, candidates were co-purified with YFP-AHG1. A GFP affinity AHG3, and HAI3 of the AHG1 subfamily members are involved column was loaded with whole-protein extracts from 3-week-old in the regulation of seed dormancy. YFP-AHG1ox or YFPox plants treated with or without ABA. Western blot analysis with an anti-GFP antibody confirmed that YFP-AHG1 or YFP were purified properly (Supplementary Identification of AHG1-interacting proteins. To further address Fig. 4a). Upon SDS-PAGE gel staining with Oriole, some visible the molecular function and regulation of AHG1, we generated bands specific to YFP-AHG1ox samples were detected (Supple- transgenic Arabidopsis plants overexpressing YFP-AHG1 (YFP- mentary Fig. 4b). Mass spectrometric analyses of three indepen- AHG1ox), HA-AHG1 (HA-AHG1ox), and YFP (YFPox) under dent samples with or without ABA treatment identified proteins the CaMV 35S promoter. Fluorescence microscopic analysis co-purified with the YFP-AHG1 (Supplementary Data 1). The showed that the YFP-AHG1 proteins localized in the nucleus specificity of the proteins purified by YFP affinity purification was (Fig. 3a) in the YFP-AHG1ox lines was consistent with the data of confirmed in parallel with experiments using the YFPox plants our transient expression analysis (Supplementary Fig. 1b). We (Supplementary Data 2). We selected the AHG1-interacting previously reported that compared to the YFPox control plants, proteins that were detected in at least three YFP-AHG1ox the YFP-ABI1ox plants were small and exhibited strong ABA- samples, but not in all of the YFPox control samples and insensitive phenotypes in seed germination, root growth, and validated their affinity to AHG1 by yeast two-hybrid assays. As a stomatal responses . The YFP-AHG1ox and HA-AHG1ox plants consequence, four AHG1-interacting proteins were identified showed ABA-insensitive phenotypes in seed germination and in a b YFP YFP-AHG1 YFP YFP-AHG1 ABA (μM) c d YFP YFP-AHG1 YFP YFP-AHG1 100 100 80 80 60 60 40 40 20 20 0 0 010 020 ABA (μM) ABA (μM) Fig. 3 Overexpression of YFP-AHG1 causes ABA insensitivity. a Morphology and subcellular localization of Arabidopsis plants overexpressing YFP (left) and YFP-AHG1 (right) at the rosette plant stage. Plants were grown for 6 weeks in soil. Scale bars, 20 μm. b Post-germination growth efficiencies of overexpressing YFP and YFP-AHG1 lines in the presence of various concentrations of ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch. c ABA-dependent root growth responses of overexpressing YFP and YFP-AHG1 lines. Seedlings were germinated and grown on hormone-free MS plates for 5 days and then transferred to MS plates with or without 10 µM ABA. Root length was measured 4 days after the transfer. Error bars show s.e.m. of three independent experiments. An asterisk indicates significant difference between the corresponding values (*P < 0.05; Tukey–Kramer test). d ABA-induced stomatal closure in overexpressing YFP and YFP-AHG1 lines. The epidermal tissues isolated from dark-adapted 4-to 6-week-old plants were incubated in basal buffer (5 mM MES-BTP pH 6.5, 50 mM KCl, and 0.1 mM CaCl ). Pre-illuminated epidermal −2 −1 −2 −1 tissues were incubated under light (red light at 50 µmol m s and blue light at 10 µmol m s ) for 2.5 h with or without 20 µM ABA. Error bars show s. e.m. of three independent experiments (35 stomata per experiment and condition). No significant difference between the corresponding values (P > 0.05; Tukey–Kramer test) 4 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | Root elongation (%) Stomatal aperture (%) Post-germination growth (%) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE (Supplementary Fig. 4c and Supplementary Tables 1,2). The (Fig. 4b), confirming the specific interaction between DOG1 and identified AHG1-interacting proteins include a known ABA- AHG1 subfamily members. To evaluate the biological relevance signaling component, ABI FIVE BINDING PROTEIN 2 (AFP2), of the interaction between DOG1 and AHG1 in plant, we gen- 39,40 which is characterized as an ABI5-interacting protein . The erated transgenic Arabidopsis plants overexpressing YFP-DOG1 AFP family proteins have been reported to form a transcriptional (YFP-DOG1ox) that was detected in the cytoplasm and nucleus co-repressor complex with TOPLESS (TPL) , which interestingly (Supplementary Fig. 5a, b). The germination and post- was identified as AHG1-interacting protein in this study. FVE/ germination growth efficiencies of the YFP-DOG1ox lines were MSI4 was also identified as AHG1-interacting protein, which apparently reduced, compared to those of the YFPox control line previously is reported to be involved in the flowering time in the presence of ABA (Fig. 4c and Supplementary Fig. 5c), regulation via epigenetic modifications . Intriguingly, we identi- indicating that DOG1 inhibits germination in an ABA-dependent fied DOG1 as AHG1-interacting protein, which has been shown manner. In contrast, the inhibitory effect of ABA on root growth to be a pivotal regulator of seed dormancy and confirmed that and stomatal responses in the YFP-DOG1ox lines were similar to DOG1 mRNA was expressed in both YFP-AHG1ox and YFPox those in the control lines (Supplementary Fig. 5d, e). In contrast, plants (Supplementary Fig. 4d). We decided to focus on DOG1 the AHG1ox lines showed ABA-insensitive phenotypes in root because we suspected that AHG1 has a specific function in seed growth (Fig. 3c and Supplementary Fig. 3b), suggesting that germination . AHG1 and DOG1 have overlapping, but distinct physiological functions. Based on the sequence alignment, DOG1 has five additional Physical interaction between AHG1 and DOG1. To clarify members in the Arabidopsis genome named DOG1-Like 1 to whether the interaction with DOG1 is specific to AHG1, we first DOG1-Like 5 (DOGL1 to DOGL5) (Supplementary Fig. 6). We tested the direct physical interaction between DOG1 and all nine examined whether AHG1 is able to interact with all DOGLs by group A PP2Cs in yeast two-hybrid assays. DOG1 could interact yeast two-hybrid assay. For DOGL5, an alternative splicing form, with all of AHG1 subfamily members, but not ABI1 subfamily DOGL5.2, with higher similarity to DOG1, was used in this study. members (Fig. 4a). To assess the yeast two-hybrid data, we per- The results showed that AHG1 interacts with DOGL3 and formed co-immunoprecipitation experiments. HA-DOG1 co- DOGL5.2 (Supplementary Fig. 7a). DOGL3 and DOGL5.2 were immunoprecipitated with YFP-AHG1 and YFP-AHG3 in an also able to interact with most of the AHG1 subfamily members, ABA-independent manner, but not detectably with YFP-ABI1 a b AD BD 1 1/10 1/100 1 1/10 1/100 AHG1 DOG1 YFP- AHG1 AHG3 ABI1 HA- DOG1 DOG1 DOG1 AHG3 DOG1 ABA –+ –+ –+ HAI1 DOG1 Anti-GFP HAI2 DOG1 IP with anti-HA HAI3 DOG1 Anti-HA ABI1 DOG1 ABI2 DOG1 Anti-GFP HAB1 DOG1 Input HAB2 DOG1 Anti-HA Empty DOG1 QDO/A DDO YFP YFP-DOG1 #1 YFP-DOG1 #2 YFP-DOG1 #3 0 0.3 ABA (μM) Fig. 4 Physical interaction of AHG1 with DOG1. a Yeast two-hybrid analysis of group A PP2Cs with DOG1. Y2H gold cells transformed with GAL4BD-DOG1 and GAL4AD-PP2Cs, as indicated, and were spotted onto DDO (Double dropout medium: SD/-Leu/-Trp) and QDO/A (Quadruple dropout medium: SD/- Ade/-His/-Leu/-Trp supplemented with Aureobasidin A) medium agar plates for 7 days after inoculation. b Interaction of DOG1 with AHG1, AHG3, and ABI1. HA-DOG1 co-immunoprecipitates with YFP-AHG1 and YFP-AHG3, but not YFP-ABI1. Total protein extracts from transformed N. benthamiana leaves were harvested 4 days after inoculation and were treated with or without 100 μM ABA for 24 h before harvesting. After co-immunoprecipitation using anti- HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies. c Post-germination growth efficiencies of overexpressing YFP-DOG1 and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 5 | | | Post-germination growth (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 but not with ABI1 subfamily members (Supplementary Fig. 7b, c). DOG1 functions upstream of AHG1 in the ABA signaling In the transgenic Arabidopsis plants overexpressing YFP-DOGL3 pathway, and led us to the idea that DOG1 directly regulates the (YFP-DOGL3ox) and YFP-DOGL5.2 (YFP-DOGL5.2ox), YFP PP2C activity of AHG1 in an ABA-dependent manner. To test fluorescence was detected in the cytoplasm and the nucleus this hypothesis, we examined the phosphatase activity of (Supplementary Fig. 8a, b). The YFP-DOGL3ox lines exhibited recombinant truncated AHG1 in the presence or absence of lower germination and post-germination growth efficiencies recombinant DOG1 or ABA using a synthetic phosphopeptide , similar to the YFP-DOG1ox line, when compared to the YFPox corresponding to the regulatory phosphorylation site of SnRK2s control line in the presence of ABA (Supplementary Fig. 8c, d). (HSQPK(pS)TVGTP) and an artificial substrate for phospha- Surprisingly, the YFP-DOGL5.2ox lines did not show an ABA- tase 2A, 2B, and 2C (RRA(pT)VA) in vitro. Interestingly, DOG1 hypersensitive phenotype in seed germination (Supplementary impaired the PP2C activity of truncated AHG1 for synthetic Fig. 8c, d), suggesting that DOGL5.2 may not function like SnRK2s phosphopeptide, but not for the artificial substrate, DOG1, even though it has the ability to interact with AHG1. regardless of ABA in our in vitro assay conditions (Fig. 5c, Supplementary Fig. 10). These findings imply that DOG1 reg- ulates the activation state of SnRK2s through the inhibition of the Genetic and functional interactions between AHG1 and PP2C activity of AHG1. DOG1. To investigate the genetic relationship between AHG1 and DOG1, we constructed ahg1-1dog1-2 and ahg1-1dog1-3 double mutants, and transgenic plants overexpressing both HA- The N-terminal portion of DOG1 interacts with AHG1.To AHG1 and YFP-DOG1 fusion proteins (Supplementary Fig. 9a). determine the regions of DOG1 required for AHG1 interaction, The germination and post-germination growth efficiencies of various deleted forms of DOG1 fused to YFP were constructed ahg1-1dog1-2 and ahg1-1dog1-3 double mutants were apparently and their ability to interact with AHG1 was examined (Supple- reduced in the presence of ABA, similar to those of ahg1-1 mentary Fig. 11a, b). Since some deleted forms of DOG1 that (Fig. 5a and Supplementary Fig. 9b). Correspondingly, the HA- excluded a conserved region were difficult to express (Supple- AHG1ox/YFP-DOG1ox double-expression line showed a strong mentary Fig. 11b; Lane 3,4,5), we could not evaluate the results in ABA-insensitive phenotype, similar to the HA-AHG1ox line those samples. HA-AHG1 co-immunoprecipitated with YFP- (Fig. 5b and Supplementary Fig. 9c). These results suggest that DOG1 , but not detectably or very weakly with YFP- Δ257-291 a b Col ahg1-1 dog1-2 dog1-3 ahg1-1dog1-2 ahg1-1dog1-3 YFP HA-AHG1 YFP-DOG1 HA-AHG1/YFP-DOG1 0 0.3 0.5 0.8 0 0.3 1 ABA (μM) ABA (μM) AHG1 AHG1 +DOG1 ΔN ΔN ABA (μM) Fig. 5 Genetic and functional interactions between AHG1 and DOG1. a Post-germination growth efficiencies of the single and double mutant lines (ahg1-1, dog1-2, dog1-3, ahg1-1dog1-2, and ahg1-1dog1-3) and wild-type were examined in the presence of various concentrations of ABA at 7 days after stratification. b Post-germination growth efficiencies of the double overexpressing (HA-AHG1ox/YFP-DOG1ox), parental (HA-AHG1ox and YFP-DOG1ox), and control YFP lines were examined in the presence of various concentrations of ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch (a, b). c The PP2C activities of the truncated AHG1 were measured with or without DOG1 or ABA using the ΔN synthetic phosphopeptide, corresponding to the regulatory phosphorylation site of SnRK2s (HSQPK(pS)TVGTP) as a substrate. Error bars show s.d. of three independent experiments 6 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | % of PP2C activity Post-germination growth (%) Post-germination growth (%) YFP-DOG1 #2 YFP YFP-DOG1 #1 YFP-DOG1 #2 Δ1–18 Δ1–18 0 0.3 NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE a b KDa Anti-GFP IP with anti-HA Anti-HA Anti-GFP Input Anti-HA ABA (μm) Fig. 6 Physical interaction of DOG1 with AHG1 affects DOG1 function. a Post-germination growth efficiencies of overexpressing YFP-DOG1 , YFP-DOG1, Δ1–18 and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch. b Interaction of the N-terminal portion of DOG1 with AHG1. HA-AHG1 co-immunoprecipitates with the deleted forms of YFP-DOG1. Total protein extracts from transformed N. benthamiana leaves were harvested 5 days after inoculation. After co-immunoprecipitation using anti-HA matrix, the input and the immunoprecipitated samples were detected with anti-GFP and anti-HA antibodies DOG1 (Supplementary Fig. 11b). In the transgenic Arabi- 84 nM for the N-terminal His -tagged DOG1 (Supplementary Δ1-18 6 dopsis plants overexpressing YFP-DOG1 (YFP-DOG1 Fig. 13f and Supplementary Table 3). Δ1-18 Δ1- ox), YFP fluorescence was detected in the cytoplasm and the To understand the secondary structure of DOG1, we nucleus (Supplementary Fig. 12a, b). The germination and post- performed Far-UV circular dichroism (CD) analysis. The apo- germination growth efficiencies of these lines were not apparently DOG1 showed a spectrum with negative peaks at 222 and 208 nm reduced, compared to those of the YFP-DOG1 line in the pre- and a positive peak at 193 nm (Supplementary Fig. 14a), sence of ABA, and were similar to those of the YFPox control line indicating that DOG1 is a typical α-helical protein. This result (Fig. 6a and Supplementary Fig. 12c), suggesting that the N- is consistent with the secondary structures predicted by Jpred4 45,46 terminal 18 residues are necessary for the DOG1 function, and and Phyre2 (Supplementary Fig. 6). The CD spectrum of the further implying that the interaction with AHG1 is important for heme-bound DOG1 is nearly superimposable with that of the DOG1 function. To further narrow the region required for apo-DOG1, suggesting that heme coordination does not affect the interaction with AHG1, additional deleted forms of YFP-DOG1 secondary structure of DOG1, but may induce tertiary structural were constructed and their abilities to interact with AHG1 were changes. examined. HA-AHG1 co-immunoprecipitated with YFP- DOG1 , YFP-DOG1 , and YFP-DOG1 , but not Δ1–6 Δ1–12 Δ7–12 Heme-binding site is essential for DOG1 function. According detectably or very weakly with YFP-DOG1 and YFP- Δ7–18 to Li et al. , five different amino acids, His, Met, Cys, Tyr, and DOG1 , like YFP-DOG1 (Fig. 6b). Thus, we con- Δ13–18 Δ1–18 Lys can preferentially function as axial ligands to heme, and cluded that the six-residue sequence of DOG1 spanning histidine is the dominant residue (ca. 80%). To confirm the position 13–18, DSYLEW, is essential for interacting with AHG1 possible involvement of histidine residues in the binding of heme, (Fig. 6b). we made mutant DOG1 proteins (H39A, H71A, H153A, H245A, and H249A) in which each of the five histidine residues were DOG1 is an α-helical heme-binding protein. To further address substituted with alanine and measured their electronic absorption the function of DOG1, we expressed the recombinant DOG1 in spectra. The wild-type and mutant DOG1 proteins in apo forms Escherichia coli under two different conditions, short-time (16 h) were incubated with an excess of hemin and passed through a gel expression in LB medium (condition I) and long-time (50–60 h) filtration column to remove the unbound hemin. This treatment expression in an enriched TB medium (condition II). Interest- produced the fully heme-bound, reddish-brown colored wild-type ingly, reddish-brown colored DOG1 was obtained from cells and mutant DOG1 proteins without free hemin. All the DOG1 under condition II, while colorless DOG1 was obtained from cells mutants, except for H245A mutant, exhibited the similar spectral under condition I, suggesting that the colored DOG1 may have characteristics as the wild-type DOG1 (Fig. 8a, b). Interestingly, H245A the potential to bind to a small chromophore molecule such as DOG1 showed drastic spectral changes with the appearance heme (Fig. 7a). The absorption spectrum of the colored DOG1 of new peaks around 390 and 645 nm (Fig. 8a). These peaks are a 3+ exhibited characteristics of heme protein complex peaks, δ peak characteristic of the pentacoordinate high-spin heme-iron Fe at 360 nm, γ (Soret) peak at 425 nm, β peak at 543 nm, and α peak form, indicating that His245 could function as an axial ligand for around 575 nm appeared as a shoulder of β peak (Fig. 7b). These bound heme in wild-type DOG1 (Fig. 8a). However, the char- peaks are consistent with a typical hexacoordinate low spin Fe acteristic peaks for the hexacoordinate low-spin heme at 360 and H245A (III) heme. Analyzing the spectrum as described in the Methods 420 nm were still observed for DOG1 , albeit at lower revealed that about 70% of the colored DOG1 expressed in E. coli intensities than the wild-type DOG1. under condition II was bound to heme. The heme binding ability According to the secondary structure predicted by the Jpred4 of DOG1 was further demonstrated by titration of hemin to the and Phyre2 programs, His249 and His245 are located close to colorless apo-DOG1 expressed in E. coli under condition I each other in the same α-helix (Supplementary Fig. 6). To test (Fig. 7c and Supplementary Fig. 13a, b). For nonlinear curve whether His249 is an axial ligand for the hexacoordinated form in H245A fitting of the data, an increase in the Soret peak against hemin the DOG1 , we measured the spectral properties of the concentration was fit to a 1:1 stoichiometric heme binding, and double amino acid substitution (H245A and H249A) DOG1 the K values were 59 nM for the untagged DOG1 (Fig. 7d) and mutant protein. The fully heme-bound, reddish-brown colored NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 7 | | | YFP-DOG1 - HA-AHG1 YFP-DOG1 - HA-AHG1 Δ1–18 YFP-DOG1 - HA-AHG1 Δ1–6 YFP-DOG1 - HA-AHG1 Δ1–12 YFP-DOG1 - HA-AHG1 Δ7–18 YFP-DOG1 - HA-AHG1 Δ7–12 YFP-DOG1 - HA-AHG1 Δ13–18 Post-germination growth (%) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 0.8 DOG1 DOGL3 a DOGL5.2 0.6 0.4 β × 2.5 0.2 0.0 300 400 500 600 700 800 Wavelength (nm) c d 1.4 1.4 1.2 1.2 γ _425 nm 1.0 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 α 0.2 0.0 0.0 300 400 500 600 700 800 05 10 15 20 Wavelength (nm) Hemin (μM) Fig. 7 Heme binding properties of DOG1. a E. coli expressed N-terminal His -tagged DOG1 (29 μM), DOGL3 (40 μM), and DOGL5.2 (40 μM) for 50-60 h in TB medium (condition II). b Electronic absorption spectra of the DOG1 (11 μM), DOGL3 (15 μM), and DOGL5.2 (15 μM) shown in a. c Electronic absorption spectra of the untagged DOG1 (9.4 μM) after the addition of hemin at the amount of up to ca. 2 mol equivalents of protein. d The γ (Soret) peak absorbance at 425 nm plotted as a function of hemin concentration. Nonlinear curve fitting of experimental data as described in the Methods produced K −1 −1 −1 −1 = 59 nM, ε = 79.4 mM cm , ε = 27.4 mM cm , and x = 0.81. ε , and ε are the extinction coefficients of the DOG1-bound heme and the free DH H DH H hemin at 425 nm, respectively, while “x” represents a fraction of the active hemin, which can be incorporated into DOG1 H245AH249A DOG1 showed a dramatically perturbed electronic Fig. 14e). These data suggest that the interaction with AHG1 of absorption spectrum, indicating that a pentacoordinate high-spin DOG1 is independent of its heme coordination. H245AH249A heme is dominant in DOG1 (Fig. 8c). This result YFP-DOG1ox and YFP-DOGL3ox lines showed a strong ABA- strongly suggests that the new axial ligand is His249 in the hypersensitive phenotype in seed germination, while YFP- H245A hexacoodinate low-spin heme of DOG1 (Supplementary DOGL5.2ox lines did not (Fig. 4c and Supplementary Fig. 8d), Fig. 14b). These data suggest that two histidine residues, His245 even though all three proteins had the ability to interact with and His249, would be located in close proximity to each other AHG1. According to the sequence alignment, histidine residues and would act as an alternative axial ligand in wild-type DOG1. His245 and His249, which bind to heme in DOG1, are conserved H245A In support of this idea, DOG1 expressed under condition II in DOGL3, but not in DOGL5.2 (Supplementary Fig. 6). H245AH249A showed a faint red color, while DOG1 was almost Recombinant DOGL3 expressed under condition II was faint colorless (Supplementary Fig. 14c). As shown in Supplementary red in color (Fig. 7a). Analysis of the electronic absorption Fig. 14d, the absorption spectra of these samples depicted in spectrum revealed that heme content in the colored DOGL3 was Supplementary Fig. 14c are largely different from those for the ca. 8% (Fig. 7b). In contrast, recombinant DOGL5.2 expressed fully heme-bound, reddish-brown colored forms of the corre- under the same condition was colorless (Fig. 7a, b) and heme sponding mutants shown in Fig. 8c. Hemin titration experiments binding was not observed, even when it was treated with excess H245A provided K values of 129 nM for DOG1 and 918 nM for hemin. These results strongly suggest that the heme-binding H245AH249A DOG1 (Supplementary Fig. 13 and Supplementary abilities of DOG1, DOGL3, and DOGL5.2 are correlated well to Table 3), further confirming their lower heme-binding affinities their abilities to ABA-hypersensitive phenotype in seed germina- than wild-type DOG1. It is noteworthy that DOG1 lacking the N- tion (Fig. 4c and Supplementary Fig. 8d). terminal 18 residues, which is required for the interaction with To see whether the heme binding via His245 and His249 is AHG1 (Fig. 6a, b), showed a reddish-brown color and spectral essential for DOG1 function in plants, we generated transgenic H245A characteristics similar to wild-type DOG1 (Supplementary Arabidopsis plants overexpressing YFP-DOG1 (YFP- H245A H245AH249A H245A- Fig. 14c, d). When HA-AHG1 was co-expressed with either DOG1 ox) and YFP-DOG1 (YFP-DOG1 H245A H245AH249A H249A YFP-DOG1, YFP-DOG1 or YFP-DOG1 protein ox) that were detected in the cytoplasm and the nucleus H245A in N. benthamiana, HA-AHG1 co-immunoprecipitated with (Supplementary Figs 15a, b and 16a, b). The YFP-DOG1 ox H245A H245AH249A YFP-DOG1, YFP-DOG1 , and YFP-DOG1 at lines exhibited lower germination and post-germination growth very similar efficiencies, regardless of hemin (Supplementary efficiencies similar to YFP-DOG1ox line, compared to the YFPox 8 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | DOG1 DOGL3 DOGL5.2 Absorbance Absorbance Absorbance (425 nm) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE ab 0.8 0.8 425 DOG1 DOG1 H245A H39A DOG1 DOG1 H71A DOG1 0.6 0.6 H153A DOG1 H249A DOG1 0.4 0.4 × 2.5 × 2.5 0.2 0.2 0.0 0.0 300 400 500 600 700 800 300 400 500 600 700 800 Wavelength (nm) Wavelength (nm) cd YFP YFP-DOG1 #2 0.8 H245AH249A YFP-DOG1 #1 H245AH249A YFP-DOG1 #2 H245A H245AH249A DOG1 YFP-DOG1 #3 H245AH249A DOG1 Hemin 0.6 0.4 × 2.5 0.2 20 0 0.3 0.0 ABA (μM) 300 400 500 600 700 800 Wavelength (nm) Fig. 8 An alternative axial ligand to heme in the wild-type DOG1. a Comparison of electronic absorption spectra of the fully heme-bound wild-type DOG1 and its H245A single mutant. b Overlay of electronic absorption spectra of the fully heme-bound wild-type DOG1 and its His-to-Ala single mutants (H39A, H71A, H153A, and H249A). c Comparison of electronic absorption spectra of the fully heme-bound H245A single mutant DOG1 and H245AH249A double mutant DOG1, as well as free hemin. The fully heme-bound wild-type and mutant DOG1 proteins were prepared by incubating their apo forms with an excess of hemin, followed by a gel filtration column to remove the unbound hemin. All the spectra shown in a–c were measured for the N-terminal His - H245AH249A tagged proteins at concentration of ca. 8 μM. d Post-germination growth efficiencies of overexpressing YFP-DOG1 , YFP-DOG1, and control YFP lines were treated with or without 0.3 μM ABA at 7 days after stratification. Error bars show s.d. of three independent experiments using the same seed batch control line in the presence of ABA (Supplementary Fig. 15c,d). To examine the AHG1-specific regulatory system, we suc- H245AH249A In contrast, the YFP-DOG1 ox lines showed similar cessfully identified four AHG1-interacting proteins (DOG1, germination and post-germination growth efficiencies to the AFP2, TPL, and FVE/MSI4) (Supplementary Fig. 4c), and char- YFPox control line in the presence of ABA (Fig. 8d and acterized the unique properties in the interactions with AHG1, Supplementary Fig. 16c). These in vivo results correlate well with AHG3, and ABI1. Among these AHG1-interacting proteins, the in vitro spectroscopic results and support the biological DOG1 was particularly enticing because AHG1 functions pri- relevance of heme binding for DOG1 function in ABA signaling. marily in the seeds . DOG1 is one of the pivotal regulators of seed dormancy and germination, although its molecular function 2–4 is largely unknown . A recent study has shown that DOG1 also regulates flowering time through microRNA, suggesting that Discussion The solved crystallographic structures of ABA-PYL/RCAR-PP2Cs DOG1 is expressed and functions in mature plants (Supple- mentary Fig. 4d) . Interestingly, AHG1 subfamily members complexes revealed that a conserved tryptophan residue in PP2Cs, excluding AHG1, inserted in the entrance of the internal interact with DOG1, while no other ABI1 subfamily members do (Fig. 4a, b). We also demonstrate the clear genetic and functional cavity of PYL/RCARs formed a water-mediated hydrogen-bond 47–49 interactions between AHG1 and DOG1 loci (Fig. 5). In addition, network with the ABA-bound PYL/RCARs complex . AHG1 removing the N-terminal portion of DOG1 required for the lacks this conserved tryptophan residue and is therefore predicted 20,50 AHG1-interaction impaired the ability of DOG1 to confer an to be unable to interact with ABA-bound PYL/RCARs . This ABA-hypersensitive phenotype in seed germination, indicating idea is supported by the fact that none of the PYL/RCAR ABA that the interaction with AHG1 is indispensable for DOG1 receptors were identified as AHG1-interacting proteins in this study (Supplementary Data 1 and Supplementary Tables 1, 2). function (Fig. 6). Therefore, it is likely that AHG1 and DOG1 constitute an alternative regulatory system, distinct from the PYL/ These observations led us to believe that AHG1 functions in a unique regulatory system that is independent of PYL/RCARs. RCAR-PP2C regulatory system, to control seed dormancy and NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 9 | | | Absorbance Absorbance Post-germination growth (%) Absorbance ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 germination. Since AHG1 and DOG1 homologs are conserved hypersensitive phenotype in germination (Fig. 8), strongly sup- 32,52,53 among higher plants , the DOG1-AHG1 regulatory system porting the importance of heme coordination for DOG1. Heme is may be common. Recently, the same region of DOG1 responsible a well-known key molecule that regulates diverse biological 10,11 for AHG1 interaction had been reported to be a self-dimerization activities . Some heme-binding proteins are reported to 54 11 site . While it remains unclear how self-dimerization is involved function as sensors for oxygen and nitric oxide (NO) , including 58,59 in DOG1 function, DOG1 might be a core protein in a large the widely studied mammalian NO sensor, Guanylate cyclase , protein complex, regulating components including itself and and FixL, a nitrogen-fixing bacteria which is a well-known oxygen AHG1. It is noteworthy that the HA-AHG1 levels seemed to be sensor . Reactive oxygen species (ROS) and NO counteract ABA less accumulated when co-expressed with the deleted forms of to change the phase transition from seed dormancy to germina- 61,62 YFP-DOG1 that can interact with HA-AHG1, in comparison tion . Similarly, heme-binding DOG1 might monitor the ROS with deleted forms of YFP-DOG1 that cannot interact with HA- and the NO levels affected by physiological or developmental AHG1 (Fig. 6b). Although we could not exclude the possibility stimuli in seeds. Indeed, DOG1 has been shown to undergo post- that this is caused by the experimental conditions, DOG1- translational modifications (PTMs) during after-ripening , interaction might affect the AHG1 level. implying that DOG1 is regulated by PTMs related to ROS and 2,62 Epistatic analysis suggested that DOG1 functions upstream of NO, such as cysteine oxidation and S-nitrosylation , and might AHG1 in ABA signaling in seed germination (Fig. 5). AHG1 was be the hub regulator integrating environmental signals. In addi- reported to interact with SnRK2.3 in vivo and regulate the acti- tion, there are reports showing a link between heme and ABA 19,20 vation state of OST1/SnRK2.6 . In our investigation, DOG1 response. The heme scavenger tryptohan-rich sensory protein could reduce the PP2C activity of AHG1, regardless of ABA (TSPO) is principally detected in dry seeds, and is accumulated in vitro with a synthetic SnRK2s phosphorylation site peptide, but under ABA treatments in Arabidopsis, and TSPO overexpression not with a conventional PP2C substrate (Fig. 5c, Supplementary lines show a weak ABA-hypersensitive phenotype in seed 8,63 Fig. 10). We thought that if DOG1 negatively regulates the AHG1 germination . PP2C activity, the HA-AHG1ox/YFP-DOG1ox line would show a The results presented here, along with the previous studies, similar phenotype to the YFP-DOG1ox line. Indeed, the PYL5/ lead us to propose a model as a working hypothesis for the reg- HAB1 double expression lines exhibited a similar phenotype to ulation of seed dormancy and germination (Fig. 9). As previously the PYL5ox expression line, indicating that the PYL5 ABA reported, ABA directly regulates PYL/RCAR-PP2C and down- receptor prohibited HAB1 function . However, the HA- stream components such as SnRK2s and ABI5, which in turn AHG1ox/YFP-DOG1ox double expression line showed an ABA- control seed dormancy and germination. The present research insensitive phenotype similar to the HA-AHG1ox line (Fig. 5b). points to the hypothesis that DOG1 and AHG1 constitute Presumably, the over-production of HA-AHG1 overcame the another parallel regulatory pathway, in which DOG1 integrates inhibitory effect of the over-produced YFP-DOG1. It is possible environmental signals or physiological conditions other than that DOG1 needs additional modifications or interactions with ABA, and the DOG1-AHG1 complex regulates downstream other components for its function. Such a modification or con- components including SnRK2s and ABI5. Presumably, the structing a large complex with multicomponents might be downstream components of these two regulatory pathways required for the proper DOG1 regulation of AHG1. overlap. In addition, there are several PP2Cs that appear to The DOG1-AHG1 complex may also regulate other interactors function in both the pathways, suggesting that both the intense of AHG1. One good candidate is ABI5, which is one of the major cross-talk and the signal integration between the two pathways 55,56 . We identified determinants of seed germination regulation are essential for the regulation of seed dormancy and germination AFP2 and TPL as AHG1-interacting proteins (Supplementary to maximize the environmental adaptability of the plant life cycle. Fig. 3c). AFP negatively regulates ABI5 controlling its protein levels through the ubiquitin-mediated system , and is also reported to interact with TPL to form a transcriptional co- Environmental and repressor complex . Recent studies indicated that DOG1 affects developmental cues the expression level of ABI5, AFPs, and some group A PP2C genes . ABI5 is one of the well-known major substrates of 22,25 subgroup III SnRK2s , and both of these components are also 19,57 reported to interact with AHG1 . Taken together, these Heme ABA observations suggest that the DOG1-AHG1 complex is likely to regulate ABI5 function through SnRK2s and AFPs-TPL repressor complex. ABA Heme Although AHG3 and other AHG1 subfamily members have PYL/RCARs DOG1 been reported to interact with some of PYL/RCAR ABA recep- tors , they all are able to interact with DOG1 (Fig. 4a). Thus these members function like AHG1 in the DOG1-dependent system. This idea is supported by the data that the triple loss-of-function AHG1/ ABI1/ mutants of these PP2Cs showed strong ABA hypersensitivity and AHG3 AHG3 increased the seed dormancy phenotypes in seeds (Fig. 2). Their ability to interact with both DOG1 and PYL/RCAR ABA recep- tors indicates that these PP2Cs can be important hub regulators connecting the two distinct regulatory pathways to integrate ABA responses multiple signals and fine-tune the regulation of seed dormancy and germination. In this study, we demonstrated that DOG1 binds to heme Seed dormancy (Fig. 7), and inferred that histidine residues function as an axial ligand for the bound heme. Conversion of histidine residues to Fig. 9 A hypothesized regulatory system of seed dormancy. See the text for alanines abolished the DOG1 activity to confer the ABA- detail 10 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE We note that, while a prior version of this manuscript was strain GV3101 carrying the gene of interest was used to transform wild-type Col-0 with plants by the floral dip method . Transgenic plants were screened for under revision, the physical and genetic interaction of DOG1 with hygromycin or bialaphos resistance, and the homozygous T3 lines were isolated. PP2Cs was reported . The conclusion of that study, that PP2Cs Subcellular localizations of YFP-AHG1, YFP-DOG1, YFP-DOGL3, and YFP- interact with and are epistatic to DOG1, is consistent with the DOGL5.2 in the root were analyzed by fluorescence microscope (Nikon Eclipse current work. The reduction of PP2C activity of AHG1 Ni). Transgenic plants were verified by immunoblotting using a 1:10,000 diluted anti-GFP antibody (in this study), followed by a 1:50,000 diluted anti-Rabbit HRP- and heme-binding data here provide further biochemical details conjugate (Pierce #31460). Detection was performed using SuperSignal West Dura regarding the function of DOG1. extended Duration Substrate (Pierce), according to the manufacturer’s instructions. The uncropped full-scan data of immunoblots and gel are shown in Supplementary Fig. 17. Methods Plant materials and growth conditions. Arabidopsis thaliana wild-type Columbia 35 33 (Col), ahg1-1 (see ref. ), ahg3-1 (see ref. ), hai1-2 (SALK_1082282), aip1-1 Yeast two-hybrid assay. The Matchmaker Gold Yeast two-hybrid system (SALK_090738), hai3-1 (SALK_033011), dog1-2 (see ref. ), and dog1-3 (Clontech) was used according to the manufacturer’s instruction. Full-length PYR1, (SALK_000867) were used in this study. Plants were grown on MS plates con- group A PP2Cs, DOG1/DOGLs, and AHG1-interacting proteins cDNAs were cloned taining 1× Murashige and Skoog salt mix, 1% sucrose, 2.5 mM MES (pH5.8), and into pACTGW-attR and pASGW-attR. Each vector was transformed into yeast 0.8% agar or on soil at 22 °C under 16 h light/8 h dark cycles. Seeds were first Y2H gold strain (Clontech). The transformed yeast cells were sprayed onto DDO stratified at 4 °C for 4 days and then transferred to a growth chamber, unless (Double dropout medium: SD/-Leu/-Trp) medium agar plates and incubated at 30 otherwise indicated. °C for 5 days. A series of tenfold serial dilutions were spotted onto DDO, QDO/A (Quadruple dropout medium: SD/-Ade/-His/-Leu/-Trp supplemented with Aur- Germination and growth assays. For germination assays, approximately 50 seeds eobasidin A), and QDO/A/ABA (Quadruple dropout medium: SD/-Ade/-His/- were sown on MS plates containing various concentration of ABA (Sigma). Ger- Leu/-Trp supplemented with Aureobasidin A and 10 μM ABA) medium agar plates mination (radicle emergence) and post-germination growth (seedling with for 7 days after inoculation. expanded green cotyledons) were scored daily for 3 days and 7 days. For seed dormancy assays, approximately 50 seeds were sown on hormone-free MS plates. Co-immunoprecipitation experiments in N. benthamiana. Full-length AHG1, Germination (radicle emergence) and post-germination growth (seedling with AHG3, ABI1, and DOG1 (WT and mutants) cDNAs were cloned into pH35YG, expanded green cotyledons) were scored daily for 7 days. For root growth inhi- and the full-length PYR1, AHG1, and DOG1 cDNAs were cloned into pEarleyGate bition assays, 12–15 seedlings were germinated and grown on hormone-free MS 201. A. tumefaciens strain GV3101 carrying the gene of interest was used and plates for 5 days and then transferred to MS plates containing various con- infiltrated at an OD of 0.5, together with p19 strain in N. benthamiana. After centration of ABA (Sigma) for 4 days. For ABA-induced stomatal closure assays, 65 4 days of infiltration, the infiltrated leaves were sprayed with water containing stomatal apertures in the abaxial epidermis were measured microscopically . The 0.01% Silwet L-77 (BMS) and either 100 μM ABA or 50 μM hemin for 24 h prior to epidermal tissues isolated from dark-adapted 4- to 6-week-old plants were incu- leaf excision. For protein extraction, N. benthamiana leaves (0.75 g) were harvested bated in basal buffer (5 mM MES-BTP pH 6.5, 50 mM KCl, and 0.1 mM CaCl ). and ground to a powder in liquid nitrogen. Ground tissues were resuspended into The epidermal tissues were pre-illuminated under light [blue light (Stick-B-32; −2 −1 1.5 mL of extraction buffer (50 mM Na-phosphate pH 7.4, 150 mM NaCl, 0.1% EYELA) at 10 μmol m s superimposed on background red light (LED-R; NP-40, 1 mM DTT and 1× protease inhibitor cocktail (Sigma)). Crude extracts −2 −1 EYELA) at 50 μmol m s ] for 2.5 h, then incubated with or without 20 µM ABA were then centrifuge at 20,000×g for 30 min at 4 °C. The supernatant was passed for 2.5 h under light (same as above). through a Miracloth (Calbiochem) and used for each immunoprecipitation as an input. The input was incubated with 40 μL anti-HA matrix (Roche #11815016001) Vector constructions. Full-length ABI1, ABI2, HAB1, and PYR1 cDNAs were for 3 h at 4 °C. Immunocomplexes were washed four times with the extraction previously cloned into entry vector (Invitrogen). Full-length AHG1, AHG3, buffer without the protease inhibitor cocktail. After washing, the matrix was HAI1, HAI2/AIP1, HAI3, DOG1, DOGL1, DOGL2, DOGL3, DOGL4, and AFP2 resupended into 50 μL 2× SDS sample buffer for 5 min at 95 °C. The protein cDNAs were cloned into pENTR vector (Invitrogen) and sequenced. The entry samples were separated by 12.5% SDS-PAGE gel (ATTO) and blotted onto an clones, which are inserted into full-length HAB2 (CIW00104), TPL (G09916), Immobilon-P membrane (Millipore), and immunodetected using 1:10,000 diluted DOGL5.2 (PENTR221-AT3G14880), FVE/MSI4 (U15529) cDNAs, were ordered anti-GFP (in this study) or anti-HA (Roche #11583816001) antibodies, followed by from ABRC and sequenced. Site-directed mutagenesis was performed with the a 1:50,000 diluted anti-Rabbit HRP-conjugate (Pierce #31460) or anti-Mouse HRP- KOD-Plus-mutagenesis kit (TOYOBO) using entry clones as templates. The entry conjugate (Pierce #31430). The uncropped full-scan data of immunoblots and gel clones were transferred to the destination vectors pACTGW-attR, pASGW-attR, are shown in Supplementary Fig. 17. pH35YG, and pEarleyGate 201 by Gateway LR clonase II enzyme mix recombi- nation reaction. Affinity column purification of YFP-AHG1 complexes. The three-week-old seedlings (10–20 gFW) grown on MS plates were incubated for 2 h in water (-ABA YFP fusion protein expression analyses. Full-length group A PP2Cs cDNAs samples) before being treated with 100 μM ABA for 24 h (+ABA samples). Sam- were cloned into pH35YG. Agrobacterium tumefaciens strain GV3101 carrying the ples were ground to a powder in liquid nitrogen and resuspended into 2× times gene of interest was used and infiltrated at an OD of 0.5 together with p19 strain extraction buffer (50 mM Na-phosphate pH 7.4, 150 mM NaCl, 0.1% NP-40, 1 mM in N. benthamiana. Mesophyll protoplasts were isolated from leaves after 5 days of DTT and 1× protease inhibitor cocktail (Sigma)). Crude extracts were then cen- infiltration, according to instructions . Fluorescence imaging was analyzed by trifuged at 20,000×g for 30 min at 4 °C. The complete supernatant was passed confocal microscopy (Nikon Eclipse TE2000-U) using 488 nm excitation and through a Miracloth (Calbiochem) and 0.45 μm syringe filter (Starlab Scientific), 500–550 nm emission filters for YFP. and loaded onto an anti-GFP (in this study) conjugated 1 mL HiTrap NHS- activated HP column (GE Healthcare). Anti-GFP affinity columns were generated according to the manufacturer’s instructions (GE Healthcare). After a 50 to 100 mL GFP expression and antiserum preparation. sGFP gene was cloned into the wash with the extraction buffer without the protease inhibitor cocktail, YFP control NdeI/XhoI sites of the expression vector pET28a (+) (Novagen) and transformed and complexed AHG1-interacting proteins were eluted into 10 mL elution buffer into E. coli. strain BL21 Rossetta2 (DE3) (Merck). The E.coli. cells carrying (0.3 M Glycine-HCl pH 3.0, 1 mM DTT and 1× protease inhibitor cocktail expression plasmid were grown at 37 °C, and protein expression was induced by (Sigma)), and were fractionated each (0.5 mL per fraction), and were TCA pre- the addition of IPTG to 1 mM at OD 0.6–0.7 in LB medium. After a 3 h cipitated. Western blot was performed using a 1:10,000 diluted anti-GFP antibody incubation, the cells were harvested by 4000×g centrifugation for 10 min at 4 °C (in this study), followed by a 1:50,000 diluted anti-Rabbit HRP-conjugate (Pierce and the pellet was resuspended in the sonic buffer (50 mM sodium phosphate #31460). Gel staining was performed using Oriole fluorescent gel stain (Bio-Rad). buffer pH 7.0, 300 mM NaCl, 10 mM imidazole, EDTA-free protease inhibitor The uncropped full-scan data of immunoblots and gel are shown in Supplementary cocktail (Roche)). Cells were sonicated on ice seven times for 30 s with regular Fig. 17. resting intervals. A supernatant was obtained after centrifugation at 20,000×g for 10 min at 4 °C and incubated with TALON metal resin (Clontech) for 30 min at 4 ° C. After being washed three times in the sonic buffer without the protease inhibitor Mass spectrometry and data analysis. Purified proteins were reduced with 5 mM cocktail, the matrix was resuspended with an elution buffer (50 mM sodium Tris (2-carboxyethyl) phosphine hydrochloride (Sigma-Aldrich) and alkylated. phosphate buffer pH 7.0, 300 mM NaCl, 250 mM imidazole) and incubated for 30 Proteins were digested for 18 h at 37 °C in 2 M urea, 100 mM Tris pH 8.5, 1 mM min. The protein was further purified with Sephacryl S-100 (GE Healthcare) using CaCl with 2 μg trypsin (Promega). Analysis was performed using an Eksigent a solution of PBS buffer. Antisera against GFP were generated by Scrum Inc. and nanopump and a Thermo LTQ Orbitrap using an in-house built electrospray were evaluated by immunoblot for further use in this study as anti-GFP antibody. stage . Protein and peptide identification and protein quantitation were done with The construction of transgenic Arabidopsis lines. Full-length AHG1, DOG1 (WT Integrated Proteomics Pipeline—IP2 (Integrated Proteomics Applications, Inc., San and mutants), DOGL3, and DOGL5.2 cDNAs were cloned into pH35YG and full- Diego, CA. http://www.integratedproteomics.com/). Tandem mass spectra were length AHG1 cDNA was cloned into pEarleyGate 201. Agrobacterium tumefaciens extracted from raw files using RawConverter and were searched against the NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 11 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 TAIR10_pep_20101214 database (https://www.arabidopsis.org/index.jsp) with spectrophotometer (VARIAN) between 250 and 800 nm at room temperature in 70,71 YFP-AHG1, YFP, and reversed sequences added using ProLuCID . The search buffer B. For the titration experiment, hemin stock solution was prepared by space included all full-tryptic and half-tryptic peptide candidates for the tryptic dissolving 4 mg hemin (Wako) in 500 μL of 0.1 M NaOH and then centrifuged at digest with static modification of 57.02146 on cysteine. Peptide candidates were 20,000×g for 15 min. Concentration of the hemin stock solution was determined filtered using DTASelect , with these parameters: -p 2 -y 1 --trypstat --pfp 0.01 −1 −1 73 using an extinction coefficient value of 58,440 M cm at 385 nm . The solution 69,72 --extra --pI -DM 10 --DB --dm -in -t 1 --brief –quiet . was diluted with 20 mM Tris-HCl buffer pH 8.5, containing 5 mM DTT just before use. The hemin titration experiments were carried out on the untagged and His - tagged DOG1, and its His -tagged H245A and H245AH249A mutants by stepwise Double and triple mutants and double-expression lines. The ahg1-1ahg3-1 addition of a 1 μL aliquot of 1 mM hemin solution to 1 mL of 10 μM apo-proteins double mutant was crossed with hai1-2, aip1-1, and hai3-1. Triple mutants were in 20 mM Tris-HCl buffer pH 8.5, containing 5 mM DTT. Samples were incubated obtained from F2 or F3 progeny using mutant-specific CAPS markers and primer for 5 min at room temperature before measuring the absorption spectra. The sets (Supplementary Table 4). The ahg1-1 mutant was crossed with dog1-2 and stoichiometry and affinity of hemin binding to apo-DOG1 were determined by dog1-3. Double mutants were obtained from F2 or F3 progeny using mutant- nonlinear curve fitting of an increase in the absorbance at 425 nm for the untagged specific CAPS markers and primer sets (Supplementary Table 4). HA-AHG1ox line H245A and His -tagged DOG1 at 418 nm for the His -tagged DOG1 and at 416 nm 6 6 was crossed with YFP-DOG1ox #2 line. The double-expression homo line was H245AH249A for the His -tagged DOG1 , as described in the Methods. To obtain the screened for hygromycin and bialaphos resistance, and obtained from homozygous absorption spectra of the fully heme-bound wild-type DOG1 and its mutants, the T3 lines. corresponding proteins in their apo forms were supplemented with 50 μM hemin dissolved in DMSO and incubated for 20 min at room temperature before gel- RNA isolation and RT-PCR experiments. Total RNA was isolated with a RNeasy filtration chromatography, which removed the unbound hemin. It is worth men- Plant mini kit (Qiagen). After treatment with RNase-free DNase I (Qiagen), first- tioning that the fully heme-bound proteins prepared hemin in NaOH and in strand cDNA was synthesized from 1 μg total RNA using Superscript III first- DMSO displayed essentially the same absorption spectra including peak positions strand synthesis super mix for qRT-PCR (Invitrogen), according to the manu- and heights. Protein concentration was estimated, as described in the Methods. facturer’s instructions. DNA fragments for AHG1, DOG1 and 18s rRNA were amplified for 20 or 30 PCR cycles by using gene specific primers (Supplementary Dissociation constant analysis. The dissociation constants, K , of heme to DOG1 Table 4). The uncropped full-scan data of gel are shown in Supplementary Fig. 18. and its H245A and H245AH249A mutants were determined by fitting the increase of absorption at 425 nm for DOG1, 418 nm for the H245A mutant, and 416 nm for the H245AH249A mutant using following equations. Cloning for protein expression. Full-length DOG1 was cloned into the NdeI/ BamHI sites of the expression vector pCold I (TaKaRa) and pET30b (+) (Nova- A ¼ ε ½DHþ ε ½H ; obs DH H gen) to obtain the N-terminal His -tagged and untagged DOG1, respectively. Full- length DOGL3 and DOGL5.2 cDNAs were inserted between NdeI/XhoI and NdeI/ BamHI sites of the pET28b (+) (Novagen), respectively, containing an N-terminal ½DH¼ ð½D þ x½H þ K  fð½D þ x½H þ K Þ d d total total total total His -tag. The catalytic PP2C domain of AHG1 (residues 104-416) cDNA was 1=2 cloned in the pET28b (+) using NdeI/BamHI sites with In-Fusion system (Clon- 4x½H ½D g Þ=2; total total tech). Site-directed mutagenesis was performed with the QuickChange mutagenesis protocol (Stratagene). All plasmid constructs were confirmed by DNA sequencing and were introduced in the Escherichia coli BL21 (DE3) strain (Merck). ½H ¼½H ½DH and ½D ¼½DHþ½D ; f total total f where [DH] represents the concentration of the heme-bound protein while [D] Protein expression and purification. To obtain apo- or heme-bound DOG1 and [H] are concentrations of the free protein and free hemin, respectively. The expression, E. coli cells carrying expression plasmid were grown at 37 °C to reach extinction coefficients of the protein-bound heme, ε , and the free hemin, ε , DH H an OD of 0.5–0.6 in LB medium or TB medium, respectively, each containing were treated as variable parameters. The “x” represents a fraction of the active antibiotics. Apo-DOG1 was induced by addition of IPTG to 0.2 mM and cell hemin, which can be incorporated into DOG1 and its relatives more easily than the growth was continued for 16 h at 16 °C in LB medium (condition I). Heme-bound aggregated hemin. We treated this “x” as a variable parameter because it is difficult DOG1 was induced with 0.1 mM IPTG at 16°C for 50–60 h in TB medium to estimate the exact value due to co-existence of the monomeric and aggregated (condition II). The harvested cells expressing His -tagged DOG1 were resuspended forms of hemin in the solution. Typically, the obtained “x” is larger than 0.8 in our in buffer A (20 mM Tris-HCl pH 8.5, 150 mM NaCl, 1 mM DTT, and 10% gly- experiments. cerol). The cells were lysed by sonication, and the debris was removed by cen- trifugation at 40,000×g for 30 min. The supernatant was purified with a 5 ml HisTrap HP column (GE Healthcare) and was further purified by HiLoad 26/60 Protein concentration. Since free heme molecules could be removed almost Superdex 75 prep grade (GE Healthcare) pre-equilibrated with buffer B (10 mM completely during the chromatographic purification processes, concentrations of Tris-HCl pH 8.5, 150 mM NaCl, and 5 mM DTT). DOGL3, DOGL5.2, and DOG1 chromatographically purified proteins, with the exception of the H245A and mutants were prepared in the same manner. The untagged DOG1-expressed cells H245AH249A mutants of DOG1, were estimated from the absorbance at 280 and were resuspended in buffer C (20 mM Tris-HCl pH 8.5, 5 mM DTT) and lysed by 425 nm using the following equations: sonication. Following centrifugation clarification, the supernatant was loaded to ½D ¼½DHþ½D ; total f HiLoad 16/10 Q sepharose HP column (GE Healthcare) equilibrated with buffer C, and eluted with linear gradient from 0 to 500 mM NaCl. Fractions containing the untagged DOG1 were applied to a HiLoad 26/60 Superdex 75 prep grade (GE ½DH¼ A =ε ð425Þ; 425 DH Healthcare) pre-equilibrated with buffer B. Further purification was performed by Mono Q 10/100 GL (GE Healthcare) equilibrated with buffer C. The untagged DOG1 was eluted in a linear gradient from 0 to 250 mM NaCl. The catalytic PP2C ½D ¼fA  ε ð280Þ½DHg=ε ð280Þ; f 280 DH D domain of AHG1 was expressed and purified by employing the same procedures used for the apo-DOG1. The purity of the expressed proteins was confirmed by where A and A are the observed absorbance at 280 and 425 nm. The values of 280 425 15% SDS-PAGE. A and A were used for the H245A and H245AH249A mutants of DOG1, 418 416 respectively. For the other proteins, the A values were used. The extinction PP2C enzyme assay. Phosphatase activity of AHG1 was measured in 50 μL coefficient of the apo-protein at 280 nm, ε (280), was evaluated from the amino reaction buffer containing 25 mM Tris-HCl (pH7.5), 10 mM MgCl , 1 mM DTT 2 acid sequence of the protein. The extinction coefficient of the heme-bound DOG1 with or without 50 μM ABA using Serine/Threonine phosphatase assay system at 425 nm, ε (425), was obtained from the K analysis of the hemin titration DH d (Promega). The phosphopeptide HSQPK(pS)TVGTP, corresponding to the reg- experiments, and then ε (280) was calculated dividing ε (425) by a factor of DH DH ulatory phosphorylation site of SnRK2s, was synthesized and purified by Scrum 1.43, the Reinheitszahl value (A /A ) determined for the fully heme-bound 425 280 Inc. Using synthetic SnRK2s phosphopeptide as substrate, the reaction buffer was DOG1. The ε (425) and ε (280) thus obtained were used for all the proteins. DH DH added, which is 100 μM SnRK2s phosphopeptide and 200 μM AHG1 with ΔΝ (1-103) Concentrations of the chromatographically purified H245A and H245AH249A or without 800 μM heme-bound DOG1. Using an artificial substrate for phos- mutants were estimated in the same manner using A and A , respectively, 418 416 phatase 2A, 2B, and 2C, reaction buffer was added, which is 25 μM RRA(pT)VA instead of A . peptide (Promega) and 50 μM AHG1 with or without 200 μM heme- ΔΝ (1-103) bound DOG1. After 30 min at 30 °C, each reaction was stopped with 50 μL Circular dichroism spectroscopy. Circular dichroism (CD) spectra of the purified molybdate dye solution. The absobance at 600 nm was measured with a plate recombinant DOG1 were recorded on a J-720W spectropolarimeter (Jasco) in the reader (GloMax-Multi + Detection system; Promega). far-UV range from 185 to 260 nm using a 0.2-mm thermostatted cell at 25 °C, using the following parameters: resolution, 0.2 nm; bandwidth, 1.0 nm; sensitivity, UV–visible absorption spectroscopy. Electronic absorption spectra of the pur- 50 mdeg; response, 1 s; speed, 100 nm/min; accumulation, 8. Solutions used to ified recombinant proteins were measured on a CARY 400 Bio ultraviolet-visible record CD spectra contained apo-DOG1 or heme-bound DOG1 dissolved in 10 12 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 ARTICLE mM sodium phosphate buffer pH 7.5, with 5 mM DTT at protein concentration of 25. Piskurewicz, U. et al. The gibberellic acid signaling repressor RGL2 inhibits 20 μM. Arabidopsis seed germination by stimulating abscisic acid synthesis and ABI5 activity. Plant Cell 20, 2729–2745 (2008). 26. Lee, S. C., Lan, W., Buchanan, B. B. & Luan, S. A protein kinase-phosphatase Data availability. The mass spectrometry proteomics data have been deposited to 74 pair interacts with an ion channel to regulate ABA signaling in plant guard the ProteomeXchange Consortium via the PRIDE partner repository with the cells. Proc. Natl Acad. Sci. USA 106, 21419–21424 (2009). dataset identifier PXD009392. All relevant data that support the findings of this 27. Geiger, D. et al. Activity of guard cell anion channel SLAC1 is controlled by study are available from the corresponding author upon reasonable request. drought-stress signaling kinase-phosphatase pair. Proc. Natl Acad. Sci. USA 106, 21425–21430 (2009). Received: 9 September 2016 Accepted: 1 May 2018 28. Vahisalu, T. et al. Ozone-triggered rapid stomatal response involves the production of reactive oxygen species, and is controlled by SLAC1 and OST1. Plant J. 62, 442–453 (2010). 29. Negi, J. et al. CO regulator SLAC1 and its homologues are essential for anion homeostasis in plant cells. Nature 452, 483–486 (2008). 30. Vahisalu, T. et al. SLAC1 is required for plant guard cell S-type anion channel References function in stomatal signalling. Nature 452, 487–491 (2008). 1. Bentsink, L. & Koornneef, M. Seed dormancy and germination. Arab. Book 6, 31. Schweighofer, A., Hirt, H. & Meskiene, I. Plant PP2C phosphatases: emerging e0119 (2008). functions in stress signaling. Trends Plant Sci. 9, 236–243 (2004). 2. Graeber, K., Nakabayashi, K., Miatton, E., Leubner-Metzger, G. & Soppe, W. J. 32. Xue, T. et al. Genome-wide and expression analysis of protein phosphatase 2C Molecular mechanisms of seed dormancy. Plant Cell Environ. 35, 1769–1786 in rice and Arabidopsis. BMC Genom. 9, 550 (2008). (2012). 33. Nishimura, N. et al. Isolation and characterization of novel mutants affecting 3. Nonogaki, H. Seed dormancy and germination-emerging mechanisms and the abscisic acid sensitivity of Arabidopsis germination and seedling growth. new hypotheses. Front. Plant Sci. 5, 233 (2014). Plant Cell Physiol. 45, 1485–1499 (2004). 4. Bassel, G. W. To grow or not to grow? Trends Plant Sci. 21, 498–505 (2016). 34. Yoshida, T. et al. ABA-hypersensitive germination3 encodes a protein 5. Shu, K., Liu, X. D., Xie, Q. & He, Z. H. Two faces of one seed: hormonal phosphatase 2C (AtPP2CA) that strongly regulates abscisic acid signaling regulation of dormancy and germination. Mol. Plant 9,34–45 (2016). during germination among Arabidopsis protein phosphatase 2Cs. Plant 6. Hubbard, K. E., Nishimura, N., Hitomi, K., Getzoff, E. D. & Schroeder, J. I. Physiol. 140, 115–126 (2006). Early abscisic acid signal transduction mechanisms: newly discovered 35. Kuhn, J. M., Boisson-Dernier, A., Dizon, M. B., Maktabi, M. H. & Schroeder, J. components and newly emerging questions. Genes Dev. 24, 1695–1708 (2010). I. The protein phosphatase AtPP2CA negatively regulates abscisic acid signal 7. Finkelstein, R. Abscisic acid synthesis and response. Arab. Book 11, e0166 transduction in Arabidopsis, and effects of abh1 on AtPP2CA mRNA. Plant (2013). Physiol. 140, 127–139 (2006). 8. Vanhee, C., Zapotoczny, G., Masquelier, D., Ghislain, M. & Batoko, H. The 36. Nishimura, N. et al. ABA-Hypersensitive Germination1 encodes a protein Arabidopsis multistress regulator TSPO is a heme binding membrane protein phosphatase 2C, an essential component of abscisic acid signaling in and a potential scavenger of porphyrins via an autophagy-dependent Arabidopsis seed. Plant J. 50, 935–949 (2007). degradation mechanism. Plant Cell 23, 785–805 (2011). 37. Nishimura, N. et al. PYR/PYL/RCAR family members are major in-vivo ABI1 9. Kobayashi, Y., Ando, H., Hanaoka, M. & Tanaka, K. Abscisic acid participates protein phosphatase 2C-interacting proteins in Arabidopsis. Plant J. 61, in the control of cell cycle initiation through heme homeostasis in the 290–299 (2010). unicellular red alga Cyanidioschyzon merolae. Plant Cell Physiol. 57, 953–960 38. Santiago, J. et al. Modulation of drought resistance by the abscisic acid (2016). receptor PYL5 through inhibition of clade A PP2Cs. Plant J. 60, 575–588 10. Mochizuki, N. et al. The cell biology of tetrapyrroles: a life and death struggle. (2009). Trends Plant. Sci. 15, 488–498 (2010). 39. Lopez-Molina, L., Mongrand, S., Kinoshita, N. & Chua, N. H. AFP is a novel 11. Li, T., Bonkovsky, H. L. & Guo, J. T. Structural analysis of heme proteins: negative regulator of ABA signaling that promotes ABI5 protein degradation. implications for design and prediction. BMC Struct. Biol. 11, 13 (2011). Genes Dev. 17, 410–418 (2003). 12. Bentsink, L., Jowett, J., Hanhart, C. J. & Koornneef, M. Cloning of DOG1,a 40. Garcia, M. E., Lynch, T., Peeters, J., Snowden, C. & Finkelstein, R. A small quantitative trait locus controlling seed dormancy in Arabidopsis. Proc. Natl plant-specific protein family of ABI five binding proteins (AFPs) regulates Acad. Sci. USA 103, 17042–17047 (2006). stress response in germinating Arabidopsis seeds and seedlings. Plant Mol. 13. Nakabayashi, K. et al. The time required for dormancy release in Arabidopsis Biol. 67, 643–658 (2008). is determined by DELAY OF GERMINATION1 protein levels in freshly 41. Pauwels, L. et al. NINJA connects the co-repressor TOPLESS to jasmonate harvested seeds. Plant Cell 24, 2826–2838 (2012). signalling. Nature 464, 788–791 (2010). 14. Dekkers, B. J. et al. The Arabidopsis DELAY OF GERMINATION 1 gene 42. Pazhouhandeh, M., Molinier, J., Berr, A. & Genschik, P. MSI4/FVE interacts affects ABSCISIC ACID INSENSITIVE 5 (ABI5) expression and genetically with CUL4-DDB1 and a PRC2-like complex to control epigenetic regulation interacts with ABI3 during Arabidopsis seed development. Plant J. 85, of flowering time in Arabidopsis. Proc. Natl Acad. Sci. USA 108, 3430–3435 451–465 (2016). (2011). 43. Fujii, H. et al. In vitro reconstitution of an abscisic acid signalling pathway. 15. Ma, Y. et al. Regulators of PP2C phosphatase activity function as abscisic acid Nature 462, 660–664 (2009). sensors. Science 324, 1064–1068 (2009). 16. Park, S. Y. et al. Abscisic acid inhibits type 2C protein phosphatases via the 44. Deana, A. D. et al. An investigation of the substrate specificity of protein PYR/PYL family of START proteins. Science 324, 1068–1071 (2009). phosphatase 2C using synthetic peptide substrates; comparison with protein phosphatase 2A. Biochim. Biophys. Acta 1051, 199–202 (1990). 17. Cutler, S. R., Rodriguez, P. L., Finkelstein, R. R. & Abrams, S. R. Abscisic acid: 45. Drozdetskiy, A., Cole, C., Procter, J. & Barton, G. J. JPred4: a protein emergence of a core signaling network. Annu. Rev. Plant Biol. 61, 651–679 (2010). secondary structure prediction server. Nucleic Acids Res. 43, W389–W394 (2015). 18. Raghavendra, A. S., Gonugunta, V. K., Christmann, A. & Grill, E. ABA 46. Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N. & Sternberg, M. J. The perception and signalling. Trends Plant. Sci. 15, 395–401 (2010). 19. Umezawa, T. et al. Type 2C protein phosphatases directly regulate abscisic Phyre2 web portal for protein modeling, prediction and analysis. Nat. Protoc. acid-activated protein kinases in Arabidopsis. Proc. Natl Acad. Sci. USA 106, 10, 845–858 (2015). 47. Melcher, K. et al. A gate-latch-lock mechanism for hormone signalling by 17588–17593 (2009). abscisic acid receptors. Nature 462, 602–608 (2009). 20. Antoni, R. et al. Selective inhibition of clade A phosphatases type 2C by PYR/ PYL/RCAR abscisic acid receptors. Plant Physiol. 158, 970–980 (2012). 48. Miyazono, K. et al. Structural basis of abscisic acid signalling. Nature 462, 609–614 (2009). 21. Fujii, H. & Zhu, J. K. Arabidopsis mutant deficient in 3 abscisic acid-activated 49. Yin, P. et al. Structural insights into the mechanism of abscisic acid signaling protein kinases reveals critical roles in growth, reproduction, and stress. Proc. Natl Acad. Sci. USA 106, 8380–8385 (2009). by PYL proteins. Nat. Struct. Mol. Biol. 16, 1230–1236 (2009). 22. Nakashima, K. et al. Three Arabidopsis SnRK2 protein kinases, SRK2D/ 50. Dupeux, F. et al. Modulation of abscisic acid signaling in vivo by an engineered receptor-insensitive protein phosphatase type 2C allele. Plant SnRK2.2, SRK2E/SnRK2.6/OST1 and SRK2I/SnRK2.3, involved in ABA Physiol. 156, 106–116 (2011). signaling are essential for the control of seed development and dormancy. Plant Cell Physiol. 50, 1345–1363 (2009). 51. Huo, H., Wei, S. & Bradford, K. J. DELAY OF GERMINATION1 (DOG1) regulates both seed dormancy and flowering time through microRNA 23. Fujita, Y. et al. Three SnRK2 protein kinases are the main positive regulators pathways. Proc. Natl Acad. Sci. USA 113, E2199–E2206 (2016). of abscisic acid signaling in response to water stress in Arabidopsis. Plant Cell Physiol. 50, 2123–2132 (2009). 52. Ashikawa, I., Abe, F. & Nakamura, S. Ectopic expression of wheat and barley 24. Brandt, B. et al. Calcium specificity signaling mechanisms in abscisic acid DOG1-like genes promotes seed dormancy in Arabidopsis. Plant Sci. 179, 536–542 (2010). signal transduction in Arabidopsis guard cells. eLife 4, e03599 (2015). NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications 13 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-04437-9 53. Graeber, K. et al. Cross-species approaches to seed dormancy and 72. Tabb, D. L., McDonald, W. H. & Yates, J. R. III DTASelect and contrast: tools germination: conservation and biodiversity of ABA-regulated mechanisms for assembling and comparing protein identifications from shotgun and the Brassicaceae DOG1 genes. Plant Mol. Biol. 73,67–87 (2010). proteomics. J. Proteome Res. 1,21–26 (2002). 54. Nakabayashi, K., Bartsch, M., Ding, J. & Soppe, W. J. Seed dormancy in 73. Dawson, R. M. C., Elliott, D. C., Elliott, W. H. & Jones, K. M. Data for Arabidopsis requires self-binding ability of DOG1 protein and the presence of biochemical research third edition, 230–231 (Oxford Science Publications, multiple isoforms generated by alternative splicing. PLoS Genet. 11, e1005737 1986). (2015). 74. Vizcaino, J. A. et al. 2016 Update of the PRIDE database and its related tools. 55. Finkelstein, R. R. & Lynch, T. J. The Arabidopsis abscisic acid response gene Nucleic Acids Res. 44(D1), D447–D456 (2016). ABI5 encodes a basic leucine zipper transcription factor. Plant Cell 12, 599–609 (2000). Acknowledgements 56. Lopez-Molina, L., Mongrand, S. & Chua, N. H. A postgermination We are grateful to Dr. Claire Delahunty and to Mr. Nathaniel Duncan for critical reading developmental arrest checkpoint is mediated by abscisic acid and requires the and comments on the manuscript. We thank Drs. Wim Soppe and Kazumi Nakabayashi ABI5 transcription factor in Arabidopsis. Proc. Natl. Acad. Sci. USA 98, for dog1-2 seeds, Dr. Paul Verslues for hai1-2, aip1-1 and hai3-1 seeds, Dr. Manabu 4782–4787 (2001). Nakayama for pACTGW-attR and pASGW-attR vectors, Dr. Taku Demura for pH35YG 57. Lynch, T., Erickson, B. J. & Finkelstein, R. R. Direct interactions of ABA- vector, Dr. Yasuo Niwa for sGFP vector, ABRC for providing plant materials or reagents, insensitive(ABI)-clade protein phosphatase(PP)2Cs with calcium-dependent Dr. Mitsuhiro Miyazawa for measuring CD spectra, and Ms. Izumi Odano for technical protein kinases and ABA response element-binding bZIPs may contribute to assistance. This research was supported by MEXT/JSPS KAKENHI (Grant Numbers turning off ABA response. Plant Mol. Biol. 80, 647–658 (2012). 23688044, 26712030, and 23119521) and MEXT as part of Joint Research Program 58. Arnold, W. P., Mittal, C. K., Katsuki, S. & Murad, F. Nitric oxide activates implemented at the Institute of Plant Science and Resources, Okayama University in guanylate cyclase and increases guanosine 3′:5′-cyclic monophosphate levels in Japan (2608, 2707, and 2805) to N.N., MEXT KAKENHI (15H05956) to T.K., National various tissue preparations. Proc. Natl Acad. Sci. USA 74, 3203–3207 (1977). Institutes of Health (GM060396) to J.I.S., the National Institute of General Medical 59. Gerzer, R., Radany, E. W. & Garbers, D. L. The separation of the heme and Sciences (8 P41 GM103533) to J.J.M and J.R.Y., and the National Institute of Agrobio- apoheme forms of soluble guanylate cyclase. Biochem. Biophys. Res. Commun. logical Sciences Strategic Fund to T.Y. 108, 678–686 (1982). 60. Gilles-Gonzalez, M. A., Ditta, G. S. & Helinski, D. R. A haemoprotein with kinase activity encoded by the oxygen sensor of Rhizobium meliloti. Nature Author contributions 350, 170–172 (1991). N.N., W.T., J.I.S., and T.Y. designed the research; N.N., W.T., J.J.M., Y.H., K.S., T.I., and 61. Liu, Y., Ye, N., Liu, R., Chen, M. & Zhang, J. H O mediates the regulation of 2 2 N.K. performed the experiments; N.N., W.T., J.J.M., Y.H., K.S., T.K., J.I.S., J.R.Y., T.H., ABA catabolism and GA biosynthesis in Arabidopsis seed dormancy and and T.Y. analyzed the data; N.N., W.T., J.J.M., J.I.S., T.H., and T.Y. wrote the manuscript. germination. J. Exp. Bot. 61, 2979–2990 (2010). 62. Arc, E., Galland, M., Godin, B., Cueff, G. & Rajjou, L. Nitric oxide implication in the control of seed dormancy and germination. Front. Plant Sci. 4, 346 Additional information (2013). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 63. Guillaumot, D. et al. The Arabidopsis TSPO-related protein is a stress and 018-04437-9. abscisic acid-regulated, endoplasmic reticulum-Golgi-localized membrane Competing interests: The authors declare no competing interests. protein. Plant J. 60, 242–256 (2009). 64. Née, G. et al. DELAY OF GERMINATION1 requires PP2C phosphatases of the ABA signalling pathway to control seed dormancy. Nat. Commun. 8,72 Reprints and permission information is available online at http://npg.nature.com/ (2017). reprintsandpermissions/ 65. Tsuzuki, T. et al. Mg-chelatase H subunit affects ABA signaling in stomatal guard cells, but is not an ABA receptor in Arabidopsis thaliana. J. Plant Res. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in 124, 527–538 (2011). published maps and institutional affiliations. 66. Yoo, S. D., Cho, Y. H. & Sheen, J. Arabidopsis mesophyll protoplasts: a versatile cell system for transient gene expression analysis. Nat. Protoc. 2, 1565–1572 (2007). 67. Clough, S. J. & Bent, A. F. Floral dip: a simplified method for Agrobacterium- Open Access This article is licensed under a Creative Commons mediated transformation of Arabidopsis thaliana. Plant J. 16, 735–743 (1998). Attribution 4.0 International License, which permits use, sharing, 68. Wolters, D. A., Washburn, M. P. & Yates, J. R. 3rd An automated adaptation, distribution and reproduction in any medium or format, as long as you give multidimensional protein identification technology for shotgun proteomics. appropriate credit to the original author(s) and the source, provide a link to the Creative Anal. Chem. 73, 5683–5690 (2001). Commons license, and indicate if changes were made. The images or other third party 69. He, L., Diedrich, J., Chu, Y. Y. & Yates, J. R. 3rd Extracting accurate precursor material in this article are included in the article’s Creative Commons license, unless information for tandem mass spectra by raw converter. Anal. Chem. 87, indicated otherwise in a credit line to the material. If material is not included in the 11361–11367 (2015). article’s Creative Commons license and your intended use is not permitted by statutory 70. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J. & Gygi, S. P. Evaluation of regulation or exceeds the permitted use, you will need to obtain permission directly from multidimensional chromatography coupled with tandem mass spectrometry the copyright holder. To view a copy of this license, visit http://creativecommons.org/ (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. licenses/by/4.0/. J. Proteome Res. 2,43–50 (2003). 71. Xu, T. et al. ProLuCID: an improved SEQUEST-like algorithm with enhanced sensitivity and specificity. J. Proteom. 129,16–24 (2015). © The Author(s) 2018 14 NATURE COMMUNICATIONS (2018) 9:2132 DOI: 10.1038/s41467-018-04437-9 www.nature.com/naturecommunications | | |

Journal

Nature CommunicationsSpringer Journals

Published: Jun 6, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off