An Arabidopsis homolog of importin β1 is required for ABA response and drought tolerance

Authors

  • Yanjie Luo,

    1. State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    2. Graduate University of Chinese Academy of Sciences, Beijing, China
    3. Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, London, Ontario, Canada
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Zhijuan Wang,

    1. State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Hongtao Ji,

    1. State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    Search for more papers by this author
  • Hui Fang,

    1. Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, London, Ontario, Canada
    2. Department of Biology, Western University, London, Ontario, Canada
    Search for more papers by this author
  • Shuangfeng Wang,

    1. State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    Search for more papers by this author
  • Lining Tian,

    Corresponding author
    1. Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, London, Ontario, Canada
    2. Department of Biology, Western University, London, Ontario, Canada
    • State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    Search for more papers by this author
  • Xia Li

    Corresponding author
    • State Key Laboratory of Plant Cell & Chromosome Engineering, Center for Agricultural Research Resources, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Shijiazhuang, Hebei, China
    Search for more papers by this author

For correspondence (e-mail lining.tian@agr.gc.ca or xli@genetics.ac.cn).

Summary

The import of proteins into the nucleus in response to drought is critical for mediating the reprogramming of gene expression that leads to drought tolerance. However, regulatory mechanisms involved in nuclear protein import remain largely unknown. Here, we have identified an Arabidopsis gene (AtKPNB1) as a homolog of human KPNB1 (importin β1). AtKPNB1 was expressed in multiple organs, and the protein was localized in the cytoplasm and nucleus. AtKPNB1 was able to facilitate nuclear import of a model protein. Null mutation of AtKPNB1 delayed development under normal growth conditions and increased sensitivity to abscisic acid (ABA) during seed germination and cotyledon development. Inactivation of AtKPNB1 increased stomatal closure in response to ABA, reduced the rate of water loss, and substantially enhanced drought tolerance. AtKPNB1 interacted with several importin α proteins, nucleoporin AtNUP62, and the Arabidopsis Ran proteins. Inactivation of AtKPNB1 did not affect the ABA responsiveness or the expression level or subcellular localization of ABI1, ABI2 or ABI5, key regulators of the ABA signaling pathway. Moreover, phenotypic analysis of epistasis revealed that AtKPNB1 modulates the ABA response and drought tolerance through a pathway that is independent of ABI1 and ABI5. Collectively, our results show that AtKPNB1 is an Arabidopsis importin β that functions in ABA signaling.

Introduction

Water availability is one of the most important environmental factors that determine plant growth, productivity and distribution. Because future climate change is predicted to increase the incidence and severity of droughts worldwide (Sheffield and Wood, 2008), improving drought tolerance of plants will be central to sustaining agriculture. Therefore, elucidation of fundamental mechanisms underlying drought tolerance represents a major challenge in plant science.

As sessile organisms, plants have evolved many adaptive mechanisms to respond to decreased water availability, and an important component of these adaptive mechanisms is the hormone abscisic acid (ABA). When induced by cellular dehydration, ABA may be perceived by ABA receptors such as PYRABACTIN RESISTANCE1 (PYR1)/PYR1-LIKE (PYL)/REGULATORY COMPONENTS OF ABA RECEPTORS (RCAR) family members. This recognition activates SNF1-related protein kinase sub-family two (SnRK2)–bZIP transcription factors that positively control the expression of many ABA-responsive genes, leading to ABA-dependent drought tolerance (Hubbard et al., 2010; Umezawa et al., 2010). Among these transcription factors, ABA INSENSITIVE 5 (ABI5), which is a bZIP transcription factor, controls the ABA-mediated response to water stress in both seeds and developing seedlings. On the other hand, the SnrK2-ABRE (ABA responsive element)/ABF (ABRE-binding factor) pathway modulates ABRE-dependent gene expression and stress responses during the vegetative stage (Fujita et al., 2011). In addition to being subject to transcriptional regulation, ABA-mediated gene expression is also regulated by protein degradation (Antoni et al., 2011). Proper subcellular and organellar localization of the proteins involved in ABA and water stress signal transduction is also necessary for drought tolerance (Kim, 2006; Bartels et al., 2010; Kuromori et al., 2010). For example, transcription factors such as ABI5, mRNA-binding proteins such as ABA HYPERSENSITIVE1 (ABH1) and type 2C protein phosphatases such as ABA INSENSITIVE1 (ABI1) must be targeted to the nucleus to activate expression of specific effector genes involved in signal transduction pathways leading to drought tolerance (Hugouvieux et al., 2001; Lopez et al., 2002). Therefore, the protein nuclear import machinery has a critical role in mediating nuclear reprogramming and stress responses.

Protein transport to the nucleus is regulated through a conserved nuclear localization signal (NLS)-mediated nuclear protein import pathway (Ström and Weis, 2001; Xu et al., 2011). In non-plant systems, the karyopherin super-family proteins importinα and importin β, mediate import of NLS-containing cargo from the cytoplasm to the nucleus. Formation of the import complex comprising importin β/importin α/cargo is initiated in the cytoplasm by the cargo adaptor, importin α, that first binds to the classical NLS of nuclear proteins and then links the complex to importin β (Conti et al., 2006). Human importin β1 (KPNB1) mediates docking of the import complex to the cytoplasmic side of nuclear pore complexes by binding to nucleoporins. The import complex is then translocated through the pore into the nucleus by an energy-requiring mechanism (Merkle, 2003; Conti et al., 2006). On the nucleoplasmic side of the nuclear pore complex, Ran-GTP binds to KPNB1, leading to dissociation of the import complex and release of the cargo and importin α from the nuclear pore complex into the nucleus. KPNB1 may also directly bind to some cargos to translocate them into the nucleus without the assistance of importin α (Jäkel and Görlich, 1998; Takizawa et al., 1999; Hill, 2009). Although the NLS-mediated nuclear protein import pathway is quite clear in yeast and animal cells, understanding of the protein nuclear import mechanism in plants is still limited.

Drought tolerance and the ABA response during seed maturation, seed germination and post-germination seedling development are complicated biological processes, requiring the nuclear import of hundreds of proteins to initiate the necessary changes in gene expression. Therefore, stage-, tissue- and organ-specific nuclear protein receptors for import of particular proteins into the nucleus are essential for the proper response to ABA and drought at each developmental stage. Little is known about the mechanisms and molecules that control targeting of the specific nuclear proteins during plant responses to drought. There are 17 Arabidopsis proteins belonging to the importin β super-family, and each family member has a homologous mammalian counterpart (Bollman et al., 2003; Merkle, 2009). SENSITIVE TO ABA AND DROUGHT2 (SAD2), which is the Arabidopsis homolog of human importin RanBP7/8, has been reported to function in ABA and osmotic stress responses (Verslues et al., 2006). However, the underlying mechanisms by which SAD2 controls the ABA and osmotic stress responses remain unknown.

In this study, we identified AtKPNB1, a member of the Arabidopsis importin β super-family, as a gene associated with ABA sensitivity during germination and early seedling development and drought tolerance. AtKPNB1 is a homolog of human KPNB1. Our studies show that the AtKPNB1 gene is expressed in various organs and cells, including guard cells of stomata. AtKPNB1 also regulates stomatal closure and functions in the ABA-mediated pathway leading to drought tolerance in Arabidopsis. The function of AtKPNB1 in ABA-mediated drought tolerance is independent of ABI1 during seed germination and cotyledon greening. Like human KPNB1, AtKPNB1 interacts with importin α proteins, nucleoporin and Ran proteins. Biochemical and genetic analysis revealed that AtKPNB1 functions in nuclear import of NLS-containing proteins, although it does not directly mediate nuclear import of ABI1 and ABI2.

Results

The atkpnb1 mutant is more drought tolerant than the wild-type

To identify importin genes that function in drought tolerance, we screened T-DNA-tagged mutants of Arabidopsis importin β family genes for the ability to survive after water was withheld. A mutant with a T-DNA insertion in At5G53480, which encodes a homolog of human KPNB1, displayed substantially increased drought tolerance (Figure S1A) and was designated atkpnb1. Because atkpnb1 plants showed delayed development (Figure S2), 21-day-old wild-type (Col-0) and 23-day-old atkpnb1 plants were of similar size and were used in the experiments. As shown in Figure 1(a), Col-0 seedlings were severely wilted 19 days after water was withheld. In contrast, atkpnb1 mutant plants showed reduced wilting symptoms in response to water stress (Figure 1a). When the wilted Col-0 and atkpnb1 plants were re-watered, most Col-0 plants died within 5 days (mean survival rate of 25%), but most atkpnb1 mutant plants rapidly recovered (mean survival rate of 80%) (Figure 1b).

Figure 1.

The atkpnb1 mutant is more drought tolerant than the wild-type. (a) Drought tolerance assay of the wild-type (WT) and atkpnb1. atkpnb1 plants (23 days old) and WT plants (21 days old) were subjected to water stress by withholding water for 19 days; they were then watered for 4 days. (b) Survival rate of the WT and atkpnb1 under water stress. Three independent experiments were performed, each with 128 plants per genotype. (c) Water loss rates of detached leaves. Values shown are representative of three independent experiments with similar results. (d) Stomatal apertures of the WT and the atkpnb1 mutant. Leaves of the WT and atkpnb1 were treated with 1 μm ABA for 2 h. CK represents leaves without ABA treatment. Representative micrographs are shown. (e) Stomatal apertures of the WT and the atkpnb1 mutant corresponding to (d). Values are mean ratios of width to length ± standard deviations of three independent experiments (= 60). Values are means ± standard deviations. Asterisks indicate a significant difference (Student's t test, < 0.05) between the WT and the mutant.

Drought tolerance is often caused by a reduced rate of water loss from leaves. Accordingly, atkpnb1 leaves lost water at a slower rate than Col-0 leaves (Figure 1c). This led us to question whether AtKPNB1 affects stomatal development or the size of stomatal apertures in response to ABA. We found that the stomatal density and development in the absence of ABA treatment were comparable in Col-0 and atkpnb1 leaves (Figure S1B,C). The stomatal apertures of atkpnb1 and Col-0 were also comparable in KCl-treated control leaves. However, under ABA treatment, the stomatal aperture was much smaller for atkpnb1 than for the wild-type (Figure 1d,e). Together, these results suggest that the improved drought tolerance of atkpnb1 was associated with an increased sensitivity of atkpnb1 stomata to ABA-induced closure.

The AtKPNB1 mutation increases the sensitivity to ABA and osmotic stress

Next, we performed a phenotypic analysis to evaluate the response of the atkpnb1 mutant to ABA during early developmental stages. Under normal conditions, seed germination and cotyledon greening were somewhat slower in atkpnb1 than in Col-0 (Figure S2A–C). However, the germination and greening delay of atkpnb1 was enhanced by ABA treatment. As shown in Figure 2 23% of atkpnb1 seeds germinated on medium containing 0.1 μm ABA within 2 days, whereas 97% of Col-0 seeds germinated under the same conditions (Figure 2b). At 7 days after stratification, the percentage of green seedlings was substantially lower for atkpnb1 than for Col-0 at all tested concentrations of ABA (Figure 2c). Using the same assays, we found that atkpnb1 also displayed increased sensitivity to NaCl and mannitol during germination and early post-germination growth (Figure 2a,d–g). The results suggest that AtKPNB1 is required for the ABA-mediated response to general osmotic stress during early development. In addition, loss of AtKPNB1 function also increased plant sensitivity to sucrose (Figure S3).

Figure 2.

The atkpnb1 mutant is hypersensitive to ABA and various stresses during early development. (a) Sensitivity of the atkpnb1 mutant to 0.3 μm ABA, 100 mm NaCl and 200 mm mannitol. Seedlings were photographed 6 days after stratification. (b,c) Germination and cotyledon greening rates of the atkpnb1 mutant and the wild-type (WT) in response to increasing concentrations of ABA. Germination was scored at 2 days, and greening was scored at 7 days after stratification. (d,e) Germination and greening rates of the WT and the atkpnb1 mutant in response to increasing concentrations of NaCl. Germination was scored at 2.5 days, and greening was scored at 5 days after stratification. (f,g) Germination and greening rates of WT and the atkpnb1 mutant in response to increasing concentrations of mannitol. Germination was scored at 2.5 days, and greening was scored at 5 days after stratification. Three independent experiments were performed, and >50 seeds for each treatment were used for each experiment. Values are means ± standard deviation.

Soil-grown atkpnb1 plants had smaller rosette leaves and shorter floral stems than Col-0 plants. Although the number of leaves at bolting was comparable in atkpnb1 and Col-0 plants, bolting was delayed by approximately 4 days in the mutant (Figure S2D). The siliques were significantly smaller for atkpnb1 than for Col-0 (Figure S2E). These observations indicate a role for AtKPNB1 protein in Arabidopsis growth and development.

Atkpnb1 is a loss-of-function mutant

The T-DNA insertion in atkpnb1 was located in the second exon of AtKPNB1 (Figure 3a). No full-length AtKPNB1 transcript was detected in atkpnb1 by RT-PCR analysis, suggesting that atkpnb1 is a loss-of-function mutant (Figure 3b). To confirm that the mutant phenotypes were caused by loss of function of AtKPNB1, we transformed the full-length cDNA of AtKPNB1 driven by the native promoter into atkpnb1, and then analyzed three independent transgenic lines with comparable levels of AtKPNB1 (Figure 3b). The delayed flowering and drought tolerance phenotypes of atkpnb1 were completely reversed in the complemented lines (Figure 3c). The sensitivity of seedlings to ABA, NaCl, mannitol and sucrose was also restored to wild-type levels in the complemented lines (Figure 3d and Figures S3 and S4), demonstrating that the AtKPNB1 gene is indeed responsible for the developmental and drought tolerance phenotypes of the mutant.

Figure 3.

The AtKPNB1 gene is responsible for the phenotypes of the atkpnb1 mutant. (a) Schematic representation of the AtKPNB1 gene. Arrows indicate the location of primers used for RT-PCR (P1 and P2) or quantitative RT-PCR (P3 and P4). UTRs are shown in white, exons in gray, and introns as thick lines. The triangle indicates the T-DNA insertion site in atkpnb1. (b) RT-PCR analysis of expression of AtKPNB1 in the wild-type (WT), atkpnb1 and three complementation lines. (c) Complementation of AtKPNB1 in the atkpnb1 mutant completely rescued the normal development and drought tolerance of atkpnb1. Three-week-old seedlings were subjected to water stress for 19 days, and were then watered. Photographs were taken 4 days after re-watering. (d) Complementation of AtKPNB1 in the atkpnb1 mutant completely rescued the hypersensitivity of atkpnb1 to ABA and osmotic stress during germination and the post-germination stage. Seeds of atkpnb1 and Col-0 were germinated on medium containing 0.3 μm ABA, 100 mm NaCl or 200 mm mannitol, and photographs were taken 9 days after stratification.

Next, we assessed the effect of AtKPNB1 over-expression on plant growth and drought tolerance (Figure S5). All of the AtKPNB1 over-expressors flowered earlier than Col-0 (Figure S5C). In media containing ABA, NaCl or mannitol, seeds of AtKPNB1 over-expressors showed significantly increased percentages of germination and greening compared to Col-0 (Figure S5B,C). Most importantly, AtKPNB1 over-expressors exhibited substantially increased sensitivity to drought compared to Col-0, confirming that AtKPNB1 is an important negative effector of drought tolerance (Figure S5E). Under normal conditions, AtKPNB1 over-expression promoted growth and development throughout the lifecycle. The percentage of seeds that germinated and the percentage of seedlings that turned green were increased by over-expression of AtKPNB1.

Bioinformatic analysis showed that AtKPNB1 is a single-copy gene, comprising three exons and two introns (Figure 3a). AtKPNB1 encodes a 96 kDa protein containing 870 amino acids, and shares 39% identity and 58% deduced amino acid sequence similarity to human KPNB1 (Figure S6A). It contains an IBN_N (importin β N-terminal) domain that is found specifically in importin β super-family members and shares 35% similarity to the corresponding domain of human KPNB1. It contains a nuclear pore-binding region (with 48% similarity to the nuclear pore-binding domain of human KPNB1), a Ran-binding (RB) region (with 46% similarity to the Ran-binding domain of human KPNB1), and an importin α binding region (Figure S6B). Phylogenetic analysis shows that importin β1 proteins are conserved in eukaryotes but not in prokaryotes (Figure S6C), supporting a role for these proteins in nuclear protein import.

Expression of AtKPNB1 and subcellular localization of AtKPNB1

Quantitative RT-PCR analysis showed that AtKPNB1 is ubiquitously expressed, with the highest expression level in seeds (Figure 4a). By analyzing AtKPNB1::GUS transgenic plants, we confirmed that AtKPNB1 was expressed in various organs and tissues such as leaves, roots and flowers (Figure 4b).

Figure 4.

Expression pattern of AtKPNB1 and subcellular localization of AtKPNB1. (a) Expression of AtKPNB1 in various organs as analyzed by quantitative RT-PCR. R, root; St, stem; I, inflorescence; fl, flower; si, silique; se, seed. (b) Histochemical analysis of AtKPNB1. The GUS gene driven by the AtKPNB1 promoter was expressed in an immature seed (a), an embryo of a mature dry seed (b), a germinating seed (c), a 5-day-old seedling (d), a young leaf (e), a root tip (f), an old leaf (g), a lateral root (h), a stem (i), a flower (j), stigma and stamens (k), and a silique (l). Scale bars = 0.1 mm (a–c,h), 0.05 mm (f), 0.3 mm (e,k) and 1 mm (d,g,i,j,l). (c) AtKPNB1 is localized in the cytoplasm and nucleus. The constructs containing 35S::AtKPNB1-GFP or 35S::AtKPNB1-YFP were transformed to N. benthaminana leaves (top panel) or Arabidopsis protoplasts (middle panel). Images in the left and middle columns show fluorescent signals in dark and bright-field views, and images in the right column are merged views. The bottom panel shows the subcellular localization of AtKPNB1 in roots of transgenic AtKPNB1pro::AtKPNB1-GFP plants. The left column shows fluorescent signals, the middle column shows the signal obtained with propidium iodide stain, and the merged signals are shown on the right. (d) AtKPNB1 is expressed in guard cells. Expression in guard cells of AtKPNB1pro::AtKPNB1-GPF (left and middle columns) and AtKPNB1pro::GUS (right column). (e) Gene expression of AtKPNB1 in response to ABA treatment. Five-day-old wild-type (WT) seedlings grown on MS plates were treated with 100 μm ABA for the indicated times. Three independent experiments were performed, each with three replicates. Values are means ± standard deviation. (f) The transcriptional level of AtKPNB1 and subcellular localization of AtKPNB1 were not affected by ABA. Transgenic plants expressing AtKPNB1pro::GFP-AtKPNB1 (top) or protoplasts transiently expressing 35S::GFP-AtKPNB1 (bottom) were treated with 100 μm ABA for 2 h or not treated with ABA (CK), and then photographed.

Next, we investigated the subcellular localization of AtKPNB1. When 35S::GFP-AtKPNB1 was transiently expressed in Nicotiana benthaminana leaves and 35S::AtKPNB1-YFP was transiently expressed in protoplasts of Col-0 leaves, fluorescence was detected in both the cytoplasm and nucleus (Figure 4c), consistent with the subcellular localization of KPNB1 in these two compartments (Lo and Hung, 2006; Frieman et al., 2007). We confirmed the subcellular localization of AtKPNB1 by analysis of the GFP signal in root tissues of AtKPNB1pro::GFP-AtKPNB1 transgenic plants (Figure 4c). AtKPNB1 protein was also detected in guard cells of the transgenic plants (Figure 4d).

Analysis of GUS expression in AtKPNB1::GUS transgenic plants showed that AtKPNB1 was relatively highly expressed in guard cells, consistent with its involvement in ABA response and drought tolerance (Figure 4d). Further results showed that the level and pattern of AtKPNB1 transcripts and the level and subcellular localization of AtKPNB1 in Col-0 seedlings were not significantly affected by ABA treatment (Figure 4e,f).

AtKPNB1 functions in NLS-mediated import of nuclear proteins

Because AtKPNB1 is the sole human KPNB1 homolog in Arabidopsis, we speculated that AtKPNB1 may play a role similar to human KPNB1 in the NLS-dependent pathway for nuclear protein import. Therefore, we performed a protoplast assay to analyze the effect of the atkpnb1 mutation on the subcellular localization of modified cytosolic chalcone synthase (CHS), which contains an in-frame GFP–NLS tag fusion and an inactivated nuclear export signal [NES(–)Rev] (Haasen et al., 1999). Exclusively nuclear-localized GFP signal was much lower in atkpnb1 than in Col-0 (Figure 5 and Table 1). To confirm this result, we detected the GFP protein by Western blotting. As shown in Figure 5(b), atkpnb1 accumulated less GFP in the nucleus but more in the cytoplasm than Col-0, indicating that NLS-mediated import of nuclear proteins is reduced in the atkpnb1 mutant.

Table 1. Subcellular localization of GFP-NLS-CHS-NES(−)Rev in wild-type and atkpnb1 protoplasts as indicated by the percentage of protoplasts emitting fluorescence from the nucleus, nucleus + cytoplasm, and cytoplasm
GenotypeNucleusNucleus + cytoplasmCytoplasm
  1. Three independent experiments were performed, and >240 protoplasts were analyzed for each experiment. Values are means ± standard deviation.

Wild-type76.2 ± 3.5%19.0 ± 1.8%4.8 ± 4.3%
atkpnb1 56.5 ± 2.9%26.1 ± 2.6%17.4 ± 3.7%
Figure 5.

NLS-mediated nuclear protein import is suppressed in the atkpnb1 mutant. (a) Arabidopsis protoplasts were transiently transformed with the GFP-NLS-CHS-NES(–)Rev construct, and fluorescence was observed 16 h after transformation. (b) Subcellular distribution of the GFP protein. Protoplast proteins were extracted, followed by subcellular fractionation and Western blotting. T, total protein; N, nuclear fraction; C, cytosolic fraction. The markers used were histone H3 (nuclei), cFBPase (cytosolic fraction) and GFP (soluble fraction and loading control). The Western blots are representative of three independent experiments. The protein levels were quantified using ImageJ software (National Institutes of Health). The ratios of total protein anti-GFP/anti-H3 or anti-GFP/anti-cFBPase values were set to 1.0, and the nuclear fraction (GFP/anti-H3) and cytosolic fraction (anti-GFP/anti-cFBPase) values were compared to total protein.

AtKPNB1 interacts with importin α, nucleoporin AtNUP62 and Ran proteins

In humans, KPNB1-dependent nuclear import of an NLS-containing cargo protein is dependent on formation of a KPNB1–importin α–NLS cargo complex, in which KPNB1 interacts directly with importin α. To demonstrate that AtKPNB1 functions as an importin β, we first detected the interaction between AtKPNB1 and importin α proteins. Among nine importin α proteins found in Arabidopsis, AtIMPA1 and AtIMPA2 shared the highest homology with human KPNA1 and KPNA2 (52.7 and 44.8% identity, respectively), which have been shown to interact with human KPNB1 during NLS nuclear protein import (Frieman et al., 2007). Yeast two-hybrid (Y2H) analyses showed that AtKPNB1 directly interacts with AtIMPA1 and AtIMPA2 (Figure 6a). Interactions between AtKPNB1 and these two proteins in plant cells were confirmed by a bimolecular fluorescence complementation (BiFC) assay (Figure 6b). YFP fluorescence was observed in the cytoplasm and nucleus of the plant cells co-expressing AtKPNB1-YFPN with IMPA1-YFPC or IMPA2-YFPC. Next, we tested the interaction between AtKPNB1 and four other putative importin α proteins using Y2H analysis, BiFC analysis and Western blotting. The results showed that AtKPNB1 directly interacted with AtIMPA4 and AtIMPA6, but not with AtIMPA3 or AtIMPA9 (Figure S7).

Figure 6.

AtKPNB1 directly interacts with imporin α, nucleoporin and Ran proteins. (a) Y2H assay of the interactions between AtKPNB1 and IMPA1, IMPA2, AtNUP62 and Ran proteins. Interaction was assayed by growth on SD plates lacking Trp, Leu, His and Ade. (b) Confirmation of positive interactions by BiFC analysis. N. benthaminana leaves were co-transformed with constructs expressing the indicated YFP N-terminal (YFPN) and YFP C-terminal (YFPC) fusions, and YFP was imaged 48 h after transformation.

In humans, formation of the trimeric KPNB1-importin α-NLS cargo complex is followed by interaction between KPNB1 and a nucleoporin such as NUP62 (Leng et al., 2007; Lott et al., 2010), which enables docking of the complex to the nuclear pore complex, and facilitates translocation of the complex into the nucleus (Merkle, 2003; Conti et al., 2006). We identified an Arabidopsis gene with highest homology to human NUP62, which is a single-copy gene that encodes a 73 kDa protein of unknown function. Using Y2H and BiFC analysis, we detected an interaction between AtKPNB1 and AtNUP62 at the nuclear periphery (Figure 6a,b).

In the nucleus, binding of Ran-GTP-Binding protein to human KPNB1 is essential for dissociation of the import complex, release of the cargo, and subsequent recycling of KPNB1 to the cytoplasm (Leng et al., 2007; Lott et al., 2010). There are four Ran proteins in Arabidopsis, of which Ran1, Ran2 and Ran3 share high homology. These three Ran proteins physically interacted with AtKPNB1 in yeast (Figure 6a). BiFC analysis showed that YFP fluorescence was clearly localized in the cytoplasm and nucleus (Figure 6b). These results indicate that AtKPNB1 functions as an importin β to transport the cargo into the nucleus.

The atkpnb1 mutation affects ABA-induced gene expression

Nuclear proteins play important roles in modulating gene expression. Because inactivation of AtKPNB1 impairs nuclear import of NLS-containing proteins and confers hypersensitivity to ABA during germination and seedling development, we compared expression of ABA-responsive genes such as ABI1, ABI2, ABI3, ABI5, Em1, Em6, RD29A, RD29B and RAB18 in germinating atkpnb1 and Col-0 seedlings. Expression levels of ABI1, ABI2, ABI3 and ABI5 were comparable in ABA-treated Col-0 versus atkpnb1 seedlings (Figure 7). The expression level of RD29A was slightly up-regulated in the atkpnb1 seedlings. In contrast, the levels of Em1, Em6, RAB18 and RD29B were more up-regulated in ABA-treated atkpnb1 seedlings than in Col-0 seedlings (Figure 7). Em1 and Em6 are direct targets of the transcription factor ABI5, while RD29A, RAB18 and RD29B are downstream ABA-responsive genes. These results suggest that AtKPNB1 does not modulate transcription of ABI1, ABI2, ABI3 or ABI5, but mediates ABA signaling through transcriptional regulation of the downstream pathway components.

Figure 7.

ABA response of typical ABA-responsive genes in wild-type and atkpnb1 seedlings. Seeds of the wild-type and atkpnb1 germinated on MS medium without or with 0.1 μm ABA. RNA was extracted from 2-day-old seedlings, and quantitative RT-RCR was performed. Three independent experiments were performed with similar results, each with three replicates. Values are means ± SE. Asterisks indicate statistically significant differences compared with wild-type (Student′s t test, **< 0.01).

atkpnb1 does not alter the pattern of subcellular localization of ABI1, ABI2 or ABI5

ABI1 and ABI2 proteins are key negative regulators of ABA signaling, and function in both the cytoplasm and the nucleus (Merlot et al., 2001; Moes et al., 2008). We hypothesized that the hypersensitivity of atkpnb1 to ABA may be due to impaired nuclear import, leading to higher cytosolic but lower nuclear levels of ABI1 and ABI2 proteins. To test this hypothesis, we transiently transformed Arabidopsis protoplasts from atkpnb1 and Col-0 with 35S::ABI1-YFP and 35S::ABI2-YFP constructs. However, ABI1 and ABI2 appeared to localize similarly to both cytoplasmic and nucleus, and the pattern of localization was not affected by inactivation of AtKPNB1 (Figure 8a).

Figure 8.

Subcellular localization of ABI1 and ABI2 in atkpnb1, and the genetic relationship between AtKPNB1 and ABI1. (a) AtKPNB1 does not regulate nuclear import of ABI1, ABI2 or ABI5. 35S::ABI1-YFP, 35S::ABI2-YFP or 35S::ABI5-YFP were transiently transformed into WT and atkpnb1 protoplasts. (b,c) Effect of ABA on germination and cotyledon greening of the wild-type and atkpnb1, abi1-2 and atkpnb1 abi1-2 mutants. Seeds of the wild-type, atkpnb1, abi1-2 and atkpnb1 abi1-2 were plated on medium containing 0.1 μm ABA. Plates were photographed (b) at 4 days after stratification. Germination and greening (c) were examined at 1.5 and 3 days after stratification, respectively.

ABI5 is a transcription factor that is required for post-germination growth arrest of seedlings in response to ABA and osmotic stress (Miura et al., 2009; Liu and Stone, 2010). The transcript level of ABI5 was induced to the same extent by ABA in Col-0 and atkpnb1 seedlings, but the transcript levels of ABI5-regulated genes were more highly induced in the mutant than in Col-0 (Figure 7), raising the possibility that AtKPNB1 may be required for nuclear import of negative factors of ABI5. We transformed atkpnb1 and wild-type protoplasts with 35S::ABI5-YFP, and found no obvious difference in the nuclear abundance of ABI5 in atkpnb1 and Col-0 protoplasts (Figure 8a). These data suggest that AtKPNB1 does not mediate transport of ABI5 into the nucleus.

To investigate the genetic relationship between AtKPNB1 and ABI1 or ABI5, we crossed atkpnb1 to abi1-2 (Saez et al., 2006) and abi5-8 (Zheng et al., 2012) to generated double mutants. In growth medium containing 0.1 μm ABA, germination and greening were more inhibited for atkpnb1 and abi1-2 mutants than for Col-0, and ABA sensitivity was lower for abi1-2 than for atkpnb1 (Figure 8b,c). In contrast, germination and greening were more severely inhibited by ABA for the atkpnb1 abi1-2 double mutant than for either single mutant (Figure 8b,c). These results suggest that AtKPNB1 and ABI1 act together in an additive manner in the ABA response. When atkpnb1 was crossed with an ABA-insensitive knockdown mutant abi5-8, the resulting atkpnb1 abi58 double mutant displayed a phenotype intermediate between that of the atkpnb1 and abi58 single mutants, regardless of the presence or absence of ABA (Figure S8A,B). Further expression analysis showed that the increased transcription levels of Em1, Em6, RAB18 and RD29A were reduced by introduction of abi58 (Figure S8C). These findings suggest that AtKPNB1 negatively mediates ABA inhibition of seed germination independently of ABI5.

Discussion

Nuclear proteins such as transcription factors, type 2C protein phosphatases and mRNA processing factors have been shown to function in ABA signaling (Hirayama and Shinozaki, 2007; Saez et al., 2008; Cutler et al., 2010; Antoni et al., 2011). However, the unresolved and important question remains as to how these proteins are selectively targeted to the nucleus in order to spatially and temporally mediate ABA signaling and stress responses. Here, we report that AtKPNB1 is a homolog of human KPNB1 that plays important roles not only in Arabidopsis development but also in the ABA-mediated osmotic stress response and drought tolerance. AtKPNB1 was found to interact with several Arabidopsis proteins, including importin α proteins, a nucleoporin and several Ran proteins that were homologous to components of the human KPNB1-dependent nuclear protein import machinery. AtKPNB1 mediated NLS-dependent nuclear protein import. These findings reveal a molecular mechanism that mediates nuclear protein import and fine-tunes expression of a set of ABA- and/or stress-induced genes for an optimal response to environmental or developmental stimuli.

An important question is whether AtKPNB1 controls nuclear import of NLS-containing proteins through the classical nuclear import pathway. In this study, we have demonstrated that loss of function of AtKPNB1 resulted in a significant reduction of nuclear accumulation of a model GFP-NLS-tagged protein (Figure 5), confirming the role of AtKPNB1 in nuclear import of NLS-containing proteins.

Human KPNB1 carries cargos through nuclear pore complexes in the nuclear protein import pathway (Merkle, 2003; Conti et al., 2006). Interaction of human KPNB1 with FG-repeat domains of nucleoporins, in which two phenylalanines (F) or a phenylalanine and a glycine (G) constitute a hydrophobic cluster, is considered important for movement of the cargo/carrier complex through nuclear pore complexes (Frey and Görlich, 2007). On the nuclear face of the nuclear envelope, human KPNB1 binds to Ran-GTP to release importin α-cargos from the nuclear pore complex. Like human KPNB1, AtKPNB1 is localized in both the cytoplasm and nucleus (Figure 4c) and contains the characteristic N-terminus (IBN_N) of importin β, a nucleoporin-binding region and a conserved Ran-binding domain for binding to Ran proteins (Figure S6B). Importantly, we showed that AtKPNB1 directly interacts with several Arabidopsis importin α proteins, AtNUP62 and Ran GTPases (Figure 6 and Figure S7), which indicates that the nuclear protein import machinery containing AtKPNB1 probably has components that are structurally and functionally similar to proteins that function in KPNB1-dependent nuclear protein import in humans. Importin α determines the specificity of cargos (Goldfarb et al., 2004). Because AtKPNB1 interacts with AtIMPA1, AtIMPA2, AtIMPA4 and AtIMPA6 (Figure 6 and Figure S7), it is plausible that regulation of cellular processes such as germination, development and the ABA-mediated stress response by AtKPNB1 may involve different importin α proteins. Human KPNB1 may also transport NLS-containing proteins into the nucleus without involvement of importin α (Merkle, 2003). It is possible that AtKPNB1 may also directly transport some cargo proteins into the nucleus.

Human KPNB1 plays a role in nucleo-cytoplasmic transport of a wide variety of cargo proteins, including histones, ribosomal proteins and signal transducers (Jäkel and Görlich, 1998). Therefore, it is not surprising that, as is the case with other importin β proteins in Arabidopsis (Bollman et al., 2003; Li and Chen, 2003; Wang et al., 2011), inactivation of AtKPNB1 (Figure S2) results in several altered phenotypes. AtKPNB1 was expressed in multiple tissues and organs from seed germination to the reproductive stage, with the highest expression in seeds and the lowest expression in hypocotyls, suggesting that this gene has different roles in different tissues or organs (Figure 4b). Over-expression of AtKPNB1 from the CaMV 35S promoter enhanced plant development and induced early flowering (Figure S5), indicating that correct levels of AtKPNB1 in different tissues or organs are required for normal development and flowering in Arabidopsis. Based on a previous report showing that nuclear import time and transport efficiency are largely determined by the concentration of importin β (Yang and Musser, 2006), we speculate that the concentration of importin β may explain why AtKPNB1 acts as a positive effector of growth and development in Arabidopsis.

The most striking phenotype of the atkpnb1 mutant is an altered response to ABA and drought (Figures 1 and 2). When germinating seeds are exposed to osmotic stress, post-germination growth arrest through an ABA-dependent pathway maintains seedlings in an embryonic state and protects them from stress damage (Kinoshita et al., 2010). This process was enhanced in atkpnb1 (Figure 2), suggesting that AtKPNB1 modulates nuclear import of proteins that negatively control ABA-mediated post-germination arrest in response to stress. Interestingly, we found that a mutation in AtKPNB1 increased the sensitivity of stomatal closure to ABA, resulting in significantly reduced water loss and increased drought tolerance (Figure 1). The enhanced drought tolerance of atkpnb1 is not due to delayed development or the size of plants (Figure 1 and Figure S1). These results suggest that AtKPNB1 plays a critical role in controlling ABA-induced stomatal closure under drought conditions by modulating the nuclear import of proteins that control stomatal movement.

The next question is which proteins are translocated from cytoplasm to nucleus via AtKPNB1? In response to osmotic stress and drought, ABA signal transduction pathways are activated, resulting in increased expression of ABA-responsive genes, leading to plant responses such as post-germination growth arrest and stomatal closure. ABI1 and ABI2 are two key upstream negative regulators in the ABA signaling pathway that function in both the cytoplasm and the nucleus (Umezawa et al., 2010). Failure of these proteins to enter the nucleus causes increased sensitivity to ABA (Moes et al., 2008). Our experimental data show that AtKPNB1 is not responsible for ABI1 or ABI2 import into the nucleus (Figure 8a). Analysis of ABA inhibition of seed germination and cotyledon greening of atkpnb1, abi1-2 and atkpnb1 abi1-2 mutants established that AtKPNB1 and ABI1 function independently in the ABA signaling pathway relative to seed germination and cotyledon greening (Figure 8b,c). Thus, we conclude that nuclear import of ABI1 is not involved in the modulation of ABA signaling by AtKPNB1, and this is supported by genetic evidence showing that ABI1 and AtKPNB1 function in independent pathways (Figure 8b,c). Perhaps AtKPNB1 is involved in nuclear translocation of other type 2C protein phosphatases that mediate ABA signaling.

ABI5 is a key positive regulator of ABA-induced post-germination growth arrest. As a transcription factor, ABI5 must be translocated into the nucleus to activate downstream genes. ABI5 over-accumulation in the nucleus may also cause ABA-hypersensitive phenotypes. However, our results demonstrate that the mutation in AtKPNB1 does not affect subcellular localization of the ABI5 protein (Figure 8a), although ABI5 contains a putative bipartite NLS at the C-terminal end (cNLS Mapper, http://nls-mapper.iab.keio.ac.jp). As there are many other importin β proteins in Arabidopsis, it is possible that nuclear translocation of ABI5 is mediated by other importin β proteins. The results of genetic epistasis analysis between atkpnb1 and abi5-8 suggest that AtKPNB1 modulates post-germinative development arrest induced by ABA independently of ABI5. The partial suppression of the ABA sensitivity of the atkpnb1 mutant by abi5-8 during seed germination and cotyledon greening (Figure S8A,B) suggests that AtKPNB1 may mediate the nuclear localization of a factor that negatively regulates ABI5 expression at the protein level.

AtKPNB1 is different from SAD2, another importin β super-family protein that may be involved in the ABA-mediated stress response. Mutation in AtKPNB1 delayed flowering time and increased sucrose sensitivity (Figures S2 and S3); in contrast, loss of function of SAD2 resulted in early flowering but no change in sucrose sensitivity (Frieman et al., 2007). Also, expression of RD29A and RAB18 in untreated seedlings was significantly higher in the sad2 mutant than in the wild-type but was unaffected or decreased in the atkpnb1 mutant (Verslues et al., 2006) (Figure 7).

Our results suggest several avenues for future research. At present, the identity of the NLS-containing proteins that are transported into nucleus by the AtKPNB1-importin α complex to mount the ABA response and confer drought tolerance remains unknown. A related question is whether AtKPNB1 participates in NLS-independent import of ABA pathway proteins to the nucleus. It will also be of interest to identify the cargos carried by AtKPNB1 during development. Further investigation is required to determine the molecular details of AtKPNB1-mediated nuclear import.

Experimental procedures

Plant materials, growth conditions and genetic analysis

Arabidopsis thaliana ecotype Col-0 was used as the wild-type in this study. The T-DNA insertion mutant atkpnb1 (GABI-Kat ID 180F05) was obtained from the Nottingham Arabidopsis Stock Centre (http://arabidopsis.info/). The seeds were germinated and grown on MS medium or MS medium supplemented with the specified concentrations of ABA, NaCl, mannitol or sucrose, as described previously (Kinoshita et al., 2010). The percentage of germinated seeds and percentage of seedlings with green cotyledons were recorded at the specified time points. Two-week-old seedlings were transplanted into soil and grown under a 16 h light/8 h dark photoperiod at 23°C.

To obtain double mutants atkpnb1 was crossed to abi1-2 and abi5-8, and the resulting F2 progeny were used to identify homozygous double mutant lines by PCR screening. Total genomic DNA was extracted from 100 mg of leaves for PCR. The primer pairs for PCR are listed in Table S1.

Construction and plant transformation

For over-expression and subcellular localization analysis, the 35S::GFP-AtKPNB1 construct, in which expression of GFP-AtKPNB1 is driven by the 35S promoter, was generated using a binary vector PEZR(K)-LC (Kaiserli and Jenkins, 2007). The AtKPNB1 coding region without the stop codon was prepared by PCR using the wild-type cDNA as template and primers GFP-F and GFP-R (Table S1). The constructs were introduced into Agrobacterium tumefaciens strain GV3101 for Col-0 transformation, or into epidermal N. benthaminana leaf cells for transient expression. For complementation, a 1397 bp promoter fragment of AtKPNB1 was amplified from Col-0 genomic DNA. This fragment was ligated into PEZR(K)-LC to replace the 35S promoter. Then the AtKPNB1 coding region without the stop codon was cloned into the construct to generate AtKPNB1pro::GFP-AtKPNB1. The construct was introduced into A. tumefaciens strain GV3101 and transformed into atkpnb1 by floral infiltration (Zhang et al., 2006). The 1397 bp promoter was also cloned into pCAMBIA1391 (Cambia lab, http://www.cambia.org/daisy/cambia/2058.html) to get AtKPNB1pro::GUS. The AtKPNB1pro::GUS construct was transformed into Col-0.

To transiently express AtKPNB1, ABI1 and ABI2 in Arabidopsis protoplasts, constructs were generated in Gateway-compatible vectors. To this end, the coding sequences of AtKPNB1, ABI1 and ABI2 without stop codons were amplified using the primer pairs listed in Table S1, cloned into the pDONR221 entry vector (Invitrogen, http://invitrogen.com), and then recombined into the P2GWY7 destination vector (Karimi et al., 2007). Protoplasts were prepared from leaves of 3- to 5-week-old WT and atkpnb1 plants grown on MS medium as described by Wu et al. (2009). Fluorescence was observed 16 h after transformation using a Leica TCS SP2 (http://www.leica-microsystems.com/) confocal microscope. For detection of transformed GFP protein, the protoplast protein was extracted according as described by Ek-Ramos et al. (2010): the transformed protoplasts were shaken gently in buffer containing 10 mm MES/HCl (pH 5.7), 1 m sucrose, 5 mm MgCl2, 2 mm β-mercaptoethanol, 10 μl ml−1 phophatase inhibitors, 10 μl ml−1 proteinase inhibitors, 10 μm MG-132 (Millipore, http://www.millipore.com/catalogue/item/474790-1mg) and 0.2% Triton X-100. After centrifugation at 5000 g for 10 min at 4°C, the pellet was washed three times (5000 g, 10 min at 4°C) in the buffer and used as the nuclei fraction, and the supernatant was used as the total protein fraction. The samples were separated by SDS-PAGE and analyzed by immunoblotting.

Protein-protein analysis

The constructs were created in two pairs of Gateway-compatible destination vectors: pGBKT7-DEST (BD) with pGBAD7-DEST (AD) and pEarleyGate201-YN (N-terminal YFP) with pEarleyGate202-YC (C-terminal YFP) (Lu et al., 2010). The coding sequences of AtIMPA genes, AtNUP62, Ran1, Ran2, Ran3 were amplified from Col-0 cDNA using the primer pairs listed in Table S1, inserted into pDONR221 and then recombined in the appropriate destination vector.

Y2H and BiFC assays were performed as previously described (Lu et al., 2010; Liu et al., 2012). For Y2H, Saccharomyces cerevisiae strain Y2HGold (Clontech, http://www.clontech.com/) was used for co-transformation of the AD and BD constructs. A series of 10 μl aliquots of diluted co-transformed Y2HGold culture was spotted onto SD plates lacking Trp and Leu or lacking Trp, Leu, His and Ade, and incubated at 30°C for 3–5 days to observe yeast growth. Plasmids pGBKT7 and pGADT7-Rec (Clotech, http:www.clotech.com) were used as negative controls. For the BiFC assay, A. tumefaciens carrying the YFP N-terminal and YFP C-terminal fusion constructs was infiltrated into N. benthaminana leaves as described by Sparkes et al. (2006). The reconstituted YFP signals were observed using confocal imaging 2 days after infiltration. Empty vectors were used as negative controls.

GUS histochemical analysis

The GUS assay was performed as previously described (Zhao et al., 2011). The seedlings or tissues were harvested and incubated in freshly prepared buffer containing 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid for 4 h at 37°C in the dark, followed by clearing with 70% ethanol. Eight independent T3 transgenic lines were analyzed under a stereomicroscope; representative images were photographed.

Drought treatment, water loss analysis, and stomatal aperture measurement

For measurement of drought tolerance, water was withheld from 21-day-old wild-type plants and 23-day-old atkpnb1 mutant plants, which were of comparable size and growing in pots. After 19 days of drought treatment, survival rates were determined, and the plants were watered; the plants were photographed 4 days after re-watering. For measurement of water loss from detached leaves, five to eight rosette leaves per plant were detached from 3-week-old watered plants and weighed at the indicated times. To analyze stomatal function, we incubated rosette leaves in a solution containing 50 mm KCl, 10 mm CaCl2 and 10 mm MES (pH 6.15) for 2 h under light. ABA was then added to the solution to a final concentration of 1 μm. After ABA treatment for 2 h, stomatal apertures were measured as described previously (Ren et al., 2010).

Gene expression analysis

Total RNA was extracted using a plant/fungi RNA purification kit (Norgen, catalog number 25800; http://www.norgenbiotek.com/). Equal amounts of RNA samples (1 μg each) were used for reverse transcription with iScript™ Reverse Transcription Supermix (Bio-Rad, http://www.bio-rad.com) according to the manufacturer's instructions. The primer pair P2/P3 was used in an RT-PCR assay to quantify the expression level of AtKPNB1 in mutant and over-expression lines. The primer pairs used for quantitative RT-PCR are listed in Table S1. ACTIN2 was used as the internal control for quantitative RT-PCR using SsoFast EvaGreen Supermix (Bio-Rad).

Accession numbers

Sequences of the genes in this paper may be found in the GenBank/EMBL database library under the following accession numbers: At5g53480 (AtKPNB1), At4g26080 (ABI1), At5g57050 (ABI2), At2g36270 (ABI5), At3g06720 (AtIMPA1), At4g16143 (AtIMPA2), At4g02150 (AtIMPA3), At1g09270 (AtIMPA4), At1g02690 (AtIMPA6), At5g03070 (AtIMPA9), At2g45000 (AtNUP62), At5g20010 (Ran1), At5g20020 (Ran2), At5g55190 (Ran3), At3g18780 (Actin2), At5g52310 (RD29A), At5g52300 (RD29B), At3g51810 (Em1), At2g40170 (Em6), At5g66400 (RAB18), and At4g37490 (cyclin B1).

Acknowledgements

We thank Yuhai Cui (Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, London, Ontario, Canada) for providing plasmids pGBKT7-DEST, pGBAD7-DEST, pEarleyGate201-YN and pEarleyGate202-YC. We thank Aiming Wang (Southern Crop Protection and Food Research Centre, Agriculture and Agri-Food Canada, London, Ontario, Canada) for providing the plasmid P2GWY7. We are grateful to Chenlong Li and Hongguang Cui for technical assistance and some valuable suggestions. We are grateful to Thomas Merkle (University of Freiburg, Institute of Biology II, Germany) for kindly providing us with the GFP-NLS-CHS-NES(-)Rev plasmid. We thank the Nottingham Arabidopsis Stock Centre and the Arabidopsis Biological Resource Center for providing T-DNA lines, and Pedro L. Rodriguez (University of Politècnica de València, Instituto de BiologíaMolecular y Celular de Plantas, Spain) for abi1-2 seeds. We would like to acknowledge support from the National Program on Key Basic Research Project (2012CB114300), the Main Direction Program of Knowledge Innovation of the Chinese Academy of Sciences (KSCX2-EW-J-5), and the National Transgenic Key Project of the Ministry of Agriculture of China (2011ZX08009-003-002). Y.J.L. was the recipient of a scholarship from the China Scholarship Council.

Ancillary