BLH1 and KNAT3 modulate ABA responses during germination and early seedling development in Arabidopsis

Authors


For correspondence (e-mail ihwang@postech.ac.kr).

Summary

The signal transduction pathway governed by the phytohormone abscisic acid (ABA) regulates not only abiotic stress responses but also early developmental programs such as seed dormancy, germination and seedling growth in response to environmental signals. Optimal plant growth and development depend on the integration of environmental stimuli and intrinsic developmental programs. Here, we show that the homeodomain transcription factors BLH1 and KNAT3, previously implicated in embryo sac development, have additional functions in ABA-mediated seed dormancy and early seedling development. The ABA-dependent induction of BLH1 and KNAT3 expression required the presence of functional PYR/PYL/RCAR receptors. The blh1 and knat3 mutants were less sensitive than the wild-type to ABA or salinity exposure during seed germination and early seedling development. In contrast, BLH1 over-expressing lines were hypersensitive to ABA and salinity, and exhibited increased expression of ABA-responsive genes, such as ABI3 and ABI5. BLH1 interacted with KNAT3 and enhanced the retention of KNAT3 in the nucleus. BLH1 and KNAT3 synergistically increased the ABA responses by binding to and subsequently activating the ABI3 promoter. Taken together, we propose that BLH1 and KNAT3 together modulate seed germination and early seedling development by directly regulating ABI3 expression.

Introduction

The phytohormone abscisic acid (ABA) is essential for various biological processes during plant growth and development, including embryo maturation, seed germination and post-germinative growth, as well as for the plant's response to abiotic stress (Leung and Giraudat, 1998; Finkelstein et al., 2002; Himmelbach et al., 2003). The canonical ABA signaling pathways involve diverse components, such as the ABA receptors PYRABACTIN RESISTANCE (PYR)/PYR1-LIKE (PYL)/REGULATORY COMPONENT OF ABA RECEPTOR (RCAR), type 2C protein phosphatases (PP2Cs), and SNF1-related protein kinases (SnRKs; Hirayama and Shinozaki, 2007; Ma et al., 2009; Park et al., 2009; Cutler et al., 2010; Nishimura et al., 2010). ABA perception by PYR/PYL/RCAR proteins releases SnRKs from the negative regulation of PP2C, which leads to activation of ABA signaling by altering the phosphorylation status and activity of downstream factors (Fujii et al., 2009; Nishimura et al., 2009; Umezawa et al., 2009). Diverse ABA-dependent responses are activated by a sub-family of bZIP proteins, including ABA INSENSITIVE 5 (ABI5) and ABA-responsive element-binding proteins (AREBs; Choi et al., 2000; Finkelstein and Lynch, 2000; Uno et al., 2000). In addition, ABI3 and ABI4, which contain the B3 and AP2 domains, respectively, mediate ABA signal transduction during seed germination (Giraudat et al., 1992; Finkelstein et al., 1998). Other transcription factors, such as MYB, MYC, NAC and WRKY factors, are also involved in the regulation of ABA- or drought-responsive genes (Abe et al., 2003; Fujita et al., 2004; Lu et al., 2007; Jiang and Yu, 2009; Ren et al., 2010).

Homeodomain transcription factors represent master regulators in developmental programming, and are ubiquitous in fungi, animals and plants. Two major classes of homeodomain proteins exist in plants: the homeodomain-leucine zipper (HD-Zip) family (Schena and Davis, 1992) and the three amino acid loop extension (TALE) homeoproteins (Burglin, 1997). TALE proteins are involved in developmental processes throughout the Arabidopsis lifecycle (Hamant and Pautot, 2010), including meristem initiation and maintenance (Clark et al., 1996; Endrizzi et al., 1996; Long et al., 1996), leaf margin formation (Kumar et al., 2007), flowering (Cole et al., 2006; Proveniers et al., 2007), internode patterning (Ragni et al., 2008), and carpel (Alonso-Cantabrana et al., 2007) and ovule (Reiser et al., 1995) development. Plant growth and development are coordinately controlled by intrinsic programs and environmental stimuli. In particular, seed germination requires mechanism(s) that sense surrounding environmental conditions, and it is critical that seeds germinate during appropriate environmental conditions, because germination is an irreversible process. ABA plays a central role in seed germination in response to environmental stress. A few HD-Zip proteins appear to be part of the ABA signaling network. ABA triggers the expression of HD-Zip genes such as ATHB6, ATHB7 and ATHB12 (Soderman et al., 1996, 1999; Lee and Chun, 1998). Transgenic lines over-expressing ATHB7 and ATHB12 are hypersensitive to ABA in root elongation assays, whereas those over-expressing ATHB6 exhibit reduced ABA sensitivity in terms of germination and stomatal closure (Himmelbach et al., 2002; Olsson et al., 2004). In addition, ATHB5 positively modulates ABA responses in developing seedlings (Johannesson et al., 2003). The expression of another type of homeodomain gene, BEL1-LIKE HOMEODOMAIN 1 (BLH1), was also induced following ABA treatment (Hoth et al., 2002). BLH1 and KNOTTED1-like homeobox (KNOX; e.g. SHOOT MERISTEMLESS (STM), KNOTTED-LIKE FROM ARABIDOPSIS THALIANA 1-7 (KNAT1-7) and KNAT MEINOX (KNATM)) proteins belong to the TALE homeodomain transcription factor family, an atypical super-family of homeodomain proteins that is conserved among eukaryotes (Burglin, 1997; Hamant and Pautot, 2010). Combinatorial interactions among TALE homeoproteins play important roles during developmental processes in fungi and animals as well as in plants (Knoepfler et al., 1997; Berthelsen et al., 1998; Bellaoui et al., 2001; Muller et al., 2001; Hamant and Pautot, 2010). In animals, the interaction between TALE homeodomain proteins leads to masking of a nuclear export signal (NES) in the homeodomain proteins, and thereby facilitates retention of the heterodimer in the nucleus (Rieckhof et al., 1997; Berthelsen et al., 1999). Arabidopsis STM, a KNOX protein that is required for meristem development, is also targeted to the nucleus through heterodimerization with a BLH partner (Cole et al., 2006). In contrast, OVATE FAMILY PROTEIN 1 (OFP1) and OFP5 interact with BLH1, which induces the cytoplasmic localization of BLH1 (Hackbusch et al., 2005). Furthermore, the subcellular localization of plant TALE homeoproteins is controlled by the nuclear export mechanism via CRM1/exportin-1 (Rutjens et al., 2009).

The finding that ABA induces BLH1 expression suggests that the encoded protein functions as a signaling mediator between a developmental program and an environmental cue; however, the roles of TALE proteins in ABA or stress signaling have not been explored. Here, we report the existence of a regulatory module of ABA signaling that is governed by two TALE homeobox proteins, BLH1 and KNAT3, during seed germination and post-germinative growth. The ABA-inducible protein BLH1 forms a heteromeric complex with KNAT3, which increases the nuclear retention of KNAT3 and results in direct activation of ABI3. We propose that the nuclear localization of the BLH1 and KNAT3 heterocomplex is important for the induction of ABA-responsive genes.

Results

BLH1 modulates ABA sensitivity during seed germination and early seedling development

To explore the functional roles of BLH1 in plant growth and development, we first determined its expression patterns in seeds, leaves, roots, stems and flowers by quantitative RT-PCR. BLH1 was widely expressed in all tested tissues, but its expression was especially enriched in seeds and stems (Figure 1a), suggesting that it has major functions in seed and vasculature development. BLH1 was found to be rapidly induced by ABA treatment in a genome-wide transcriptome analysis (Hoth et al., 2002). We confirmed this finding using quantitative RT-PCR analysis (Figure 1b). Endogenous levels of BLH1 increased 2.7-fold within 3 h of ABA treatment in the wild-type. In contrast, the activation was severely impaired in the pyr1 pyl1 pyl2 pyl4 quadruple PYR/PYL/RCAR receptor mutant and completely blocked in the abi1-1 mutant (Figure 1b), implying that BLH1 expression is regulated by the PYR/PYL/RCAR–ABI-mediated ABA signaling pathway.

Figure 1.

BLH1 modulates ABA sensitivity during seed germination and early seedling development.

(a) Expression of BLH1 in Arabidopsis tissues. Total RNAs were isolated from the indicated tissues. Leaves/roots and stems/flowers were obtained from 2- and 5-week-old plants, respectively. The transcript levels of BLH1 were normalized to those of UBQ1, and the relative level of BLH1 transcript in the roots was set to 1. (b) The expression of BLH1 was induced by ABA. Five-day-old wild-type (Ler), pyr1 pyl1 pyl2 pyl4 mutant and abi1-1 mutant seedlings were grown in the presence or absence of 100 μm ABA for 3 h. Transcript levels of BLH1 were normalized to those of UBQ1, and the relative level of BLH1 transcripts in mock-treated wild-type seedlings was set to 1.

(c) BLH1 positively modulates the ABA response during seed germination. The germination rates of Col-0, blh1, 35S:BLH1 and pyr1 pyl1 pyl2 pyl4 were recorded from 1–5 days after stratification in the presence of 0.3 μm ABA (top) or 100 mm NaCl (bottom).

(d) BLH1 affects the opening of cotyledons. Cotyledon opening of Col-0, blh1 and 35S:BLH1 in the presence of 0.5 μm ABA was determined, and photographs were taken 7 days after stratification. The numbers in parentheses represent the number of seedlings with open cotyledons per total number of seeds examined. (e,f) BLH1 is positively involved in the ABA response during primary root growth. Three-day-old Col-0, 35S:BLH1 and blh1 seedlings were transferred to growth medium containing 0, 5 or 15 μm ABA, grown for another 14 days, and then photographed. Representative seedlings are shown in (e), and the primary root lengths are shown in (f). Values are means and SE for = 3 (a,b), = 150 (c) and = 10 (f). Asterisks indicate statistically significant differences (*< 0.05, **< 0.01; Student's t test).

The finding that BLH1 expression is ABA-inducible suggests that this gene has a physiological function in ABA signaling during early developmental processes. To test this possibility, we measured the rate of seed germination in BLH1 over-expressing (35S:BLH1) plants, knockout blh1 plants, and the pyr1 pyl1 pyl2 pyl4 plants. Over-expression of the BLH1 transgene in 35S:BLH1 and the silencing of BLH1 in the homozygous blh1 mutant isolated from the SALK T-DNA insertion mutant collection (SALK_089095) were confirmed by semi-quantitative RT-PCR (Figure S1). The germination rates of 35S:BLH1 and blh1 were similar to those of the wild-type in the absence of ABA (Figure S2a). However, in ABA-containing medium, the seed germination rate of blh1 was significantly higher than that of the wild-type, but not as high as that of the pyr1 pyl1 pyl2 pyl4 mutant, while 35S:BLH1 seeds germinated more slowly than those of the wild-type (Figure 1c and Figure S2a). The cotyledon opening rates for blh1 and 35S:BLH1 did not differ from those of the wild-type in the absence of ABA (Figure S2b), whereas blh1 showed reduced ABA sensitivity and 35S:BLH1 showed increased ABA sensitivity in the presence of ABA (Figure 1d). We further evaluated whether the altered ABA sensitivity of blh1 or 35S:BLH1 affected the response to salinity during seed germination. NaCl treatment significantly reduced the germination rate of 35S:BLH1 seeds, whereas the germination of blh1 seeds was less inhibited by NaCl than that of the wild-type, but remained lower than that of the pyr1 pyl1 pyl2 pyl4 mutant (Figure 1c and Figure S2a).

We next examined whether the blh1 and 35S:BLH1 lines also showed impaired sensitivity to ABA during post-germinative growth by analyzing the root growth phenotypes of plants treated with ABA, which is known to inhibit root growth (Figure 1e,f). In the absence of ABA, the primary roots of blh1 and 35S:BLH1 grew at similar rates to those of the wild-type. However, the inhibition of root elongation in response to ABA treatment was much more pronounced in 35S:BLH1 plants than in wild-type plants. In contrast, the differences in root length between blh1 and the wild-type were not significant, which indicates the presence of a possible functional overlap between members of the BLH1 gene family.

BLH1 modulates the expression of a subset of ABA-responsive genes

To establish whether the ABA-related phenotypes of blh1 and 35S:BLH1 lines are directly associated with ABA signaling pathways, we determined the expression patterns of ABA-inducible genes in these lines. During seed maturation, ABA accumulates and positively regulates expression of late embryogenesis-abundant genes, including Em1 and Em6, and the ABA signaling modulator genes ABI3 and ABI5 (Bensmihen et al., 2002; Lopez-Molina et al., 2002). RD29A and RD29B are typical marker genes induced by ABA and abiotic stresses (Yamaguchi-Shinozaki and Shinozaki, 1993). We monitored the expression of these genes in 5-day-old seedlings in the presence or absence of exogenous ABA (Figure 2). The ABA-induced expression of these genes was markedly increased in the 35S:BLH1 plants, whereas the induction of RD29B and ABI3 was attenuated in blh1. In the absence of ABA, the expression of RD29B, Em1 and Em6 in 35S:BLH1 was higher than in the wild-type, whereas ABI3 expression in blh1 was lower. Together, these data suggest that BLH1 positively regulates the expression of ABA-induced genes, including ABI3, during early seedling development.

Figure 2.

BLH1 regulates the expression of ABA-inducible genes. The expression patterns of ABA signaling-related RD29A, RD29B, ABI3 and ABI5, and embryogenesis-related Em1 and Em6 were determined in 1-week-old Col-0, blh1 and 35S:BLH1 seedlings treated with 100 μm ABA for 2 h. The transcript level of each gene was normalized to that of UBQ1, and the relative level of each gene in mock-treated Col-0 was set to 1. The inset graphs show expression of the indicated genes in mock-treated seedlings. Values are means and SE (= 4). Asterisks indicate statistically significant differences compared to Col-0 (*< 0.05; Student's t test).

KNAT3, an interacting partner of BLH1, modulates ABA sensitivity during seed germination

As BLH1 was reported to interact with KNAT3, KNAT5 and KNAT6 proteins in a yeast two-hybrid assay (Hackbusch et al., 2005), we tested whether these interactions affected the ABA response during seed germination. Hemagglutinin (HA)-tagged BLH1 and FLAG-tagged KNAT3, KNAT5 or KNAT6 were co-expressed in protoplasts, and the KNOX proteins were pulled down from the plant extracts using anti-FLAG antibodies. As expected, BLH1–HA was clearly detected in the immunoprecipitates with KNAT3–FLAG, KNAT5–FLAG and KNAT6–FLAG (Figure S3). These interactions suggest that KNOX proteins in the BLH1 complexes also modulate ABA signaling. To test this possibility, we performed germination and cotyledon opening assays using the knat3, knat5 and knat6 mutants. The homozygous knat3, knat5 and knat6 mutants isolated from the SALK T-DNA insertion mutant collection (SALK_136464, SALK_000339 and SALK_054482) were confirmed by semi-quantitative RT-PCR (Figure S1). In the presence of ABA, knat3 showed reduced ABA sensitivity in terms of seed germination and cotyledon opening, whereas knat5 and knat6 displayed normal ABA sensitivities (Figure 3a,b and Figure S4). These data indicate that KNAT3 may be involved in ABA-mediated seed germination and early seedling development. The knat3 mutant, but not knat5 and knat6, also showed hyposensitivity to salinity stress during germination (Figure 3a). Furthermore, among these three KNOX genes, KNAT3 expression was exclusively induced 2.6-fold by exogenous ABA in the wild-type, but this induction was suppressed in the pyr1 pyl1 pyl2 pyl4 mutant and the abi1-1 mutant (Figure 3c). As the ABA response is similar in blh1 and knat3, this result suggests that the BLH1/KNAT3 complex may regulate ABA responses. Interestingly, in 35S:BLH1 knat3 lines (Figure S1), the sensitivity of the 35S:BLH1 mutant to ABA during seed germination was partly compromised by the loss-of-function knat3 mutation (Figure 3d and Figure S5). Furthermore, we determined the expression patterns of ABA-inducible genes such as RD29A, RD29B and ABI3 in Col-0, 35S:BLH1 and 35S:BLH1 knat3 lines. As expected, ABA-induced expression of these genes was markedly decreased in 35S:BLH1 knat3 compared to 35S:BLH1 (Figure S6). To further examine the action of the BLH1/KNAT3 complex in ABA signaling, we determined the expression of RD29A and RD29B in blh1 protoplasts transfected with KNAT3 RNAi plasmids (blh1 KNAT3RNAi). Transformation with the KNAT3 RNAi plasmid decreased endogenous KNAT3 expression by approximately 52% compared to the control plasmid in the absence of ABA (Figure S7). In the presence of ABA, the reduction of RD29A and RD29B expression in the blh1 mutant compared to wild-type was further fortified by the suppression of KNAT3 expression, although the reduced level in blh1 KNAT3RNAi was not as great as in the pyr1 pyl1 pyl2 pyl4 mutant. These data suggest that KNAT3 controls ABA signaling together with BLH1 during seed germination.

Figure 3.

KNAT3 regulates the ABA response during seed germination.

(a) KNAT3 is positively involved in the ABA response during seed germination. The germination rates of Col-0, knat3, knat5 and knat6 were recorded from 1–5 days (ABA) or at 5 days (NaCl) after stratification in the presence of 0.3 μm ABA or the indicated concentration of NaCl.

(b) KNAT3 affects cotyledon opening. Cotyledon opening of Col-0, knat3, knat5 and knat6 seedlings in the presence of 0.3 μm ABA was determined, and photographs were taken 5 days after stratification. The numbers in parentheses represent the number of seedlings with open cotyledons per total number of seeds.

(c) The expression of KNAT3 was induced by ABA. Five-day-old wild-type (Ler), pyr1 pyl1 pyl2 pyl4 mutant and abi1-1 mutant seedlings were grown in the presence or absence of 100 μm ABA for 3 h. The transcript level of each gene was normalized to that of UBQ1, and the relative level of each gene in mock-treated wild-type seedlings was set to 1.

(d) The hypersensitivity of 35S:BLH1 to ABA during seed germination was partially resolved by a loss-of-function knat3 mutation. The germination rates of Col-0, knat3, 35S:BLH1 and 35S:BLH1 knat3 were recorded from 1–5 days after stratification in the presence of 0.3 μm ABA.

Values are means and SE for = 150 (a,d) and = 3 (c) in the bottom panel of Figure 3a, asterisks indicate significant differences compared to Col-0 (*< 0.05, **< 0.01; Student's t test).

BLH1 and KNAT3 synergistically increase the ABA response in protoplasts

We next investigated how the interaction between BLH1 and KNAT3 affects ABA sensitivity. Because two conserved domains, KNOX1 and 2, of the KNOX family are known to interact with other TALE proteins (Bellaoui et al., 2001), we first generated deletion mutants of KNAT3, KNAT3ΔKNOX1 and KNAT3ΔKNOX2, that lacked each sub-domain (Figure 4a). HA-fused BLH1 and FLAG-fused KNAT3, KNAT3ΔKNOX1 or KNAT3ΔKNOX2 were co-expressed in protoplasts, and the KNAT3 proteins were then pulled down using anti-FLAG antibodies. BLH1 proteins were pulled down by full-length KNAT3, but not by KNAT3ΔKNOX1 or KNAT3ΔKNOX2 (Figure 4b), indicating that the two KNOX sub-domains of KNAT3 are essential for the interaction with BLH1. We then determined the effect of the BLH1/KNAT3 complex on transcription of ABA-responsive genes in a transient expression assay using a luciferase gene as a reporter under the control of the RD29B promoter. The activity of the RD29B promoter was induced by ABA, and co-expression of BLH1 further enhanced the promoter activity of RD29B to 3.9 times in the absence of ABA and 2.5 times in the presence of ABA compared to a control without effector plasmids (Figure 4c). Co-expression of KNAT3 also increased RD29B expression, to 1.5 times in the absence of ABA and 1.9 times in the presence of ABA compared to a control without effector plasmids. Interestingly, the co-expression of BLH1 and KNAT3 further enhanced the ProRD29B:LUC reporter activity to 8.9 times in the absence of ABA and 6.2 times in the presence of ABA compared to a control without effector plasmids, indicating that these two transcription factors have a synergistic effect on induction of ABA-responsive genes. KNAT3ΔKNOX2 activated RD29B expression as much as full-length KNAT3 did, but co-expression did not further increase the reporter activity. Interestingly, KNAT5 and KNAT6 also slightly activated RD29B expression, but co-expression did not have any synergistic effect on the induction of RD29B expression. These data indicate that the specific interaction between BLH1 and KNAT3 enhances ABA-mediated activation of expression of the ABA-responsive reporter gene.

Figure 4.

BLH1 and KNAT3 synergistically increase the ABA response.

(a) Schematic diagram of KNAT3 deletion mutants. Full-length KNAT3 contains four putative functional domains: KNOX1, KNOX2, ELK and the homeodomain (HD). KNAT3ΔKNOX1 and KNAT3ΔKNOX2 indicate deletion of the KNOX1 and KNOX2 domain, respectively. Amino acid numbers are indicated above the diagram.

(b) BLH1 interacted with KNAT3, but not with KNAT3ΔKNOX1 and KNAT3ΔKNOX2. BLH1–HA was co-transfected with KNAT3–FLAG, KNAT3ΔKNOX1–FLAG or KNAT3ΔKNOX2–FLAG into mesophyll protoplasts. Co-immunoprecipitation assays were performed using monoclonal anti-FLAG antibodies in total protein extracts from protoplasts, and proteins were detected using anti-HA or anti-FLAG antibodies.

(c) BLH1 and KNAT3 synergistically activate ProRD29B. Protoplasts were co-transfected with 10 μg ProRD29B:LUC as a reporter and the indicated effector plasmids, and treated with 100 μm ABA for 1 h. The amount of DNA used in each experiment was 10 (+) or 20 μg (++), as indicated. Renilla luciferase was used as an internal control. The relative luciferase activity in mER7-GFP (control plasmid)-transfected cells not treated with ABA was set to 1. Values are means and SE (= 3). Asterisks indicate statistically significant differences (**< 0.01; Student's t test).

BLH1 promotes nuclear localization of KNAT3

Sequence analysis revealed that Arabidopsis KNOX proteins have both a putative nuclear localization signal in the homeodomain and an NES in the KNOX1 domain (Figure 5a), which suggests nucleo-cytoplasmic shuttling of these KNOX proteins. We thus examined whether the localization of BLH1 and KNOX are altered by their interactions, using BLH1–GFP and KNOX–RFP proteins in protoplasts. BLH1 localized mainly to the nucleus, whereas KNAT3 proteins, surprisingly, were located mainly in the cytoplasm (Figure 5b). Following treatment with leptomycin B, an exportin inhibitor (Wolff et al., 1997), most of the KNAT3 protein accumulated in the nucleus, implying that KNAT3 is actively transported to the cytoplasm via an exportin. Consistently, a lysine 173 to serine mutation in the NES led to nuclear accumulation of KNAT3 (Figure 5b). Furthermore, in the presence of ABA, the activity of the RD29B promoter was induced 2.4-fold by KNAT3L173S compared to the control without effectors, but only 1.6-fold by wild-type KNAT3 (Figure S8). This finding suggests that nuclear localization of KNAT3 is necessary for the efficient activation of ABA-responsive genes. Interestingly, when BLH1 and KNAT3 were expressed together, a greater proportion of the KNAT3 proteins resided in the nucleus (Figure 5c). KNAT3ΔKNOX1 proteins were localized mainly to the nucleus regardless of whether or not BLH1 was present, which may be due to the lack of an NES in the mutant. In contrast, KNAT3ΔKNOX2 mutants localized mainly to the cytoplasm, similar to full-length KNAT3; however, only few KNAT3ΔKNOX2 proteins appeared to be retained in the nucleus after their co-expression with BLH1, and the others still remained in the cytoplasm (Figure 5b,c). Taken together, these data suggest that the interaction between KNAT3 and BLH1 via its KNOX domains facilitates the nuclear retention of KNAT3. KNAT5 was constitutively localized in the nucleus, regardless of whether or not BLH1 was present, while the localization of KNAT6 was similar to that of KNAT3 (Figure S9). Even though all KNOX proteins have a putative nuclear localization signal and an NES, this protein-specific localization suggests the existence of an as yet undetermined mechanism that distinguishes the functions of individual KNOX proteins.

Figure 5.

BLH1 promotes the nuclear localization of KNAT3. (a) Amino acid sequence alignment of KNOX1 domains from multiple KNOX proteins in Arabidopsis. The conserved sequences of the nuclear export signal (NES) in the KNOX1 domain are indicated by boxes. The NES consensus sequence is shown below. (b) Localization of BLH1–GFP and KNAT3–RFP. Top row: protoplasts expressing BLH1–GFP or KNAT3–RFP. Middle row: protoplasts expressing KNAT3–GFP in the presence of leptomycin B or KNAT3L173S–RFP. Samples were treated with leptomycin B, an exportin inhibitor, for 2 h. Bottom row: protoplasts expressing KNAT3ΔKNOX1–RFP or KNAT3ΔKNOX2–RFP. Bright-field (left) and fluorescence (right) images were taken 6 h after transfection. (c) KNOX1 and KNOX2 domains are necessary for the BLH1-mediated nuclear retention of KNAT3. Top row: protoplasts expressing both BLH1–GFP and KNAT3–RFP. Middle row: protoplasts expressing both BLH1–GFP and KNAT3ΔKNOX1–RFP. Bottom row: protoplasts expressing both BLH1–GFP and KNAT3ΔKNOX2–RFP. Bright-field, GFP, RFP and merged images (left to right) were taken 6 h after transfection. Scale bars = 10 μm (b,c).

BLH1/KNAT3 binds directly to the ABI3 promoter

The BLH1-induced expression of ABI3 (Figure 2), the ABA-insensitive phenotypes during germination of blh1, knat3 and abi3 (Figures 1c and 3a; Giraudat et al., 1992; Nakashima et al., 2006), and the identification of multiple BLH1 binding sites in the ABI3 promoter (Staneloni et al., 2009) suggest that the BLH1/KNAT3 complex may be directly recruited to the ABI3 promoter. We tested this hypothesis by performing chromatin immunoprecipitation (ChIP) assays using protoplasts transfected with BLH1–HA or KNAT3–HA in the presence of ABA. The cross-linked DNA was pulled down using anti-HA antibodies and quantified by quantitative RT-PCR using three primer sets covering the putative BLH1 binding sites (R1–R4) in the ABI3 promoter. As shown in Figure 6(a), BLH1 bound to R1 (3.3-fold compared to the control), and binding was further increased by co-expression with KNAT3 (7.8-fold compared to the control). However, KNAT3 did not associate with R1 by itself. In contrast, BLH1 or KNAT3 itself did not bind to R2, R3, and R4. However, the BLH1/KNAT3 complex bound strongly to R2 (18.5-fold compared to control), but not in R3 and R4. These data indicate that KNAT3 enhances the binding affinity of BLH1 to the ABI3 promoter. To examine whether direct binding of the BLH1/KNAT3 complex to the ABI3 promoter affects the expression of ABI3, we performed a transient expression assay using ProABI3:LUC as a reporter plasmid (Figure 6b). The activity of the ABI3 promoter was almost insensitive to ABA (Figure 2). However, the promoter activity was enhanced by BLH1 even in the absence of ABA (3.2-fold compared to control), and was further greatly induced (122.5-fold compared to control) within 3 h of exposure to ABA, whereas KNAT3 did not affect the expression of ABI3. Interestingly, co-expression of BLH1 and KNAT3 further enhanced the expression of ABI3 to 8.5-fold in the absence of ABA and 246.1-fold in the presence of ABA compared to control. These results indicate that the BLH1/KNAT3 heterodimer has a higher DNA-binding affinity for the target promoter than does either of the monomers individually. Interestingly, BLH1 over-expression in wild-type protoplasts induced the expression of RD29A, RD29B and Em6 in the presence of ABA, but this BLH1-dependent induction of ABA-responsive genes was abolished in the abi3-8 loss-of-function mutant (Figure S10). Taken together, these data suggest that BLH1 regulates the expression of ABA-responsive genes by directly targeting ABI3.

Figure 6.

The BLH1/KNAT3 complex increases the activity of the ABI3 promoter.

(a) BLH1 strongly binds to the ABI3 promoter in the presence of KNAT3. Upper panel: schematic diagram of the ABI3 promoter showing BLH1-binding TGGA motifs (white circles) and the regions (R1–R4) analyzed by ChIP. Lower panel: protoplasts were transfected with 40 μg BLH1–HA (black bars), 40 μg KNAT3–HA (light gray bars) or 20 μg BLH1–HA and 20 μg KNAT3–HA (dark gray bars). The chromatin associated with BLH1, KNAT3 or BLH1 and KNAT3 was immunoprecipitated using anti-HA antibodies, and the regions of the ABI3 promoter indicated in the upper panel were amplified by quantitative RT-PCR using specific primers. The fold enrichment was calculated and normalized to the values of the protoplasts without antibodies (control). Values are means and SE (= 5 or 6). Asterisks indicate statistically significant differences compared to the control (*< 0.05; Student's t test).

(b) BLH1 and KNAT3 synergistically activate the ABI3 promoter. Protoplasts were co-transfected with 5 μg ProABI3:LUC as a reporter plasmid and 5 μg of each effector plasmid as indicated. Renilla luciferase was used as an internal control. The relative luciferase activity in mER7-GFP-transfected cells without ABA was set to 1.

(c) Hypothetical model of BLH1/KNAT3 action during ABA signaling. At low concentrations of ABA, low levels of BLH1 and KNAT3 proteins are produced. KNAT3 monomers are translocated to the cytosol via exportin, and thus BLH1 weakly induces ABI3 expression. Under stress conditions (e.g. high levels of ABA), expression of BLH1 and KNAT3 is enhanced. The increased amount of BLH1 promotes retention of KNAT3 in the nucleus by masking the NES. The resulting BLH1/KNAT3 complex strongly binds to the ABI3 promoter and enhances the ABA response.

Discussion

Plant seeds need to recognize and respond to environmental factors in order to germinate and resume their lifecycles at the proper time and in a suitable space. Even after seed germination has already occurred, seedlings monitor changes in the surrounding conditions and inhibit growth in a process called post-germination developmental arrest (Lopez-Molina et al., 2001). ABA is known to balance the early stages of development and growth inhibition by targeting a large number of genes in response to environmental stimuli (Leung and Giraudat, 1998; Finkelstein et al., 2002; Himmelbach et al., 2003). Here, we showed that two homeodomain transcription factors, BLH1 and KNAT3, form a dimer that modulates ABA-mediated developmental processes by directly targeting ABI3. The biological roles of these proteins have been unclear, as loss-of-function mutants of BLH1 and KNAT3 were found to exhibit wild-type phenotypes (Truernit et al., 2006; Pagnussat et al., 2007). We also observed that the blh1 and knat3 mutants do not display any noticeable changes in plant growth and development under normal conditions. However, they showed altered seed germination and early seedling development after treatment with ABA or high salinity. A gain-of-function allele of BLH1, eostre-1, has been shown to disrupt the establishment of cell fate in the mature embryo sac, and the knat3 mutation suppressed the eostre-1 phenotype (Pagnussat et al., 2007), Although this defect was caused by ectopic expression of BLH1, this finding supports the relevance of BLH1 and KNAT3 in seed development. The roles of homeoproteins have been investigated and have mainly focused on developmental procedures; however, we have highlighted here a previously unrevealed function of homeoproteins, i.e. coordination of an endogenous developmental process with inputs of environmental stimuli.

BLH1 and KNAT3 formed a positive feedback loop with ABA; their expression was induced by PYR/PYL/RCAR–ABI-mediated ABA signaling (Figures 1b and 3c), which in turn stimulated the expression of several ABA-responsive genes, including ABI3 and ABI5, which encode important regulators of germination and post-germinative growth (Figure 2). BLH1 promotes its own expression (Staneloni et al., 2009), which is expected to accelerate this feedback loop. This positive regulatory system reinforces the ABA signaling network that suppresses germination and early seedling development. ABA-induced BLH1 expression was completely inhibited and the induction of KNAT3 expression was significantly suppressed in the abi1-1 mutant (Figures 1b and 3c); however, expression of BLH1 and KNAT3 was still induced in the pyr1 pyl1 pyl2 pyl4 mutant, suggesting that the other PYR/PYL/RCAR receptors and ABA signaling components may play a role in transcriptional regulation of BLH1 and KNAT3.

It is known that BEL1-like (BELL) homeoproteins with a conserved NES interact with CRM1/exportin-1, but KNOX proteins such as STM does not bind to CRM1/exportin-1 (Rutjens et al., 2009). Here, we found that KNAT3 was localized mainly to the cytoplasm, and its cytoplasmic re-localization from the nucleus was apparently controlled in an active NES- and exportin-dependent manner (Figure 5b). However, KNAT3 proteins were retained in the nucleus through an interaction with BLH1 (Figure 5c), suggesting that BLH1 interrupted the access of KNAT3 to exportin. The KNOX1 domain of KNAT3 contains an NES and is necessary for the interaction with BLH1 (Figures 4b and 5b). Cytoplasmic KNAT3 reduced ABA-mediated activation of gene expression in the absence of its BLH partners. The finding that KNAT3 is recruited to the nucleus in association with BLH1 and that expression of ABA-responsive genes is subsequently regulated provides insight into the regulation of ABA signaling, and adds a level of complexity to the regulation of the ABA signaling network in response to environmental stimuli. Dimerization of the transcription factors, which is a well-conserved regulatory mechanism in all eukaryotes, increases their binding affinity and the specificity of their transcriptional regulation, highlighting the importance of protein–protein interactions in transcriptional regulation (Hai and Curran, 1991; Lamb and McKnight, 1991; Smith et al., 2002). Our data suggest that BLH1 is able to function on its own in the ABA response; however, its interaction with KNAT3 cooperatively strengthens the transcriptional activity of BLH1 (Figures 3d and 4c). As expression of both BLH1 and KNAT3 is induced by ABA under physiological conditions, the BLH1/KNAT3 complex may play a bigger role in mediating ABA responses than BLH1 alone. It is plausible that BLH1 and KNAT3 bind to different cis-elements of a target gene in tandem, and collaboratively activate the target gene. Alternatively, BLH1 and KNAT3 may share a DNA target site and stabilize the DNA–protein complex via dimerization. In contrast to the cytoplasmic localization of KNAT3 and KNAT6, BLH1 and KNAT5 proteins were localized to the nucleus despite of their conserved NES signals (Figure 5a,b and Figure S9). As BLH1 and KNAT5, but not KNAT3 and KNAT6, form homodimers (Hackbusch et al., 2005), the nuclear localization of BLH1 and KNAT5 may be due to masking of the NES by homodimerization, which further emphasizes the importance of the interactions among TALE proteins in determining their subcellular localization, and explains the activation of the ABI3 promoter by BLH1 itself.

Taken together, we have found that the BLH1/KNAT3 heterodimer modulates ABA signaling during early developmental stages, including germination, and that ABA signaling outputs may be modulated by altering the dimerization status of the two proteins. Our hypothesis predicts that, under normal conditions with low levels of ABA, a few BLH1 and KNAT3 proteins will be generated (Figure 6c). KNAT3 monomers may move from the nucleus to the cytoplasm, and thus BLH1 weakly induces the transcription of ABA-related genes. Under stress conditions, during which the ABA levels increase, more BLH1 and KNAT3 proteins are available in the nucleus, and these form a heterodimer that retains KNAT3 in the nucleus. The resulting BLH1/KNAT3 complex binds to the promoter of ABA-related genes, such as ABI3, and facilitates ABA signaling. A genome-wide analysis of the target genes and binding sites of the BLH1/KNAT3 complex would further clarify the BLH1/KNAT3-based regulatory network.

Experimental Procedures

Plant materials and growth conditions

Arabidopsis thaliana ecotype Col-0 or Ler was used as the wild-type control. Mutants of blh1 (SALK_089095), knat3 (SALK_136464), knat5 (SALK_000339), knat6 (SALK_054482), abi1-1 and abi3-8 were obtained from the Arabidopsis Biological Resources Center (http://abrc.osu.edu/). A transgenic plant expressing BLH1 under the control of the CaMV 35S promoter in the Col-0 background was provided by Jorge J. Casal (Fundación Instituto Leloir, Instituto de Investigaciones Bioquímicas Buenos Aires, Consejo Nacional de Investigaciones Cientificas y Técnicas and Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, C1405BWE Buenos Aires, Argentina) (Staneloni et al., 2009). The pyr1 pyl1 pyl2 pyl4 quadruple mutants in the Col– and Ler background (Q32) were provided by Sean R. Cutler (Department of Botany and Plant Sciences, University of California at Riverside, Riverside, CA 92521, USA) (Park et al., 2009). For protoplast isolation, 7-day-old seedlings transferred to autoclaved soil and grown at 22°C under short-day conditions (10 h light/14 h dark) for 3 weeks or 7-day-old seedlings on half-strength MS medium were used.

Seed germination assay

All genotypes were grown at 22°C under long-day conditions (16 h light/8 h dark) for 2 months, and mature seeds of each genotype were collected at the same day. Surfaced-sterilized seeds were plated on half-strength MS medium containing 0.5% sucrose and 0.8% agar (pH 5.7), supplemented with various concentrations of NaCl (0, 50, 100, 150 and 200 mm) or ABA (0, 0.3, 0.6 and 0.9 μm). Plates were kept at 4°C in the dark for 2 days, and transferred to a culture room set at 22°C under long-day conditions. Radicle emergence was scored after the indicated time intervals.

Primary root length measurement

Arabidopsis seeds were germinated on half-strength MS medium containing 0.5% sucrose and 1.5% agar (pH 5.7) and grown in a vertical position. Three-day-old seedlings were transferred to fresh medium with or without ABA (0, 5 and 15 μm), and incubated for an additional 2 weeks before the primary root length was measured.

Protoplast transient expression assay and fluorescence microscopy

BLH1 cDNA was introduced into plant expression vectors containing hemagglutinin (HA) or GFP tags and the 35S-C4PPDK promoter. The cDNAs of KNAT3, KNAT5 and KNAT6 were fused to FLAG or monomeric red fluorescent protein (mRFP) tags in the same vectors. The isolated protoplasts were transfected as previously described (Hwang and Sheen, 2001). For transient expression assays, the ProRD29B:LUC or ProABI3:LUC reporter gene was co-transfected with BLH1 or KNAT3 constructs into protoplasts. Transfected protoplasts were incubated for 4 h at room temperature and then treated with 100 μm ABA for 1 or 3 h. Renilla luciferase (Rluc) was used as an internal transfection control. To investigate subcellular localization, the GFP and RFP fluorescence signals from protoplasts were observed using a fluorescence microscope (Zeiss, http://www.zeiss.com) and a confocal microscope (Zeiss LSM 510 META). For the transient RNAi assay, a 300 bp KNAT3 cDNA fragment was cloned in two opposite orientations in the pHANNIBAL vector (Wesley et al., 2001), and this construct was transfected into protoplasts. Transfected protoplasts were incubated for 12 h at room temperature, and then treated with 100 μm ABA for 3 h.

Co-immunoprecipitation assays

HA-tagged BLH1 was transfected into protoplasts together with FLAG-tagged KNAT3, KNAT5 or KNAT6. After a 6 h incubation at room temperature, total proteins were extracted using a buffer containing 50 mm Tris/HCl (pH 7.5), 100 mm NaCl, 5 mm EDTA, 1 mm dithiothreitol, protease inhibitor cocktail (Roche, http://www.roche.com) and 1% Triton X-100. Protein complexes were bound to a monoclonal anti-FLAG antibody (Sigma, http://www.sigmaaldrich.com), and co-precipitated using Protein G Plus/Protein A-Agarose beads (Calbiochem, http://www.millipore.com/calbiochem). Precipitated proteins were detected using horseradish peroxidase-conjugated anti-HA antibody (Roche).

Real-time RT-PCR

Five-day-old seedlings were cultured in liquid MS medium (pH 5.7), and then treated with 100 μm ABA for 2 or 3 h. Total RNAs were extracted with TRIzol (Invitrogen, http://www.invitrogen.com) according to the manufacturer's instructions. DNA contamination was removed using DNAfree (Ambion, http://www.invitrogen.com/ambion), and cDNA was synthesized using the ImProm-II reverse transcription system (Promega, http://www.promega.com). Gene expression was quantified using gene-specific primers and SYBR Premix Ex Taq (Takara, http://www.takara-bio.com) in a LightCycler 2.0 (Roche).

Chromatin Immunoprecipitation (ChIP)

ChIP was performed as previously described (Nelson et al., 2006; Saleh et al., 2008) with minor modifications. Chromatin was immunoprecipitated using anti-HA antibodies (Roche) and salmon sperm DNA/protein A agarose beads (Millipore, http://www.millipore.com). DNA was purified using phenol/chloroform/isoamyl alcohol, sodium acetate (pH 5.2) and linear acrylamide (Ambion) and then used as a template for PCR. Primer sequences used in this study are listed in Table S1.

Accession numbers

The Arabidopsis Genome Initiative (AGI) numbers for the genes referred to in this paper are At2 g35940 (BLH1), At5 g52310 (RD29A), At5 g52300 (RD29B), At3 g24650 (ABI3), At2 g36270 (ABI5), At3 g51810 (Em1), At2 g40170 (Em6), At3 g52590 (UBQ1), At1 g62360 (STM), At4 g08150 (KNAT1), At1 g70510 (KNAT2), At5 g25220 (KNAT3), At5 g11060 (KNAT4), At4 g32040 (KNAT5), At1 g23380 (KNAT6) and At1 g62990 (KNAT7).

Acknowledgements

This work was supported by the Advanced Biomass R&D Center of the Global Frontier Project funded by the Korean Ministry of Education, Science and Technology (ABC-2010-0029720), a National Research Foundation of Korea grant funded by the Korean Ministry of Education, Science and Technology (2012R1A2A2A02014387) and the World Class University (WCU) program (R31-10105). H.R. was supported by a grant from the Next-Generation BioGreen 21 Program (PJ009516).

Ancillary

Advertisement