ABSTRACT
Seed germination is a crucial checkpoint for plant survival under unfavorable environmental conditions. Abscisic acid (ABA) and its signaling play a vital role in integrating environmental information to regulate seed germination. MCM1/AGAMOUS/DEFICIENS/SRF (MADS)-box transcription factors are mainly known as key regulators of seed and flower development in Arabidopsis. However, their functions in seed germination are still poorly understood. Here we report that MADS-box transcription factor AGL21 negatively modulates seed germination and post-germination growth by controlling the expression of ABA-INSENSITIVE 5 (ABI5) in Arabidopsis. AGL21 responds to multiple environmental stresses and plant hormones. The AGL21-overexpressing plants are hypersensitive to ABA, salt and osmotic stresses during seed germination and early post-germination growth, whereas agl21 mutants are less sensitive. AGL21 positively regulates ABI5 expression in seeds. Genetic analyses reveal that AGL21 is epistatic to ABI5 in controlling seed germination. Chromatin immunoprecipitation assays further demonstrate that AGL21 could directly bind to the ABI5 promoter in plant cells. Taken together, our results suggest that AGL21 acts as a surveillance integrator that incorporates environmental cues and endogenous hormonal signals into ABA signaling to regulate seed germination and early post-germination growth.
INTRODUCTION
Seed dormancy is a vital trait for plants to adapt to varieties of habitats and climates. Seed germination is arrested under adverse conditions and resumed when the conditions are favorable. Generally, seed germination commences with three phases of water uptake (Bewley, 1997), followed by embryo expansion and finished with radicle emergence. The optimal level of seed dormancy and germination has important repercussions on agricultural production. Therefore, it is essential to study the underlying molecular mechanisms that control seed dormancy and germination.
To sessile organisms, the ambient environment is vital for their survival. A variety of environmental factors, including nutrients, water availability, temperature, light, oxygen, and soil salinity, affect seed germination (Finkelstein et al., 2008; Holdsworth et al., 2008). Plants have evolved an array of strategies to constantly monitor the changing environmental conditions to decide when to germinate (Finkelstein et al., 2008). The environmental cues perceived by seeds can be incorporated into endogenous hormonal signaling pathways to regulate germination. Among various phytohormones, abscisic acid (ABA) and gibberellic acid (GA) are regarded as the primary regulators of transition from dormancy to germination (Seo et al., 2006; Shu et al., 2016). ABA is required for seed dormancy maintenance, while GA acts antagonistically to release dormancy and initiate seed germination (Shu et al., 2013). Environmental factors regulate the ABA:GA balance and the sensitivity to these hormones by modifying the biosynthetic and catabolic pathways, as well as signaling pathways, thus modulating seed dormancy and germination (Finch-Savage and Leubner-Metzger, 2006).
During germination, ABA content is decreased rapidly, and ABA signaling must be actively repressed. Through screening for ABA-insensitive (ABI) mutants, several ABA signaling components control seed germination have been identified in Arabidopsis (Finkelstein et al., 2002; Nambara and Marion-Poll, 2003). In contrast to ABI1/2, ABI3, ABI4 and ABI5 are key positive regulators of ABA signaling that modulate seed germination and post-germination development (Giraudat et al., 1992; Parcy et al., 1994; Finkelstein et al., 1998; Lopez-Molina et al., 2002). ABI5 is one of the 13 members of the group-A bZIP TF subfamily, and can directly bind to the ABA-RESPONSE ELEMENT (ABRE) cis-element in the promoter sequence of ABA-responsive genes, such as Arabidopsis EARLY METHIONINE-LABELED 1 (AtEm1), AtEm6, and RD29B to modulate their expression (Carles et al., 2002; Finkelstein et al., 2005; Nakashima et al., 2006). ABI5 interacts with ABI3 and acts downstream of ABI3 to execute an ABA-dependent growth arrest during germination (Lopez-Molina et al., 2002). ABI5 protein level and activity are tightly regulated post-transcriptionally. In stress conditions, ABA triggers phosphorylation and activation of ABI5 (Lopez-Molina et al., 2001; Dai et al., 2013), and KEEP ON GOING (KEG) E3 ligase is rapidly degraded (Stone et al., 2006), promoting the accumulation of high levels of ABI5. In favorable conditions, ABI5 is dephosphorylated by FyPP/PP6 (for Phytochrome-associated serine/threonine protein phosphatase/Ser/Thr-specific phosphoprotein phosphatase 6), and then the inactive ABI5 is degraded by the 26S proteasome (Dai et al., 2013). ABI5 can also be modified by S-nitrosylation and sumoylation, and is rapidly degraded, which is facilitated by ABI FIVE BINDING PROTEIN (AFP) and KEG (Miura et al., 2009; Albertos et al., 2015).
The MADS-box gene family in higher plant is a large family with more than 100 members (De Bodt et al., 2005). These TFs are involved in almost every developmental process in plants (Smaczniak et al., 2012). However, their roles in seed germination are largely unexplored. Only two MADS-box genes were found regulating seed germination till now. One is FLOWERING LOCUS C (FLC)/AGL25, which is involved in temperature-dependent seed germination through influencing ABA catabolic pathway and GA biosynthetic pathway (Chiang et al., 2009). Another MADS-box gene AGL67 may act as a repressor of seed germination, for knockout of it decreasing seed dormancy (Bassel et al., 2011). In our previous study, we found MADS-box TF AGL21 is involved in lateral root development and growth mediated by various environmental and physiological signals (Yu et al., 2014). In this study, we report that AGL21 also modulates seed germination by regulating ABA signaling. Overexpression of AGL21 conferred hypersensitive seed germination to ABA, high NaCl and mannitol, while the AGL21 knockout mutants showed the opposite phenotypes. Further analyses showed that AGL21 was induced by various stresses, and highly expressed in dry seeds, but decreased quickly after imbibition and germination, which is coincident with that of ABI5. Genetic analysis showed that AGL21 was involved in ABA signaling, acting downstream of ABI1/2 and upstream of ABI5. Taken together, our results suggest that AGL21 integrate environmental signals and internal hormonal signals to ABA signaling by directly modulating ABI5 to fine-tune seed germination and post-germination growth.
RESULTS
AGL21 negatively regulates seed germination and post-germination in response to ABA, salt and osmotic stress
AGL21 was reported to be expressed in developing embryos (Burgeff et al., 2002), and our previous study showed that AGL21 also had high expression levels in siliques and dry seeds besides in roots (Yu et al., 2014). These data indicates that AGL21 may play some roles in seed development or seed germination. To study the function of AGL21 in seed germination, we obtained 35S::AGL21 (OX 1-6 and OX 3-5), 35S::AGL21-HA (OX 29-3) transgenic Arabidopsis lines, and two T-DNA insertion mutants: CS118325 (agl21-1) and GK_157C08 (agl21-2) from ABRC (Yu et al., 2014). Gene expression analyses by quantitative real-time polymerase chain reaction (qRT-PCR) showed that the expression of AGL21 was abolished in the mutants while highly up-regulated in the AGL21-overexpressing plants (Fig. S1). To evaluate the function of AGL21 in seed germination, we germinated the seeds of AGL21-overexpressing plants, agl21 mutants and wild-type (WT) plants on MS media with or without ABA. In the absence of ABA, the seed germination rates of different genotypes were similar (Fig. 1A, B). In the presence of different concentrations of ABA, agl21-1 and agl21-2 seeds were more resistant to ABA inhibition. On the contrary, AGL21-overexpressing seeds were more sensitive to ABA during germination (Fig. 1A-D). In line with the germination rate data, agl21 mutant plants also showed higher cotyledon-greening percentages than WT plants, while the three AGL21-overxpressing lines exhibited significant lower cotyledon-greening rates after 8 days germination (Fig. 1E). These results suggested that AGL21 acts as a negative regulator of seed germination and post-germination.
Since AGL21 modulates ABA-regulated seed germination, we tested whether AGL21 affects the seed germination response to salt stress. The seeds were sown on MS medium supplemented with 150 mM NaCl. Compared with WT seeds, germination and cotyledon-greening of AGL21-overexpressing seeds were more severely inhibited by NaCl, whereas the agl21 mutant seeds showed much higher germination and cotyledon-greening ratios (Fig. 2A, C and E). To identify whether AGL21 is involved in salt-specific or general osmotic responses, we further tested seed germination on medium containing osmotic reagent mannitol. Similarly, on MS medium containing 300 mM mannitol, germination of AGL21-overexpressing seeds were much more sensitive while the agl21 mutant seeds were insensitive to the inhibition effects of mannitol compared with WT seeds (Fig. 2A, D and E), indicating AGL21 negatively regulates seed germination in response to osmotic stress.
AGL21 is responsive to multiple stresses during seed germination
AGL21 responds to multiple endogenous and exogenous signals, such as ABA, methyl jasmonate (MeJA), indole-3-acetic acid (IAA), nitrogen (N) and sulfur (S) starvation during root development (Yu et al., 2014). Our results in this study show that AGL21 is involved in the inhibition of germination in response to ABA, NaCl and osmotic stress. Thus we wanted to know whether AGL21 also responded to these stress signals during germination stage. We analyzed the AGL21 expression levels in seed germination stages on media with ABA, NaCl or mannitol. The results show that all these stress treatments significantly induced AGL21 expression in germinating seed. Moreover, AGL21 was also markedly induced by N deficiency, MeJA and IAA treatments during seed germination (Fig. 3A). These data support the surveillance function of AGL21 in seed germination in response to ABA, salt and osmotic stresses, and maybe other stresses.
AGL21 is involved in ABA signaling
In order to study whether AGL21 is involved in ABA signaling, we checked the AGL21 expression levels in abi mutants in response to ABA. As shown in Figure 3B and 3C, ABA and mannitol treatments could significantly induce AGL21 expression in abi3-8, abi4-1, abi5-7 background plants as that in WT background plants. However, in abi1-1 and abi2-2 background plants, no significant induction of AGL21 transcription was observed. These data suggested that AGL21 might function in ABA signaling pathway, and may act downstream of ABI1 and ABI2, and upstream or in parallel with ABI3, ABI4 and ABI5.
We further analyzed the expression levels of several ABA signal pathway genes in germinating seeds of AGL21-overexpressing, knockout and WT plants germinated on MS medium or MS medium with ABA. We found that AGL21 positively regulated some downstream ABA signal pathway genes, such as ABI5, AtEM6, RD29B and ABA RESPONSE ELEMENT-BINDING FACTOR 2 (AREB2)/AREB BIND FACTOR 4 (AREB2/ABF4) (Fig. 4A-H). However, gene expression levels of upstream ABA signal pathway genes, such as ABI1, ABI2, SNF1-RELATED PROTEIN KINASE 2.2 (SnRK2.2) and SnRK2.3, did not change significantly in AGL21-overexpressing, knockout and WT plants (Fig. 4I-P). Together, these results indicate that AGL21 is involved in ABA signaling, and may act downstream of TYPE 2C PROTEIN PHOSPHATASES (PP2Cs) and SnRKs and upstream of ABI5 and other downstream ABA signal components to regulate seed germination and post-germination growth.
AGL21 does not affect ABA and GA biosynthesis
ABA and GA antagonistically govern seed germination. ABA promotes seed dormancy, while GA stimulates seed germination. To investigate whether AGL21 regulates seed germination by altering ABA or GA content, we analyzed the expression levels of many key genes in ABA biosynthesis pathway, such as ABA DEFICIENT 2 (ABA2), ABA3, ARABIDOPSIS ALDEHYDE OXIDASE 3 (AAO3), 9-CIS-EPOXYCAROTENOID DIOXYGENASE (NCED3), and key ABA catabolic enzyme family genes CYP707A1–CYP707A4. Our data show that all of these genes had similar expression levels in day 2 and day 3 germinating seeds of different genotypes both on MS and ABA media (Fig. S2). Moreover, we tested the ABA contents in dry seeds, and found there was no significant change in ABA contents between the WT, agl21-1 and AGL21-overexpressing seeds (Fig. S3).
We also analyzed the expression levels of several key genes in GA biosynthesis pathway, including ENT-COPALYL DIPHOSPHATE (CPS), ENT-KAURENE SYNTHASE (KS), ENT-KAURENOIC ACID OXIDASE 2 (KAO2), GA20-OXIDASE 1 (GA20OX1), GA20OX2, GA20OX3, GA REQUIRING 3 (GA3), GA3-OXIDASE 1 (GA3OX1), GA3OX2 and GA3OX3, in day 3 germinating seeds. qRT-PCR analysis revealed that the expression levels of the examined GA metabolism genes were largely unaltered in the AGL21-overexpressing, knockout and WT seeds grown on MS medium with or without ABA (Fig. S4). Taken together, these results show that AGL21 does not alter ABA or GA content during seed germination and post-germination growth.
AGL21 has similar expression pattern to ABI5 and regulates ABI5 expression and protein accumulation in seeds
To uncover the molecular networks underlying AGL21 regulation of seed germination, we monitored the expression levels of genes associated with ABA dependent seed germination, such as ABI1, ABI2, ABI3, ABI4, ABI5, MYB DOMAIN PROTEIN 96 (MYB96), ACYL-COENZYME A-BINDING PROTEIN 1 (ACBP1), ARABIDOPSIS HISTIDINE KINASE 1 (AHK1), RING-H2 FINGER A2A (RHA2a), RELATED TO ABI3/VP1 1 (RAV1) (Tran et al., 2007; Bu et al., 2009; Du et al., 2013; Feng et al., 2014; Lee et al., 2015), in day 3 germinating seeds in the present of ABA. Among these genes, ABI3, ABI4, ABI5 were found significantly up-regulated in AGL21-overexpressing plants. However, their expression was not significantly down-regulated in agl21-1, except ABI5 (Fig. 5A), indicating that ABI5 might be a target of AGL21.
Moreover, we found high levels of AGL21 transcripts accumulated in WT dry seeds, but the levels gradually dropped after 12-72 hours imbibition and continually decreased to a very low level after 1-3 days germination (Fig. 5B). However, when germinated on medium with ABA, AGL21 expression was significantly induced by ABA during germination. Similar expression pattern of AGL21 in AGL21-overexpressing seeds was also observed, except small increases during 1-3 days germination on MS medium (Fig. 5B), which were probably due to the higher expression of AGL21 in developing roots. The pattern of AGL21 expression is similar to that of ABI5 (Okamoto et al., 2010), and coincides with the content changes of ABA, which is also high in dry seeds but reduces rapidly after imbibition (Ali-Rachedi et al., 2004). We then compared the expression levels of ABI5 in dry seeds, imbibed seeds and germinating seeds of AGL21-overexpressing and WT plants (Fig. 5C). ABI5 showed a very similar expression pattern to AGL21. More importantly, expression of ABI5 was significantly higher in AGL21-overexpressing seeds compared with that in WT seeds in most time-points checked, especially in seeds imbibed for 0-48 h and seeds germination on ABA medium for 24-72 h (Fig. 5C). In addition, one of ABI5 target genes, AtEM6, also had a similar expression pattern to ABI5 and AGL21, and also had increased expression levels in AGL21-overexpressing imbibed seeds, seeds germination on MS medium for 24 h and 72 h and seeds grown on ABA medium for 72 h (Fig. 5D). Furthermore, we also compared the expression profiles of AGL21 and ABI5 in seed using public expression data extracted from the Arabidopsis eFP browser (Winter et al., 2007). As data showed in Figure 5E, AGL21 and ABI5 have similar expression patterns in different seed development stages, with much higher expression levels of both genes at the late stages of seed development. Both of these two genes have the highest expression levels in dry seeds and dramatically decrease after 3-24 h imbibition (Fig. 5F). These data are consistent with our qRT-PCR analyses, implying that AGL21 may directly regulate ABI5 expression in seeds.
Since ABI5 protein contents directly affect ABA-dependent seed germination and post-germination growth, we examined ABI5 protein levels in seeds of WT, AGL21-overexpressing and agl21-1 in the present or absence of exogenous ABA. In seeds germinated on MS medium for 3 days, no clear differences of ABI5 protein level were detected between the seed of different genotypes. However, in the present of ABA, significantly higher ABI5 protein levels accumulated in AGL21-overexpressing seeds, while significantly decreased in agl21-1 seeds compared with WT seeds (Fig. 5G and H). All these results suggest that AGL21 modulates ABI5 at both the transcriptional and protein levels.
ABI5 is one of the target genes of AGL21 in regulating ABA-dependent seed germination
The expression pattern of AGL21 shows high similarity to that of ABI5 in seeds and its function in seed germination also similar to ABI5. To test whether AGL21 acts in the same pathway to ABI5 in seed germination, we generated 35S::AGL21 plants in the genetic background of the abi5-7 mutant (abi5-7/AGL21OX 1-6). As shown in ABA response assays, AGL21OX 1-6 seeds were hypersensitive, however, the abi5-7/ AGL21OX 1-6 plants showed an ABA insensitive phenotype as that of abi5-7 mutant (Fig. 6), suggesting that AGL21 act upstream of ABI5 in ABA signaling pathway to regulate seed germination.
MADS-box TFs can modulate their target genes expression by specifically binding as homo- or heterodimers to the flexible CArG-box (C-[A/T]rich-G) cis-element (Riechmann et al., 1996). Sequence analysis found that the ABI5 promoter contains six putative binding sites (Fig. 7A), which contained a core sequence that meets the patterns for any of the possible three CArG-box motifs, C(A/T)8G, C(C/T)(A/T)6(A/G)G, or C(C/T)(A/T)G(A/T)4(A/G)G (de Folter and Angenent, 2006; Ito et al., 2008; Fujisawa et al., 2011). The presence of CArG-box motifs led us to examine whether AGL21 is targeted to the ABI5 promoter.
We performed chromatin immunoprecipitation (ChIP) assays using 4-day-old 35S::AGL21-HA transgenic plants grown on MS medium or MS medium containing 0.15 μM ABA. DNA fragments bound to epitope-tagged AGL21 proteins were analyzed by qPCR assays. The ChIP-qPCR analysis demonstrated that the P5 DNA fragment was significantly enriched by AGL21 both in plants grown on MS medium or MS medium containing ABA, with much higher enrichment in the plants grown in the present of ABA (Fig. 7B). However, other four genomic fragments containing putative CArG-box motif were not enriched. In addition, none of the five genomic fragments was found enriched by AGL21 in WT plants (Fig. 7B). All these data support the specific interaction of AGL21 with the P5 region of the ABI5 promoter.
To verify that AGL21 directly regulates ABI5 expression, we checked the expression levels of ABI5-targeted genes AtEM1 and AtEM6 in seeds sown on MS media or MS media with 0.2 μM ABA for three days. As shown in Figure 7C, on MS medium, AtEM6 was up-regulated in AGL21-overexpressing seeds, and significantly down-regulated in agl21-1, AGL21OX 1-6/abi5-7, and abi5-7 compared with WT. As for AtEM1, only significantly decreased expression level in abi5-7 seeds was observed on MS medium (Fig. 7D). On MS medium supplemented with ABA, the expression levels of both AtEM1 and AtEM6 markedly increased in AGL21-overexpressing seeds, while significantly reduced in agl21-1 seeds. However, when AGL21 was overexpressed in abi5 background (AGL21OX 1-6/abi5-7), similar expression levels of AtEM1 and AtEM6 were observed to that of abi5-7 seeds (Fig. 7C and D), indicating that AGL21-upregulated AtEM1 and AtEM6 expression is ABI5-dependent. These results confirm that ABI5 is a target gene of AGL21.
DISCUSTION
There are more than 100 MADS-box genes found in Arabidopsis genome (Gramzow and Theissen, 2013), with functions in the morphogenesis of almost all plant organs and throughout the whole life cycle (Smaczniak et al., 2012). However, their functions in seed dormancy and germination are largely unknown. So far, only two MADS-box gene, FLC and AGL67, were reported involved in seed germination (Chiang et al., 2009; Bassel et al., 2011). Recently, gene expression profiling analysis identified three MADS-box genes, including AGL21, were differentially expressed between imbibed dormant and after-ripened ecotype C24 seeds, suggesting putative functions of these genes in seed dormancy and germination (Barrero et al., 2010). In this study, we discovered that AGL21 regulated ABA-mediated seed germination and post-germination growth by modulating ABA signaling.
We found that AGL21 acts as a negative regulator in seed germination. AGL21 is primarily expressed in root, but also has high expression levels in siliques and dry seeds, and it responds to multiple environmental and internal signals both in germinating seeds and roots (Fig. 3) (Yu et al., 2014). Overexpression of AGL21 results in hypersensitivity of seed germination to ABA, high salt, and osmotic stress, while knockout AGL21 confers opposite phenotypes during germination (Fig. 1 and Fig. 2). Seed germination is tightly regulated by ABA:GA balance and ABA signaling in seed (Finch-Savage and Leubner-Metzger, 2006). We found that AGL21 does not affect ABA and GA biosynthesis or catabolism (Fig. S2-S4), indicating that AGL21-regulated seed germination is not through affecting ABA or GA content. Our further analyses of AGL21 expression in abi mutants (Fig. 4) and ABA signal pathway genes expression in agl21-1, AGL21-overexpression, and WT seeds (Fig. 5) indicate that AGL21 is involved in ABA signal pathway to regulate seed germination and may acts downstream of PP2Cs and SnRKs, but upstream of ABI5 and other ABA-responsive genes.
As a key player in ABA-triggered arrest of germination and post-germination growth, ABI5 is regulated at both transcriptional and post-transcriptional levels. Several genes, such as WRKY2, RAV1, MYB7, SALT- AND DROUGHT-INDUCED RING FINGER1 (SDIR1), SDIR1-INTERACTING PROTEIN1 (SDIRIP1), HY5, B-BOX21 (BBX21), DELAY OF GERMINATION 1 (DOG1), NUCLEAR FACTOR-Y C- RGA-LIKE 2 (NF-YC–RGL2), were reported to regulate seed germination by modulating ABI5 expression directly or indirectly (Zhang et al., 2007; Chen et al., 2008; Jiang and Yu, 2009; Feng et al., 2014; Xu et al., 2014; Kim et al., 2015; Zhang et al., 2015; Dekkers et al., 2016; Liu et al., 2016). Our current study implicates that AGL21 can directly regulate ABI5 expression during seed germination. Firstly, we found AGL21 and ABI5 had similar expression patterns during seed development and germination. ABI5 is expressed throughout seed development, reaching the highest transcript level at mature seed stage, but dropping during imbibition and germination unless exposed to stresses (Fig. 5C) (Brocard et al., 2002). Here, we found that the expression pattern of AGL21 in seeds is in accord with that of ABI5 (Fig. 5B- 5F) and ABA content changes in seeds during the same period. These results imply a possible regulation between these two genes. Further qRT-PCR analyses found that among 10 seed germination related genes, only ABI5 had significantly increased expression level in AGL21-overexpressing germinating seeds, with significantly reduced expression in agl21-1 seeds (Fig. 5A). Then we found ABI5 and its target gene AtEM6 had markedly elevated expression in AGL21-overexpressing dry seeds, imbibed seeds, germinating seeds grown on MS or ABA media (Fig. 5C and D). Moreover, more ABI5 protein was accumulated in AGL21-ovexpression germinating seeds, while down-regulated in agl21-1 seeds on ABA media at the same time (Fig. 5G and H). In addition, genetics evidence found that AGL21-regulated seed germination is depend on ABI5. When overexpressing AGL21 in abi5 background (abi5-7/AGL21OX1-6), the hypersensitive seed germination to ABA was abolished, instead abi5-7/AGL21OX1-6 seeds showed similar sensitivity to abi5 (Fig. 6). Taken together, these data suggest that AGL21 acts upstream of ABI5, and may directly modulate ABI5 transcription.
To verify whether ABI5 is the direct target of AGL21, ChIP-qPCR assay was carried out. We found one promoter DNA fragment containing a putative CArG-box cis-element of ABI5 was significantly enriched in 4-day-old plants grown on MS or MS medium containing ABA (Fig. 7A and B), indicating that AGL21 directly binding to ABI5 promoter to regulate its expression. AtEM1 and AtEM6, two target gene of ABI5 (Carles et al., 2002), were markedly up-regulated in AGL21-overexpressing plants and down-regulated in agl21-1 plants as that of ABI5. However, once ABI5 was mutated in AGL21-overexpressing background, the induction of AtEM1 and AtEM6 expression was blocked (Fig. 7C and D). Together, these data support the idea that AGL21 regulates seed germination by directly modulating ABI5. Although we also found other seed germination regulator genes, such as ABI3, ABI4 and ABF4, were up-regulated in AGL21-overexpressing plants (Fig. 5H and Fig. 5A), the relationships of these genes to AGL21 await further investigation.
Timing of germination is a complex biological process that is regulated through intricate signaling pathways integrating diverse environmental signals, such as light, temperature, water, soil salinity and nutrition, into internal developmental programs, such as endogenous hormone signaling (Finkelstein et al., 2008; Jiang et al., 2016). For example, both osmotic and salinity stresses inhibit seed germination by affecting ABA signal pathway (Llanes et al., 2016). NO3− can act as a seed germination enhancer by decreasing the level of ABA in the seed (Finkelstein et al., 2008; Osuna et al., 2015; Yan et al., 2016). Sulfate availability affects germination response to ABA and salt stress in Arabidopsis by regulating ABA biosynthesis (Cao et al., 2014). GA, ethylene, brassinosteroids and cytokinin have been shown to promote seed germination, while ABA, auxin, JA, SA inhibit seed germination (Kucera et al., 2005; Finkelstein et al., 2008; Wang et al., 2011; Liu et al., 2013; Shu et al., 2016). All these different phytohormones modulate seed germination most likely by regulating the ABA/GA balance at either the signaling or biogenesis levels, with ABI5 as one of the pivots involved in hormone crosstalk (Shu et al., 2016). Thus, the molecular links that incorporate external signals into internal plant hormonal signaling are crucial for seeds to germinate at proper time in the changing environment. We have demonstrated in this paper that AGL21 is a negative regulator of seed germination. AGL21 responds to a variety of internal and external signals, such as ABA, MeJA, IAA, osmotic stress, salt stress, N and S deficiency (Fig. 3A), which are known to affect seed germination. Therefore, we propose that AGL21 may serve as environmental surveillance for seed germination. It controls seed germination by regulating ABI5 according to its environmental surveillance, which prevents seed germination under adverse conditions, an adaptive mechanism for plant survival.
METHODS
Plant Material and Growth Conditions
Arabidopsis Columbia-0 (Col-0) ecotype was used in this study. AGL21-overexpressing lines, such as AGL21OX 1-6 and AGL21OX 3-5, and mutants, such asagl21-1 and agl21-2 were reported previously (Yu et al., 2014). 35S::AGL21-HA transgenic line AGL21OX 29-3 was obtained from Agrobacterium tumefaciens-mediated transformation with 35S:AGL21-HA construct. To get 35S::AGL21-HA construct, the coding region of AGL21 was amplified and cloned into pDONR207 with the primers AGL21-HA LP and AGL21-HA RP, and then shuttled it into pCB2004 vector (Lei et al, 2007). The AGL21OX 1-6/abi5-7 double mutant was generated by genetic cross of AGL21OX 1-6 and abi5-7 mutant. All plants were grown at 220 under long-day condition (16-h light/8-h dark cycles).
Seed germination assays
Seeds were collected at the same time were used for germination assays. Harvested seeds were aired dried at room temperature at least 3 weeks before the germination assays. For seed germination assays, seeds of each genotype were surface sterilized with 10% bleach for 12 min and washed five times with sterile water, and then were stratified at 40 for 2 days in darkness before sown on MS medium (1% sucrose, 0.5 % agar, pH 5.8) or MS medium supplemented with ABA, NaCl or mannitol. Seeds were germinated at 220 under 16-h light/8-h dark cycles. Germination (emergence of radicles) and post-germination growth (green cotyledon appearance) were scored at the indicated time points.
qRT-PCR assay
Total RNA of seedlings was extracted with Trizol reagent (Invitrogen) and RNAs from dry seeds, imbibed seeds and germinating seeds were isolated using a TRIzol-based two-step method as previously described (Meng and Feldman, 2010). Total RNA samples were pretreated with DNase I (RNase Free) and 1.5 ug of total RNA was used for reverse transcription with oligo (dT)18 to synthesize first-strand cDNA. qRT-PCR was performed with a StepOne Plus Real Time PCR System by using a TaKaRa SYBR Premix Ex Taq II reagent kit as described previously (Yu et al, 2013). All primers used are listed in Supplemental Table S1.
Quantification of ABA
Dry seeds were ground in liquid nitrogen and ABA contents were measured by the ABA immunoassay kit as described (Yang et al, 2001).
ChIP-qPCR assay
The ChIP assay was performed as reported previously (Cai et al., 2014). 35S::AGL21-HA transgenic plants (OX 29-3), anti-HA antibodies (Abmart), and salmon sperm DNA/protein A agarose beads (Millipore, USA) were used for ChIP experiment. DNA was purified using phenol/ chloroform (1:1, v/v) and precipitated. The enrichments of DNA fragments was quantified by qPCR using specific primers (Supplemental Table S1). Enriched values were normalized with the level of input DNA.
Western blot
For western blot analysis, germinating seeds were powdered with liquid nitrogen and proteins were extracted with RIPA buffer [50 mmol/L Tris-HCl, pH 8.0, 0.1% Nonidet P-40,150 mmol/L NaCl, 1% sodium dodecyl sulfate, 0.5% sodium deoxycholate, and protease inhibitor cocktail tablets (Roche)]. Protein contents were determined using the Bradford method. Proteins were separated on 12% SDS-PAGE and electrotransferred to nitrocellulose membrane (Immobilon-P, MILLIPORE Corporation, USA). Anti-ABI5 antibody (Abiocode) at 1:1000 dilution was used for protein immunoblotting as previously described (Liu and Stone, 2010). The results were detected using a CCD camera system (Image Quant LAS 4000) using Super Signal West Femto Trial Kit (Thermo, USA).
Statistical analysis
Statistically significant differences were computed based on the Student’s t-tests.
SUPPLEMENTARY DATA
Figure S1. Expression level analyses of AGL21-overexpressing and knockout mutants.
Figure S2. Expression levels of ABA biosynthetic and catabolic pathway genes.
Figure S3. ABA contents in dry seeds of WT, AGL21-overexpressing and knockout mutant plants.
Figure S4. Expression levels of GA biosynthetic pathway genes.
Table S1. Primers used for PCR.
FUNDING
This work was supported by the China National Natural Science Funds for Distinguished Young Scholar (grant no. 31500231), China Postdoctoral Science Foundation, No.9 Special Fund (grant no.2016T90577).
AUTHOR CONTRIBUTIONS
L.-H.Y. and C.-B.X. designed the experiments; L.-H.Y. performed experiments and data analysis, and wrote the manuscript; J.W., Z.-Q.M., P.-X.Z., and Z.W. contributed to assist in performing part of the experiments; C.-B.X. supervised the project and revised the manuscript.
ACKNOWLEDGEMENTS
We thank Dr. Chuan-You Li (Institute of Genetics and Developmental Biology, Chinese Academy of Sciences) for kindly providing abi5-7 mutant. We also thank the ABRC for providing T-DNA insertion lines used in this study.