Differential control of seed primary dormancy in Arabidopsis ecotypes by the transcription factor SPATULA

Abstract

Freshly matured seeds exhibit primary dormancy, which prevents germination until environmental conditions are favorable. The establishment of dormancy occurs during seed development and involves both genetic and environmental factors that impact on the ratio of two antagonistic phytohormones: abscisic acid (ABA), which promotes dormancy, and gibberellic acid, which promotes germination. Although our understanding of dormancy breakage in mature seeds is well advanced, relatively little is known about the mechanisms involved in establishing dormancy during seed maturation. We previously showed that the SPATULA (SPT) transcription factor plays a key role in regulating seed germination. Here we investigate its role during seed development and find that, surprisingly, it has opposite roles in setting dormancy in Landsberg erecta and Columbia Arabidopsis ecotypes. We also find that SPT regulates expression of five transcription factor encoding genes: ABA-INSENSITIVE4 (ABI4) and ABI5, which mediate ABA signaling; REPRESSOR-OF-GA (RGA) and RGA-LIKE3 involved in gibberellic acid signaling; and MOTHER-OF-FT-AND-TFL1 (MFT) that we show here promotes Arabidopsis seed dormancy. Although ABI4, RGA, and MFT are repressed by SPT, ABI5 and RGL3 are induced. Furthermore, we show that RGA, MFT, and ABI5 are direct targets of SPT in vivo. We present a model in which SPT drives two antagonistic “dormancy-repressing” and “dormancy-promoting” routes that operate simultaneously in freshly matured seeds. Each of these routes has different impacts and this in turn explains the opposite effect of SPT on seed dormancy of the two ecotypes analyzed here.

Plants have evolved to alter how they grow and develop in response to signals from their environment. A good example of this is the process of seed germination, which marks the start of growth following a period of quiescence or primary dormancy. This dormant state enables seeds to survive in the soil until conditions are favorable for growth. In Arabidopsis and many other species, dormancy can be released by either low temperature of imbibed seeds (stratification) or by an extended period of dry seed storage (after-ripening). Regulation of germination results from a balance between levels of and sensitivity to the antagonistic phytohormones abscisic acid (ABA) and gibberellic acid (GA) (1). ABA is synthesized during seed development leading to induction of primary dormancy, which inhibits precocious germination. ABA acts through ABA-response transcription factors, such as ABA-INSENSITIVE3 (ABI3), ABI4, and ABI5, which are all known to affect seed dormancy and germination (2–9). Opposite to the action of ABA, GA promotes germination by triggering 26S-proteosome degradation of the growth-repressing DELLA transcription factors. In Arabidopsis (Arabidopsis thaliana) there are five different DELLA genes and four of them [GA-INSENSITIVE (GAI), REPRESSOR-OF-GA, (RGA), RGA-LIKE2 (RGL2), and RGL3] have a role in repressing germination (10–15). The mechanism for establishing primary dormancy during seed maturation remains poorly understood with only a few genes, such as DELAY-OF-GERMINATION (DOG1) being identified that are not directly involved in seed maturation or phytohormone metabolism (16–22). Environmental conditions during seed set influence dormancy status. For example, low temperature conditions result in increased primary dormancy of mature Arabidopsis seeds (23–25) by inducing expression of genes associated with dormancy and influencing ABA and GA levels (24).

The phosphatidyl ethanolamine-binding protein (PEBP) family is an evolutionary conserved group of proteins present from bacteria to animals and plants. In Arabidopsis, the better characterized PEBP-family members so far are FLOWERING-LOCUS (FT), TERMINAL-FLOWER1 (TFL1), and MOTHER-OF-FT-AND-TFL1 (MFT). Although FT and TFL1 are antagonistic regulators of flowering time control (26–28), MFT is involved in regulating seed germination (29, 30), localizing in the nucleus and acting as a transcription factor that directly binds to the ABI5 promoter to repress its expression in nondormant seeds (30). SPATULA (SPT) is a transcription factor belonging to the basic helix–loop–helix (bHLH) subfamily, which also includes PHYTOCHROME-INTERACTING-(LIKE)-FACTORS (PIFs and PILs) (31). SPT was originally described as a regulator of septum, style, and stigma growth during gynoecium development (32–34) and promotes seed dispersal (35). SPT also plays a role integrating day time and temperature signaling to repress growth of vegetative tissue, such as hypocotyls, cotyledons, and leaves (36–39). We showed previously that SPT is also involved in germination of freshly matured (but not after-ripened) seeds by regulating expression of genes involved in GA biosynthesis in imbibed seeds (38). Interestingly, expression of SPT peaks during seed development (40) as well as seed germination (38), which suggested to us that it may be involved in the poorly understood process of primary dormancy establishment.

In the present study, we investigated further the role of SPT in seed dormancy/germination using mutant and overexpressing lines in two different Arabidopsis ecotype backgrounds (Landsberg erecta, Ler, and Columbia, Col). Unexpectedly, we found that SPT has opposite effects in germination in the two ecotypes analyzed: although in Ler SPT represses dormancy, in Col SPT promotes it. We also investigated the role of SPT in gene expression and found that it promotes ABI5 and RGL3 and represses ABI4, RGA, and MFT in freshly matured seeds before germination in both Ler and Col ecotypes. We also found that MFT promotes primary dormancy and germination in freshly matured and after-ripened seeds, respectively. We propose a model in which the role of SPT on gene expression explains the opposite spt mutant phenotypes in the Ler and Col ecotypes.

Results and Discussion

SPT Has Opposite Effects on Seed Primary Dormancy in Ler and Col.

We previously described the role of SPT in germination of freshly matured Ler seeds (38). Here we extend our phenotypic analysis to the Col ecotype. As reported earlier (38), Ler seeds carrying the spt-2 mutation, resulting in an R²⁰⁹K amino acid change in the DNA-binding domain, are extremely dormant (Fig. 1A). However, seeds from SPT overexpressing (SPToe) lines exhibit increased germination (Fig. 1A). In contrast, Col seeds of the spt-11 and spt-12 T-DNA insertion loss-of-function mutants are less dormant and SPToe seeds are more dormant than WT under unstratified conditions (Fig. 1C). After stratification, all of the genetic backgrounds analyzed in Col germinated at 100% (Fig. 1C). Comparison of Ler and Col nucleotide sequences determined that the germination differences between these ecotypes are not because of allelic variation at the SPT locus. The spt-12 T-DNA insertion null mutant was introgressed from Col into Ler and in this ecotype spt-12 seeds are more dormant than WT (Fig. 1B). To assess the effect of spt-2 in Col, we took advantage of the fact that it is a semidominant mutation (38) by overexpressing the spt-2 cDNA (SPT2oe) in Col plants: SPT2oe seeds are less dormant than WT-Col controls (Fig. 1D). Taken together, these results underline the importance of background ecotype in determining the role played by SPT, which results in more germination in Ler and less in Col. These results led us to retest the experimental materials we used to report on the role of SPT in germination (38). Consistent with the results shown in Fig. 1C, we found that the SPToe data reported previously was derived from a Col rather than a Ler background. Furthermore, spt-10, which was reported as a loss-of-function allele in Ler because of a transposon insertion near the 5′ end of the first exon, actually results in increased transcript levels (Fig. S1). The similarity of the spt-10 and Ler overexpressor germination phenotypes lead us to conclude that this particular allele acts as a SPT2oe at this stage of development.

Fig. 1.

Different dormancy effects of SPT depending on ecotype background. Germination assays of: (A) SPToe, WT, and spt-2 in Ler (unstratified, u; or stratified, s); (B) spt-12 retrogressed into Ler (stratified, s; two independent retrogressions are shown); (C) SPToe, WT, spt-11 and spt-12 in Columbia (Col) (unstratified, u; or stratified, s); (D) two independent SPT2oe lines in Col (unstratified, u); (E) WT and spt-2 in Ler (stratified) supplemented with GA (100 μM), Norflurazon (Norf; 50 μM), or both; and (F) WT, spt-2 and aba1-1 single-mutants, spt-2 aba1-1 double-mutant, gai-6 rga-2 rgl1-1 rgl2-1 quadruple-mutant (della4), and spt-2 della4 quintuple-mutant in Ler (stratified). Assays were performed on freshly matured seeds; germination was counted 7 d after imbibition in continuous light. (G–J) ABA and GA in freshly matured dry seed (DS) and 1- and 2-d after imbibition (DAI) of WT and spt mutants in Ler and Col. Relative phytohormone abundance (Rel. phytohorm. abund.) values were obtained by dividing absolute amounts of all samples by the WT DS amount, which was defined as the reference point. In all panels error bars represent standard deviation (SD) of at least three determinations. Asterisks and letters above the bars indicate statistically significant differences (P < 0.05). See Experimental Procedures for details.

To gain more information about the mode of action of SPT, exogenously applied GA and the ABA biosynthesis inhibitor Norflurazon (100 and 50 µM, respectively) were used in germination assays. Both these compounds promote germination of WT seeds and rescue the extreme spt-2 dormancy phenotype (Fig. 1E). Furthermore, the germination promoting effect of these compounds on spt-2 is synergistic resulting in close to 100% germination (Fig. 1E). A series of genetic analyses were also carried out by introducing spt-2 into aba1-1 (disrupted in ABA biosynthesis) and a quadruple rgl1-1/rgl2-1/gai-6/rga-2 DELLA mutant (della4; having a constitutive GA response). Germination assays determined that the extreme spt-2 dormancy phenotype is alleviated by decreasing endogenous ABA levels or constitutively activating the GA response pathway (Fig. 1F).

To better characterize the spt mutants, endogenous levels of ABA and GA were measured in freshly matured dry and imbibed seeds. Because stratification abolishes the spt mutant phenotype in Col (Fig. 1C), the analyses were performed on unstratified samples. The ABA and GA levels are presented relative to the respective WT dry-seed samples (for absolute values, see Table S1): In spt-2, ABA levels are significantly higher in dry and imbibed seeds (Fig. 1G); GA levels are significantly lower in dry seeds but unchanged in imbibed seeds (Fig. 1H). In spt-12, ABA levels are lower throughout the time course (Fig. 1I) and GA abundance is higher in dry seeds but unchanged in imbibed seeds (Fig. 1J). Although the differences in ABA and GA accumulation in spt mutant backgrounds compared with their respective WTs is generally less than twofold, the same trends were observed in independent experiments. Thus, ABA and GA levels in dry seeds are consistent with the dormancy status of the spt mutants in Ler and Col (Fig. 1 A and C). These observations, together with the fact that SPT is expressed during seed development (Fig. S2), led us to further investigate its possible role at this developmental stage.

ABI5 and RGA Are Direct Targets of SPT.

To find possible direct targets of SPT, we assessed expression of selected genes involved in ABA and GA biosynthesis [ABA1, GA-3-OXIDASE1 (GA3OX1), and GA3OX2], ABA signaling (ABI3, ABI4, and ABI5), GA signaling (GAI, RGA, RGL2, and RGL3), and seed dormancy (DOG1) that are expressed during seed development and germination (Fig. S2). We measured transcript abundance of these genes in freshly matured dry seeds as an indicator of expression during late stages of seed development (Fig. 2 A and B). ABA1 expression is increased in spt-2 and decreased in spt-12, consistent with ABA levels in these two mutants, but there is no difference in GA3OX1 and GA3OX2 expression despite GA levels being altered. Regarding the ABA-signaling genes, ABI3 expression is unchanged, ABI4 is increased, and ABI5 decreased in both Ler and Col spt mutants. Of the GA-signaling genes, the expression of GAI and RGL2 are unchanged in the mutant backgrounds, RGL3 expression is repressed in both Ler and Col spt mutants, and RGA transcript levels are increased twofold in Ler spt-2 but unchanged in Col spt-12. DOG1 transcript accumulation is unchanged in spt mutant irrespective of ecotype.

Fig. 2.

ABI5, RGA, and MFT are direct targets of SPT. (A and B) Relative transcript abundance (Rel. transc. abund.) in Ler and Col freshly matured dry seeds of selected genes involved in ABA and GA biosynthetic and signaling pathways, DOG1 and MFT. Error bars represent SD of three determinations. (C) Schematic representation of the 2,500-bp region upstream of the ATG start codon of the indicated genes: Vertical thick black bars represent G-box motifs; vertical thick gray bars represent E-box motifs in G-box free promoters; light gray rectangles indicate the presence of an upstream gene; horizontal thin black lines represent amplicons assayed for SPT binding. (D) Relative ChIP-qPCR enrichment (Rel. ChIP enrich.) of the indicated promoter regions compared with the negative MLS control. Assays were performed on silique samples from a SPTmyc epitope-tagged transgenic line (Ler) with (+Ab) or without (−Ab) anti-MYC antibodies. Error bars represent standard error (SE) of three replicates (similar results were obtained in independent experiments). Asterisks above the bars indicate statistically significant differences (P < 0.05). See Experimental Procedures for details.

The bHLH-type transcription factors bind to the E-box (CANNTG) motif present in target gene promoters (41). The members of the PIF/PIL subfamily, to which SPT belongs, preferentially bind to the G-box (CACGTG) motif (42–44). We screened the promoter regions of the SPT-regulated genes for these motifs. Although ABI5 and RGA promoters have five and three G-boxes, respectively, ABA1, ABI4, and RGL3 promoters only contain E-boxes (Fig. 2C and Table S2). To test whether SPT binds directly to these promoters in vivo, ChIP followed by quantitative PCR (qPCR) analyses were performed on fully expanded green siliques of a MYC-epitope tagged SPT line (SPTmyc) in Ler. Primers spanning different regions of each of the promoters (Fig. 2C) were used, as well as an internal control region of the unrelated G-box free MALATE SYNTHASE (MLS) gene promoter. These experiments show enrichment of only the ABI5 and RGA G-Box specific amplicons (Fig. 2D). We found that the promoters of both ABI5 and RGA are also direct targets of SPT in Col (Fig. S3). Interestingly, despite the in vivo SPT–RGA interaction in Col, we did not find RGA expression to be altered in spt-12 (Fig. 2B and Fig. S3).

During seed imbibition, SPT expression coincides with that of GA3OX1 and GA3OX2 (Fig. S2). Expression of both these genes is altered by SPT in stratified imbibed germinating Ler seeds (38) (Fig. S1) and their promoters each contain a G-box motif (Fig. S3 and Table S2). To test whether SPT binds directly to these promoters in imbibed seeds, ChIP-qPCR analyses were performed but failed to detect any SPT-GA3OX1 or SPT-GA3OX2 interactions, suggesting that these genes are not primary targets of SPT (Fig. S3).

Transcriptomic and ChIP Analyses Reveal MFT as a Direct Target of SPT.

To find other SPT targets, we analyzed transcriptomic data available in our laboratory from freshly matured stratified Ler seeds sampled 1 d after imbibition, which is the time point when SPT expression is highest in germinating Ler seeds (38). Two sets of transcriptomic comparisons were carried out: (i) SPToe vs. WT vs. spt-2, and (ii) aba1-1 vs. spt-2 aba1-1. The first comparison identifies many genes that are up- or down-regulated depending on levels of SPT, but because this material shows large differences in germination capacity (Fig. 1A), many of these genes will not be direct targets of SPT. In the second comparison germination capacity is similar, because aba1-1 rescues the strong dormancy phenotype of spt-2 (Fig. 1F). We found that only 10 genes had their profiles altered in the same pattern between the two comparison sets (Table S3). Six of these genes contain G-boxes within their promoters (Tables S2 and S3) with one of these, MFT, which is repressed by SPT, having three (Fig. 2C and Table S2). Interestingly, the MFT expression profile parallels that of SPT during seed development (Fig. S2). Transcript abundance in freshly matured dry seeds and ChIP-qPCR analyses on silique material showed that MFT expression increases in the absence of SPT (Fig. 2 A and B) and SPTmyc binds to the three MFT promoter regions assessed in Ler and Col backgrounds (Fig. 2D and Fig. S3). However, these SPT-MFT interactions were not observed in material from germinating seeds (Fig. S3). Of the other G-box containing putative primary SPT targets described in Table S3, ChIP-qPCR analyses indicate that only the At4g33980 gene is a direct target of SPT (Fig. S3) and although the gene function is not known, it is interesting to note that it is mainly expressed during seed development (Fig. S2).

MFT Promotes Primary Dormancy.

MFT was recently shown to be involved in seed germination (29, 30). However, although MFT has been found to negatively regulate ABA signaling and promote germination in Arabidopsis (30), in wheat (Triticum aestivum) an MFT homolog (Ta-MFT) has been shown to be a key factor in promoting dormancy (29). The Arabidopsis study used after-ripened nondormant seeds but the wheat study used freshly matured dormant seeds. These differences prompted us to assess whether MFT plays a role in regulating primary dormancy in Arabidopsis. We found that mft-2 Ler and Col seeds are less dormant than the corresponding WTs (Fig. 3 A and C). We also analyzed sensitivity to exogenously applied ABA (10 µM) on 2-mo-old after-ripened seeds and found that, as previously reported for Col (30), mft-2 nondormant Ler seeds are hypersensitive to the phytohormone (Fig. 3 B and D). We therefore conclude that MFT operates to promote (i) dormancy during seed development and (ii) germination in after-ripened imbibed seeds with exogenous ABA. Surprisingly, given the less-dormant phenotype of mft-2, relative endogenous levels of ABA in freshly matured dry seeds are approximately twofold higher in the mft-2 mutants compared with the Ler and Col WTs, and GA levels are similar (Fig. 3 E and F; for absolute values see Table S1). The role of MFT in promoting germination in after-ripened seeds is mediated at least in part by directly repressing ABI5 expression (30). Hence, we assessed ABI5 expression (as well as other ABA and GA signaling related-genes) in freshly matured dry seeds. We found that of all these genes only RGA expression is altered, showing an increase in mft-2 in both ecotype backgrounds (Fig. 3 G and H). As with the elevated ABA levels, the increase in RGA expression is unexpected given that this gene promotes dormancy but in mft-2 dormancy levels are decreased. These results suggest that to promote dormancy, MFT acts downstream with or parallel to the ABA and GA response pathways. We reported recently that the pxa1 mutant (that is disrupted in a peroxisomal ABC transporter activity) has low levels of seed germination despite accumulating higher GA levels than WT (45). Hence, the high ABA and GA levels in mft and pxa1 mutants, respectively, may be a consequence of compensatory mechanisms attempting to rescue germination back to WT levels.

Fig. 3.

MFT regulates seed primary dormancy. (A and C) Germination assays of freshly mature (FM) unstratified (u) and stratified (s) WT and mft-2 Ler and Col seeds. (B and D) Germination assays of after-ripened (AR) stratified WT and mft-2 Ler and Col seeds supplemented with 10 μM ABA (+ABA) or without (−ABA). (E and F) ABA and GA in freshly matured dry seed of WT and mft mutants in Ler and Col. Absolute amounts were divided by the reference (WT DS) to obtain relative phytohormone abundance (Rel. phytohorm. abund.). (G and H) Relative transcript abundance (Rel. transc. abund.) of selected genes involved in ABA and GA signaling pathways in freshly matured WT and mft-2 Ler and Col dry seeds. (I and J) Germination assays of freshly matured unstratified WT, spt, and mft single mutants and the spt mft double-mutants Ler and Col seeds. Error bars represent SD of at least three determinations. Asterisks and letters above the bars indicate statistically significant differences (P < 0.05). See Experimental Procedures for details. (K) Model of SPT-mediated regulation of seed primary dormancy. Red and green lines from SPT denote repression and activation, respectively, by this transcription factor of ABI4, MFT, RGA, ABI5, and RGL3 gene expression. These red and green lines also represent the dormancy-repressing and dormancy-promoting routes, respectively. MFT and ABI5 are direct SPT targets (solid lines); RGA is a direct SPT target in Ler and Col but gene expression is only altered in Ler (dashed line); ABI4 and RGL3 are indirect SPT targets (dotted lines). Red horizontal line from MFT denotes the repressive action on RGA expression (whether RGA is a direct or indirect target of MFT is still unknown). Solid black lines from ABI4, MFT, RGA, ABI5, and RGL3 denote the promotion of primary seed dormancy by these transcription factors.

To gain insight into the combined action of SPT and MFT, we obtained the spt-2 mft-2 (in Ler) and spt-12 mft-2 (in Col) double-mutants and assessed germination of freshly matured seeds. We found the effect of the two mutations to be additive in Col as expected from the single-mutant phenotypes (Fig. 3J). However, in Ler, the two mutations are not additive and the double-mutant is less dormant than the mft-2 single-mutant despite the strong dormancy phenotype of spt-2 in this ecotype (Fig. 3I).

Arabidopsis seed primary dormancy is promoted by low-temperature conditions during seed development (23–25). Low temperature also increases seed dormancy in wheat and this correlates with an increase in Ta-MFT gene expression (29). Once primary dormancy is released, secondary dormancy can be induced depending on environmental conditions (46). A recent study has demonstrated that secondary dormancy of Arabidopsis Cvi ecotype seeds correlated with low seasonal soil temperatures and high MFT expression (47). Thus, MFT appears to be associated with the induction of both primary and secondary dormancy in Arabidopsis. Another recent study has described a temperature-mediated promotion of FT expression by PIF4 (48). Interestingly, PIF4 and FT belong to the same protein families as SPT and MFT, respectively. Our demonstration that SPT and MFT control the transition from seed to seedling parallels the control of the transition from vegetative to floral development by PIF4 and FT. Allelic variation at the PIF4 locus is known to be associated with several ecological traits, including flowering time (49). Here we find ecotype differences in SPT function between Ler and Col in the absence of allelic variation at the SPT locus.

Antagonistic Routes Mediated by SPT Control Primary Dormancy.

Integrating the data reported here, we propose the following model for the role of SPT in setting seed primary dormancy (Fig. 3K). During seed development, SPT controls expression of transcription factors encoding genes that promote seed dormancy in both Ler and Col: SPT promotes ABI5 and represses MFT and RGA expression by binding to their respective promoters. This process creates two antagonistic “dormancy-repressing” and “dormancy-promoting” routes that operate in both ecotypes simultaneously. Both routes are reinforced by the indirect action of SPT in repressing ABI4 and promoting RGL3. These routes have different outcomes in Ler and Col seeds and this, in turn, influences ABA1 expression and ABA/GA ratios in dry seeds (Figs. 1 G and I, and 2 A and B). RGA enhances RGL2 function and they are both known to stimulate ABA biosynthesis (50), but it remains to be established whether this is through a direct effect on ABA1 expression. We propose that the low germination rates of spt mutant Ler seeds (Fig. 1 A and B) result from higher dormancy levels because of the elevated expression of MFT, RGA, and ABI4, which override the low expression of ABI5 and RGL3 (Fig. 2A). In our model the Ler-specific increase of RGA expression in the spt mutant background (Fig. 2 A and B) may be a factor reinforcing the dormancy-repressing route in this ecotype. This role for RGA is supported by the rescue of the strong dormant phenotype of spt-2 when crossed into della4 (Fig. 1F). In contrast to what happens in Ler, the high germination rates of spt mutant Col seeds (Fig. 1 C and D) result from lower dormancy levels because of the low ABI5 and RGL3 expression, which are dominant over the increased MFT and ABI4 expression (Fig. 2B). Our model is backed-up by the outcome of the spt mft double-mutant analyses. In Col the additive effect of the double-mutant is a consequence of the removal of these two dormancy-promoting factors. Furthermore, the fact that in Ler the spt mft double-mutant is not additive (the double-mutant germinates even more that the single mft-mutant and overrides the strong dormant phenotype of spt-2), suggests that MFT is a major component of the dormancy-repressing route in this ecotype.

Apart from its role in dormancy/germination, SPT is also involved in fruit development (32–34) and repression of hypocotyl elongation and cotyledon and leaf size (36–38). Hence, the role of SPT as a growth repressor of vegetative developmental stages parallels its effect in promoting seed dormancy (i.e., repressing growth) in Col seeds. The nonseed-related roles of SPT do not appear to vary between ecotypes. The fact that perturbation of SPT has different outcomes in terms of seed dormancy in Ler and Col perhaps reflects adaptation to different environmental conditions. Our double-mutant and gene-expression analysis implicate MFT and RGA as candidates for the ecotype differences. Genetic variation in seed-germination sensitivity of Ler and Col to the GA-biosynthesis inhibitor paclobutrazol has been mapped to three quantitative trait loci, but none of these colocalize with MFT or RGA (51). This study highlights the central role played by the transcription factor SPT in controlling seed dormancy and demonstrates how fine-tuning of five other transcription factors can lead to very different outcomes depending on ecotypes.

Experimental Procedures

All mutant lines used in this study were described previously: spt-2 (32); spt-11 and spt-12 (36); abi3-4 (52); aba1-1 (53); rgl1-1 rgl2-1 gai-6 rga-2 (della4) (10); spt-2 della4 (37); mft-2 (30). The SPToe and SPT2oe transgenic lines were obtained by transforming plants with the pK2GW7 and pH2GW7 vectors carrying the SPT and spt-2 coding sequences, respectively, downstream of the 35S promoter. SPTmyc containing Ler plants were produced by transformation with a pGREENII based binary vector carrying SPT-12xMYC and Col plants with a derivative of the pBI121 binary vector in which the GUS sequence was replaced with the SPT-9xMYC. In both ecotypes the SPTmyc genes are driven by the 35S promoter.

Plants were grown in the departmental greenhouse facilities supplemented with artificial light to give a photoperiod of 16-h light at a temperature of ∼20–22 °C. Germination assays were performed as described previously (38). Stratification was performed during imbibition of seeds for 2–3 d in the dark at 4 °C. RNA extractions from 1-d after-imbibition seeds and ABA and GA (GA4) measurements were performed as described previously (38, 45). For RNA extractions from developing seeds the method was scaled down to use ∼200 seeds and 15% of the buffer volumes. cDNA synthesis was performed using standard methods. qPCR was performed using iQ SYBR GreenSupermix and the MyiQ Real-Time PCR detection system (Bio-Rad) according to the manufacturer’s instructions. Expression of ACTIN2 was used for normalization. Transcriptomic analyses and ChIP assays were performed as described previously (42, 45). Public domain Affymetrix ATH1 data sets were obtained from NascArrays (http://affymetrix.arabidopsis.info/narrays/experimentbrowse.pl) and TAIR (www.arabidopsis.org) and used to produce Fig. S2. A MLS specific amplicon was used as negative control. Sequence of primers used in RT-qPCRs and ChIP-qPCRs are listed in Table S4. Relative positions of the G-box and E-box motifs are indicated in Table S2.

In relevant figures, asterisks indicate results of two tailed Student t test analyses performed in Microsoft Excel, comparing overexpressors, mutants, or treatments to corresponding controls. Grouped matching letters indicate nonsignificant differences from ANOVA followed by Tukey’s honestly significant difference tests performed in R (v2.15.2; www.R-project.org).