The phytohormone ABA and the transcription factor ABSCISIC ACID INSENSITIVE 3 (ABI3)/VIVIPAROUS 1 (VP1) function in protecting embryos during the desiccation stage of seed development. In a similar signaling pathway, vegetative tissue of the moss Physcomitrella patens survives desiccation by activating downstream genes (e.g. LEA1) in response to ABA and ABI3.
We show that the PpLEA1 promoter responds to PpABI3 primarily through the ACTT-core element (5′-TCCACTTGTC-3′), while the ACGT-core ABA-responsive element (ABRE) appears to respond to ABA alone. We also found by yeast-two-hybrid screening that PpABI3A interacts with PpNF-YC1, a subunit of CCAAT box binding factor (CBF)/nuclear factor Y (NF-Y). PpNF-YC1 increased the activation of the PpLEA1 promoter when incubated with PpABI3A, as did NF-YB, NF-YC, and ABI3 from Arabidopsis.
This new response element (ACTT) is responsible for activating the ABI3-dependent ABA response pathway cooperatively with the nuclear factor Y (NF-Y) complex. These results further define the regulatory interactions at the transcriptional level for the expression of this network of genes required for drought/desiccation tolerance.
This gene regulatory set is in large part conserved between vegetative tissue of bryophytes and seeds of angiosperms and will shed light on the evolution of this pathway in the green plant lineage.
The phytohormone ABA is required to protect embryos during the desiccation stage of seed development by activating downstream genes in concert with the transcription factor ABSCISIC ACID INSENSITIVE 3 (ABI3)/VIVIPAROUS 1 (VP1; Finkelstein et al., 2002). Seeds of abi3/vp1 mutants are characterized by reduced amounts of storage proteins, sensitivity to desiccation, and precocious germination (Ooms et al., 1993; Nambara et al., 1994; Parcy et al., 1994, 1997).
The ABI3/VP1 protein has three functional domains, B1, B2 and B3, which are highly conserved among various plant species (Hill et al., 1996; Suzuki et al., 1997), including Physcomitrella patens ABI3 (PpABI3; Marella et al., 2006). The B1 and B2 domains are involved in a complex with the bZIP transcription factor ABI5 (Nakamura et al., 2001) and with the ABA-responsive element (ABRE; Hill et al., 1996), respectively. Additionally, the B2 domain also regulates nuclear localization (Marella & Quatrano, 2007), while the B3 domain binds DNA, specifically the RY/Sph (RY) promoter element (CATGCA; Suzuki et al., 1997; Ezcurra et al., 2000). These three domains of ABI3/VP1 along with specific cis-elements and transcription factors such as bZIPs, regulate the expression of seed-specific genes, for example, the wheat Em gene (Vasil et al., 1995) and the maize C1 gene (Hattori et al., 1992).
In P. patens also, ABA treatment regulates a large set of genes as demonstrated by microarray analysis (Machuka et al., 1999; Cuming et al., 2007; Richardt et al., 2010). Indeed, the ABA-signaling pathway was reported to be present in P. patens, based on the activation of the wheat Em promoter and the PpLEA1 (the P. patens homolog of Em) promoter by exogenous ABA treatment (Knight et al., 1995; Kamisugi & Cuming, 2005) and its activation is performed through an ACGT-core ABRE (Kamisugi & Cuming, 2005; Sakata et al., 2010). PpABI3A was shown to activate synergistically the PpLEA1 promoter in concert with ABA (Marella et al., 2006; Khandelwal et al., 2010) and to partially complement the phenotypes of the Arabidopsis abi3-6 mutant such as ABA insensitivity of germination, the color of seeds, and the expression of seed maturation genes (MAT), but it failed to restore the expression of the LEA genes through the ACGT-core ABRE (Marella et al., 2006). On the other hand, a previous report has shown that the napA (the Brassica napus homolog of MAT genes) promoter is activated not only through an ACGT-core ABRE but also an ACTT-core ABRE (5′-GCCACTTGTC-3′), called distB for the transactivation by ABI3 both with and without ABA (Ezcurra et al., 2000).
LEAFY COTYLEDON1 (LEC1) was also found to play central roles in the regulation of embryo maturation in seeds (Meinke et al., 1994; West et al., 1994), similar to ABI3. Ectopically overexpressing LEC1 can induce embryo development in vegetative cells (Lotan et al., 1998). LEC1 encodes a homolog of yeast HAP3 subunit of the CCAAT box binding factor (CBF)/nuclear factor Y (NF-Y), which is an evolutionarily conserved transcription factor that contains three subunits, HAP2 (CBF-B/NF-YA), HAP3 (CBF-A/NF-YB), and HAP5 (CBF-C/NF-YC; Li et al., 1992). The complex forms through an initial interaction between HAP3 and HAP5, which then allows the formation of a heterotrimer with HAP2. The heterotrimer then binds to a CCAAT element with very high specificity and affinity (Romier et al., 2003).
Each subunit of the NF-Y complex is encoded by gene families in angiosperms (Stephenson et al., 2007; Siefers et al., 2009). For example, the Arabidopsis genome encodes 10 NF-YA, 13 NF-YB, and 13 NF-YC paralogs (Siefers et al., 2009). Several other NF-Y subunits have been implicated in specific developmental processes in angiosperms, including nodule formation (Combier et al., 2006), flowering-time regulation (Ben-Naim et al., 2006; Wenkel et al., 2006; Cai et al., 2007; Kumimoto et al., 2010; Li et al., 2011), ABA/blue light response (Warpeha et al., 2007; Yamamoto et al., 2009), stress response (Chen et al., 2007; Nelson et al., 2007; Li et al., 2008; Liu & Howell, 2010), and chloroplast development (Kusnetsov et al., 1999; Miyoshi et al., 2003). These observations indicate that a variety of molecular functions are attributed to the NF-Y subunit family members in angiosperms, but no function has been assigned to those in P. patens.
In this report, we identified a new response element (ACTT) responsible for activating the ABI3-dependent ABA response pathway cooperatively with the nuclear factor Y (NF-Y) complex. Coexpression of PpABI3A along with PpNF-YC1 and PpNF-YB4 synergistically activated the ACTT-core element of the PpLEA1 promoter, but not the ACGT-core ABRE. The ability of NF-Y proteins to functionally enhance and support the expression of the ABI3-dependent ABA response pathway provides a framework for understanding the mechanism of how ABI3 interacts with transcriptional complexes to activate gene sets. These results further define the regulatory interactions at the transcriptional level for the expression of gene sets required for drought/desiccation tolerance and will shed light on the evolution of this pathway in the green plant lineage.
Materials and Methods
Plant materials and growth conditions
Physcomitrella patens ssp. patens (Gransden) was used as the wildtype strain. The triple deletion mutant of the PpABI3 (Δabi3) was described previously (Khandelwal et al., 2010). Protonemal tissues were grown on PpNH4 medium (Schaefer & Zryd, 1997) at 25°C under continuous light as described previously (Marella et al., 2006).
Nucleic acid extractions and cDNA synthesis
Genomic DNA and total RNA were extracted from 1-wk-old protonemal tissues using the Nucleon PhytoPure system (GE Healthcare UK Ltd, Little Chalfont, Buckinghamshire, UK) and the RNeasy Plant Mini Kit (Qiagen), respectively, according to the manufacturer's manual. For cDNA synthesis, 1 μg of total RNA was first treated with DNaseI (Sigma-Aldrich) for 30 min at room temperature, and the enzyme was inactivated by heating at 70°C for 10 min. Reverse transcription was performed with the ThermoScript RT-PCR system (Invitrogen) according to the manufacturer's manual. Synthesized cDNAs were purified using the QIAquick Gel Extraction kit (Qiagen) according to the manufacturer's manual.
The PpLEA1-GUS construct (Kamisugi & Cuming, 2005) harboring the 1452 bp upstream region of the PpLEA1 gene (GenBank accession no. AY870926) fused to the β-glucuronidase (GUS) gene was used as a backbone construct to generate promoter variants used in this study.
To create 5′-deletion constructs of the PpLEA1 promoter, the fragments spanning from −790 to +159, −540 to +159, −230 to +159, −190 to +159, and −180 to +159 (relative to the transcription start site) were PCR-amplified using primer sets (Supporting Information, Table S1; nos. 1–6), which created PstI and BamHI sites at its 5′- and 3′ ends, respectively. The amplified fragments were first cloned into the Zero Blunt TOPO vector (Invitrogen) and then fused with the GUS sequence by insertion into pBI221 (Clontech, Mountain View, CA, USA) at the PstI and BamHI sites, replacing the 35S promoter with the PpLEA1 sequences. Resulting constructs were designated as Δ-790-GUS, Δ-540-GUS, Δ-230-GUS, Δ-190-GUS, and Δ-180-GUS.
To create point mutations in the PpLEA1 promoter, the GeneTailor™ Site-Directed Mutagenesis System (Invitrogen) was used according to the manufacturer's manual. Two DNA fragments overlapping at the mutation sites were PCR-amplified using a pair of the mutagenic and end primers (Table S1; nos. 7–22). This procedure created mABRE-GUS, m190-180-GUS, mBIHD-GUS, mDOF-GUS, mEBOX-GUS, mCACT-GUS, mdistB1-GUS, mdistB2-GUS, and mACTT-GUS. All of the promoters were fused to a GUS reporter gene with a 35S-terminator or Nos-terminator sequence.
In addition to previously described effector constructs, Act::ABI3, Act::PpABI3A, and Ubi-LUC (Sakata et al., 2010), we generated new effector constructs used in this study: Act::PpNF-YB1, Act::PpNF-YB2, Act::PpNF-YB3, Act::PpNF-YB4, Act::PpNF-YB5, Act::PpNF-YB6, Act::PpNF-YC1, and Act::PpNF-YA2. Coding sequences of PpNF-YB1 (Phypa_27666), PpNF-YB2 (Phypa_96394), PpNF-YB3 (Phypa_53068), PpNF-YB4 (Phypa_118490), PpNF-YB5 (Phypa_36907), PpNF-YB6 (Phypa_53069), PpNF-YC1 (Phypa_124620), and PpNF-YA2 (Pp1s31_299V6) were PCR-amplified from cDNA or genomic DNA using primer sets (Table S1; nos. 23–38). To create the Arabidopsis thaliana effector construct the Act::LEC1 (AtNF-YB9/At1 g21970), we PCR-amplified the fragment from the cDNA clone of LEC1, provided by Professor John J. Harada at the University of California, Davis, using the primer set (Table S1; nos. 39, 40). To create another effector construct, the Act::AtNF-YC1 (At3 g48590), we digested the cDNA clone (pda07541) of AtNF-YC1, provided by the RIKEN Bio Resource Center (BRC, Tsukuba, Ibaraki, Japan), with SfiI. All of the DNA fragments were first cloned into the Zero Blunt TOPO vector (Invitrogen) or by digesting a cDNA clone with SfiI, and then subcloning downstream of the rice Actin 1 promoter (McElroy et al., 1990) with a 35S-terminator sequence using NotI or SfiI restriction sites.
The cDNAs were prepared from P. patens protonemal tissue treated with 10 μM ABA for 24 h and cloned into yeast pGADT7 vector (Clontech) using the BD Matchmaker™ Library Construction & Screening Kit (Clontech), according to the manufacturer's manual. PpABI3A cDNA PCR-amplified using the primer set (Table S1; nos. 41, 42) was cloned into pGBKT7 (Clontech) using the EcoRI and BamHI sites.
PpNF-YC1 was PCR-amplified using the primer set (Table S1; nos. 43, 44), and then inserted into pGAD424 (Clontech) or pGBTK (Yoshimura et al., 2004), respectively, at the EcoRI and BamHI sites. This procedure created pGAD424-PpNF-YC1 and pGBTK-PpNF-YC1. To create the pGAD424-PpABI3A construct, the pGBKT7-PpABI3A construct was digested by EcoRI and BamHI, and then inserted into pGAD424 at the EcoRI and BamHI sites.
Yeast-two-hybrid screen was performed using the BD Matchmaker Library Construction & Screening Kit (Clontech) according to the manufacturer's manual. Briefly, the bait construct and the cDNA library were cotransformed into yeast strain AH109 and plated on synthetic medium lacking leucine, tryptophan, and histidine (SD–LWH). Positive colonies observed were then streaked onto SD medium lacking leucine, tryptophan, histidine, and adenine (SD–LWHA) containing 25 mM 3-amino-1,2, 4-triazole (3-AT) to repress autoactivation.
To confirm the interaction between PpABI3A and PpNF-YC1, both bait construct (pGBTK-PpNF-YC1) and prey constructs (pGAD424-PpNF-YC1 and pGAD424-PpABI3A) were individually transformed into yeast strains PJ69-4A-a and PJ69-4A-α, respectively. After mating, interactions between fusion proteins in yeast cells were analyzed on the SD–LWHA containing 70 mM 3-AT after 2 wk. β-Galactosidase assay was performed using O-nitrophenyl-β-d-galactopyranoside (ONPG; Sigma) to measure the strength of interactions according to the Yeast Protocols Handbook (Clontech). Clones harboring both pGBTK and pGAD424 were cultured in SD–LW liquid medium to an OD600 of c. 1.0. Aliquots of 500 μl of culture supernatant were mixed with 700 μl of Z buffer containing 2.7 ml l−1 β-mercaptoethanol and 160 μl of Z buffer containing 4 mg ml−1 ONPG. The reaction mixes were incubated at 30°C for 1 h before stopping the reaction with the addition of 400 μl of 1 M Na2CO3. The β-galactosidase activity was calculated as the ratio of OD420 : OD600.
Physcomitrella patens transformation and analysis
Transient DNA delivery to protonemal tissues of P. patens was performed as described previously (Marella et al., 2006). In this study, we used 0.8 μg of each reporter construct (PpLEA1-GUS/the mutant variants), effector construct, and Ubi-LUC to prepare gold particles for four shots. One-wk-old protonemal tissues were used for bombardment and then incubated on PpNH4 agar medium with or without 10 μM ABA for 24 h as described in the plant material section. GUS and luciferase (LUC) activities were measured as described previously (Marella et al., 2006). GUS activity was normalized by the LUC activity and represented as relative GUS activity ± SE. All experiments consisted of four replicates.
The ACGT core (-91) is not essential for ABI3-mediated activation of PpLEA1
The PpLEA1 promoter requires both ABI3 and ABA for its activation (Marella et al., 2006; Khandelwal et al., 2010) and at least the ACGT core of the ABRE was reported to be an essential component for the ABA induction (Kamisugi & Cuming, 2005). However, the detailed ABI3-mediated interaction with the PpLEA1 promoter has not been identified. To further investigate the activation mechanism by which ABA and ABI3 interact in the regulation of the PpLEA1 promoter, we performed a transient promoter analysis of the PpLEA1 promoter using the triple deletion mutant of PpABI3 genes (Δabi3; Khandelwal et al., 2010). When the PpLEA1-GUS construct (Fig. 1a) was introduced by particle bombardment into Δabi3, no transient expression was observed in the presence of ABA unless ABI3 was also introduced as an effector (Fig. 1b). When the mABRE construct (Fig. 1a) was used, activation of gene expression with PpABI3A and ABA was reduced but not eliminated, indicating that the ABRE is not the only element for ABI3-mediated activation (Fig. 1b).
The ACTT-core (-182) sequence is essential for ABI3-mediated activation of the PpLEA1 promoter
To determine the promoter elements required for the ABI3-mediated activation, we deleted sequences in the promoter sequentially from the 5′ end (Fig. 2a). ABA and PpABI3A still significantly activated the Δ-190 construct in Δabi3, but deletion of an additional 10 bp (Δ-180) completely abolished the activation (Fig. 2b). This 10 bp region contained the sequence 5′-TCCACTTGTC-3′, which is closely related to the distB ABRE (5′-GCCACTTGTC-3′) in the napA gene promoter of B. napus (Ezcurra et al., 1999, 2000; see Table 1). In fact, insertion of the NotI recognition sequence (5′-gcggccgc-3′) at different locations within the 10 bp region (Fig. S1a) showed that the 4 bp ACTT sequence was critical for the ABI3-driven activation (Fig. S1b). Fig. 3(a) shows that in the presence of ABA, two types of only a 3 or 4 bp substitution in the putative ACTT core in the distB (mdistB) element resulted in a drastic reduction of the PpLEA1 promoter activity in Δabi3 (Fig. 3b), suggesting that the ACTT sequence is essential for ABI3-mediated activation of the PpLEA1 promoter in P. patens. We tested this by mutating the ACTT core of the distB element to the ACGT core, which converted the distB to the canonical ABRE (Fig. S2a). This conversion resulted in activation of the PpLEA1 promoter by ABA in the absence of ABI3 (Fig. S2b). The powerful wheat Em promoter, which contains two tandem ABREs, was also ABA-induced in the Δabi3 mutant. These data indicate that in P. patens, the ABI3-mediated ABA-responsive gene expression is regulated through the ACTT core, while the ACGT core functions for ABI3-independent ABA induction.
Table 1. Alignment of ABA-responsive elements (ABREs) in ABA-responsive genes
Name or (position)
Pp, Physcomitrella patens; Bn, Brassica napus; Ta, Triticum aestivum; Os, Oryza sativa; At, Arabidopsis thaliana; Zm, Zea mays; Hv, Hordeum vulgare. The ACTT core and the ACGT core are underlined.
PpABI3A interacts with PpNF-YC1 in the yeast-two-hybrid system
To identify proteins that function in the transcriptional complex with ABI3, we employed the Gal4 based yeast-two-hybrid screening assay. A P. patens cDNA library cloned in a yeast pGADT7 vector was used as prey, while PpABI3A cloned in pGBKT7 was used as bait. Six positive interacting clones were identified but one clearly showed similarity to a putative regulatory protein, that is the NF-YC gene (PpNF-YC1), which is a subunit of CBF/NF-Y. Reproducibility of the interaction was confirmed by a growth assay to see the expression of the markers HIS and ADE as well as by β-galactosidase assay to show quantitatively the expression of the marker LacZ (Fig. 4a,b).
Coexpression of PpABI3A and PpNF-YC1 strongly activates the PpLEA1 promoter
We evaluated the possible functional interaction between PpABI3A and PpNF-YC1 in P. patens by a transient assay using Δabi3. The effector construct Act::PpNF-YC1 alone had no effect on the PpLEA1 promoter, even in the presence of ABA, but when introduced with Act::PpABI3A, a greater than twofold activation compared with Act::PpABI3A alone was observed (Fig. 5). This strongly suggests that the interaction of PpNF-YC1 increases the effectiveness of PpABI3A in driving the expression of the PpLEA1 promoter.
Expression analysis of NF-Y in P. patens
Since PpNF-YC1 has a promoting effect on the expression of ABI3-dependent activation of the PpLEA1 promoter, we investigated the expression levels of the PpNF-Y subunits using RNA blot analysis (Fig. S3). Three clear results were apparent: PpNF-YC1 mRNA showed the strongest accumulation level compared with other PpNF-YCs; mRNA expression levels of all of the subunits appeared not to be under the regulation of PpABI3; and PpNF-YC5 mRNA was the only one showing a requirement for ABA (Fig. S3). These data indicate that the mRNA levels of the subunits do not appear to be responsible for the effects we have observed in the ABI3-mediated ABA response.
The PpLEA1 promoter is activated by both NF-YC and NF-YB subunits in an ABA-dependent manner
To evaluate the role of other NF-Y subunits in the ABA-dependent activation of the PpLEA1 promoter, we cloned genes encoding NF-YAs (PpNF-YAs) and NF-YBs (PpNF-YBs) from P. patens (Xie et al., 2008) and performed a transient assay in Δabi3. Neither the Act::PpNF-YC1 nor each of the Act::PpNF-YB effector constructs alone had any effect on the PpLEA1 promoter (Fig. 5, and I. Yotsui, Y. Sakata, unpublished data); however, the Act::PpNF-YC1, in combination with each of the six Act::PpNF-YB constructs, significantly activated the PpLEA1 promoter in the presence of ABA in P. patens (Fig. 6). This result suggests that the regulatory combination of NF-YB and NF-YC mediates ABA-dependent gene expression in P. patens. By contrast, coexpression of PpNF-YA with PpNF-YB and PpNF-YC had no effect on activation of the PpLEA1 promoter; rather we observed reproducible reduction of the promoter activity (Fig. S4). Although endogenous PpNF-YA is supplied in the target protonemal tissues of Δabi3 (Fig. S3), its involvement in the activation of the PpLEA1 promoter is yet to be elucidated.
Activation of the PpLEA1 promoter by PpNF-YC1 and PpNF-YB4 depends on the ACTT ABRE
When the combination of PpNF-YC1 and PpNF-YB4 was introduced in Δabi3, they activated the PpLEA1 promoter in the presence of ABA (Figs 6, 7b). To evaluate whether activation of the PpLEA1 promoter by PpNF-YC1 and PpNF-YB4 required the ACTT-core sequence, we introduced mdistB1 (Fig. 7a) to the transient expression assay in Δabi3 (Fig. 7b) and WT (Fig. 7c). The combination of Act::PpNF-YC1 and Act::PpNF-YB4 could not activate mdistB1 in either Δabi3 or WT (Fig. 7b,c). These data indicate that the ACTT core is also essential for synergistic activation of the PpLEA1 promoter by PpNF-YC1 and PpNF-YB4.
PpNF-YC1 and PpNF-YB4 act synergistically with PpABI3A to activate the PpLEA1 in the presence or absence of ABA
When PpABI3A was coexpressed together with PpNF-YC1 and PpNF-YB4, the PpLEA1 promoter was synergistically activated even in the absence of ABA (Fig. 8a). This degree of activation was greater than that by the combination of PpNF-YC1 and PpNF-YB4 alone in the presence of ABA. We also tested whether Arabidopsis NF-YC1 (AtNF-YC1) and LEC1 could function similarly in P. patens. AtNF-YC1 alone had no effect on the PpLEA1 promoter, but when LEC1 was coexpressed with AtNF-YC1, they activated the promoter only in the presence of ABA (Fig. 8b). When AtABI3 was coexpressed together with AtNF-YC1 and LEC1, the PpLEA1 promoter was synergistically activated regardless of ABA application (Fig. 8b). This degree of activation was much greater than those of AtABI3 alone or with coexpression of AtNF-YC1-LEC1 without ABA.
Cis-element analysis in the PpLEA1 promoter
Most of the ABA-responsive genes expressed in angiosperms are regulated through an ACGT-core ABRE (Guiltinan et al., 1990; Skriver et al., 1991; Pla et al., 1993; Hattori et al., 1995; Shen & Ho, 1995; de Bruxelles et al., 1996; Ono et al., 1996; Ishige et al., 1999; Carles et al., 2002). In P. patens, it has been shown that the ACGT core in the PpLEA1 promoter is required for activation by ABA (Kamisugi & Cuming, 2005). However, this study has revealed that the ACGT is not the sole element of the full activation in the PpLEA1 promoter. A new ACTT-core element appears to be essential for the ABI3-mediated ABA-responsive expression. The sequence containing the ACTT core (5′-TCCACTTGTC-3′) is closely related to the distB ABRE (5′-GCCACTTGTC-3′) in the napA gene promoter of B. napus (Ezcurra et al., 1999, 2000; Table 1). This distB ABRE appears to regulate the transactivation by ABI3 both with and without ABA (Ezcurra et al., 2000). We further delineated the relationships between the ACGT- and ACTT-core elements by mutating the ACTT core of the distB element to the ACGT core (Fig. S2a). This conversion clearly showed that the ACTT core alone could not support ABA activation without ABI3, but the ACGT core was responsive to ABA alone (Fig. S2b). The Em promoter, which has the canonical ABRE, also showed ABI3-independent activation by ABA (Fig. S2b). These data appear to indicate that, in P. patens, the ABI3-mediated ABA-responsive gene expression is at least partly regulated through the ACTT core, while the ACGT-core ABRE confers ABI3-independent ABA induction. Recently, Mönke et al. (2012) reported a set of 98 ABI3 target genes (ABI3 regulon). They also searched the G-box-like (GBL) motifs in the promoter of the members of the ABI3 regulon. Most of the GBL motifs were ACGT core; however, we found that promoters of seven genes of the ABI3-target genes (At3 g21370, At1 g73190, At1 g17810, At4 g27140, At4 g27160, At4 g27150, and At4 g36700) contained the ACTT core as GBL motifs. These observations suggest that the ACTT element plays a role in ABI3-dependent gene activation not only in B. napus but also in Arabidopsis.
Functional interaction between PpABI3 and PpNF-Y
In this report, we demonstrate that PpABI3A cooperatively interacts with PpNF-YC1 (Fig. 4), which is a subunit of CBF/NF-Y. In contrast to mammals and fungi in which a single gene encodes each subunit of the trimeric NF-Y, those of land plants are encoded by multigene families. For example, the Arabidopsis genome encodes 10 NF-YA, 13 NF-YB, and 13 NF-YC paralogs, and their expression patterns are tissue- and developmental stage-specific (Siefers et al., 2009). This suggests that there is great variety in molecular functions among the NF-Y subunit family members. Previous reports showed that some plant NF-Y subunits function in specific developmental processes or in response to environmental stimuli (Miyoshi et al., 2003; Ben-Naim et al., 2006; Combier et al., 2006; Wenkel et al., 2006; Nelson et al., 2007; Warpeha et al., 2007; Li et al., 2011; Mu et al., 2013). Interactions between the NF-YB/C complex and DNA-binding proteins have been reported. For example, a rice MADS-box protein, OsMADS18, either with or without a heterodimerizing MADS-box partner, interacts with OsNF-YB1. OsNF-YB1 is capable of heterodimerizing with OsNF-YC but not OsNF-YA (Masiero et al., 2002). Arabidopsis CONSTANS (CO) interacts with a complex of NF-YB and NF-YC through their plant-specific CCT domain (Wenkel et al., 2006). Yamamoto et al. (2009) reported that Arabidopsis LEC1 and L1L, which encode NF-YBs and control embryo maturation, activate the CRC promoter with NF-YC in an ABA-dependent manner through physical interaction with a seed-specific bZIP67 that mediates binding to the ABRE (Yamamoto et al., 2009). These various observations indicate that NF-Y proteins in plants established a broader range of functions by making different NF-YA/B/C combinations among the NF-Y subunit family members or by incorporating another transcription factor(s), which contrasts with the paradigms in fungi and mammals. Indeed, our data showed that there is no CCAAT box-related sequence in the PpLEA1 Δ-190 promoter that was shown to be a minimal promoter for ABA/ABI3-mediated activation (Fig. 2). It appears that PpABI3 and the NF-Y complex share the ACTT-core element to up-regulate the target gene. Moreover, coexpression of PpNF-YA with PpNF-YB and PpNF-YC had no effect on PpLEA1 activation (Fig. S3). We speculate that NF-YB/C may not act as a conventional trimer with NF-YA in this system and that PpNF-YB and PpNF-YC interact with an as-yet-unknown DNA-binding protein to the ACTT-core element.
We demonstrated that PpNF-YC1 and PpNF-YB4 activated the PpLEA1 promoter via ACTT ABRE in Δabi3 (Fig. 7b), suggesting that NF-YB/C can form the transcriptional complex on the ACTT core without PpABI3. Moreover, PpABI3A strongly activated the PpLEA1 promoter when it was coexpressed with PpNF-YC1 and PpNF-YB4 (Fig. 8a). These observations may suggest that a transcriptional complex including PpNF-YB/C acts as a scaffold for PpABI3, which mediates full activation of the PpLEA1 promoter via the ACTT core. Of interest, NF-YB, NF-YC, and ABI3 derived from Arabidopsis also synergistically activated the PpLEA1 promoter, implicating the evolutionarily conserved transcriptional complex for activation of ABRE-mediated regulation of ABA-responsive genes between vegetative tissue of bryophytes and seeds of angiosperms, which share the feature of desiccation tolerance.
The authors thank RIKEN Bio Resource Center (BRC) and Dr John J. Harada (University of California, Davis, USA) for providing the cDNA clones, and Dr Andrew Cuming (University Leeds, UK) for providing the PpLEA1-GUS construct. We also thank Dr Pierre-François Perroud (Washington University, St Louis, MO, USA) for his help in editing the final manuscript. This work was funded by Strategic Young Researcher Overseas Visits Program for Accelerating Brain Circulation of the Japan Society for the Promotion of Science (JSPS) (no. S2306), and the Advanced Research Project of Tokyo University of Agriculture (to Y.S.).