Histone Deacetylases Regulate MORE AXILLARY BRANCHED 2-Dependent Germination of Arabidopsis thaliana

Abstract

Under specific conditions, the germination of Arabidopsis thaliana is dependent on the activation of the KARRIKIN INSENSITIVE 2 (KAI2) signaling pathway by the KAI2-dependent perception of karrikin or the artificial strigolactone analogue, rac-GR24. To regulate the induction of germination, the KAI2 signaling pathway relies on MORE AXILLARY BRANCHED 2- (MAX2-)dependent ubiquitination and proteasomal degradation of the repressor protein SUPPRESSOR OF MAX2 1 (SMAX1). It is not yet known how the degradation of SMAX1 proteins eventually results in the regulation of seed germination, but it has been hypothesized that SMAX1-LIKE generally functions as transcriptional repressors through the recruitment of co-repressors TOPLESS (TPL) and TPL-related, which in turn interact with histone deacetylases. In this article, we show the involvement of histone deacetylases HDA6, HDA9, HDA19 and HDT1 in MAX2-dependent germination of Arabidopsis, and more specifically, that HDA6 is required for the induction of DWARF14-LIKE2 expression in response to rac-GR24 treatment.

Introduction

Seed germination is a key process in the plant life cycle that is essential for survival and reproduction. The decision to germinate is tightly regulated and integrates several environmental and endogenous signals, thereby ensuring that germination takes place under favorable conditions, such as the right temperature and light quality (Oh et al. 2004, Kendall et al. 2011). The capability of seeds to respond to these favorable conditions depends on the depth of seed dormancy that is regulated by two phytohormones, abscisic acid (ABA) and gibberellic acid (GA), that are required for germination inhibition and induction, respectively (Kucera et al. 2005, Holdsworth et al. 2008, Lee et al. 2010).

For some plant species, another environmental condition favoring seed germination is the occurrence of a fire that releases important resources for seedling establishments, such as light and nutrients (Van Staden et al. 2000). Several pyrophilic or fire-following plant species germinate in response to the availability of compounds found in plant-derived smoke, among which karrikins (KARs) are the most studied (Flematti et al. 2004). Six different KAR analogues have been identified in smoke water (KAR₁ to KAR₆), each with a different degree of germination-inducing activity (Flematti et al. 2009, 2010). Interestingly, seed germination of the non-fire follower Arabidopsis thaliana was found to be induced by KARs as well, of which KAR₁ and KAR₂ had the strongest effect (Nelson et al. 2009).

In Arabidopsis, the KAR receptor is an α/β hydrolase protein, designated KARRIKIN INSENSITIVE 2/HYPOSENSITIVE TO LIGHT (KAI2/HTL) (Sun and Ni 2011, Waters et al. 2012). Additionally, KAI2 signaling can also be activated by the synthetic strigolactone (SL) analogue rac-GR24, although its activity is lower than that of KAR₁ or KAR₂ (Nelson et al. 2009). Arabidopsis mutants lacking the functional KAI2 were not only insensitive to KARs and rac-GR24 but also displayed an increased seed dormancy in the absence of these compounds (Waters et al. 2012). This, taken together with the observation that non-fire followers, such as Arabidopsis, have also been found to be responsive to KARs, has led to the hypothesis that the KAI2 pathway is conserved in all land plants and its primary function is the detection of a currently unknown endogenous ligand, designated KAI2 ligand (KL) (Conn and Nelson 2016).

When KAI2 perceives KARs, rac-GR24 or KL, it recruits an Skp, Cullin, F-box containing (SCF) complex with F-box protein MORE AXILLARY BRANCHED 2 (MAX2) that, in turn, will ubiquitinate specific members of the SUPPRESSOR OF MAX2 (SMAX)1-LIKE (SMXL) family, marking them for proteasomal degradation (Nelson et al. 2011, Stanga et al. 2013, 2016, Khosla et al. 2020, Wang et al. 2020b). Both SMAX1 and SMXL2 are known targets of KAI2-MAX2 signaling involved in the regulation of several processes in plant development, including hypocotyl elongation (Stanga et al. 2013, Wang et al. 2020b), cotyledon expansion (Stanga et al. 2013, 2016), lateral root development (Villaécija-Aguilar et al. 2019), and root hair formation and elongation (Villaécija-Aguilar et al. 2019). However, for the induction of seed germination, KAI2-MAX2 signaling appears to mainly rely on the degradation of SMAX1, but currently the role of SMXL2 in this process has not been evidenced (Stanga et al. 2013, 2016).

Although the core KAI2-MAX2 signaling pathway has been resolved, it is still unclear how precisely the degradation of SMAX1 and SMXL2 eventually results in the control of developmental processes. A lot has been speculated based on the similarities with SL signaling, in which other members of the SMXL family, SMXL6, SMXL7 and SMXL/8 (SMXL6/7/8), act as transcriptional corepressors by interacting with TOPLESS (TPL) and TPL-related (TPR) proteins (Wang et al. 2015, 2020a). SMAX1 and SMXL2 also have an ETHYLENE-RESPONSE FACTOR Amphiphilic Repression motif, allowing association with TPL/TPR (Soundappan et al. 2015), presumably to repress an unknown transcription factor. TPL/TPR proteins are known to act through the recruitment of histone deacetylases (HDACs) HDA6 and HDA19 that repress gene expression by removing acetylation from histone tails (Krogan et al. 2012, Wang et al. 2013a, Ryu et al. 2014). According to the current hypothesis, the TPL/TPR-dependent interaction with HDACs would allow SMAX1 to suppress downstream responses on a transcriptional level. Proteasomal degradation of SMAX1 would then release HDACs from the repressed loci, thereby activating KAI2-related responses and leading to germination induction.

In accordance with a possible transcriptional output, the expression of several dormancy- and germination-related genes has been reported to be controlled by KAI2-MAX2 signaling. The ABA-catabolic genes CYTOCHROME P450 monooxygenase 707A 1 (CYP707A1) and CYP707A3 are induced by the addition of rac-GR24 in dark-imbibed seeds and together with CYP707A2 seem to redundantly regulate rac-GR24-dependent germination (Brun et al. 2019). Additionally, the GA biosynthesis genes GIBBERELLIC ACID 3 OXIDASE 1 (GA3ox1) and GA3ox2 are upregulated in seeds of Arabidopsis, ecotype Landsberg erecta upon treatment with KAR₁ (Nelson et al. 2009, 2010). Transcriptome analysis indicates that treatment with either rac-GR24 or GA triggers partially overlapping transcriptional changes in Arabidopsis seeds that express ShHTL7, a divergent KAI2 homologue from Striga hermonthica (Bunsick et al. 2020). Finally, KAI2-MAX2 signaling also seemingly regulates the expression of light-responsive genes, because a putative binding site for ELONGATED HYPOCOTYL5 (HY5) has been reported to be enriched among the KAR-responsive genes. Moreover, activated KAI2-MAX2 signaling has been reported to induce GA, ABA and ethylene responses in dark-imbibed seedlings that are normally associated with the response to light (Bunsick et al. 2022).

In Arabidopsis, 18 HDACs were identified that can be divided into three main types: 12 type I yeast reduced potassium deficiency 3 (RPD3)/HDA1-like HDACs, four type II HD-tuins (HDTs) and two type III sirtuins (reviewed by Hollender and Liu 2008). HDACs that are involved in a myriad of processes throughout plant development often require interaction with specific cofactors (Wang et al. 2014, Kumar et al. 2021, Tahir and Tian 2021). The chemical HDAC inhibitor, trichostatin A (TSA), has pleiotropic effects on plant development (Xu et al. 2005, Tanaka et al. 2008, Li et al. 2014, van Zanten et al. 2014), but its effect on seed germination depends on the concentration used and the exact conditions to which seeds are subjected (Tanaka et al. 2008, Perrella et al. 2013, van Zanten et al. 2014).

Indeed, eight HDACs have been implicated in the regulation of several overlapping aspects of seed dormancy and germination. Of the germination-involved RPD3/HDA1-like HDACs, HDA6 and HDA19 are the best characterized. Both repress embryogenesis-related genes in germinating seeds (Tanaka et al. 2008) and reduce their sensitivity to ABA and NaCl by interacting with the scaffolding protein HDC1 (Perrella et al. 2013). For HDA6, the impact on the ABA sensitivity correlated with the repression of ABA response genes (Chen et al. 2010). In addition, HDA19 can stimulate germination when recruited by HIGH-LEVEL EXPRESSION OF SUGAR-INDUCIBLE GENE2-LIKE 1 (HSL1) through the repression of seed maturation genes (Zhou et al. 2013). However, when HDA19 interacts with SWI-INDEPENDENT3-LIKE 1 (SNL1), it is seemingly involved in the inhibition, rather than the promotion of seed germination (Wang et al. 2013b). In contrast to the roles of HDA6 and HDA19, recent reports show that HDA9 stimulates the ABA sensitivity of germinating seeds, potentially by interacting with ABA INSENSITIVE 4 (ABI4) and repressing the ABA-catabolic genes CYTOCHROME P450 monooxygenase 707A 1 (CYP707A1) and CYP707A2 (Baek et al. 2020). Mutant analysis of the HDA9 gene suggests a role in the maintenance of seed dormancy and the suppression of seedling traits in dry seeds because the hda9 mutant shows increased seed germination, reduced expression of the dormancy genes DELAY OF GERMINATION 1 (DOG1) and HISTONE MONOUBIQUITINATION 1 (HUB1) and increased expression of seedling-associated genes (van Zanten et al. 2014). Furthermore, both silencing and overexpression of HDA7 resulted in a minor reduction in the germination percentage (Cigliano et al. 2013), but this might be an indirect effect because HDA7 silencing also leads to misregulation of HDA6, HDA9 and HDA19 (Cigliano et al. 2013). Finally, HDA15 was shown to interact with PHYTOCHROME INTERACTING FACTOR 1 (PIF1) and has been proposed to repress seed germination in the dark by suppressing light-responsive germination genes (Gu et al. 2017).

HDACs of the HD-tuin family have been reported to be involved in the regulation of seed dormancy and germination as well. Mutants of HDT1 and HDT3 have opposing phenotypes, namely increasing and reducing seed germination, respectively, hinting at an antagonistic role for HDT1 and HDT3 (Colville et al. 2011). Additionally, whereas the hdt1 mutant seemed to be less sensitive to the germination-inhibiting effects of ABA, NaCl and glucose than the wild type, the hdt3 mutant displayed the opposite response (Colville et al. 2011). Additionally, HDT2 was shown to promote germination through the regulation of GA levels, as seeds of an HDT2 overexpression line displayed upregulation of GA3ox1 and GA3ox2 and, consequently, increased GA accumulation (Yano et al. 2013).

By combining the proposed role of TPL/TPR-mediated HDAC recruitment in downstream KAI2-MAX2 signaling with the observation that several HDACs regulate the seed responses to internal and external stimuli, we hypothesized that at least one of these HDACs might be involved in the MAX2-controlled seed germination. Here, we demonstrate the importance of several HDACs in MAX2-dependent germination and reveal the particular role of HDA6 in the regulation of rac-GR24-induced transcriptional changes.

Results

TSA treatment inhibits MAX2-controlled germination in Arabidopsis

TSA is often applied to study the role of HDACs in specific processes (Xu et al. 2005, Tanaka et al. 2008, Li et al. 2014, van Zanten et al. 2014). To evaluate the possible involvement of HDACs in rac-GR24-dependent seed germination, we used TSA in a thermo-inhibition assay, in which seeds were imbibed under elevated temperature and in the absence of light, resulting in a dependence on KAI2-MAX2 signaling for successful germination (Toh et al. 2014). Under these conditions, Arabidopsis accession Columbia-0 (Col-0) seeds were treated without (mock) or with 10 µM rac-GR24, and 0, 5, 25 or 50 µM TSA. Treatment with TSA seemed to impede germination in a concentration-dependent manner because the germination of both rac-GR24- and mock-treated seeds was significantly reduced at concentrations of 25 and 50 µM TSA (Fig. 1A). TSA also affected seed germination under continuous light and could partially override the increased germination phenotype of smax1, indicating that TSA has a broader effect on germination than restricting MAX2-dependent germination (Fig. 1B; Supplementary Fig. S1). Moreover, the impact of TSA was seemingly not restricted to seed germination, because seedling development was slowed down under both dark and light conditions (Fig. 1). To summarize, these data provide a first indication that HDACs are important for rac-GR24-controlled germination, among other processes.

HDA6, HDA9, HDA19 and HDT1 are involved in rac-GR24-dependent seed germination

To isolate the specific HDACs of which the inhibition by TSA could account for its effect on rac-GR24-dependent germination, we screened the literature for HDACs known to play a role in seed germination and checked the phenotypes of the corresponding knockout mutants in our thermo-inhibition assay (Table 1).

Interestingly, for six of the eight selected hdac mutants, seed germination was affected both without and in the presence of rac-GR24 (Figs. 2A, 3A; Supplementary Fig. S2). The germination of the hda6, hda19 and hdt1 mutants was significantly lower under mock treatment, similar to max2, and seeds were significantly less sensitive to rac-GR24 than those of the wild type (Fig. 2A). On the contrary, although hda15, hdt2 and hdt3 also showed a significantly reduced germination under mock conditions, germination of these mutants in the presence of rac-GR24 was similar to that of Col-0 (Supplementary Fig. S2). Interestingly, hda9 seemed to be the only mutant with an increased, rather than a decreased, germination rate both in the presence and absence of rac-GR24 (Fig. 3A). To check the possible involvement of HDA7, we used a gain-of-function mutant (Cigliano et al. 2013) because both loss-of-function and gain-of-function mutants of HDA7 displayed reduced germination (Cigliano et al. 2013). In our assay, the germination percentage displayed by hda7 was not significantly different from that of Col-0 (Supplementary Fig. S2).

KAI2-MAX2 signaling has also been shown to be involved in the control of photomorphogenesis. Therefore, we assessed the hypocotyl elongation phenotype of the selected hdac mutants (Supplementary Fig. S3). For hda6, hda7 and hda15, hypocotyls were significantly elongated under mock conditions compared to those of the wild type, consistent with the phenotypes of the kai2 and max2 mutants (Shen et al. 2007, Sun and Ni 2011, Waters et al. 2012). Only for hda15, however, this elongation was not significantly different from that of max2, indicating that MAX2-dependent photomorphogenesis and seed germination are regulated by different sets of HDACs, and that HDA6 and HDA15 are involved in both KAI2-related responses.

As the impact on the MAX2-controlled germination was the highest in the hda6, hda9, hda19 and hdt1 mutants, we selected them for further research. Notably, besides the reduced seed germination phenotype in the dark, none of the tested mutants displayed germination defects under continuous light or the inhibited post-germination growth observed after TSA treatment (Fig. 1). Thus, HDA6, HDA19 and HDT1 might be the HDACs that upon inhibition by TSA give rise to the reduced germination phenotype in the dark, whereas the other TSA effects might be caused by the inhibition of other HDACs. Whereas the predicted impact of the TSA-mediated inhibition of HDA9 would be an increase in germination, it would probably be masked by the TSA effect on the other HDACs. In conclusion, although TSA has additional pleiotropic effects, HDA6, HDA9, HDA19 and HDT1 seem to be specifically involved in MAX2-mediated germination rather than seed germination in general.

KAI2-MAX2 pathway marker genes respond to rac-GR24 in seeds

As the main function of HDACs is transcriptional regulation through the deacetylation of histones, we hypothesized that HDA6, HDA9, HDA19 and HDT1 influence seed germination by affecting the expression of KAI2-MAX2 responsive genes. If HDACs regulate the primary targets of the KAI2-MAX2 pathway, the most interesting genes would be those that respond to rac-GR24 in a relatively short time. In a transcriptomics analysis of Arabidopsis seeds treated with KAR₁, both early (3–6 h) and late (12–24 h) response genes had been found (Nelson et al. 2010). To confirm that the later response was due to the treatment only and not to a need for the seeds to reach a later developmental stage through extended imbibition, we designed a quantitative real-time polymerase chain reaction (qRT-PCR) time course experiment. Wild-type seeds were first imbibed for 1, 24 or 48 h and subsequently treated for 1 or 6 h with 10 µM rac-GR24 under the same conditions used in our germination assay. This experiment allowed the distinction between genes that need a prolonged rac-GR24 treatment, which are probably not primary targets of the pathway and those that need a long imbibition before responding to rac-GR24. We selected three of the early KAR₁ response genes, GA3ox1, KARRIKIN UP-REGULATED F-BOX 1 (KUF1) and SALT TOLERANCE HOMOLOG 7 (STH7) and two late response genes, HY5 and FIDDLEHEAD (FDH) (Nelson et al. 2009, 2010). In addition, we included the DWARF14-LIKE2 (DLK2) gene, known to be highly responsive to KAR₂ and rac-GR24 in seedlings (Waters et al. 2012, Waters and Smith 2013).

Both DLK2 and STH7 were significantly upregulated after rac-GR24 incubation (Fig. 4A). For DLK2, this response occurred after 6 h of rac-GR24 treatment and seemed to be independent of the imbibition time. The expression of STH7 was significantly upregulated after 48 h of imbibition only and 6 h of treatment, although a nonsignificant response to rac-GR24 was observed when seeds were imbibed for 1 and 24 h as well (Fig. 4A). For KUF1 and GA3ox1, no significant response to rac-GR24 was detected (Supplementary Fig. S4A). The late response gene FDH was also not affected by the rac-GR24 treatment, indicating that its upregulation indeed requires longer treatment and not only an extended imbibition period (Supplementary Fig. S4B). By contrast, the other late response gene HY5 showed a small, but significant, induction by rac-GR24 after 1 h of imbibition and 1 h of treatment. Regardless of their responses to rac-GR24, all marker genes seemingly show transcriptional changes throughout the imbibition period from 1 to 48 h. For instance, for STH7, this change in baseline expression might illustrate how imbibition affects the degree to which the gene is responsive to rac-GR24 (Fig. 4B).

HDA6 is required for the transcriptional response to rac-GR24

To verify whether HDAC genes are needed for the transcriptional regulation during MAX2-controlled germination, we performed a qRT-PCR experiment and examined the expression of the early response genes DLK2, KUF1, STH7 and GA3ox1 in seeds after 1 h of imbibition and subsequent treatment with rac-GR24 for 6 h (Fig. 4B and Supplementary Fig. S4C). Similar to previous results, in Col-0, rac-GR24 significantly induced DLK2, but not the KUF1, STH7 and GA3ox1 genes. Interestingly, whereas the DLK2 expression was similar in Col-0 and the hdac mutants under mock conditions, the upregulation by rac-GR24 was significantly reduced and completely abolished in the hda9 and hda6 mutants, respectively. Although no statistically significant rac-GR24 response was seen for STH7 in Col-0 or the mutants, the basal expression of this gene seemingly increased in both hda6 and hda19 (Fig. 4B), but the response to rac-GR24 in hdt1 appears to be, nonsignificantly, lower than that in Col-0. A nonsignificant increase in GA3ox1 could be observed in hda9, both in the presence and the absence of rac-GR24 (Supplementary Fig. S4C).

Several genes involved in ABA and GA signaling are unresponsive during MAX2-controlled germination

Although the modified responsiveness of DLK2 in the hdac mutants hints at a possible role of HDACs in the transcriptional response to rac-GR24, it does not reveal how these HDACs eventually regulate seed germination. To gain more insights into the mechanism by which HDAC-dependent transcriptional changes regulate rac-GR24-induced germination, we selected a number of genes involved in ABA and GA metabolism that are also known to be regulated by HDA6, HDA9, HDA19 or HDT1 (Table 2). To assess whether these genes respond to rac-GR24 in seeds imbibed in the dark, we performed a time course qRT-PCR as previously described.

Many of these genes showed transcriptional changes during imbibition; most notably, the ABA signaling genes ABI3 and ABI4 were upregulated independently of the rac-GR24 treatment when the 6-h time point after 1-h imbibition was compared with the 0-h time point (Supplementary Fig. S5A) as well as the GA biosynthesis genes GA3ox1, GA3ox2, GA20ox1 and GA20ox3 and the DELLA protein-encoding gene REPRESSOR OF GA (RGA) (Supplementary Fig. S5B). By contrast, the dormancy master regulator DOG1 and the ABA-catabolic gene CYP707A1 showed downregulation at this time point (Supplementary Fig. S5A, C). Opposite to the KAI2-MAX2 pathway marker genes DLK2 and STH7, none of the selected genes had a significant response to rac-GR24 under the tested conditions (Supplementary Fig. S5). As such, while these genes play a role in seed germination in general, they seem not to be involved in the rac-GR24-dependent induction of germination. Nevertheless, these genes might still be under the control of the HDACs under study and might be important in the eventual germination phenotypes observed. For instance, in hda9, the DOG1 expression was apparently reduced in seeds imbibed for 1 h when compared with the wild-type Col-0 (Supplementary Fig. S5D).

HDA gene expression is unaffected by the rac-GR24 treatment

To further explore the exact role of HDA6, HDA9, HDA19 and HDT1 in rac-GR24-dependent seed germination, we checked the expression of these genes in our time course qRT-PCR. None of the four HDA genes showed a significant response to rac-GR24 (Supplementary Fig. S6). In addition, although HDA6, HDA9 and HDA19 did not display clear transcriptional changes during imbibition, the HDT1 expression was distinctly upregulated between 1 and 6 h of imbibition, seemingly corresponding to the time point of the first detected response to rac-GR24 (Fig. 4; Supplementary Fig. S6). Thus, the HDAC genes appear not to be transcriptionally regulated by rac-GR24 signaling, whereas HDT1 might be differentially expressed at different imbibition stages.

Discussion

Despite their importance as repressors of KAR and SL signaling, the precise mechanisms by which SMXL proteins regulate the output of these pathways is still unknown. SMXLs have been proposed to control the expression of downstream target genes through interaction with TPL/TPR corepressors, which, in turn, recruit HDACs (Krogan et al. 2012, Wang et al. 2013a, Ryu et al. 2014). Indeed, SMXL6/7/8 have been shown to regulate the expression of the transcription factors BRANCHED 1, TCP DOMAIN PROTEIN 1 and PRODUCTION OF ANTHOCYANIN PIGMENT 1 in a TPL-dependent manner, thereby controlling shoot branching, leaf shape and anthocyanin production, respectively (Wang et al. 2020a). However, thus far, no target genes that control seed germination have been identified for SMAX1. Here, we propose a function for several HDACs in the regulation of seed germination in the dark and under increased temperature conditions. Under these conditions, seed germination depends on activated KAI2-MAX2 signaling.

The reduced germination phenotype of hda6, hda19 and hdt1, resembling that of a max2 mutant, indicates that these HDACs might play a role as positive, rather than negative, regulators of MAX2-dependent germination (Fig. 2A). In accordance with the max2-like germination phenotype of hda6, the rac-GR24-induced upregulation of the KAI2-MAX2 signaling marker DLK2 in seeds was inhibited (Fig. 4), suggesting that HDA6 might regulate downstream transcriptional responses. However, the reduced transcriptional induction in the hda6 mutants suggests that the HDA6 activity would induce gene expression, rather than inhibit it, which would be the expected result of histone deacetylation. A role for HDACs in transcriptional activation in plants has been described previously in jasmonate (JA) signaling, in which response genes depend on HDA6 and HDA19 for their induction (Zhou et al. 2005, Wu et al. 2008). Also in yeast and mammalian cells, a role of HDACs as transcriptional activators was observed (reviewed by Nusinzon and Horvath 2005), possibly to be attributed to different mechanisms. First, the regulation of DLK2 might be indirect, even though we made efforts to capture the earliest transcriptional responses to rac-GR24. Second, for certain genes, transcriptional activation might require a delicate balance of histone acetylation and deacetylation (Nusinzon and Horvath 2005). In these cases, both histone acetyl transferase and HDAC activity are necessary for optimal transcriptional induction. Third, HDACs can deacetylate and, thus, influence the activity of other proteins, possibly playing a role in transcriptional regulation. Most notably, in JA signaling, HDA6 induces target gene expression by deacetylating TPL and thereby abolishing the interaction with the corepressor protein NOVEL-INTERACTOR-OF-JAZ (NINJA) (An et al. 2022). A similar mechanism might also occur in KAI2-MAX2 signaling, in which the interaction between SMAX1 and TPL could be influenced by HDA6-dependent TPL deacetylation. Alternatively, HDACs could induce rac-GR24-dependent transcriptional changes through the recruitment or the regulation of the still unknown cofactors.

Besides HDA6, the other HDACs also appeared important during MAX2-dependent germination (Figs. 2A, 3A; Supplementary Fig. S2). The max2 phenotype of hda19 corresponds to the reported HDA19 role in stimulating germination, but no changes in sensitivity to rac-GR24 were observed on the level of DLK2 expression (Fig. 4B; Supplementary Fig. S4B), implying that HDA19 and HDA6 could regulate germination through the control of a distinct gene set. The same is true for HDT1, because, in hdt1, the genes studied appeared to respond to rac-GR24 to an extent similar to that in Col-0 (Fig. 4; Supplementary Fig. S4). Interestingly, the reduced germination of hdt1 in the thermo-inhibition assay contrasted with the reported HDT1 role as germination inhibitor (Fig. 2) (Colville et al. 2011), underlining that HDACs could play varying roles, depending on the tested conditions. At least for HDA19, both an inhibitory role and a stimulatory role in seed germination were described (Zhou et al. 2013, Wang et al. 2013b). Similarly, HDT1 might play an inhibitory role in seeds treated with ABA, NaCl and glucose, but its role might be opposite under the conditions of the thermo-inhibition assay (Colville et al. 2011). Also for HDA6 and HDA19, their role in seed germination is often been demonstrated through a decreased sensitivity to certain germination-impeding treatments or conditions (Chen et al. 2010, Chen and Wu 2010, Perrella et al. 2013, Zhou et al. 2020). However, under optimal conditions, HDA6, HDA19 and HDT1 might not be absolutely required or might work redundantly with other HDACs. This suggestion is in accordance with the observation that hda6, hda19 and hdt1 germinated perfectly in light and at 21°C, in contrast with their reduced germination under less favorable conditions (Fig. 2B).

By contrast, hda9 displayed an increased germination phenotype, corresponding to a role for HDA9 in inhibiting germination reported previously (van Zanten et al. 2014). Based on this phenotype, HDA9 could be a candidate corepressor mediating the SMAX1 role as a transcriptional inhibitor of the KAI2-MAX2 signaling output. However, none of the rac-GR24-responsive genes tested conclusively showed an increased expression in hda9 compared to Col-0, besides a nonsignificant increase in GA3ox1 expression. Moreover, the expression of DLK2 in rac-GR24-treated seeds even seems to be reduced in hda9. Interestingly, the expression of the dormancy master regulator DOG1 is reduced in hda9, also in agreement with a previous report (van Zanten et al. 2014). As DOG1 does not respond to rac-GR24 in our assay and, thus far, its involvement in MAX2-dependent germination has not been reported, insight into the role of HDA9 in MAX2-dependent germination is still lacking.

The elongated hypocotyl phenotype of hda6, resembling that of max2 under mock conditions, indicates that HDA6 might also be required for MAX2-dependent reduction of hypocotyl elongation (Supplementary Fig. S3). Previously, an opposite role for HDA6 has been described, based on the observation that hypocotyls of dark-grown hda6 seedlings were shorter than those of Col-0 (Hao et al. 2016). The possible reason for this discrepancy might be the different light conditions. For HDA15, light treatment can strongly influence its role in light-regulated processes, such as seed germination and hypocotyl elongation (Gu et al. 2017, Alinsug and Deocaris 2023).

Besides our examination of HDACs, we also provide a new perspective on time course experiments, particularly when the effect of a given treatment in time is assessed. A simple time course, in which the treatment is administered at a single ‘0 h’ time point, can reveal the effect of a treatment in time when compared with the untreated control. However, when the response to the treatment is delayed, this set-up does not allow the determination of whether the delay is caused by an indirect response to the treatment or by a direct response triggered later in time. In the context of seeds, for instance, a transcriptional response to rac-GR24 can be delayed either because it is a secondary response, or because a certain developmental stage has to be reached before the induction of a particular response. After all, early imbibition is a very dynamic process, marked by rapid physiological changes (Weitbrecht et al. 2011). By administering rac-GR24 at different time points during imbibition, we could assess whether a short rac-GR24 treatment was effective when seeds are allowed to imbibe for a longer period of time. For instance, the late KAI2-MAX2 signaling marker genes HY5 and FDH had previously been reported to have the strongest response to the KAR₁ treatment after 24 and 48 h, respectively (Nelson et al. 2010). In our time course, however, FDH was not responsive to a short rac-GR24 treatment after 24 or 48 h imbibition, indicating that at least for this gene, a lengthened treatment seems indeed required. By contrast, HY5 responded significantly to rac-GR24 after 1 h of imbibition followed by 1 h of treatment, which is earlier than expected. Similarly, for KUF1 and GA3ox1, the convincing induction reported previously could not be detected (Nelson et al. 2009, 2010). Probably, discrepancies in transcriptional responses in the seeds can be attributed to the use of a different Arabidopsis ecotype, the utilization of primary dormant seeds instead of thermo-inhibited seeds, or variations in the conditions under which the seeds were imbibed or treated, including temperature, light or solid instead of liquid growth medium (Nelson et al. 2009, 2010).

Moreover, our results indicate that certain transcriptional responses to rac-GR24 are modulated by the imbibition time. In general, the final gene expression level is determined by the baseline expression, which is influenced by imbibition time, and the changes of this baseline by rac-GR24 treatment. For instance, whereas DLK2 seems equally responsive to rac-GR24 after different imbibition periods, changes in the basal expression of STH7 during imbibition apparently modulate how strongly it responds to rac-GR24. Similarly, the high GA3ox1 baseline expression at the 6 h time point after 1 h of imbibition seemingly, but nonsignificantly, increased to even enhanced levels upon rac-GR24 treatment.

One final observation drawn from our time course experiments is the apparent scarcity of rac-GR24 responsiveness of the tested genes. A possible explanation might be that, at least under the conditions used here, the set of genes that respond to rac-GR24 within 6 h of treatment is quite limited, albeit not absent, as indicated by the response of DLK2 and HY5 (Fig. 4; Supplementary Fig. S4). Thus, the earliest transcriptional response to the rac-GR24 treatment may not encompass genes related to GA and ABA metabolism or signaling, as assumed during the selection of possible target genes. At least for CYP707A1, CYP707A2 and CYP707A3, no expression was induced before the 12 h time point (Brun et al. 2019). Additionally, under certain conditions, the KAI2-MAX2 pathway had been shown to work in parallel to, rather than upstream of, GA (Bunsick et al. 2020). Recently, KAI2-MAX2 signaling has been reported to at least partially complement the absence of light by inducing the transcriptional program that is also regulated by light (Bunsick et al. 2022). This observation is in agreement with the results of previous transcriptome analysis, in which KAR-responsive genes were enriched for genes with a putative HY5-binding site (Nelson et al. 2010). A role in the response to light, or rather its absence, could indeed combine the two genes HY5 and DLK2 that respond early to rac-GR24. Whereas the role of HY5 in the response to is well established, that of DLK2 is still unclear (Ang and Deng 1994, Ang et al. 1998, Chattopadhyay et al. 1998). However, light-dependent induction of DLK2 by rac-GR24 treatment has been shown, as well as a role in KAI2-MAX2-related responses in Arabidopsis seeds and seedlings under low light conditions (Végh et al. 2017, Bunsick et al. 2022). As such it appears DLK2 could be involved in seed germination under specific conditions, and the regulation of DLK2 may in fact possibly be a way through which HDA6 regulates MAX2-dependent germination. Potentially, new knowledge on the exact role of DLK2 and the additional players therein might help discover additional target genes of KAI2-MAX2 signaling in dark-imbibed seeds.

To conclude, this study provides insights into a possible role for HDACs in MAX2-dependent regulation of seed germination. Further biochemical and genetic analyses could be carried out to elucidate the exact function of these HDACs and the mechanism by which KAI2-MAX2-induced gene expression is regulated. Additionally, transcriptomics analysis could aid in clarifying which transcriptional changes can link the HDACs to their effect on rac-GR24-controlled seed germination.

Materials and Methods

Plant material

Arabidopsis thaliana (L.) Heynh. accession Columbia-0 (Col-0) was used as wild type. The alleles for hda6 (SALK_201895), hda7 [SALK_002192 (Cigliano et al. 2013)], hda9 [SALK_007123 (Kim et al. 2013)], hda15 [SALK_004027 (Liu et al. 2013)], hda19 [SALK_139445 (Kim et al. 2008)], hdt1 [GABI_758H10 (Zhang et al. 2019)], hdt2 [SAIL_1247A02 (Li et al. 2017)] and hdt3 [SALK129799 (Luo et al. 2012)] in Col-0 background were obtained from the Nottingham Arabidopsis Stock Center (University of Nottingham, Sutton Bonington Campus, UK). The max2-1 (Stirnberg et al. 2002) and htl-3 (Toh et al. 2014) alleles have been described previously.

Thermoinduced seed germination assay

Before use in the assay, harvested seeds were stored under ambient conditions for at least 3 weeks. Seeds were sterilized with 25% NaClO, rinsed with sterile water and resuspended in seed resuspension solution [1 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.5), 0.1% (v/v) Preservative plant mixture (PPM; Plant Cell Technology, Washington, DC, USA)]. Resuspended seeds were pipetted in a 96-well plate and were treated with 10 µM rac-GR24 or 0.1% (v/v) acetonitrile as a control. Additionally, for assessment of the TSA effect, seeds were treated with 5, 25 or 50 µM TSA, or with 1% (v/v) dimethyl sulfoxide as a control. The seeds were incubated for 4 days at 24°C in the dark or under continuous light. Germination was indicated by the emergence of the radicle. For each combination of treatment, genotype and repeat, 36 wells, divided over six plates and each containing 15–40 seeds, were counted.

Hypocotyl elongation assay

Seeds were surface-sterilized [70% (v/v) ethanol and 0.05% (w/v) sodium dodecyl sulfate (SDS) for 5 min and 95% (v/v) ethanol for 5 min] and sown in square plates with half-strength Murashige and Skoog (½MS) medium, supplemented with 1 µM rac-GR24 or 0.01% (v/v) acetonitrile as a control. After stratification for 2 days at 4°C, plates were sequentially transferred to continuous white light at 21°C for 3 h, continuous darkness at 21°C for 21 h and continuous red light for 4 days at 21°C. Finally, hypocotyl length was measured. For each combination of treatment, genotype and repeat, 60 hypocotyls, divided over four plates, were measured.

mRNA extraction and cDNA synthesis

The mRNA extraction protocol (Townsley et al. 2015) was optimized to be applied to seeds. In short, 25 mg of seeds for each sample were imbibed in resuspension solution [1 mM HEPES, pH 7.5, 0.1% (v/v) preservative plant mixture (Plant Cell Technology)] supplemented with 10 µM rac-GR24 or 0.1% (v/v) acetonitrile as a control. Before crushing the samples in a bead mill (2 min at 20 Hz), they were supplemented with 400 µl lysis/binding buffer (LBB) [100 mM Tris-HCl, pH 8, 1 M LiCl, 10 mM EDTA, 1% (w/v) SDS, 5 mM dithiotreitol (DTT), 1.5% (v/v) Antifoam A (Sigma Aldrich, Saint Louis, MO, USA) and 0.5% (v/v) 2-mercaptoethanol]. After 10-min incubation at room temperature, two consecutive centrifugation steps were done to purify the lysate (10 min at 20,000×g). To capture mRNA, 200 µl of lysate was added to 1 µl of 12.5 µM biotinylated polyT-oligonucleotides (5′-Biotin-ACAGGACATTCGTCGCTTCCTTTTTTTTTTTTTTTTTTTT-3′) and incubated on a shaker for 10 min at room temperature. The mRNA was extracted from the lysate by adding 20 µl of streptavidin-coated magnetic beads (New England BioLabs, Ipswich, MA, USA) that were prewashed once with LBB. Again the mixture was incubated on a shaker for 10 min at room temperature. Using a 96-well magnetic separator, the bead-bound mRNA was pulled down from the lysate and washed consecutively with 200 µl wash buffer A [10 mM Tris-HCl, pH 8, 150 mM LiCl, 1 mM EDTA, 0.1% (w/v) SDS], 200 µl Wash buffer B (10 mM Tris-HCl, pH 8, 150 mM LiCl and 1 mM EDTA), and 200 µl Low-Salt buffer (20 mM Tris-HCl, pH 8, 150 mM NaCl, 1 mM EDTA). Finally, mRNA was eluted from the beads by the addition of 16 µl elution buffer (10 mM Tris-HCl, pH 8 and 1 mM 2-mercaptoethanol) and incubation at 80°C for 2 min. The cDNA was prepared using the qScript cDNA SuperMix (Quantabio, Beverly, MA, USA) according to the manufacturer’s instructions.

Quantitative real-time PCR

The qRT-PCR experiments were done on a Lightcycler 480 (Roche Diagnostics, Basel, Switzerland) and detected with SYBR Green (Roche Diagnostics). All reactions were done in triplicate and averaged. Cycle threshold values were obtained with the accompanying software and data were analyzed with the 2^−ΔΔCt method (Livak and Schmittgen 2001). The relative expression was normalized against the constitutively expressed AT4G34270 and AT4G12590, which has been proposed as suitable reference genes for Arabidopsis seeds (Dekkers et al. 2012). Sequences of primers used in this study are provided in Supplementary Table S1.

Supplementary Data

Supplementary data are available at PCP online.

Data Availability

The data underlying this article will be shared upon reasonable request to the corresponding author.

The Arabidopsis Genome Initiative identifiers for the genes described in this work are as follows: ABA2, AT1G52340; ABI3, AT3G24650; ABI4, AT2G40220; CYP707A1, AT4G19230; CYP707A2, AT2G29090; CYP707A3, AT5G45340; DLK2, AT3G24420; DOG1, AT5G45830; FDH, AT2G26250; GA2ox2, AT1G30040; GA3ox1, AT1G15550; GA3ox2, AT1G80340; GA20ox1, AT4G25420; GA20ox3, AT5G07200; HDA6, AT5G63110; HDA7, AT5G35600; HDA9, AT3G44680; HDA15, AT3G18520; HDA19, AT4G38130; HDT1, AT3G44750; HDT2, AT5G22650; HDT3, AT5G03740; HY5, AT5G11260; KAI2, AT4G37470; KAO1, AT1G05160; KUF1, AT1G31350; RGA, AT2G01570; SMAX1, AT5G57710; STH7, AT4G39070; TIP41, AT4G34270; Unknown protein, AT4G12590.

Funding

This work was supported by the Research Foundation-Flanders [1S91021N to A.T.].

Acknowledgments

The authors thank Martine De Cock for help in preparing the manuscript.

Author Contributions

A.T., S.S., and S.G. involved in designing the experiments. A.T. helped in performing the experiments. A.T. and S.S. involved in writing the manuscript; F.-D.B involved in the synthesis of SL analogues. S.S. and S.G. involved in conceptualization and funding acquisition. All the authors have read and agreed to the published version of the manuscript.

Disclosures

The authors have no conflicts of interest to declare.

References

Author notes