lifting della repression of arabidopsis seed germination · from juvenile to adult plant...

Lifting DELLA Repression of Arabidopsis Seed Germinationby Nonproteolytic Gibberellin Signaling1[C][W][OPEN]

Tohru Ariizumi2,3, Amber L. Hauvermale2, Sven K. Nelson, Atsushi Hanada,Shinjiro Yamaguchi, and Camille M. Steber*

United States Department of Agriculture-Agricultural Research Service, Wheat Genetics, Quality, Physiology,and Disease Research Unit (C.M.S.), Department of Crop and Soil Science (T.A., C.M.S.), and Molecular PlantSciences Program (A.L.H., S.K.N., C.M.S.), Washington State University, Pullman, Washington 99164–6420;and Graduate School of Life Sciences, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980–8577,Japan (A.H., S.Y.)

ORCID ID: 0000-0001-6255-7670 (C.M.S.).

DELLA repression of Arabidopsis (Arabidopsis thaliana) seed germination can be lifted either through DELLA proteolysis by theubiquitin-proteasome pathway or through proteolysis-independent gibberellin (GA) hormone signaling. GA binding to theGIBBERELLIN-INSENSITIVE DWARF1 (GID1) GA receptors stimulates GID1-GA-DELLA complex formation, which in turntriggers DELLA protein ubiquitination and proteolysis via the SCFSLY1 E3 ubiquitin ligase and 26S proteasome. AlthoughDELLA cannot be destroyed in the sleepy1-2 (sly1-2) F-box mutant, long dry after-ripening and GID1 overexpression can relievethe strong sly1-2 seed dormancy phenotype. It appears that sly1-2 seed dormancy results from abscisic acid (ABA) signalingdownstream of DELLA, since dormant sly1-2 seeds accumulate high levels of ABA hormone and loss of ABA sensitivity rescuessly1-2 seed germination. DELLA positively regulates the expression of XERICO, an inducer of ABA biosynthesis. GID1boverexpression rescues sly1-2 germination through proteolysis-independent DELLA down-regulation associated with increasedexpression of GA-inducible genes and decreased ABA accumulation, apparently as a result of decreased XERICO messengerRNA levels. Higher levels of GID1 overexpression are associated with more efficient sly1 germination and increased GID1-GA-DELLA complex formation, suggesting that GID1 down-regulates DELLA through protein binding. After-ripening results inincreased GA accumulation and GID1a-dependent GA signaling, suggesting that after-ripening triggers GA-stimulated GID1-GA-DELLA protein complex formation, which in turn blocks DELLA transcriptional activation of the XERICO inhibitor of seedgermination.

Gibberellins (GAs) are tetracyclic diterpenoid planthormones that stimulate many developmental transi-tions in the plant life cycle, including the transitionfrom seed dormancy to germination, the phase changefrom juvenile to adult plant development, stem elon-gation, and the transition to flowering and floral de-velopment (for review, see Sun and Gubler, 2004). Therole of GA in these processes has been establishedthrough studies of GA biosynthesis mutants, like themutation in copalyl synthetase in ga1-3. The failure to

synthesize GA results in failure to germinate, extremedwarfism, delayed transition to flowering, and poorfertility (Koornneef and van der Veen, 1980). All of thesephentoypes are reversed by GA application, dem-onstrating the direct connection between GA signalingand physiological responses. This study examines therole of Arabidopsis (Arabidopsis thaliana) GIBBERELLIN-INSENSITIVE DWARF1 (GID1) GA receptors in sleepy1(sly1)-independent GA signaling with a focus on seedafter-ripening and germination.

Seeds are an important source of human nutrition and avehicle for plant propagation and dispersal. Making thedecision to germinate at the appropriate time is essentialboth to agriculture and to species survival. Seeds can bedormant when they are first released from the motherplant so that they fail to germinate under favorable con-ditions (for review, see Bewley and Black, 1994; Finkelsteinet al., 2008). Seed dormancy prevents germination on themother plant or out of season, reduces competition be-tween siblings, and allows species to survive naturalcatastrophes as ungerminated seeds in the soil. Dormantseeds eventually acquire the ability to germinate througha period of dry storage called after-ripening and/orthrough a period of moist chilling called cold stratification.

Seed dormancy is associated with higher levels ofthe plant hormone abscisic acid (ABA), whereas theability to synthesize GA is associated with dormancy-

1 This work was supported by the National Science Foundation(grant no. 0850981 to C.M.S.) and by the RIKEN Plant Science Center(to S.Y.).

2 These authors contributed equally to the article.3 Present address: University of Tsukuba, Gene Research Center,

1-1-1 Tenno-dai, Tsukuba 305–8572, Japan.* Corresponding author; e-mail [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Camille M. Steber ([email protected]).

[C] Some figures in this article are displayed in color online but inblack and white in the print edition.

[W] The online version of this article contains Web-only data.[OPEN] Articles can be viewed online without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.113.219451

Plant Physiology�, August 2013, Vol. 162, pp. 2125–2139, www.plantphysiol.org � 2013 American Society of Plant Biologists. All Rights Reserved. 2125

Dow

nloaded from https://academ

ic.oup.com/plphys/article/162/4/2125/6110907 by guest on 19 June 2021

http://orcid.org/0000-0001-6255-7670mailto:[email protected]://www.plantphysiol.orgmailto:[email protected]://www.plantphysiol.org/cgi/doi/10.1104/pp.113.219451

breaking treatments (Karssen et al., 1983; Jacobsenet al., 2002; Yamauchi et al., 2004; Lefebvre et al.,2006; Okamoto et al., 2006). ABA accumulation preventspremature germination during embryo maturationand establishes and later maintains seed dormancy(for review, see Finkelstein et al., 2008). Loss of dor-mancy through after-ripening is associated with ABAturnover (Okamoto et al., 2006). GA, on the otherhand, accumulates during cold stratification and is re-quired for Arabidopsis seed germination (Koornneefand van der Veen, 1980; Yamauchi et al., 2004).According to the hormone balance theory, the antag-onism between these two hormones regulates seeddormancy, cold stratification, after-ripening, and ger-mination (Karssen and Lacka, 1986; Finch-Savageet al., 2007). Seeds of the GA biosynthesis mutantsga1, ga2, and ga3 cannot germinate and do not after-ripen sufficiently to germinate without GA applica-tion (Koornneef and van der Veen, 1980). It is difficultto distinguish whether these seeds fail to germinatebecause they are highly dormant or because of a physicalrequirement for GA in seed germination. The GA-insensitive sly1 mutants also show increased seed dor-mancy, and the degree of seed dormancy is dependenton the severity of the sly1 allele. The seed dormancyphenotype of sly1-2 can be rescued by the ABA-insensitive mutation abi1-1 by removing the seed coatand by long after-ripening for 1 to 2 years (Steber et al.,1998; Ariizumi and Steber, 2007). The fact that the GA-insensitive sly1-2 mutant shows increased seed dor-mancy, but regains the ability to germinate with longafter-ripening, provides a useful tool to examine theroles of GA and ABA in seed dormancy relief.

At the molecular level, GA hormone induces GAsignal transduction by triggering proteasomal degra-dation of DELLA repressors of GA responses. DELLA(for Asp-Glu-Leu-Leu-Ala) family proteins are definedby the presence of the highly conserved D-E-L-L-A andV-H-Y-N-P amino acid sequences in the N-terminalDELLA regulatory domain and by the C-terminalfunctional domain with homology to the GRAS (forGAI, RGA and Scarecrow) family of putative tran-scription factors (for review, see Sun and Gubler,2004). Mutations in the N-terminal DELLA regulatorydomain result in varying degrees of gain-of-functionGA-insensitive phenotypes (i.e. dwarfism), whereasmutations in the GRAS functional domain result inloss-of-function GA-hypersensitive phenotypes (i.e.increased height). The Arabidopsis genome containsfive DELLA genes, RGA (for REPRESSOR OF GA1),GAI (for GA-INESENSITIVE), RGL1 (for RGA-LIKE1),RGL2, and RGL3 with partly overlapping functions.The roles of Arabidopsis DELLA genes were definedbased on double mutant analysis with ga1-3 and de-pend largely on the location and timing of gene ex-pression (Cheng et al., 2004; Tyler et al., 2004; Caoet al., 2005; Piskurewicz and Lopez-Molina, 2009;Gallego-Bartolomé et al., 2010). RGA, GAI, and RGL1are the main repressors of stem elongation, whereasRGL2, RGA, and RGL1 repress the transition to flowering.

RGL2 is the main repressor of seed germination, althoughRGA, GAI, and RGL3 also contribute. Mutations in RGL2rescue the germination phenotypes of the ga1-3 biosyn-thesis mutant and of sly1 alleles (Lee et al., 2002; Tyleret al., 2004; Ariizumi and Steber, 2007).

GA lifts the DELLA repression of GA responses bytriggering DELLA proteolysis through the ubiquitinproteasome pathway (Wang et al., 2009). The GA re-ceptor GID1 undergoes a conformational change whenit binds GA that stimulates DELLA protein binding toform the GID1-GA-DELLA complex (Murase et al.,2008; Shimada et al., 2008; Ueguchi-Tanaka et al., 2008;Hirano et al., 2010). GID1 binding to DELLA proteinrequires the N-terminal DELLA regulatory domain(Dill and Sun, 2001; Ueguchi-Tanaka et al., 2007). GID1-GA-DELLA complex formation stimulates DELLA rec-ognition by the F-box protein SLY1, leading to DELLApolyubiquitinaiton, which in turn targets DELLA fordestruction by the 26S proteasome (McGinnis et al.,2003; Dill et al., 2004; Fu et al., 2004). SLY1 is the F-boxsubunit of the SCF (for Skp, Cullin, F-box) E3 ubiquitinligase that is the main SCF responsible for the recog-nition and proteasomal targeting of DELLA (Ariizumiet al., 2011). There are three GID1 homologs in Arab-idopsis, GID1a, GID1b, and GID1c, all of which bindGA, form stable complexes with DELLAs, and showoverlapping function in GA signaling (Griffiths et al.,2006; Willige et al., 2007). GID1b has a higher bindingaffinity for bioactive GA and shows limited ability tobind DELLA in the absence of GA (Nakajima et al.,2006; Suzuki et al., 2009; Yamamoto et al., 2010). Triplegid1a gid1b gid1c mutants and loss of sly1 function re-sult in GA-insensitive phenotypes associated with aninability to degrade DELLA protein, indicating thatthese genes are positive regulators of GA signalingbecause they are negative regulators of DELLA repressors.Both sly1 and gid1a gid1b gid1cmutants result in eitherincreased seed dormancy or a complete failure to ger-minate (Griffiths et al., 2006; Ariizumi and Steber, 2007;Iuchi et al., 2007; Willige et al., 2007). However, singleand double gid1a, gid1b, and gid1c mutants cause no orlittle change in germination capacity (Willige et al., 2007).

The GA receptor GID1 can down-regulate DELLArepressors of GA signaling via both proteolysis-dependent and proteolysis-independent mechanisms(Ariizumi et al., 2008; Ueguchi-Tanaka et al., 2008). IfGA signaling is controlled only by DELLA degradation,then we would expect the level of DELLA accumulationto positively correlate with the severity of GA signalingphenotypes. In contrast, the sly1 F-box mutants ofArabidopsis have weaker GA-insensitive seed dormancy,plant height, and infertility phenotypes than ga1-3 andgid1a gid1b gid1c mutants, but they accumulate higherlevels of DELLA proteins (McGinnis et al., 2003; Griffithset al., 2006; Willige et al., 2007; Ariizumi et al., 2008). Thissuggests that some GA signaling can occur in sly1 mu-tants even in the presence of DELLA repressors. This sly1and proteolysis-independent GA signaling requires GAand the DELLA domain needed for the formation of theGID1-GA-DELLA complex. GID1 overexpression and

2126 Plant Physiol. Vol. 162, 2013

Ariizumi et al.

Dow



GA application partly rescue sly1 dwarf and infertilityphenotypes with no reduction in DELLA protein levels(Ariizumi et al., 2008). Thus, there appears to be amechanism for GA signaling that does not require DELLAproteolysis but does require DELLA-GID1 protein-proteininteraction.This study examines the GA response mechanisms,

especially those controlling germination, in the sly1mutant background, where DELLA protein is not regu-lated by the ubiquitin-proteasome pathway. The sly1-2mutant can be rescued by long after-ripening, abi1-1 mu-tations, andGID1 overexpression without any decrease inDELLA RGL2 protein levels. Rescue of sly1-2 by after-ripening and GID1 overexpression is associated withdecreased ABA accumulation, suggesting that decreasedABA signaling is one mechanism allowing the rescue ofsly1-2 by after-ripening and GID1 overexpression. After-ripening also leads to a marked increase in GA hormoneaccumulation. Both after-ripening and GID1 overexpressionlead to an increase in GA-regulated gene expression,suggesting that proteolysis-independent GA signalingis also involved in sly1-2 rescue by after-ripening andGID1 overexpression. This notion is further supportedby the fact that gid1mutations interfere with the abilityof sly1-2 to after-ripen as well as with other aspects ofGA signaling in the sly1-2 mutant. Finally, this studyfinds a correlation between the capacity to germinateand GID1-GA-DELLA complex formation, providing,to our knowledge, the first in planta evidence that GID1relieves seed dormancy through the down-regulation ofDELLA via direct protein-protein interaction.

RESULTS

GA Receptor GID1 Overexpression and the abi1-1 MutationPartially Rescue sly1 Mutant Seed Germination withoutTriggering DELLA Proteolysis

Previous research showed that GID1 overexpressioncan rescue plant height and flowering phenotypes ofsly1 mutants (Ariizumi et al., 2008; Ueguchi-Tanakaet al., 2008). To explore the role of GID1 in controllingseed germination, we examined whether GID1 over-expression could rescue sly1 seed germination. The effectof HA:GID1a, HA:GID1b, and HA:GID1c overexpressionon the constitutive cauliflower mosaic virus 35S pro-moter (lines called GID1a-OE, GID1b-OE, GID1c-OE) onseed germination was examined in sly1-2 seeds in theLandsberg erecta (Ler) background (Supplemental Fig.S1A). All seeds were after-ripened for 2 weeks. Dor-mant sly1-2 seeds showed very low germination over9 d of imbibition, whereas GID1 overexpression re-sulted in partial restoration of seed germination (Fig. 1A).GID1b-OE in sly1-2 resulted in 81% seed germination, andGID1a-OE and GID1c-OE resulted in 61% and 46% ger-mination, respectively. The rescue of sly1-2 seedgermination by GID1b-OE was similar to rescueresulting from the abi1-1 ABA-insensitive mutation(84%; Fig. 1A; Steber et al., 1998). Previous research

showed that abi1-1 causes increased germination ofsly1-2 in the presence of ABA (Steber et al., 1998). GID1b-OEdid not allow sly1-2 to germinate on 0.3mMABA, suggestingthat abi1-1 and GID1b-OE rescue sly1-2 germination bydifferent mechanisms (Supplemental Fig. S1B).

Long after-ripening can rescue sly1-10 and sly1-2seed germination without causing a decrease inDELLA RGL2 or RGA protein accumulation (Ariizumiand Steber, 2007). Protein-blot analysis was used todetermine whether the rescue of sly1-2 seed germina-tion by GID1 overexpression or by the abi1-1 mutationwas associated with DELLA RGL2 protein disappear-ance during a time course of seed imbibition (Fig. 1B;Supplemental Fig. S1B). No decrease in DELLA RGL2protein accumulation was observed in the sly1-2 abi1-1 lineor in sly1-2 lines overexpressing HA:GID1a, HA:GIDb, andHA:GIDc over 4 d of seed imbibition. This result suggeststhat, like after-ripening, abi1-1 and GID1 overexpressionsuppression of sly1 dormancy is independent of DELLAproteolysis.

Figure 1. Rescue of sly1 germination by GID1 overexpression is notassociated with decreased RGL2 protein levels. A, Germination of0.5-month-old wild-type (WT), sly1-2, sly1-2 abi1-1, sly1-2 GID1a-OE,sly1-2 GID1b-OE, and sly1-2 GID1c-OE seeds imbibed on 0.53MS-moistened filter paper at 22˚C. Percentage germination was scoreddaily. B, Accumulation of RGL2 protein during the time course of seedimbibition. Seeds were imbibed on filter paper at 22˚C and harvesteddaily over 4 d for protein-blot analysis of RGL2 accumulation. Cor-responding germination rates are shown.

Plant Physiol. Vol. 162, 2013 2127

Nonproteolytic Gibberellin Signaling in Seed Germination

Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

An ABA dose-response experiment was used toexamine the effect of GID1b-OE on the germinationpotential of wild-type Ler and gai-1, the gain-of-functionDELLA mutant that lacks the 17-amino acid DELLAdomain required for interaction with GID1 proteins.While the gai-1 mutant shows 82% germination in theabsence of ABA, it is hypersensitive to 0.3 and 0.6 mMABA (Fig. 2). GID1b overexpression caused a significantincrease in the ability to Ler to germinate in the presenceof 0.6 mM ABA (20% versus 59% germination) and at1.2 mM ABA. In contrast, GID1b-OE caused no increasein the germination of gai-1 in the presence of ABA. Thisresult suggests that the increased germination potentialresulting from GID1b-OE is not specific to the sly1-2mutant background and that this increased germinationpotential depends, at least in part, on the DELLA do-main of GAI.

Bioactive GA and ABA Hormone Measurementsin Imbibing Seeds

If the increased seed dormancy of sly1-2 resultsmainly from elevated ABA levels, then dormant sly1-2seeds (2 weeks after-ripened) should show higher ABAlevels than sly1-2 after-ripened for 2 years or sly1-2seeds (2 weeks after-ripened) whose germinationcapacity is restored by GID1 overexpression. Massspectrometry was used to examine endogenous levelsof ABA and bioactive GA (Fig. 3; Supplemental Fig.S2). Seeds of wild-type Ler, dormant and after-ripenedsly1-2, and sly1-2 GID1b-OE were imbibed for 60 h at4°C, followed by 12 h at 22°C, and then harvested forhormone measurement. Dormant sly1-2 seeds havehigher endogenous ABA levels than the wild type (P,0.001). ABA levels are lower in after-ripened sly1-2 andin sly1-2 GID1b-OE seeds than in dormant sly1-2 seeds(P , 0.05). This is a small, but statistically significant,decrease based on a two-sided Student’s t test, sug-gesting that the lower ABA levels contribute to the

increased seed germination capacity (SupplementalFig. S2). Dormant sly1-2 seeds accumulated signifi-cantly higher GA4 and GA1 levels than did wild-typeLer (P , 0.05). After-ripening for 2 years results in ahighly significant increase in GA4 levels comparedwith dormant sly1-2 (P , 0.0001), suggesting thatincreased GA signaling contributes to the rescue ofsly1-2 germination by after-ripening. GID1b-OE didnot result in a significant change in GA hormonelevels compared with dormant sly1-2. Thus, it appearsthat sly1-2 seed germination can be partly rescuedthrough decreased ABA signaling, through increasedGA signaling, or through both.

After-ripening and Germination of sly1-2 SeedsRequire GID1a

If after-ripening of sly1-2 results mainly from de-creased ABA signaling, then sly1-2 after-ripening shouldbe independent of the presence of the GA receptor GID1.Thus, we examined the effect of gid1 mutations on theseed dormancy phenotype of sly1-2 in the Columbia

Figure 2. Overexpression of GID1b in the gai1-1 background does notreduce the ABA sensitivity of gai-1. An ABA dose-response curve wasperformed using wild-type Ler, Ler GID1b-OE, gai-1, and gai-1 GID1b-OEseeds. Seeds were incubated on MS-agar containing the indicated ABAconcentrations at 4˚C for 3 d followed by 96 h at 22˚C. Error bars indicate SEfor three replicates (n = 30). [See online article for color version of this figure.]

Figure 3. Hormone measurement of imbibing seeds. Hormone mea-surements were conducted using seeds after-ripened for 2 weeks ex-cept for sly1-2(AR), which was after-ripened for 28 months. All seedswere imbibed on MS-agar at 4˚C for 60 h followed by 22˚C for 12 h.ABA and GA4 levels are averages of six to 10 samples. Error bars showSE. DW, Dry weight.


Ariizumi et al.

Dow




(Col-0) ecotype. Introduction of gid1a-1 and gid1c-2mutations into sly1-2 resulted in decreased seed ger-mination (Fig. 4A; Supplemental Fig. S3). While sly1-2and sly1-2 gid1c-2 germination improved with 20 monthsof after-ripening, sly1-2 gid1a-1 and sly1-2 gid1a-1 gid1c-2seed germination did not. The germination of these andother dormant seeds was rescued only by cutting the seedcoat. The sly1-2 gid1b-1 double mutant also failed togerminate but was not analyzed carefully due to lowfertility. Next, we examined whether the increaseddormancy of the sly1-2 gid1a-1 double mutants wasassociated with increased accumulation of the DELLARGL2 repressor. The increased seed dormancy of sly1-2gid1a-1 and sly1-2 gid1a-1 gid1c-2 compared with sly1-2was associated not with increased, but with decreased,

levels of the DELLA RGL2 repressor (Fig. 4B). Thissituation, where a high degree of seed dormancy isaccompanied by lower DELLA protein accumulation,resembles that in ga1-3 and the gid1a gid1b gid1c triplemutant (Fig. 4B; Willige et al., 2007). The reducedRGL2 protein levels in sly1-2 gid1a-1 and in sly1-2 gid1c-2are associated with reduced RGL2 transcript accumula-tion, suggesting that reduced mRNA levels are at leastpartly responsible. It appears that while there is lessDELLA RGL2 in sly1-2 gid1a-1 than in sly1-2, thisDELLA protein is more active as a repressor of GAsignaling and seed germination.

GA Signaling in sly1 Shoots and Flowers Requires GID1

Although sly1 mutants have higher levels of DELLArepressor protein than either ga1-3 or the gid1a gid1bgid1c triple mutant, sly1 mutants show less extremeGA-insensitive phenotypes, suggesting that some GAsignaling can occur in the absence of DELLA destruc-tion (Ariizumi et al., 2008). If GA signaling in sly1mutants results from a GID1-dependent mechanism,then we would expect gid1 mutations to increase theseverity of sly1 plant height, flowering, and fertilityphenotypes. To examine this, sly1-2 gid1a-1, sly1-2gid1b-1, and sly1-2 gid1c-2 double and triple mutantslines were constructed in the Col-0 ecotype (Fig. 5).Both the sly1-2 gid1a-1 and sly1-2 gid1c-2 double mu-tants showed a significant decrease in final plantheight and fertility compared with sly1-2, with sly1-2gid1a-1 showing the strongest dwarfism of the doublemutants (Fig. 5A; Supplemental Fig. S4). The gid1a-1 mu-tation significantly reduced the plant height and fertility ofsly1-2, even in the heterozygous condition (sly1-2/sly1-2GID1a+/gid1a-1), indicating that gid1a can be codominant(Supplemental Fig. S4). Even stronger plant height andinfertility phenotypes were observed in triple mutantcombinations. It appears that GID1b also contributes tostem elongation, since the sly1-2 gid1a-1 gid1b-1 triplemutant shows the most extreme dwarfism (Fig. 5A;Supplemental Fig. S5B).

Introduction of the gid1b-1 allele into sly1-2 resultedin the strongest infertility and floral developmentphenotypes. The sly1-2 gid1b-1 double mutant was al-most completely infertile, and the sly1-2 gid1a gid1band sly1-2 gid1b gid1c triple mutants were completelyinfertile (Supplemental Fig. S4). The increasing infer-tility of the sly1-2 gid1b-1 double and triple mutantswas associated with increasingly defective floral de-velopment (Fig. 5C). Compared with sly1-2, the sly1-2gid1b-1 and sly1-2 gid1b-1 gid1c-2 lines had smallerflowers with underdeveloped carpels, smaller petals,shorter filaments, and small anthers showing failedpollen dehiscence. Interestingly, the sly1-2 gid1b-1 andsly1-2 gid1b-1 gid1c-2 mutants were slightly taller thansly1-2 at maturity (Supplemental Fig. S5A). Possibly,these multiple mutants continued to grow taller andmake flowers to compensate for extreme infertility.The sly1-2 gid1a-1 gid1b-1 triple mutant is similar to thegid1a-1 gid1b-1 gid1c-2 triple mutant in that it did not

Figure 4. GID1a is required for after-ripening and controls RGL2protein accumulation in the sly1 mutant background. A, GID1a isneeded for after-ripening of sly1-2. Germination is shown for seedsafter-ripened for 1 and 20 months incubated on MS-agar plates for 3 dat 4˚C followed by 7 d at 22˚C. B and C, Seeds after-ripened for 20months were incubated on MS-moistened filter paper for 3 d at 4˚Cfollowed by 48 h at 22˚C and then used to examine RGL2 protein (B)and RGL2 mRNA (C) accumulation by protein-blot and RT-PCRanalyses, respectively. Total protein (60 mg) was loaded, and Ponceaustaining was used as a loading control.



Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

transition to flowering within 20 weeks of growth.Thus, it appears that GID1b makes a strong contribu-tion to GA signaling in flowers.

Since DELLAs are negative regulators of GA responses,one would expect stronger GA-insensitive phenotypes tobe associated with higher DELLA protein accumulation.Consistent with this, GA rescue of ga1-3 seed germinationis associated with DELLA RGL2 protein disappearance(Supplemental Fig. S5C; Tyler et al., 2004; Ariizumi andSteber, 2007). Paradoxically, the ga1-3 and gid1a gid1bgid1c mutants accumulate less DELLA protein thansly1 mutants in spite of the fact that sly1 mutants areless infertile and less severely dwarfed (McGinniset al., 2003; Griffiths et al., 2006; Willige et al., 2007;Ariizumi et al., 2008). Mutations in gid1 and sly1-2 hadadditive effects on DELLA RGA protein accumulation.

The sly1 gid1 multiple mutants had higher levels of RGAprotein accumulation than the gid1a gid1b gid1c and ga1-3mutants but lower RGA levels than sly1-2 (Fig. 5B). Thisreduction was more apparent in sly1-2 gid1a-1 gid1b-1 andin sly1-2 gid1a-1 gid1c-2 than in sly1-2 gid1a-1, sly1-2 gid1b-1,and sly1-2 gid1c-2. Reduced RGA protein levelsroughly correlated with decreased RGA mRNAaccumulation, suggesting that the changes in proteinlevel may result from changes in transcript level(Fig. 5B). Introduction of gid1 mutations into sly1-2also resulted in decreased DELLA RGL2 and RGAprotein accumulation in flowers (Fig. 5D). The RGL2protein level was especially sensitive to the gid1a-1 mu-tation. Thus, gid1 mutations are partly epistatic to sly1for the DELLA protein accumulation phenotype (Fig. 5,B and D).

Figure 5. Additive effects of gid1 and sly1-2 mutations on plant height, fertility, and associated DELLA protein levels. A, Sixty-five-day-old wild-type (WT) and homozygous mutant plants with and without 100 mM GA4 treatment (every 3 d). B, VegetativeDELLA RGA protein accumulation appeared to depend on the gid1 genotype in protein-blot analysis of 40 mg of total proteinextracted from 28-d-old whole plants. Two exposures of the same blot are shown with Ponceau staining as a loading control.Normalized RGA mRNA accumulation based on the mean values from three independent qRT-PCR experiments is plotted onthe histogram. Error bars show SE. C, Flower morphology at anthesis. D, DELLA RGA and RGL2 protein accumulation in theseflowers depended on the presence of GID1 genes. [See online article for color version of this figure.]


Ariizumi et al.

Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

The Level of GID1-DELLA Interaction Correlates withthe Level of GID1 Suppression

GID1 overexpression may rescue sly1 seed germina-tion through two nonproteolytic mechanisms: byGA-stimulated inactivation of DELLA repressors throughdirect protein-protein interaction or by bypassing theneed for DELLA destruction through direct down-regulation of ABA signaling. If GID1 blocks DELLArepression of sly1 seed germination through protein-protein interaction, then lines expressing higher levelsof GID1b should show better rescue of seed germina-tion associated with increased GID1-DELLA protein-protein interaction. To examine this, three sly1-10 35S:FLAG:GID1b overexpression lines were identified showinga range of seed germination potentials when incubated onMurashige and Skoog (MS)-moistened filter paper for5 d at 22°C (Fig. 6A). Each line showed partial rescueof sly1-10 germination: line A showed the highestgermination at 93% at 5 d, ranging down to line C at21% germination after 5 d. Higher germination po-tential was correlated with higher FLAG:GID1b accu-mulation when FLAG:GID1b was immunoprecipitatedfrom protein extracts of lines A through C (Fig. 6B).The lines showing higher germination potential andhigher FLAG:GID1b accumulation also showed highercoimmunoprecipitation (co-IP) of DELLA RGA protein(Fig. 6B). This correlation suggests that GID1b over-expression rescues seed germination through increas-ing the fraction of DELLA RGA protein in complexwith GID1 protein.To examine whether this is also true for rescue of

the sly1 plant height phenotype, we compared two in-dependent sly1-2 HA-GID1b-OE lines that expressHA-GID1b at different levels. Line 386 shows higher levelHA-GID1b protein accumulation and faster vegetativeplant growth than line 388 (Fig. 7; Supplemental Fig.S5D). The higher levels of HA-GID1b expression inline 386 also correlated with co-IP of more DELLARGA protein from 12-d-old plants (Fig. 7B). The cor-relation of plant height and germination phenotypeswith DELLA co-IP supports the model that GID1boverexpression suppresses sly1 phenotypes via directGID1-DELLA protein-protein interaction. When thegermination phenotype of the sly1-2 HA-GID1b lineswas examined, higher level accumulation of HA:GID1bwas associated with higher sly1-2 germination potential(Supplemental Fig. S6A). At day 7, the high line reached66% germination, the low line reached 49%, and un-transformed sly1-2 seeds reached 1% seed germination.After-ripening of sly1-2 seeds for 24months resulted in 39%germination at day 7, indicating that germination rescuedue to HA:GID1b overexpression was even more effectivethan after-ripening in rescuing sly1-2 seed germination.The higher level of HA:GID1b protein accumulation wasalso correlated with higher co-IP of DELLA RGL2(Supplemental Fig. S6B). Thus, it appears that GID1boverexpression rescues sly1-2 and sly1-10 seed germi-nation through increasing the fraction of DELLA RGL2and RGA protein in complex with GID1 protein.

Suppression of the sly1-2 Germination Phenotype IsAssociated with Increased GA-Responsive GeneExpression and Decreased GA20ox andXERICO Expression

To examine whether GID1b overexpression rescuessly1-2 seed germination via increased GA signaling,the expression patterns of six GA-inducible genes wereexamined in seeds imbibed in the presence and ab-sence of 10 mM GA3 by semiquantitative reverse tran-scription (RT)-PCR (Supplemental Fig. S7A), while the

Figure 6. Higher germination potential correlates with higher accu-mulation of FLAG-GID1b and formation of the GID1b-DELLA com-plex. A, Germination rates of sly1-10 (black squares) and threeindependent sly1-10 FLAG-GID1b seed lines (A, B, and C) sown onMS-moistened filter paper and incubated 5 d at 22˚C. Under theseconditions, sly1-10 does not germinate. B, A co-IP experiment per-formed using protein extracted from the indicated sly1-10 FLAG-GID1b seeds imbibed for 48 h at 22˚C. Total protein was incubatedwith anti-FLAG matrix agarose. Immunoprecipitated FLAG-GID1b wasdetected by immunoblot with anti-FLAG, and DELLA RGA in complexwith GID1b was detected with anti-RGA. Total protein (60 mg) wasloaded (input). A Ponceau-stained membrane provided a loadingcontrol.



Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

expression of the GA-induced gene XTH31 was ex-amined by quantitative reverse transcription (qRT)-PCR (Fig. 8; Herzog et al., 1995; Ogawa et al., 2003;Yamauchi et al., 2004; Cao et al., 2006; Ariizumi andSteber, 2007). Seeds of ga1-3, ga1-3 rgl2-1, and ga1-3abi1-1 were included as controls showing GA-responsive gene expression resulting from GA treat-ment, loss of DELLA RGL2 repression, and ABAinsensitivity, respectively (Steber et al., 1998; Cao et al.,2005; Ariizumi and Steber, 2007). All of the seed stockscompared were after-ripened for 2 weeks, except forthe after-ripened sly1-2 seed stock, which was after-ripened for 28 months. Dormant sly1-2 seeds failto show GA-induced gene expression, and increasedGA-inducible mRNA accumulation is associated witha loss of sly1-2 seed dormancy (Ariizumi and Steber,2007; Supplemental Fig. S7A). After-ripening, GID1boverexpression, and the abi1-1 mutation all result inincreased expression of GA-inducible genes in sly1-2without exogenous GA3 application (Figs. 7A and 8).GA3 treatment results in increased XTH31 mRNA ac-cumulation only in after-ripened sly1-2 seeds (Fig. 8).Thus, it appears that all three conditions rescue sly1-2germination, at least in part, through increasing theexpression of GA-inducible genes.

The observation that the abi1-1 mutation results inincreased accumulation of GA-inducible transcripts in

ga1-3 and in sly1-2 raised the question of whether ABAcould also repress the expression of these GA-induciblegenes. To examine this, GA-inducible mRNA accumu-lation was compared in the same after-ripened sly1-2,sly1-2 abi1-1, and sly1-2 GID1b-OE seeds imbibed withand without 10 mM ABA (Supplemental Fig. S7B). ABAtreatment repressed GA-inducible transcript accumu-lation in after-ripened sly1-2 and sly1-2 GID1b-OE butnot in sly1-2 abi1-1. Thus, ABA and GA have opposingeffects on the expression of these GA-responsive genes.

If changes in sly1 seed germination result fromchanges in DELLA repression, then we would expectthis to result in changes in DELLA-regulated geneexpression. The positive regulator of ABA biosynthe-sis, XERICO, and the GA biosynthesis gene GA20ox2were identified as DELLA RGA-activated genes invegetative tissue by chromatin immunoprecipitation(Zentella et al., 2007). Thus, the effect of after-ripening,GID1b-OE, and the abi1-1 mutation on the expressionof these genes was examined by qRT-PCR in sly1-2 seeds(Fig. 8). Both GA treatment and the DELLA rgl2-1 mu-tation significantly decreased XERICO and GA20ox2mRNA accumulation in ga1-3, indicating that DELLARGL2 also activates the expression of these transcriptsin seeds (Supplemental Fig. S8A). GA20ox2 and XERICOtranscript levels are highest in dormant sly1-2 seeds anddecrease significantly as a result of after-ripening (3.5-and 2-fold), GID1b overexpression (5- and 1.2-fold), andthe abi1-1 mutation (15.5- and 37-fold; Fig. 8; for Pvalues, see Supplemental Fig. S8). The abundance ofGA20ox2 transcript in dormant and after-ripened sly1-2seeds likely contributes to higher bioactive GA accu-mulation (Figs. 3 and 8). GA20ox2 and XERICO mRNAlevels in dormant sly1-2 seeds are both reduced by GAtreatment, suggesting that DELLA down-regulationdoes not require the SLY1 gene. XERICO and GA20ox2mRNA levels in sly1-2 are lowered by conditions thatrescue germination, including after-ripening, GID1boverexpression, and the abi1-1 mutation. Collectively,these data suggests that all three mechanisms thatrescue sly1 seed germination interfere with the abilityof DELLA to activate XERICO gene expression. More-over, since this rescue does not require sly1, it appearsthat this involves down-regulation of the DELLA acti-vator of XERICO and GA20ox2 transcription withoutDELLA proteolysis.

DISCUSSION

Nonproteolytic Regulation of DELLA Repressorsof GA Signaling

This study examines the role of GID1-GA-DELLAprotein complex formation in the nonproteolytic down-regulation of DELLA repressors of GA signaling. In theclassic model of GA signaling, DELLA destruction bythe ubiquitin-proteasome pathway lifts DELLA repres-sion, thereby stimulating GA responses, including stemelongation, flowering, and seed germination. Based on

Figure 7. HA-GID1b protein levels and HA-GID1b-RGA complexformation correlate with the suppression of sly1-2 dwarfism. A, sly1-2plants expressing higher (line 386) or lower (line 388) levels of theHA-GID1b fusion protein during stem elongation. B, HA-GID1bwas immunoprecipitated from protein extracted from 12-d-old sly1-2plants expressing either high (line 386) or low (line 388) levels ofHA-GID1b protein. Coimmunoprecipitated RGA was detected byprotein-blot analysis. Protein extract was incubated with HA matrixagarose in the presence of 0.1% ethanol (mock), 1 mM GA3, and 100 mMGA3 and loaded on an SDS-PAGE gel. Protein-blot analysis was per-formed using anti-RGA and anti-HA. Protein extract (40 mg) was loaded(input). [See online article for color version of this figure.]


Ariizumi et al.

Dow




this model, higher DELLA protein levels should alwayscorrelate with stronger GA-insensitive phenotypes. Incontrast, the sly1 mutants accumulate more DELLAprotein than ga1-3 or gid1a gid1b gid1cmutants but haveweaker GA-insensitive phenotypes (Ariizumi et al.,2008). Previous research showed that some vegetativeGA signaling occurs in sly1mutants of Arabidopsis andin the corresponding gid2 mutants of rice (Oryza sativa)in the absence of DELLA protein destruction (Ariizumiet al., 2008; Ueguchi-Tanaka et al., 2008). GA signalingalso occurs during the rescue of sly1 seed germinationin the absence of DELLA destruction (Figs. 1 and 8). Toour knowledge, this is the first study to demonstrateGID1-DELLA complex formation in an F-box mutantand the first study to show that the degree of sly1 rescuecorrelates with the degree of GID1b-DELLA complexformation (Figs. 6 and 7; Supplemental Fig. S6). More-over, to our knowledge, this is the first study to showthat after-ripening is associated not only with a decreasein ABA accumulation but with an increase in GA ac-cumulation using the sly1-2 background (Fig. 3).Seed germination was used as a system to examine

nonproteolytic GA signaling, because the sly1-2 seeddormancy phenotypes are partly rescued not only byGID1 overexpression but by dry after-ripening and theabi1-1 mutation (Fig. 1). In all three cases, sly1 seedgermination occurs in the presence of high levels of theDELLA RGL2 repressor of seed germination. It appears

that the seed dormancy in sly1-2 seeds results fromdownstream ABA signaling, because dormant sly1-2seeds have elevated ABA hormone levels, whereasnonproteolytic rescue of sly1-2 germination by GID1overexpression and after-ripening are associated withreduced ABA levels (Fig. 3). The model that after-ripening relieves sly1-2 seed dormancy through in-creased GID1-GA-DELLA complex formation is basedon the observations that GID1a is required for sly1after-ripening and that GA levels rise with after-ripening (Fig. 9A). It appears that ABA signaling isat least partly responsible for sly1 dormancy, sincegermination rescue by after-ripening and GID1b-OE isassociated with reduced ABA levels (Fig. 3). SinceGID1a is required for sly1 after-ripening, and since GAlevels rise with after-ripening, it appears that increasedGID1-GA-DELLA complex formation may result innonproteolytic rescue of sly1 germination throughafter-ripening (Fig. 9A). Increasing GA levels shouldstimulate GID1-GA-DELLA complex formation. Theimportance of GID1-GA-DELLA complex formation inlifting sly1 seed dormancy is demonstrated by the factthat the degree of GID1-GA-DELLA complex forma-tion correlates with the degree of seed germinationrescue (Fig. 6; Supplemental Fig. S6). Moreover, rescueof sly1-2 germination by after-ripening and GID1b-OEoccurs via GA signaling, since rescue is associated withincreased GA-inducible mRNA levels (Fig. 8; Supplemental

Figure 8. Effect of GID1 overexpression on geneexpression in seeds. The expression of the ABAbiosynthesis gene, XERICO,GA20ox2, and XTH31was examined by qRT-PCR. For each experiment,total was RNA extracted from seeds imbibed onMS-agar for 60 h at 4˚C followed by 24 h at 22˚Cand then incubated for a further 12 h after theaddition of a mock treatment or 10 mM GA3. Allseeds were dry after-ripened for 2 weeks, includingdormant sly1-2 (D), except for sly1-2 after-ripened(AR) for 28 months For each condition, there werethree biological replicates, and each replicate wassubsampled three times. Error bars represent SD. Sta-tistical significance was determined with an ANOVAusing Tukey’s adjustment. P values are given inSupplemental Figure S8.



Dow




Fig. S7). The high DELLA protein level in sly1-2 seedsis associated with high levels of the DELLA-inducedtranscripts XERICO and GA20ox2. Since XERICO is apositive regulator of ABA biosynthesis (Ko et al.,2006), this suggests that the strong dormancy in sly1seeds results from DELLA-induced ABA biosynthe-sis as well as a lack of GA-inducible gene expres-sion. Nonproteolytic rescue of sly1 germination byafter-ripening and GID1b-OE is associated with de-creased XERICO mRNA and ABA accumulation, sug-gesting that DELLA activation of XERICO can beprevented through GID1-GA-DELLA complex forma-tion (Fig. 9B).

GID1 and GA Signaling in sly1-2

Examination of sly1 gid1 multiple mutants providedinformation about the functional roles of the threeArabidopsis GID1 genes in proteolysis-independentGA signaling during multiple stages of development(Figs. 4 and 5). Proteolysis-independent GA signalingin sly1 can be augmented by GID1 overexpression,resulting in partial suppression of sly1 phenotypes(Fig. 1; Ariizumi et al., 2008). This suggests that DELLArepressors can be inactivated even when they cannot bedegraded. Mutations in all of the GA receptors had anadditive effect with sly1-2, resulting in increased seeddormancy, dwarfism, infertility, and underdevelopedflowers (Figs. 4 and 5). This is consistent with theanalysis of plant height in the gid1 gid2 double mutantsof rice (Ueguchi-Tanaka et al., 2008). Based on sly1 gid1double mutant analysis, it appears that GID1b plays animportant role in regulating floral development andfertility, while GID1a plays an important role in stim-ulating stem elongation in the absence of DELLAproteolysis (Fig. 5; Supplemental Fig. S4). Paradoxi-cally, the increased severity of GA-insensitive pheno-types in sly1 gid1 double mutants was associated notwith increased DELLA repressor accumulation butwith decreased DELLA RGA or RGL2 protein ac-cumulation in whole plants, flowers, and seeds (Figs. 4and 5). The reduction in DELLA protein accumulationcorrelated with the number of gid1mutant alleles (Fig. 5, Band D). The gid1a-1 gid1b-1 gid1c-2 triple and single mu-tants show far less DELLA protein accumulation than thesly1 mutants (Willige et al., 2007; Fig. 5D). The correlationbetween reduced DELLA protein and mRNA levelssuggests that GID1 is needed for the high DELLAmRNA and protein levels seen in sly1 mutants (Fig.5B). Down-regulation of DELLA protein activity byGID1 and GA may be followed by up-regulation ofDELLA gene transcription as a feedback response.

The sly1-2 allele has increased seed dormancy andcan require 2 years to after-ripen well (Ariizumi andSteber, 2007). Both the sly1-2 gid1a and sly1-2 gid1cdouble mutants show higher seed dormancy than thesly1-2 mutant (Fig. 4). Whereas sly1-2 typically after-ripens within 15 to 24 months, sly1-2 gid1a completelyfailed to germinate after 20 months of after-ripening.While the gid1a-1 gid1b-1 gid1c-2 triple mutant fails togerminate, all of the gid1 single and double mutantcombinations are able to germinate (Griffiths et al.,2006; Willige et al., 2007). Thus, it is only in the contextof a sly1 mutant that a single gid1a allele blocks after-ripening. Although the role of GID1b in sly1-2 after-ripening could not be well characterized due to sly1-2gid1b-1 infertility, HA-GID1b overexpression stimulatedsly1-2 seed germination more strongly thanHA-GID1a orHA-GID1c (Fig. 1). It is known that GID1b protein hashigher affinity for DELLA than GID1a and GID1c, evenin the absence of GA (Suzuki et al., 2009; Yamamotoet al., 2010). The fact that sly1-2 after-ripening requiresGID1a suggests that GID1 overexpression and after-ripening rescue sly1-2 seed germination by similar

Figure 9. Model for the rescue of sly1 seed germination by GID1-GA-DELLA complex formation. A, After-ripening of sly1-2 stimulates seedgermination by increasing endogenous GA levels that promote GID1-GA-DELLA complex formation. GID1 overexpression also increasesGID1-GA-DELLA formation, thereby inactivating DELLA repression ofGA-responsive (GA-R) gene expression and responses. B, GID1 canprevent DELLA induction of XERICO transcription either via SLY1-directed DELLA degradation or directly through GID1-GA-DELLAcomplex formation. Reduced XERICO expression results in decreasedABA accumulation in after-ripened sly1 and sly1 GID1 overexpressionlines.


Ariizumi et al.

Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

GID1-dependent mechanisms. Additional evidence sup-ports a role for GID1 in seed germination: the gid1b singlemutant shows decreased GA sensitivity during seedgermination, while the gid1c, gid1a gid1b, and gid1agid1c mutants show an apparent increase in ABA sen-sitivity and slower seed germination (Iuchi et al., 2007;Voegele et al., 2011).

After-ripening of sly1-2 Seeds Is Associated with ReducedABA Levels Associated with Reduced XERICO Expression

Rescue of sly1-2 seed germination by both after-ripening and GID1b-OE is associated with decreasedaccumulation of the DELLA-activated mRNA XERICOand decreased ABA hormone levels (Figs. 8 and 9). Itappears that the seed dormancy in GA-insensitivesly1-2 seeds partly results from increased ABA signal-ing, because imbibing dormant sly1-2 seeds accumulatesignificantly more ABA than wild-type Ler (Fig. 3).Rescue of sly1-2 seed germination by after-ripeningand GID1b-OE is associated with decreased ABAhormone levels, suggesting that ABA acts downstreamof SLY1, GID1, and DELLA. This is consistent withabi1-1 rescue of sly1 seed germination and with the factthat lack of ABA biosynthesis rescues the germinationof GA biosynthesis mutants (Fig. 1; Koornneef et al.,1982). Previous research showed that DELLA RGAand GAI are activators of XERICO transcription in veg-etative tissue (Zentella et al., 2007; Gallego-Bartoloméet al., 2011) and that mutations in DELLA RGL2 result indecreased ABA accumulation in seeds (Piskurewiczet al., 2008). This study shows that both GA and anrgl2 mutation decrease XERICO mRNA levels in thega1-3 mutant, indicating that DELLA RGL2 is a posi-tive regulator of XERICO in seeds. Thus, high ABAlevels in dormant sly1-2 are likely a direct consequenceof high DELLA protein levels, resulting in increasedtranscription of the inducer of ABA biosynthesis, XERICO(Figs. 8 and 9). After-ripening and GID1b-OE result indecreased XERICO mRNA and ABA levels, likely due toproteolysis-independent inactivation of DELLA as a pos-itive regulator of XERICO (Figs. 8 and 9B).

Rescue of sly1-2 Seed Germination Is Associated withIncreased GA Signaling

The GA signaling in sly1 mutants may result from in-creased accumulation of bioactive GA. The GA-insensitivesly1-2 mutant has higher bioactive GA levels than thewild type, consistent with the increased expression ofGA biosynthesis genes such as GA3ox1 (McGinniset al., 2003) and GA20ox2 (Fig. 8). A similar increasein bioactive GA levels is seen in the GA-insensitiveDELLA gain-of-function mutant gai-1 and is likely afeedback response to decreased GA sensitivity (Talonet al., 1990). The elevated levels of GA20ox2 in dormantsly1-2 may result from the increased accumulation ofactive DELLA protein in sly1, since GA20ox2 was

identified as a possible DELLA target by chromatinimmunoprecipitation (Zentella et al., 2007). Our datalend support to this finding, as we see a significantdecrease in GA20ox2 mRNA in ga1-3 rgl2-1, after-ripened sly1-2, s1y-2 abi1-1, and sly1-2 GID1b-OE (Fig.8; Supplemental Fig. S8). After-ripening resulted in afurther increase in GA4 and GA1 hormone levels inimbibing sly1-2 seeds (Fig. 3; Supplemental Fig. S2).This is, to our knowledge, the first evidence showingan increase in GA accumulation with after-ripening ofArabidopsis seeds. An increase in precursors of bio-active GA was observed with after-ripening of im-bibed barley (Hordeum vulgare) grain, indicating thatthis phenomenon is likely not limited to Arabidopsissly1 mutants (Jacobsen et al., 2002). Previous researchshowed that GA application mildly stimulates thegermination of dormant sly1-2 seeds without DELLAdegradation (Ariizumi and Steber, 2007). The increasein GA levels with after-ripening may lift the DELLArepression of germination by stimulating GID1-GA-DELLA complex formation (Fig. 9A; Griffiths et al.,2006; Ariizumi et al., 2008).

Rescue of sly1-2 seed germination by after-ripeningand GID1b-OE is associated with increased expression ofa diverse set of GA-inducible genes (Fig. 8; SupplementalFig. S7A). Thus, both GID1 overexpression and after-ripening appear to act via increased GA signaling. Itis interesting that with the exception of SCS10, thelevels of all of the GA-inducible transcripts examinedincreased in sly1-2 abi1-1 compared with dormant sly1-2.Since ABA treatment of after-ripened sly1-2 and sly1-2GID1b-OE seeds resulted in decreased expression ofGA-inducible transcripts, it appears that many, but notall, GA-inducible genes of Arabidopsis are also ABArepressed (Supplemental Fig. S7B). This competitionbetween GA and ABA at the level of gene expression isconsistent with previous research showing that ABAand GA have opposing effects on a-amylase gene ex-pression in the barley aleurone (for review, see Sunand Gubler, 2004).

Evidence That GID1 Interaction with DELLA ProteinsRescues sly1-2 Phenotypes

We propose a model in which after-ripening andGID1 overexpression both rescue sly1 seed germinationby stimulating GID1-GA-DELLA complex formation,which in turn down-regulates DELLA protein activity(Fig. 9A). Rescue of sly1 dwarfism and infertility byGID1 overexpression depends upon GA biosynthesisand the presence of the DELLA protein domain, bothrequired for GID1-DELLA protein-protein interaction(Ariizumi et al., 2008). GID1b-OE can increase thegermination of wild-type Ler, but not of the DELLAdeletion allele gai-1, on ABA (Fig. 2). This suggests thatthe ability of GID1b to stimulate seed germination ismediated by the DELLA domain required for GID1-DELLA protein interaction. Higher levels of HA:GID1b and FLAG:GID1b protein expression are associated



Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

with more efficient sly1 germination and increasedFLAG:GID1b-GA-RGA and HA:GID1b-GA-RGL2 com-plex formation, suggesting that GID1 down-regulatesDELLA RGA and DELLA RGL2 through direct proteinbinding (Fig. 6; Supplemental Fig. S6). To our knowl-edge, this is the first proof that GID1-DELLA complexformation can stimulate seed germination without DELLAproteolysis. We suggest that after-ripening also inactivatesDELLA by stimulating GID1-GA-DELLA complex for-mation, thereby blocking DELLA activation of XERICOtranscription, leading to lower ABA hormone accumula-tion (Fig. 9B). GID1 binding to DELLAmay block DELLAbinding to other transcription factors to regulate genetranscription. For example, DELLA has been shown toregulate dark-induced gene expression through inter-action with the transcription factors PHYTOCHROME-INTERACTING FACTOR3 (PIF3) and PIF4 (de Lucaset al., 2008; Feng et al., 2008). Future work will need toexamine DELLA interaction with transcription factorsregulating seed dormancy and germination. DELLAcan negatively regulate GA signaling by stimulatingthe expression of negative regulators of GA signalinglike XERICO (Zentella et al., 2007). Consistent withthis, both rice and Arabidopsis DELLA proteins activatetranscription when fused to a DNA-binding domain inthe yeast two-hybrid system, and GID1-GA-DELLAcomplex formation blocks DELLA SLR1 transcriptionalactivation in vegetative tissue (Fu et al., 2004; Hiranoet al., 2012).

This study suggests that GA signaling functions toimprove seed germination potential during after-ripening. Decreased ABA levels can be detected withafter-ripening of dry Arabidopsis and barley seeds, butchanges in bioactive GA accumulation are difficult todetect until after seeds are imbibed (Jacobsen et al.,2002; Okamoto et al., 2006; Yamauchi et al., 2007;Barrero et al., 2009). Nevertheless, after-ripening resultsin increased GA sensitivity, suggesting that GA sig-naling elements are involved (Karssen et al., 1989). Thatsly1-2 seeds undergo GID1a-dependent after-ripeningsuggests that this process does not require DELLAproteolysis but does require the GA receptor and GAsignaling. Other lines of evidence suggest that GID1and DELLA proteins regulate seed after-ripening andgermination. DELLA appears to act in the coveringtissues of the seed coat and endosperm to inhibit cellwall loosening and stimulate ABA production, sincethe DELLA rgl2 mutation results in reduced ABAhormone levels and an RGL2 promoter-GUS fusion isexpressed in imbibed Arabidopsis seed coats (Piskurewiczet al., 2008). Analysis of GID1-GUS fusions showed thatGID1a and GID1c are also expressed in the endospermand in the embryo, where they may down-regulateDELLA RGL2 (Voegele et al., 2011).

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Arabidopsis (Arabidopsis thaliana) accessions were cultivated in Convirongrowth chambers according to McGinnis et al. (2003). The infertility

and dwarfism of ga1-3 were rescued through the application of 100 mM GA4every 3 d (Fig. 5). After harvest, seeds were stored in open tubes for 2 weeks atroom temperature and then stored in closed tubes either at room temperaturefor after-ripening or at 220°C to preserve dormancy. The following lines wereconstructed in the Ler ecotype: sly1-2, sly1-10, gai-1, sly1-2 abi1-1, ga1-3 rgl2-1,ga1-3 abi1-1 sly1-2 GID1a-OE, sly1-2 GID1b-OE, sly1-2 GID1c-OE, gai-1 GID1b-OE,and wild-type Ler GID1b-OE as described previously (Steber et al., 1998; Steberand McCourt, 2001; Cao et al., 2005; Ariizumi et al., 2008). The gid1a-1 gid1b-1,gid1b-1 gic1c-2, and gid1a-1 gid1c-2 double mutants were in the Col-0 ecotype(gift of C. Schwechheimer; Willige et al., 2007). The ga1-3 allele in Col-0 was a giftof T. Sun (Tyler et al., 2004; Fig. 5). The sly1-2mutant allele, originally isolated inLer (Steber et al., 1998), was moved into the Col-0 background through threecrosses to wild-type Col-0. The BC3F4 sly1-2 was crossed to the gid1a-1 gid1b-1,gid1b-1 gic1c-2, and gid1a-1 gid1c-2 double mutants in order to derive the sly1-2gid1a-1, sly1-2 gid1b-1, and sly1-2 gid1c-1 doublemutants and the sly1-2 gid1a-1 gid1b-1,sly1-2 gid1b-1 gid1c-2, and sly1-2 gid1a-1 gid1c-2 triple mutants, respectively. Thegenotypes containing single and double gid1 mutations in the sly1-2 backgroundwere selected in the F2 and F3 generations by PCR analysis (Supplemental Table S1;McGinnis et al., 2003; Willige et al., 2007). Triple mutants sly1-2 gid1a-1 gid1b-1 andsly1-2 gid1b-1 gid1c-2were maintained as heterozygotes due to complete infertility. F3plants were used in this study (Fig. 5). 35S:FLAG-GID1b and 35S:HA-GID1 plasmidand strain construction was described previously by Ariizumi et al. (2008).

Germination Experiments

For germination experiments (Figs. 1, 2, 4, and 6; Supplemental Figs. S1Cand S6A), 30 to 100 seeds were sterilized with 10% (v/v) bleach and0.01% (w/v) SDS for 15 to 20 min and plated on 0.53MS salts (Sigma-Aldrich) and5 mM MES, pH 5.5, and 0.8% (w/v) agar (referred to as MS-agar plates). Per-centage germination was determined following 3 d of incubation at 4°C fol-lowed by 3 to 10 d of incubation under constant light at 22°C. The averagegermination rate was calculated using at least three independent replicates.For Figure 2, (+/2)-ABA (Phytotechnology) was added to 0.53 MS-agarmedium. For protein preparation, seeds were sown on filter paper moistenedwith 0.53 MS salts buffered with 5 mM MES to pH 5.5. Based on the previouscharacterization of the sly1 germination phenotype, we know that conditionsresulting in reduced availability of water, such as sowing on filter paper,PEG8000, or soil, reduce the germination efficiency of sly1 mutants (Ariizumiand Steber, 2007). Consistent with this, sly1 mutant seeds germinated onMS-moistened filter paper in this study show lower germination potential thanthose on MS-agar plates (Figs. 1 and 6; Supplemental Fig. S6A). For experi-ments in which the effect of GID1 overexpression on sly1 mutant seed ger-mination was examined (Figs. 1 and 6; Supplemental Fig. S6A), seedsof similar age from plants grown side by side were used, including theT4 generation of sly1-2 GID1a-OE, sly1-2 GID1b-OE, and sly1-2 GID1c-OE andT3 sly1-10 FLAG-GID1b.

RT-PCR Analysis of mRNA Expression

RT-PCR analysis was used to analyze the mRNA levels of RGL2, RGA, andGA-regulated genes. For quantitative and semiquantitative analysisof GA-regulated gene expression during seed imbibition (Fig. 8; Supplemental Fig.S7), seeds were incubated for 60 h at 4°C to stimulate germination and thenincubated at 22°C for 24 h, at which time seeds were treated with 10 mM GA3,10 mM (+/2)-ABA, or a mock treatment and then incubated for a further 12 h.No seed germination was observed at this time point. GA-induced transcriptsbased on previous reports (Herzog et al., 1995; Ogawa et al., 2003; Yamauchiet al., 2004; Cao et al., 2006) were analyzed using published primer pairs(Supplemental Table S1; Ariizumi and Steber, 2007).

For semiquantitative RT-PCR (Supplemental Fig. S7), RNA was isolatedfrom imbibing seeds using a modified SDS-phenol extraction according toMartin et al. (2005; H. Nonogaki, personal communication). Genomic DNAcontamination was removed using the DNA-Free RNA kit (ZYMO Research).Complementary DNA (cDNA) was generated from 5 mg of total RNA usingthe First-Strand cDNA Synthesis Kit (GE Healthcare). cDNA was diluted 1:10and then used as a template for PCR amplification with Mango-Taq DNApolymerase (Bioline) for 30 to 40 cycles of denaturation for 1 min at 94°C,annealing for 1 min at 60°C, and extension for 1 min at 72°C, followed by finalextension for 5 min at 72°C. Agarose gels were stained with SYBR gold.

For qRT-PCR analysis (Fig. 8), total RNA was isolated from imbibing seedsusing the phenol-chloroform extraction method of Oñate-Sánchez and Vicente-Carbajosa (2008) with an additional chloroform extraction and the use of Phase


Ariizumi et al.

Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

Lock (5-PRIME) to prevent contamination of the aqueous phase. cDNA wassynthesized using 1 mg of total RNA (ProtoScript M-MuLV First Strand cDNASynthesis Kit; New England Biolabs), and qRT-PCR analysis was performedusing the Roche LightCycler FastStart DNA Master SYBR Green I kit. PCRconditions were as follows: 10 min of denaturation at 95°C, followed by 45 cyclesof 10 s of denaturation at 95°C, 5 s of annealing at 60°C, and a 10-s extension at72°C. RGA and RGL2 mRNA levels were expressed relative to the constitutiveGAPC mRNA (Griffiths et al., 2006). XERICO, GA20ox2, and XTH31 expressionchanges were calculated using the Delta-Delta Ct method relative to the con-stitutive IAP-LIKE PROTEIN1 mRNA control (Livak and Schmittgen, 2001;Graeber et al., 2011). An ANOVA with Tukey’s comparison adjustment wasconducted using the SAS 9.3 statistical package.

Co-IP Experiment

Co-IP was performed to detect the interactions FLAG-GID1b andHA-GID1b with DELLA RGA as described previously (Figs. 6 and 7; Ariizumiet al., 2011). For Figure 6, T3 seeds of sly1-10 FLAG-GID1b were imbibed on0.53 MS/MES-moistened filter paper for 48 h and then harvested and groundunder liquid nitrogen with extraction buffer C (20 mM Tris, 150 mM NaCl,0.5% (v/v) Triton, and 13 complete proteinase inhibitor [Roche]). T3 sly1-2 35S-HA-GID1bwas used for detection of the HA-GID1b interaction with RGA proteinusing total protein extracted from 12-d-old seedlings (Ariizumi et al., 2008;Supplemental Fig. S6). Protein-blot analysis was used to detect immunopre-cipitated FLAG-GID1b, HA-GID1b, and coimmunoprecipitated RGA.

The following method was devised for co-IP of RGL2 (Supplemental Fig.S6B). Imbibing T4 sly1-2 HA-GID1b seeds were ground under liquid nitrogenand resuspended in extraction buffer (50 mM potassium phosphate buffer,pH 7.0, 13 protease inhibitor cocktail [Sigma-Aldrich], and GA as indicated).Crude extracts were incubated on ice for 30 min. Total protein (500 mg) wasincubated with 20 mL of anti-HA magnetic beads (MBIOL) for 24 h at 4°C.Beads were washed three times with extraction buffer, and the protein wasresuspended in 30 mL of extraction buffer plus 30 mL of 23 loading buffer(Bio-Rad). Immunoprecipitated HA:GID1b and RGL2 after co-IP were detec-ted by protein-blot analysis. Samples were also probed with anti-CUL1 toconfirm the specificity of the HA:GID1b and RGL2 interaction (SupplementalFig. S6, B and C).

Protein-Blot Analysis

Total protein was extracted from imbibed seeds by grinding in liquid nitrogenwith extraction buffer A (50 mM potassium phosphate buffer, pH 6.8, with 13protease inhibitor cocktail [Roche]). The protein concentration was determinedusing the Bradford protein assay (Bio-Rad). Total protein (60 mg) was separatedon an 8% (w/v) SDS-PAGE gel and transferred onto a polyvinylidene difluoridemembrane. Time points for protein extraction were taken at 24-h intervalsduring imbibition as indicated. Primary antibodies included rabbit polyclonalantibodies to RGL2 (1:10,000 [gift of J. Peng]; Hussain et al., 2005), RGA(1:10,000; Silverstone et al., 2001), anti-CUL1 (1:10,000 [gift of X.W Deng]; Wanget al., 2002), HA (1:5,000; Immuno Consultants Laboratory), and monoclonalantibody to FLAG (1:5,000; Sigma-Aldrich). The specificity of the RGA andRGL2 antibodies was confirmed using sly1 rga-24 and sly1 rgl2 double mutants(Ariizumi and Steber, 2007). Protein detection was performed using anti-rabbitIgG-horseradish peroxidase as a secondary antibody (1:200,000; GE Health-care) and the ECL Advance chemiluminescence system (GE Healthcare).

Hormone Measurement

For hormone measurements, 200 mg of seeds was incubated at 4°C for 60 hfollowed by 12 h at 22°C and then ground in liquid N2 and lyophilized for 48 h(Fig. 3; Supplemental Fig. S2). ABA, GA4, and GA1 plant hormone levels weredetermined by liquid chromatography-tandem mass spectrometry using 100to 150 mg dry weight for each biological replicate (n = 6–10) as described pre-viously (Oh et al., 2007; Varbanova et al., 2007) A two-tailed Student’s t test wasperformed using the R statistics package (R Development Core Team, 2010).

Arabidopsis Genome Initiative locus identifiers for the genes in this studyare as follows: GID1a (At3g05120), GID1b (At3g63010), GID1c (At5g27320),SLY1 (At4g24210), GAI (At1g14920), RGA (At2g01570), RGL2 (At3g03450),ABI1 (At4g26080), XTH5 (At5g13870), XTH31 (At3g44990), CP1 (At4g36880),SCS10 (At2g27920), GASA4 (At5g15230), EXP1 (At1g69530), GAPC (At3g04120),and IAP-LIKE PROTEIN1 (At1g17210).

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure S1. Effects of the 35S:HA-GID1 overexpressionconstruct.

Supplemental Figure S2. Bioactive GA1 levels and Student’s t testP values.

Supplemental Figure S3. Lack of GID1c slows down sly1-2 after-ripening.

Supplemental Figure S4. Effect of gid1 mutations on final plant height andfertility of sly1-2.

Supplemental Figure S5. Phenotypes of sly1-2 gid1 multiple mutants,DELLA RGL2 expression in ga1-3 and the gid1a gid1b gid1c triple mutant,and the specificity control for RGA co-IP with FLAG-GID1b.

Supplemental Figure S6. HA-GID1b protein levels and RGL2 co-IP corre-late with seed germination efficiency.

Supplemental Figure S7. Effect of GID1 overexpression on gene expres-sion in seeds.

Supplemental Figure S8. P values for qRT-PCR analysis in Figure 7.

Supplemental Table S1. Primers used in this study.

ACKNOWLEDGMENTS

We thank members of the Steber laboratory for helpful comments on themanuscript. We also thank Dr. T. Sun for providing access to the polyclonalRGA antibody, Dr. J. Peng for providing the RGL2 antibody, and Dr. C.Schwechheimer for providing gid1 mutant lines.

Received April 10, 2013; accepted June 24, 2013; published July 1, 2013.

LITERATURE CITED

Ariizumi T, Lawrence PK, Steber CM (2011) The role of two F-box pro-teins, SLEEPY1 and SNEEZY, in Arabidopsis gibberellin signaling. PlantPhysiol 155: 765–775

Ariizumi T, Murase K, Sun T-P, Steber CM (2008) Proteolysis-independentdownregulation of DELLA repression in Arabidopsis by the gibberellinreceptor GIBBERELLIN INSENSITIVE DWARF1. Plant Cell 20: 2447–2459

Ariizumi T, Steber CM (2007) Seed germination of GA-insensitive sleepy1mutants does not require RGL2 protein disappearance in Arabidopsis.Plant Cell 19: 791–804

Barrero JM, Talbot MJ, White RG, Jacobsen JV, Gubler F (2009) Ana-tomical and transcriptomic studies of the coleorhiza reveal the impor-tance of this tissue in regulating dormancy in barley. Plant Physiol 150:1006–1021

Bewley DJ, Black M (1994). Seeds: Physiology of Development and Ger-mination. Plenum Press, New York

Cao D, Cheng H, Wu W, Soo HM, Peng J (2006) Gibberellin mobilizesdistinct DELLA-dependent transcriptomes to regulate seed germinationand floral development in Arabidopsis. Plant Physiol 142: 509–525

Cao D, Hussain A, Cheng H, Peng J (2005) Loss of function of four DELLAgenes leads to light- and gibberellin-independent seed germination inArabidopsis. Planta 223: 105–113

Cheng H, Qin L, Lee S, Fu X, Richards DE, Cao D, Luo D, Harberd NP,Peng J (2004) Gibberellin regulates Arabidopsis floral development viasuppression of DELLA protein function. Development 131: 1055–1064

de Lucas M, Davière J-M, Rodríguez-Falcón M, Pontin M, Iglesias-PedrazJM, Lorrain S, Fankhauser C, Blázquez MA, Titarenko E, Prat S (2008)A molecular framework for light and gibberellin control of cell elonga-tion. Nature 451: 480–484

Dill A, Sun T-P (2001) Synergistic derepression of gibberellin signaling byremoving RGA and GAI function in Arabidopsis thaliana. Genetics 159:777–785

Dill A, Thomas SG, Hu J, Steber CM, Sun T-P (2004) The Arabidopsis F-boxprotein SLEEPY1 targets gibberellin signaling repressors for gibberellin-induced degradation. Plant Cell 16: 1392–1405



Dow



http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1http://www.plantphysiol.org/cgi/content/full/pp.113.219451/DC1

Feng S, Martinez C, Gusmaroli G, Wang Y, Zhou J, Wang F, Chen L, Yu L,Iglesias-Pedraz JM, Kircher S, et al (2008) Coordinated regulation ofArabidopsis thaliana development by light and gibberellins. Nature 451:475–479

Finch-Savage WE, Cadman CSC, Toorop PE, Lynn JR, Hilhorst HWM(2007) Seed dormancy release in Arabidopsis Cvi by dry after-ripening,low temperature, nitrate and light shows common quantitative patternsof gene expression directed by environmentally specific sensing. Plant J51: 60–78

Finkelstein RR, Reeves W, Ariizumi T, Steber CM (2008) Molecular as-pects of seed dormancy. Annu Rev Plant Biol 59: 387–415

Fu X, Richards DE, Fleck B, Xie D, Burton N, Harberd NP (2004) TheArabidopsis mutant sleepy1gar2-1 protein promotes plant growth by in-creasing the affinity of the SCFSLY1 E3 ubiquitin ligase for DELLA pro-tein substrates. Plant Cell 16: 1406–1418

Gallego-Bartolomé J, Alabadí D, Blázquez MA (2011) DELLA-inducedearly transcriptional changes during etiolated development in Arabi-dopsis thaliana. PLoS ONE 6: e23918

Gallego-Bartolomé J, Minguet EG, Marín JA, Prat S, Blázquez MA,Alabadí D (2010) Transcriptional diversification and functional con-servation between DELLA proteins in Arabidopsis. Mol Biol Evol 27:1247–1256

Graeber K, Linkies A, Wood ATA, Leubner-Metzger G (2011) A guidelineto family-wide comparative state-of-the-art quantitative RT-PCR anal-ysis exemplified with a Brassicaceae cross-species seed germination casestudy. Plant Cell 23: 2045–2063

Griffiths J, Murase K, Rieu I, Zentella R, Zhang Z-L, Powers SJ, Gong F,Phillips AL, Hedden P, Sun T-P, et al (2006) Genetic characterizationand functional analysis of the GID1 gibberellin receptors in Arabidopsis.Plant Cell 18: 3399–3414

Herzog M, Dorne AM, Grellet F (1995) GASA, a gibberellin-regulated genefamily from Arabidopsis thaliana related to the tomato GAST1 gene.Plant Mol Biol 27: 743–752

Hirano K, Asano K, Tsuji H, Kawamura M, Mori H, Kitano H, Ueguchi-Tanaka M, Matsuoka M (2010) Characterization of the molecularmechanism underlying gibberellin perception complex formation in rice.Plant Cell 22: 2680–2696

Hirano K, Kouketu E, Katoh H, Aya K, Ueguchi-Tanaka M, Matsuoka M(2012) The suppressive function of the rice DELLA protein SLR1 is de-pendent on its transcriptional activation activity. Plant J 71: 443–453

Hussain A, Cao D, Cheng H, Wen Z, Peng J (2005) Identification of theconserved serine/threonine residues important for gibberellin-sensitivity of Arabidopsis RGL2 protein. Plant J 44: 88–99

Iuchi S, Suzuki H, Kim Y-C, Iuchi A, Kuromori T, Ueguchi-Tanaka M,Asami T, Yamaguchi I, Matsuoka M, Kobayashi M, et al (2007) Mul-tiple loss-of-function of Arabidopsis gibberellin receptor AtGID1s completelyshuts down a gibberellin signal. Plant J 50: 958–966

Jacobsen JV, Pearce DW, Poole AT, Pharis RP, Mander LN (2002) Abscisicacid, phaseic acid and gibberellin contents associated with dormancyand germination in barley. Physiol Plant 115: 428–441

Karssen CM, der Brinkhorst-van Swan DLC, Breekland AE, KoornneefM (1983) Induction of dormancy during seed development by endoge-nous abscisic acid: studies in abscisic acid deficient genotypes of Arab-idopsis thaliana (L.) Heynh. Planta 157: 158–165

Karssen CM, Laçka E (1986). A revision of the hormone balance theory ofseed dormancy: studies on gibberellin and/or abscisic acid-deficient mutantsof Arabidopsis thaliana. In M Bopp, ed, Plant Growth Substances 1985:Proceedings of the 12th International Conference on Plant Growth Sub-stances. Springer-Verlag, New York, pp 315–323

Karssen CM, Zagorski S, Kepczynski J, Groot SPC (1989) Key role forendogenous gibberellins in the control of seed germination. Ann Bot (Lond)63: 71–80

Ko JH, Yang SH, Han KH (2006) Upregulation of an Arabidopsis RING-H2gene, XERICO, confers drought tolerance through increased abscisic acidbiosynthesis. Plant J 47: 343–355

Koornneef M, Jorna ML, der Brinkhorst-van Swan DLC, Karssen CM(1982) The isolation of abscisic acid (ABA) deficient mutants by selectionof induced revertants in non-germinating gibberellin sensitive lines ofArabidopsis thaliana (L.) Heynh. Theor Appl Genet 61: 385–393

Koornneef M, van der Veen JH (1980) Induction and analysis of gibberellinsensitive mutants in Arabidopsis thaliana (L.) Heynh. Theor Appl Genet58: 257–263

Lee S, Cheng H, King KE, Wang W, He Y, Hussain A, Lo J, Harberd NP,Peng J (2002) Gibberellin regulates Arabidopsis seed germination viaRGL2, a GAI/RGA-like gene whose expression is up-regulated followingimbibition. Genes Dev 16: 646–658

Lefebvre V, North H, Frey A, Sotta B, Seo M, Okamoto M, Nambara E,Marion-Poll A (2006) Functional analysis of Arabidopsis NCED6 andNCED9 genes indicates that ABA synthesized in the endosperm is in-volved in the induction of seed dormancy. Plant J 45: 309–319

Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression datausing real-time quantitative PCR and the 2(-Delta Delta C(T)) method.Methods 25: 402–408

Martin RC, Liu P-P, Nonogaki H (2005) Simple purification of small RNAsfrom seeds and efficient detection of multiple microRNAs expressed inArabidopsis thaliana and tomato (Lycopersicon esculentum) seeds. Seed SciRes 15: 319–328

McGinnis KM, Thomas SG, Soule JD, Strader LC, Zale JM, Sun T-P,Steber CM (2003) The Arabidopsis SLEEPY1 gene encodes a putativeF-box subunit of an SCF E3 ubiquitin ligase. Plant Cell 15: 1120–1130

Murase K, Hirano Y, Sun T-P, Hakoshima T (2008) Gibberellin-inducedDELLA recognition by the gibberellin receptor GID1. Nature 456:459–463

Nakajima M, Shimada A, Takashi Y, Kim Y-C, Park S-H, Ueguchi-Tanaka M, Suzuki H, Katoh E, Iuchi S, Kobayashi M, et al (2006)Identification and characterization of Arabidopsis gibberellin receptors.Plant J 46: 880–889

Ogawa M, Hanada A, Yamauchi Y, Kuwahara A, Kamiya Y, Yamaguchi S(2003) Gibberellin biosynthesis and response during Arabidopsis seedgermination. Plant Cell 15: 1591–1604

Oh E, Yamaguchi S, Hu J, Yusuke J, Jung B, Paik I, Lee H-S, Sun T-P,Kamiya Y, Choi G (2007) PIL5, a phytochrome-interacting bHLH pro-tein, regulates gibberellin responsiveness by binding directly to the GAIand RGA promoters in Arabidopsis seeds. Plant Cell 19: 1192–1208

Okamoto M, Kuwahara A, Seo M, Kushiro T, Asami T, Hirai N, KamiyaY, Koshiba T, Nambara E (2006) CYP707A1 and CYP707A2, which en-code abscisic acid 89-hydroxylases, are indispensable for proper controlof seed dormancy and germination in Arabidopsis. Plant Physiol 141:97–107

Oñate-Sánchez L, Vicente-Carbajosa J (2008) DNA-free RNA isolationprotocols for Arabidopsis thaliana, including seeds and siliques. BMC ResNotes 1: 93

Piskurewicz U, Jikumaru Y, Kinoshita N, Nambara E, Kamiya Y, Lopez-Molina L (2008) The gibberellic acid signaling repressor RGL2 inhibitsArabidopsis seed germination by stimulating abscisic acid synthesis andABI5 activity. Plant Cell 20: 2729–2745

Piskurewicz U, Lopez-Molina L (2009) The GA-signaling repressor RGL3represses testa rupture in response to changes in GA and ABA levels.Plant Signal Behav 4: 63–65

R Development Core Team (2010). R: A Language and Environment forStatistical Computing. http://www.R-project.org/ (February 3, 2012)

Shimada A, Ueguchi-Tanaka M, Nakatsu T, Nakajima M, Naoe Y,Ohmiya H, Kato H, Matsuoka M (2008) Structural basis for gibberellinrecognition by its receptor GID1. Nature 456: 520–523

Silverstone AL, Jung HS, Dill A, Kawaide H, Kamiya Y, Sun TP (2001)Repressing a repressor: gibberellin-induced rapid reduction of the RGAprotein in Arabidopsis. Plant Cell 13: 1555–1566

Steber CM, Cooney SE, McCourt P (1998) Isolation of the GA-responsemutant sly1 as a suppressor of ABI1-1 in Arabidopsis thaliana. Genetics149: 509–521

Steber CM, McCourt P (2001) A role for brassinosteroids in germination inArabidopsis. Plant Physiol 125: 763–769

Sun T-P, Gubler F (2004) Molecular mechanism of gibberellin signaling inplants. Annu Rev Plant Biol 55: 197–223

Suzuki H, Park S-H, Okubo K, Kitamura J, Ueguchi-Tanaka M, Iuchi S,Katoh E, Kobayashi M, Yamaguchi I, Matsuoka M, et al (2009) Dif-ferential expression and affinities of Arabidopsis gibberellin receptorscan explain variation in phenotypes of multiple knock-out mutants. Plant J60: 48–55

Talon M, Koornneef M, Zeevaart JAD (1990) Accumulation of C19-gibberellins in the gibberellin-insensitive dwarf mutant gai of Arabidopsisthaliana (L.) Heynh. Planta 182: 501–505

Tyler L, Thomas SG, Hu J, Dill A, Alonso JM, Ecker JR, Sun T-P (2004)Della proteins and gibberellin-regulated seed germination and floraldevelopment in Arabidopsis. Plant Physiol 135: 1008–1019


Ariizumi et al.

Dow



http://www.R-project.org/

Ueguchi-Tanaka M, Hirano K, Hasegawa Y, Kitano H, Matsuoka M (2008)Release of the repressive activity of rice DELLA protein SLR1 by gib-berellin does not require SLR1 degradation in the gid2 mutant. Plant Cell20: 2437–2446

Ueguchi-Tanaka M, Nakajima M, Katoh E, Ohmiya H, Asano K, Saji S,Hongyu X, Ashikari M, Kitano H, Yamaguchi I, et al (2007) Molecularinteractions of a soluble gibberellin receptor, GID1, with a rice DELLAprotein, SLR1, and gibberellin. Plant Cell 19: 2140–2155

Varbanova M, Yamaguchi S, Yang Y, McKelvey K, Hanada A, BorochovR, Yu F, Jikumaru Y, Ross J, Cortes D, et al (2007) Methylation ofgibberellins by Arabidopsis GAMT1 and GAMT2. Plant Cell 19: 32–45

Voegele A, Linkies A, Müller K, Leubner-Metzger G (2011) Members ofthe gibberellin receptor gene family GID1 (GIBBERELLIN INSENSITIVEDWARF1) play distinct roles during Lepidium sativum and Arabidopsisthaliana seed germination. J Exp Bot 62: 5131–5147

Wang F, Zhu D, Huang X, Li S, Gong Y, Yao Q, Fu X, Fan L-M, Deng XW(2009) Biochemical insights on degradation of Arabidopsis DELLA pro-teins gained from a cell-free assay system. Plant Cell 21: 2378–2390

Wang X, Kang D, Feng S, Serino G, Schwechheimer C, Wei N (2002)CSN1 N-terminal-dependent activity is required for Arabidopsis devel-opment but not for Rub1/Nedd8 deconjugation of cullins: a structure-

function study of CSN1 subunit of COP9 signalosome. Mol Biol Cell 13:646–655

Willige BC, Ghosh S, Nill C, Zourelidou M, Dohmann EMN, Maier A,Schwechheimer C (2007) The DELLA domain of GA INSENSITIVEmediatesthe interaction with the GA INSENSITIVE DWARF1A gibberellin receptor ofArabidopsis. Plant Cell 19: 1209–1220

Yamamoto Y, Hirai T, Yamamoto E, Kawamura M, Sato T, Kitano H,Matsuoka M, Ueguchi-Tanaka M (2010) A rice gid1 suppressor mutantreveals that gibberellin is not always required for interaction between itsreceptor, GID1, and DELLA proteins. Plant Cell 22: 3589–3602

Yamauchi Y, Ogawa M, Kuwahara A, Hanada A, Kamiya Y, Yamaguchi S(2004) Activation of gibberellin biosynthesis and response pathways bylow temperature during imbibition of Arabidopsis thaliana seeds. PlantCell 16: 367–378

Yamauchi Y, Takeda-Kamiya N, Hanada A, Ogawa M, Kuwahara A, SeoM, Kamiya Y, Yamaguchi S (2007) Contribution of gibberellin deacti-vation by AtGA2ox2 to the suppression of germination of dark-imbibedArabidopsis thaliana seeds. Plant Cell Physiol 48: 555–561

Zentella R, Zhang Z-L, Park M, Thomas SG, Endo A, Murase K, Fleet CM,Jikumaru Y, Nambara E, Kamiya Y, et al (2007) Global analysis of DELLAdirect targets in early gibberellin signaling in Arabidopsis. Plant Cell 19: 3037–3057



Dow



lifting della repression of arabidopsis seed germination · from juvenile to adult plant...

Documents