SKP1 is involved in abscisic acid signalling to regulate seed germination, stomatal opening and root growth in Arabidopsis thaliana

Authors

  • CHIJUN LI,

    1. Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China
    2. Graduate University of the Chinese Academy of Sciences, Beijing 100049, China
    Search for more papers by this author
    • Equal contributors to this work.

  • ZUOJUN LIU,

    1. Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China
    2. Graduate University of the Chinese Academy of Sciences, Beijing 100049, China
    Search for more papers by this author
    • Equal contributors to this work.

  • QIRUI ZHANG,

    1. Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China
    2. Graduate University of the Chinese Academy of Sciences, Beijing 100049, China
    Search for more papers by this author
  • RUOZHONG WANG,

    1. College of Bioscience and Biotechnology, Hunan Agricultural University, Changsha 410128, China
    Search for more papers by this author
  • LANGTAO XIAO,

    1. College of Bioscience and Biotechnology, Hunan Agricultural University, Changsha 410128, China
    Search for more papers by this author
  • HONG MA,

    1. State Key Laboratory of Genetic Engineering, Institute of Plant Biology, School of Life Sciences, Fudan University, Shanghai 200433, China
    Search for more papers by this author
  • KANG CHONG,

    1. Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China
    Search for more papers by this author
  • YUNYUAN XU

    Corresponding author
    1. Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China
    Search for more papers by this author

Y. Xu. Fax: +86 10 82594821; e-mail: xuyy@ibcas.ac.cn

ABSTRACT

Abscisic acid (ABA) regulates many aspects of plant development, including seed dormancy and germination, root growth and stomatal closure. Plant SKP1 proteins are subunits of the SCF complex E3 ligases, which regulate several phytohormone signalling pathways through protein degradation. However, little is known about SKP1 proteins participating in ABA signalling. Here, we report that the overexpression of Triticum aestivum SKP1-like 1 (TSK1) in Arabidopsis thaliana (Arabidopsis) resulted in delayed seed germination and hypersensitivity to ABA. The opening of stomatal guard cells and the transcription of several ABA-responsive genes were affected in transgenic plants. In contrast, Arabidopsis skp1-like 1 (ask1)/ask1 ASK2/ask2 seedlings exhibited reduced ABA sensitivity. Furthermore, the transcription of ASK1 and ASK2 was down-regulated in abi1-1 and abi5-1 mutants compared with that in wild type. ASK1 or ASK2 overexpression could rescue or partially rescue the ABA insensitivity of abi5-1 mutants, respectively. Our work demonstrates that SKP1 is involved in ABA signalling and that SKP1-like genes may positively regulate ABA signalling by SCF-mediated protein degradation.

INTRODUCTION

S-phase kinase-associated protein 1 (SKP1) is a component of a Skp1-Cullin1-F-box (SCF) complex (Hotton & Callis 2008) that facilitates ubiquitin-mediated protein degradation in eukaryotes. The SCF-type E3 ligase is composed of four major subunits: Cullin (CDC53 in yeast), SKP1, a RING finger protein (RBX1/HRT1/ROC1) and an F-box protein. Among them, SKP1 acts as an adaptor to link Cullin1 (Cul1) to F-box protein, which specifically recognizes different target proteins via a variable C-terminal domain and interacts with SKP1 through an N-terminal structural motif (Bai et al. 1996; Connelly & Hieter 1996).

SKP1 homologues have been identified in eukaryotes since the discovery of P19, the first SKP1, in humans (Zhang et al. 1995). In yeast, analysis of skp1 mutants indicated that SKP1 is vital for both the transition from G1 to S and from G2 to M phases (Bai et al. 1996; Connelly & Hieter 1996). In plants, Arabidopsis Skp1-Like 1 (ASK1) was the first SKP1 homologue isolated. Mutation of ASK1 resulted in male sterility and some defects in organ development (Yang et al. 1999). ASK1 interacts with unusual floral organs (UFO) to regulate the expression of the floral organ identity genes APETALA3 (AP3) and PISTILLATA (PI) (Zhao et al. 1999) and with reduced male fertility (RMF) to regulate tapetum degeneration (Kim, Jung & Park 2010). The root growth of the ask1-1 mutant was insensitive to exogenous auxin treatment, suggesting that ASK1 is involved in auxin response (del Pozo et al. 2002). ASK2, another SKP1 homologue in Arabidopsis, was able to substitute for ASK1 during male meiosis (Zhao et al. 2003a). Interestingly, ASK1 and ASK2 together were essential for normal embryo development and seedling viability according to the analysis of ask1/ask2 double mutants (Liu et al. 2004). There were 21 predicted SKP1 homologues in the Arabidopsis genome; SKP1-like genes were found in other plant species (Kong et al. 2007), including wheat (Li et al. 2006) and rice (Sasaki et al. 2003).

Many F-box proteins, such as UFO and RMF, have been found to interact with ASK1 or ASK2, suggesting that they could form various SCF complexes (Gagne et al. 2002; Zhao et al. 2003a). SCF complexes have been demonstrated to be involved in various hormone signalling pathways, such as the auxin, gibberellin, jasmonate, ethylene and brassinosteroid pathways (Dreher & Callis 2007; Somers & Fujiwara 2009). Among the four components of the SCF complex, the F-box proteins Arabidopsis thaliana TUBBY-like protein 9 (AtTLP9) and drought tolerance repressor (DOR) were found to be involved in abscisic acid (ABA) signalling and interact with ASK1 (Lai et al. 2004; Dreher & Callis 2007) and ASK14 (Zhang et al. 2008), respectively. However, no direct evidence indicated that SKP1 protein is involved in ABA signalling.

ABA is a key plant hormone that regulates many physiological processes, including the maintenance of seed dormancy, inhibition of seed germination and seedling growth, control of stomatal closure and protection against various environmental stresses. Recent data demonstrated that ABA is transported into the cytosol by ATP-binding cassette (ABC) transporters localized at the plasma membrane (Kang et al. 2010; Kuromori et al. 2010). Significant progress was made by the discovery of a soluble ABA receptor in 2009 (for review, see Raghavendra et al. 2010). RCARs/ PYLs were identified as the ABA receptor that can bind to ABA and interact with group A protein phosphatase 2C (PP2C) to inhibit the activity of the phosphatase (Ma et al. 2009; Park et al. 2009). Several group A PP2Cs have been found to be the negative regulators of early ABA signalling including ABI1 and ABI2 (Leung, Merlot & Giraudat 1997), HAB1 and HAB2 (Leonhardt et al. 2004; Saez et al. 2004), and AHG1 and AHG3 (Yoshida et al. 2006; Nishimura et al. 2007).

In the presence of ABA, the inactivation of PP2Cs allows the release of the SNF1-related kinase 2 (SnRK2), which phosphorylates the downstream targets, including the basic leucine zipper (bZIP) transcription factors and ion channels (Geiger et al. 2009; Sato et al. 2009). The regulation of SnRK2 activity by PP2Cs is the upstream step in ABA signal transduction (Nakashima et al. 2009). The transcription factors playing a positive role include ABI3 of B3-containing proteins (Giraudat et al. 1992), bZIP-type transcription factors (ABI5, ABF1, ABF2, ABF3 and ABF4) (Finkelstein & Lynch 2000; Kang et al. 2002; Choi et al. 2005; Finkelstein et al. 2005; Fujita et al. 2005), the AP2-type transcription factor ABI4 (Finkelstein et al. 1998), the bHLH-type member AtMYC2 and the MYB family member AtMYB2 (Abe et al. 2003). Constitutive expression of ABI3, ABI4 and ABI5 confers hypersensitivity to ABA (Finkelstein et al. 1998; Lopez-Molina, Mongrand & Chua 2001; Lopez-Molina et al. 2002). These transcription factors regulate downstream ABA-regulated genes such as RD29B containing ABRE cis elements (Himmelbach, Yang & Grill 2003; Nakashima et al. 2006) and SOMONS containing RY cis elements (Park et al. 2011).

Another ABA receptor, ABAR/CHLH, directly interacts with the transcription factor WRKY40 at high level of ABA to recruit WRKY40 from the nucleus to the cytosol and induce ABI5 expression to trigger downstream physiological processes (Shen et al. 2006; Pandey, Nelson & Assmann 2009). ATHB6 from the homeodomain-leucine zipper transcription factor family (Himmelbach et al. 2002) and several WRKY family transcription factors (Shang et al. 2010) are negative regulators of ABA signalling.

Although ASK1 and ASK2 are required for the signal pathways of several plant hormones, the role of SKP1 in ABA signalling remains unclear. In this study, we investigated the relationship between SKP1 and ABA signalling based on the ABA-related phenotypes, taking a clue from the transgenic Arabidopsis lines overexpressing Triticum aestivum SKP1-like 1 (TSK1), an SKP1 homologue from wheat. We also examined the functions of ASK1 and ASK2 in the ABA signalling pathway using the Arabidopsis mutants.

MATERIALS AND METHODS

Plant materials

The TSK1-overexpressing transgenic lines and ask1-1 mutant were in the A. thaliana C24 and Landsberg erecta ecotypes, respectively. The ASK1- and ASK2-overexpressing transgenic lines were in the Col-0 ecotype. The ASK1/ask1 ask2/ask2 plants and TSK1-overexpressing transgenic Arabidopsis were previously described (Liu et al. 2004; Li et al. 2006). The ASK1/ask1 ask2/ask2 plant was crossed with the wild type, and seeds from the ASK1/ask1 ASK2/ask2 plants were collected for further analysis. The abi1-1 and abi5-1 mutants were described previously (Leung et al. 1997; Finkelstein & Lynch 2000). Arabidopsis seeds were maintained at 4 °C for 3 d in the dark to break residual dormancy, and were grown in a greenhouse at 22 °C under a 16 h light/8 h dark cycle.

PCR-based genotyping in mutants

Primers were designed to amplify ASK1, ask1-1, ASK2 and ask2-1 alleles according to previous reports (Xu et al. 2002; Liu et al. 2004; Ni et al. 2004) with slight modifications. The wild-type ASK1 allele was confirmed by A1 (5′-ATG TCT GCG AAG AAG ATT GTG-3′) and oMC383 (5′-GAA GAT AGT CAT GAT TCA TGA AG-3′) primers. The Ds insertion in the ask1-1 mutant was determined by A1 and oMC490 (5′-CGT TCC GTT TTC GTT TTT TAC C-3′). The ASK2 wild-type allele was amplified by B1 (5′-ATG TCG ACG GTG AGA AAA ATC-3′) and oMC593 (5′-AAA TGG GTC GAG GAC ATG AC-3′). The T-DNA insertion in ask2-1 mutant was confirmed by B1 and 320T (5′-CAT TTT ATA ATA ACG CTG CGG ACA TCT AC-3′). The abi5-1 mutant was genotyped by dCAPS (Neff et al. 1998) with the primers for bZIPFOR (5′-CAA TCA ACA ACA AGC AGC AG-3′) and bZIPREV (5′-TCT CTC CAC TAC TTT CTC CAC-3′) (Supporting Information Fig. S2). The wild-type allele produces three DNA fragments of 686, 114 and 26 bp after AvaII digestion of the PCR products, whereas the abi5-1 allele only produces two bands of 800 and 26 bp.

Transformation vectors and construction of transgenic plants

To generate ASK1 and ASK2 overexpressing transgenic plants, the full-length cDNAs of ASK1 (At1g75950) and ASK2 (At5g42190) were amplified, and the fragments were cloned to the BamHI-KpnI site of the pSN1301 vector (Li et al. 2006), in which the transgene expression is under the control of the CaMV 35S promoter. Transformation of the Arabidopsis ecotype Col-0 was performed by the floral dip method using Agrobacterium tumefaciens strain C58 (Clough & Bent 1998). For the phenotypic analysis, the T3 or T4 homozygous lines were used. T3 homozygous transgenic lines were selected through screening seeds harvested from T2 plants on the 1/2 MS medium containing 20 µg mL−1 hygromycin. To generate the 35S::ASK1/abi5-1 and 35S::ASK2/abi5-1 plants, T2 ASK1 and ASK2 overexpressing transgenic lines were crossed with the abi5-1 mutant, respectively, and F2 plants were selected by hygromycin to obtain homozygous plants of 35S::ASK1/abi5-1 and 35S::ASK2/abi5-1. The abi5-1 mutant was identified mentioned above. Homozygous plants of 35S::ASK1/abi5-1 and 35S::ASK2/abi5-1 were used for phenotypic analysis.

Germination assay and the response of the root to ABA

Seeds were collected at the same time. For aseptic growth, seeds were treated with 0.2% Triton X-100 and 10% sodium hypochlorite for 10 min, washed six times with sterile water and put on plates with 1/2 MS medium solidified with 0.8% agar. After incubation at 4 °C for 3 d, the plates were transferred to the greenhouse for germination. The fully emerged radicle tip or the fully opened cotyledon and turning green was used as a standard for identifying the germinated seeds.

Seeds germinated on 1/2 MS plates for 4 d in the greenhouse. The surviving seedlings were transferred to the MS medium containing different concentrations of ABA, and the position of each primary root tip was marked. After 3 or 4 d, the root length was measured. The relative root length of each genotype was the root elongation on 1/2 MS medium containing ABA divided by that without ABA.

Measurement of stomatal aperture

The stomatal aperture was measured on epidermal strips isolated from the abaxial side of leaves of 4–5-week-old plants. Methods for incubation and ABA treatment were described previously (Ephritikhine et al. 1999). Photographs were taken under a Zeiss microscope, and the stomatal aperture was measured by use of ImageJ 1.34 s (http://rsb.info.nih.gov/ij/). Leaves of 10 different plants of the wild-type and transgenic lines were used in each experiment.

Measurement of ABA content

The ABA content in 3-week-old whole plants was measured by high-performance liquid chromatography (HPLC). Fresh materials were carefully weighed, frozen in liquid nitrogen and dried by speed vacuum. The freeze-dried samples were homogenized in 80% methanol in the dark and incubated at 4 °C for 15 h. Samples were centrifuged at 3000 g for 20 min, and the supernatant was collected. The pellet was re-suspended in 2 mL of 80% methanol, stirred for 5 min and centrifuged again at 3000 g for 20 min. The two supernatants were combined, lyophilized, dissolved in 8 mL of 0.1 m ammonium acetate (pH 9.0) and centrifuged at 27 000 g for 20 min. The ABA content in the supernatant was purified by a PVP, a DEAE Sephadex A-25 and a C18 Sep-pak column (Classical; Waters, Milford, MA, USA), and then washed with 50% methanol. The extract was subjected to HPLC (Waters Spherisorb, 4.6 × 25 cm) analysis. The fraction was eluted at a speed of 1 mL min–1 and detected at 262 nm.

Quantitative real-time PCR

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA). For real-time PCR, 2 µg of total RNA was used for reverse transcription with M-MLV Reverse Transcriptase (Promega, Madison, WI, USA). The cDNA samples were diluted 50–100-fold. For each RT-PCR reaction, 15 µL of total volume with gene-specific primers was used with 7.5 µL of SYBR GreenMaster mix (Applied Biosystems, Foster City, CA, USA), 0.25 µL of primers (10 µm) and 5 µL of diluted cDNA sample. Triplicate quantitative assays were performed under an ABI7900HT Fast-Real Time system according to the manufacturer's protocol (Applied Biosystems). The delta-delta Ct method (Yoshida et al. 2003) was used to evaluate quantitative variation among replicates. The amplification of Tubulin was used as an internal control to normalize all data. Primers are listed in Table 1.

Table 1.  Primers for qRT-PCR
Gene nameSequence (5′ to 3′)
  1. F, forward primer; R, reverse primer.

TSK1F: CCAGACTGTTGCTGACATGA
R: GGCGTAAAGTCGTTCTTGATG
ABF3F: CTTTGTTGATGGTGTGAGTGAG
R: GTGTTTCCACTATTACCATTGC
ABF4F: AGAAAACGAATCAAGAACTGCA
R: GCTTCGACGTTTCTTTCAGC
ABI1F: CGGTTCTCAGGTAGCGAAC
R: TCACCATCGCAGAGCATC
ABI2F: CACCGGCAGCTTCTATCCTC
R: ATCACAAACCGGATAGGGATGA
ABI3F: CATGGAGATTCCATTAGACAG
R: GGTGTCAAAGAACTCGTTGCTATC
ABI5F: GTGGAATTGGAAGCTGAACT
R: TCAAAATACTGTTGCTTCCTC
ADH1F: TCGATGCAAAGCTGCTGT
R: TGTTTCTGCGGTGGAGC
ATHB6F: TCTTTATGTCCTACAACTTCCACA
R: ATCCTTCAAGCATCGACTGA
RAB18F: GAACATGGCGTCTTACCAGA
R: TCGTCATACTGCTGCTGGAT
ERA1F: TTTCGGTTGCAAGCATCCTA
R: CCCTCCAATGCCACCTTC
RD29BF: CTTTGGAAAATGGAGTCACAGT
R: TTCTGGATGGTGAATTCTGATT
ASK1F: GAACTCATTCTGGCTGCTAATT
R: TCTTCTGGAGTCTTTCCTTTGA
ASK2F: GCTGCGTCGGGTTCTAG
R: GCAGCCAGGATAAGATCG
TubulinF: ATCCGTGAAGAGTACCCAGAT
R: AAGAACCATGTACTCATCAGC

RESULTS

Overexpressing TSK1 Arabidopsis altered ABA sensitivity regarding germination and root growth

Our previous results indicated that TSK1, an SKP1 homologous gene from wheat, partially rescued the meiotic defect in the ask1-1 mutant, and TSK1 overexpressing transgenic Arabidopsis seedlings exhibited an altered response to auxin (Li et al. 2006).

As known, ABA controls seed dormancy and germination. The inhibition of seed germination assay demonstrated that the germination of TSK1-overexpressing transgenic lines was delayed compared with that of the wild type. Within 16 h, 80% of the wild-type seeds had germinated, whereas less than 20% of the transgenic seeds of five independent transgenic lines had germinated. At the extremeness, few germinated seeds were observed in lines 21 (L21) and 82 (L82) (Fig. 1a,b). This deficient germination suggested that TSK1 overexpression interferes with the response to ABA. To gain more insights into the possible roles of TSK1 in ABA signalling, we examined the responses of TSK1-overexpressing lines to exogenous ABA. Seed germination was significantly inhibited in transgenic lines compared with that in the wild type (Fig. 1c,d). For example, in the presence of 0.3 µm ABA, less than 30% of transgenic seedlings had fully opened cotyledons after 4 d. In contrast, 97% of wild-type seedlings had fully opened cotyledons after 4 d. In the presence of 0.9 µm ABA, the percentage of transgenic seedlings with fully opened cotyledons was less than 10%, compared with approximately 40% for the wild type (Fig. 1d).

Figure 1.

Abscisic acid (ABA) sensitivity of TSK1-overexpressing lines. (a) Representative germination at 16 and 20 h. (b) Germination rates of wild type and TSK1-overexpressing lines. Seeds of C24 wild type and lines L21, L41, L42, L81 and L82 were plated on 1/2 MS medium and incubated at 4 °C for 3 d, after which the germinated seeds were counted (fully emerged radicle tip) at various times. Each data point represents the mean of triplicate experiments (n = 50–200 each). Some standard errors are smaller than the symbols. (c) Germination of representative plants on media containing 0, 0.3 or 0.6 µm ABA. (d) ABA dose–response of germination. Seeds were incubated at 4 °C for 3 d and germinated on 1/2 MS medium containing various concentrations of ABA, and seedlings with fully expanded cotyledons were counted after 4 d. Experiments were performed at least three times (n > 50 each). The bars represent standard errors. (e) Growth of transgenic plants on MS medium containing 0 or 0.3 µm ABA. The representative seeds were germinated and grown for 5 d. Bar = 0.5 cm. (f) Response of root growth to ABA. Seeds were germinated on 1/2 MS without ABA for 4 d, and then seedlings (n > 20) were transferred to 1/2 MS containing 0.3 µm ABA. The relative root length was measured 4 d later. The experiments were repeated at least three times. Bars represent standard errors.

Likewise, the root growth of TSK1 transgenic seedlings was also more inhibited by exogenous ABA than the wild type (Fig. 1e). When the 4-day-old seedlings were transferred to a medium containing 0.3 µm ABA, the relative root length of wild-type seedlings was not affected by growth for 4 d, whereas that of both L21 and L82 was suppressed to less than 30% of the control (Fig. 1f). These results indicated that TSK1-overexpressing transgenic Arabidopsis was hypersensitive to exogenous ABA treatment during seed germination and root growth.

Endogenous ABA biosynthesis was not affected in TSK1-overexpressing lines

To test whether the biosynthesis of ABA was affected in TSK1-overexpressing lines, 10 or 30 µm abamine, an inhibitor of ABA biosynthesis (Han et al. 2004), was added to the medium for seed germination. As shown in Fig. 2a, the seed germination rate of the wild type increased approximately sevenfold within 12 h under abamine treatment. Germination was also promoted in L21 and L82, but was still lower than that of the wild type. For example, in the presence of 30 µm abamine, the germination rate was about 100% in wild type within 18 h. In contrast, less than 30% of seeds germinated in the transgenic lines. Likewise, when 30 µm abamine and 0.6 µm exogenous ABA were added together in the medium, approximately 70% of wild-type seeds had germinated after sowing for 4 d, compared with approximately 31% in L21 and 25% in L82 (Fig. 2b). The wild type, L21 and L82 exhibited similar levels of ABA (0.680 ± 0.049, 0.780 ± 0.044 and 0.745 ± 0.015 ng g−1 fresh weight, respectively) (Fig. 2c), which was not reached to significant difference level in statistics between wild type and the TSK1 overexpression lines. The data suggested that the overexpression of TSK1 affects the ABA signalling pathway rather than the endogenous ABA biosynthesis.

Figure 2.

Abscisic acid (ABA) synthesis in TSK1-overexpressing transgenic lines. (a) Abamine treatment in transgenic lines. Seeds of the wild type and transgenic lines L21 and L82 were germinated on 1/2 MS plates containing 10 or 30 µm abamine. After 3 d of cold treatment, the germination was scored (fully emerged radicle tip) at various times. Each data point represents the mean of triplicate experiments (n = 50–100 each). (b) Seeds were germinated on 1/2 MS plates containing 0.6 µm ABA and 30 µm abamine. Seeds germinated for 4 d after incubation at 4 °C for 3 d. The seedlings with fully emerged cotyledons were counted after 4 d. Experiments were performed at least three times (n = 100 each). Bars represent standard errors. (c) Seeds from wild type and transgenic plants were geminated on MS medium. The ABA content of 3-week-old seedlings was measured (ng g–1 fresh weight). Each data point represents mean ± SE. Experiments were performed three times.

Stomatal aperture was hypersensitive to ABA treatment in TSK1 transgenic lines

It is well known that ABA can induce stomatal closure. To test for the possible effects of TSK1 overexpression on the ABA response of stomatal cell, epidermal strips were first floated on incubation solution for 2 h under light to fully open the stomata before adding ABA. As shown in Fig. 3a, the mean stomatal width in the transgenic lines (3.86 ± 0.24 µm in L21 and 3.75 ± 0.17 µm in L82) was smaller than that of the wild type (5.12 ± 0.38 µm) in the light. The stomatal aperture of the wild type and transgenic plants decreased with increasing exogenous ABA concentrations, and stomatal aperture response in the transgenic plants was more sensitive to the ABA treatment (Fig. 3b). In addition, transgenic Arabidopsis lines displayed enhanced tolerance to drought stress (Fig. 3c). These data suggested that the overexpression of TSK1 resulted in an enhanced stomatal closure response to ABA and enhanced drought tolerance in transgenic plants, and that TSK1 transgenic Arabidopsis was hypersensitive to ABA.

Figure 3.

The stomata opening in wild type and TSK1-overexpressing transgenic lines. (a) Stomatal guard cells of the wild type and transgenic lines (L21 and L82). Arrows indicate the guard cells. Bar = 50 µm. (b) Effect of ABA on the closure of stomata in the wild type and transgenic lines. Epidermal strips were treated with different ABA concentrations under light. The results are expressed as relative stomatal aperture ± SE of 100 stomata. The experiments were repeated three times. The mean aperture width before adding ABA (taken as 100%) was 5.12 ± 0.38 µm in wild-type plants, 3.86 ± 0.24 µm in L21 plants and 3.75 ± 0.17 µm in L82 plants. (c) Drought tolerance of the wild type and transgenic lines. Top, water was withheld for 2 weeks; bottom, after a 2 week drought treatment, the seedlings were re-watered, and photographs were taken after 7 d.

Overexpression of TSK1 altered expression patterns of ABA-responsive genes in transgenic lines

To further investigate the effect of TSK1 in ABA signalling, the expression of various ABA/stress-responsive genes in L21 and L82 was determined by quantitative real-time PCR (qPCR) (Fig. 4). ABI1 and ABI2 encode PP2Cs and negatively regulate ABA signalling (Leung et al. 1997). The qPCR assay indicated that the relative expression level of ABI1 was increased by 2.6- and 8.5-fold in L21 and L82, respectively, compared with that in wild-type C24. The expression level of ABI2 was increased by more than 10-fold in both L21 and L82. The homeodomain protein ATHB6, a negative regulator of the ABA signalling pathway, is a target of ABI1 and acts downstream of ABI1 (Himmelbach et al. 2002). In L21 and L82, the relative expression levels of ATHB6 were reduced to 0.07- and 0.22-fold, respectively, of that in wild type. ERA1 is another negative regulator of ABA signal transduction, and its mutation resulted in ABA hypersensitivity regarding stomatal closing and seed germination (Brady et al. 2003). The qPCR result demonstrated that the ERA1 transcription level was decreased to 0.5- and 0.25-fold in L21 and L82, respectively, of that in the wild type.

Figure 4.

Expression of abscisic acid (ABA)-responsive genes in TSK1-overexpressing transgenic lines. Transcription levels of ABA-responsive genes were determined by quantitative RT-PCR with total RNA from 2-week-old seedlings growing in soil. Tubulin was used as an internal control. The data represent results of experiments performed in triplicate. Error bars represent ± SE.

We also detected several key transcription factors in the ABA signalling pathway, including ABI3, ABI5, ABF3 and ABF4, the constitutive expression of which in Arabidopsis results in hypersensitivity to ABA (Lopez-Molina et al. 2001, 2002; Kang et al. 2002). Compared with the expression in the wild type, the relative expression level of the four genes was increased by more than twofold in the transgenic lines. Our results suggested that TSK1 affected the transcription of ABA signalling components.

To investigate the changes of known ABA up-regulated genes in transgenic lines, RD29B, RAB18 and ADH1 were selected. RD29B is induced by ABA and abiotic stress (Yamaguchi-Shinozaki & Shinozaki 1994). RAB18 and ADH1 encode a dehydrin and an alcohol dehydrogenase, respectively (Lang & Palva 1992; de Bruxelles et al. 1996). Compared with the expression in the wild type, expression level of RD29B was increased by 26.7- and 107.6-fold in L21 and L82, respectively. The transcription levels of RAB18 and ADH1 were also increased, as expected, in the TSK-overexpressing lines L21 and L82. Thus, the overexpression of TSK1 in Arabidopsis resulted in an alteration in the expression of ABA/stress-responsive genes.

Knockdown of SKP1 led to reduced ABA sensitivity in the ask1 and ask2 double mutant

Previous studies demonstrated that both ASK1 and ASK2 are required for Arabidopsis floral organ and embryo development, and ASK2 can partly substitute for ASK1 in male meiosis (Zhao et al. 2003a; Liu et al. 2004). These results indicated a functional overlap and redundancy of ASK1 and ASK2.

To further elucidate a possible role of SKP1 in ABA signalling, we investigated whether the loss of function of ASK1 and ASK2 affected the sensitivity to ABA. Because of the possible functional redundancy between ASK1 and ASK2 and the lethality of the loss of function of both ASK1 and ASK2, the seeds harvested from ASK1/ask1 ASK2/ask2 plants (Ni et al. 2004) were used for ABA response analysis (Fig. 5). As shown in Fig. 5a,b, the genotype of each seedling was determined by PCR with four sets of primers as previously reported (Liu et al. 2004; Ni et al. 2004). Approximately 6% of the mutant seeds did not germinate or produce roots after 4 d of growth on MS medium without ABA. These seeds were genotyped as ask1/ask1 ask2/ask2 double mutants by PCR. Four-day-old seedlings were transferred to MS medium containing 10 µm ABA, and root elongation was measured after 3 d. The seedlings of wild type, ASK1/ask1, ask2/ask2, ASK2/ask2 and ASK1/ask1 ASK2/ask2 exhibited no significant difference in root elongation after ABA treatment. However, the double mutant ask1/ask1 ASK2/ask2 exhibited less ABA sensitivity in relative root length (∼56%) than did the wild type (∼33%) (Fig. 5c). These data suggested that homozygous mutations of ASK1 and heterozygous mutations of ASK2 result in reduced ABA sensitivity in Arabidopsis.

Figure 5.

Sensitivity of progeny from ASK1/ask1 ASK2/ask2 plants to ABA. (a) The schematic map of mutations in ask1-1 and ask2-1, and the primers for genotyping ask1-1 and ask2-1 alleles. (b) PCR identification of ASK1, ask1-1, ASK2 and ask2-1 alleles in seedlings. (c) Response of root elongation to ABA. Seeds obtained from wild-type and ASK1/ask1 ASK2/ask2 plants were germinated on 1/2 MS plates for 4 d. The surviving seedlings were transferred to MS containing 10 µm ABA. After 3 d, the relative root length was measured. Each data point represents the mean of experiments performed in triplicate (n = 150–200 each).

ASK1 and ASK2 act downstream of ABI5

The Arabidopsis ABI5, which encodes a bZIP transcription factor, is involved in ABA and stress responses from the late embryonic stage through early seedling development (Finkelstein & Lynch 2000). The decreased ABA sensitivity of the ask1/ask1 ASK2/ask2 double mutant allowed us to address whether ASK1 and ASK2 expression is regulated by ABA signalling. The expression of ASK1 and ASK2 in the ABA-insensitive mutants abi1-1 and abi5-1 was detected by qPCR. Compared with the expression in the wild type, the expression levels of ASK1 and ASK2 were decreased to 0.6- and 0.4-fold, respectively, in the abi1-1 mutant, and 0.1- and 0.08-fold, respectively, in the abi5-1 mutant (Fig. 6a). This result suggested that ASK1 and ASK2 acts downstream of ABI5 to mediate ABA signalling.

Figure 6.

ASK1 and ASK2 act downstream of ABI5. (a) The transcription levels of ASK1 and ASK2 are down-regulated in abi1-1 and abi5-1 mutants. Tubulin was used as an internal control. The data represent results of experiments performed in triplicate. Error bars represent ± SE. (b) Germination rates of ASK1- or ASK2-overexpressing lines in the wild-type or abi5-1 background. Seeds with emerged radicles were counted after growing 5 d in plates containing different concentrations of abscisic acid (ABA); data are expressed as the means of three independent experiments with n ≈ 100 seeds for each genotype. (c) Cotyledon turning green rates of ASK1- or ASK2-overexpressing lines in the wild-type or abi5-1 background. Seeds with fully expanded and greened cotyledons were counted after growing 5 d in plates containing different concentrations of ABA; data are expressed as the means of three independent experiments with n ≈ 100 seeds for each genotype. (d) Growth of ASK1- or ASK2-overexpressing lines in either the wild-type or abi5-1 background on 1/2 MS plates containing 0.5 µm ABA. The picture was taken after seeds were germinated and allowed to grow for 5 d. (e) Root elongation response of ASK1- or ASK2-overexpressing lines in the wild-type or abi5-1 background to ABA. Seeds were germinated on 1/2 MS plates for 4 d. The seedlings were then transferred to 1/2 MS containing 1% sucrose and 10 µm ABA. After 3 d, the relative root length was measured. Each data point represents the mean of experiments performed in triplicate (n = 30–50 each). (f) Response of root elongation to ABA. Seeds obtained from wild-type and ASK1/ask1 ASK2/ask2 abi5/abi5 plants were germinated on 1/2 MS plates for 4 d. The surviving seedlings were transferred to 1/2 MS containing 1% sucrose and 10 µm ABA. After 3 d, the relative root length was measured. Each data point represents the mean of experiments performed in triplicate (n = 30–50 each).

It can be expected that the overexpression of ASK1 or ASK2 may rescue the ABA-insensitive phenotype of the abi5-1 mutant. The ASK1- and ASK2-overexpressing Arabidopsis were generated in both the wild-type and abi5-1 backgrounds. However, unlike TSK1 transgenic Arabidopsis with ABA hypersensitivity, 35S::ASK1 and 35S::ASK2 transgenic Arabidopsis exhibited similar ABA sensitivity as the wild type regarding germination (Fig. 6b), cotyledon expanding, and turning green (Fig. 6c,d), as well as root growth (Fig. 6e). In addition, the expression pattern of ABA-responsive genes was not changed significantly in 35S::ASK1 transgenic lines (Supporting Information Fig. S3). The ABA sensitivity of 35S::ASK1/abi5-1 and 35S::ASK2/abi5-1 was also examined. In the presence of 3 µm ABA, the germination rates of abi5-1, 35S::ASK1/abi5-1, 35S::ASK2/abi5-1 and wild type were 92, 33, 51 and 31%, respectively (Fig. 6b). In the presence of 0.3 µm ABA, the cotyledon expanding and turning green percentages of abi5-1, 35S::ASK1/abi5-1, 35S::ASK2/abi5-1 and wild type were 95, 67, 78 and 65%, respectively (Fig. 6c,d). The relative root lengths of abi5-1, 35S::ASK1/abi5-1, 35S::ASK2/abi5-1 and wild type were 77, 56, 65 and 53%, respectively (Fig. 6e). These results suggested that overexpression of ASK1 can rescue the ABA insensitivity of the abi5-1 mutant, and the overexpression of ASK2 can partially rescue the ABA insensitivity of the abi5-1 mutant. Our results indicated that ABI5 is required for the expression of ASK1 and ASK2, and ASK1 and ASK2 act downstream of ABI5 to mediate ABA signalling for biological processes.

To determine whether ASK1 and ASK2 were acting in the same pathway with ABI5, the ASK1/ask1 ASK2/ask2 abi5/abi5 genotype was generated, and its progeny seeds were used to examine the ABA response (Fig. 6f). The genotypes for the double mutant ask1/ask1 abi5/abi5 and ask2/ask2 abi5/abi5 and the triple mutant ask1/ask1 ASK2/ask2 abi5/abi5, as well as the wild type, were identified by PCR (Fig. 5b and Supporting Information Fig. S2). When treated with ABA, the relative root length of the abi5-1 mutant was 86%, compared with 47% in the wild type (Fig. 6f), suggestive of stronger resistance of abi5-1 to ABA. In contrast, the relative root lengths of the double mutant ask1/ask1 abi5/abi5 and ask2/ask2 abi5/abi5 were 50% and 50%, respectively, under the same treatment, and this was similar to the length of the wild type. The triple mutant ask1/ask1 ASK2/ask2 abi5/abi5 exhibited increased ABA resistance regarding relative root length (71%) compared with the wild type, but the relative root length did not equal that of the abi5-1 mutant. The relative root length of the ask1/ask1 ASK2/ask2 double mutant was 23% more than that of the wild type (Fig. 5c), and that of the ask1/ask1 ASK2/ask2 abi5/abi5 triple mutants was 24% more than that of the wild type (Fig. 6f). Therefore, introducing ABI5 mutations into the ask1/ask1 ASK2/ask2 background did not significantly affect the root elongation under ABA treatment compared with that in ask1/ask1 ASK2/ask2 plants. This result indicated that the relationship between ASK1 (ASK2) and ABI5 was complicated, and that SKP1 and ABI5 were not in the parallel pathway to regulate ABA signalling pathway.

DISCUSSION

SKP1-like genes encoding one of the components of SCF complexes have been found to regulate many hormone signalling pathways in plants via protein degradation, such as auxin, gibberellin, ethylene and jasmonic signalling pathways (del Pozo & Estelle 1999; Xu et al. 2002; Guo & Ecker 2003; Sasaki et al. 2003; Dharmasiri, Dharmasiri & Estelle 2005; Kepinski & Leyser 2005). SDIR1, a RING finger E3 ligase, was found to positively regulate ABA signalling (Zhang et al. 2007). Our data suggested that SKP1, a component of the SCF complex, participates in ABA signalling to regulate a series of biological processes including germination, stomatal aperture closure and root growth.

Our results demonstrated that TSK1 overexpression in Arabidopsis can lead to ABA hypersensitivity regarding seed germination, root growth and stomatal closure. The expression levels of transcription factors, including the B3 domain gene ABI3 and the bZIP transcription factor genes ABI5, ABF3 and ABF4, are increased in TSK1-overexpressing Arabidopsis. The ABA up-regulated genes RD29B, RAB18 and ADH1 are also up-regulated in transgenic Arabidopsis. Meanwhile, the expression of ERA1 and ATHB6, two negative regulators of the ABA signalling pathway, is reduced. The expression patterns of these aforementioned genes are consistent with the ABA hypersensitivity phenotype of TSK1-overexpressing transgenic Arabidopsis lines. The endogenous ABA levels are not significantly changed in TSK1-overexpressing Arabidopsis lines. These results indicate that the overexpression of TSK1 enhanced ABA signalling and might function as a positive regulator of ABA signalling.

The expression levels of two members of the PP2C family, ABI1 and ABI2, are increased in TSK1-overexpressing transgenic Arabidopsis lines. ABI1 and ABI2 are the negative regulators of the early step of the ABA signalling pathway (Leung et al. 1997). The transcription level of PP2Cs was down-regulated in both the ABA-insensitive mutant abi5-1 and the triple mutant snrk2.2/snrk2.3/snrk2.6 of the SnRK2s protein kinase, a positive regulator of ABA signalling (Nakashima et al. 2009). Furthermore, the expression of ABI1 and ABI2 was up-regulated in ABF3- and ABF4-overexpressing Arabidopsis (Kang et al. 2002). In other words, enhancing ABA signalling up-regulates PP2C gene transcription, whereas blocking ABA signalling down-regulates PP2C gene transcription. This hypothesis is supported by the increased transcription levels of ABI1 and ABI2 in TSK1-overexpressing lines. These data indicate that the negative regulator PP2C is a key switch that can be regulated by downstream ABA signalling to maintain the ABA signal steady state.

How dose the SKP1 regulate ABA signalling in Arabidopsis? The SCF complex functions in ubiquitin-mediated proteolysis to degrade the specific substrates. In fact, SCF complex was identified to participate in early hormone response (Dreher & Callis 2007; Somers & Fujiwara 2009). For example, the F-box protein TIR1 was identified as the auxin receptor, which can interact with ASK1 or ASK2 to form a SCF complex. Aux/IAA protein is the substrate of TIR1. Aux/IAA inhibits the transcription activity of the ARFs transcription factors in auxin signalling transduction pathway. DELLA proteins act as transcription repressors of GA signalling pathway and are recognized by the F-box protein SLY1, which can interact with ASK1 or ASK2. Is there a similar regulation in ABA signalling pathway? Based on the positive effects of ASK1 and ASK2 on ABA signalling, it can be speculated that the negative regulators of ABA pathway might be candidates of the SCF-mediated targeting substrate. The group A PP2Cs were identified as negative regulators in the early step of ABA signalling pathway, including ABI1 and ABI2 (Leung et al. 1997), HAB1 and HAB2 (Leonhardt et al. 2004; Saez et al. 2004), and AHG1 and AHG3 (Yoshida et al. 2006; Nishimura et al. 2007). It is possible that one member of the group A PP2Cs is the substrate. ABI5 is unlikely to be the substrate because ABI5 is a positive regulator of ABA signalling pathway and ABA insensitivity phenotype of abi5-1 mutant can be rescued by overexpression of ASK1. Based on present results, it is difficult to determine which member of PP2C is the real target protein of ASK1- or ASK2-participated SCF complex and further biochemical tests are needed to confirm the target.

Phylogenetic studies demonstrated that members of the SKP1 family in angiosperms have evolved at highly heterogeneous rates (Kong et al. 2004). SKP1 homologues from eudicots and monocots formed two separate clades, suggesting that they are derived from a single ancestral gene in the most recent common ancestor. The very low dN/dS (ratios of non-synonymous to synonymous nucleotide substitutions) for ASK1/2 in Arabidopsis and OSK1 in rice evolved slowly during plant evolution, suggesting that they are highly similar in sequence and critical functions (Kong et al. 2004, 2007). Sequence alignments of SKP1 homologues indicate that some amino acids are highly conserved in humans, plants and other species, and these conserved amino acids are identical in ASK1, ASK2, OSK1 and human SKP1 proteins (Schulman et al. 2000; Kong et al. 2004). Analysis of ASK1/2, TSK1 and OSK1 sequences (Supporting Information Fig. S1) indicates that excluding Leu 114 in ASK1 or Leu 125 in ASK2, which is replaced by Pro 129 in TSK1, all of the other conserved residues in TSK1 are identical to those in ASK1 and ASK2, and thus TSK1 may have functions similar to those of ASK1 and ASK2. Our previous study also revealed some conserved functions between TSK1 and ASK1 in male meiosis and auxin signalling (Li et al. 2006). Because the ask1-1 mutant is male sterile (Yang et al. 1999), the ask1/ask1 ask2/ask2 double mutant is lethal in embryogenesis (Liu et al. 2004), and because of the functional redundancy of ASK1 and ASK2 (Yang et al. 1999; Zhao et al. 1999, 2001, 2003a), ask1/ask1 ASK2/ask2 plants were used to study their functions in ABA signalling. Our results indicated that the ask1/ask1 ASK2/ask2 plant was ABA insensitive. Furthermore, the transcription levels of ASK1 and ASK2 are reduced in abi5-1 mutant compared with the wild-type levels, and the overexpression of ASK1 and ASK2 in the abi5-1 mutant can rescue or partially rescue the ABA insensitivity of the abi5-1 mutant, indicating that ASK1 and ASK2 act downstream of ABI5. Therefore, ASK1 and ASK2 are involved in the ABA response and are consistent with ABA-hypersensitive phenotypes in TSK1-overexpressing Arabidopsis lines.

Our results also reveal some functional divergence between TSK1 and ASK1 or ASK2. We generated ASK1- and ASK2-overexpressing lines in a Col-0 background and investigated the ABA response, including the seed germination and post-germination growth of ASK1- and ASK2-overexpressing transgenic lines. In contrast to the ABA-hypersensitive phenotypes in TSK1-overexpressing Arabidopsis lines, we found no significant differences between 35S::ASK1 and 35S::ASK2 lines, and wild type (Fig. 6b–e). The transcription levels of the ABA-responsive genes were not changed compared with those of the wild type (Supporting Information Fig. S3). We propose that the endogenous ASK1 and ASK2 levels under normal conditions are necessary to maintain the ABA response, but excess ASK1 and ASK2 proteins may not be sufficient for ABA-hypersensitive responses in Arabidopsis, which differs in the effect of TSK1 overexpression. In Arabidopsis, approximately 700 F-box genes (Gagne et al. 2002) and 21 SKP1 genes (Initiative 2000; Risseeuw et al. 2003; Zhao et al. 2003b) were identified. It is possible that one SKP1 protein can interact with more F-box proteins in different tissues or developmental stages. Some of the amino acid residues in SKP1 are important for its interaction with F-box proteins, and some slight differences in amino acid sequence in SKP1 might lead to different F-box protein-binding affinities (Schulman et al. 2000; Zheng et al. 2002). Crystal structure analysis of human SCF complex revealed that the last four α-helices in the C-terminal domain of SKP1 were directly attached to the F-box protein (Schulman et al. 2000; Zheng et al. 2002). It may be speculated that TSK1 and ASK1 might interact with different F-box proteins as a result of difference of amino acid between TSK1 and ASK1 and result in different sensitivity to ABA between TSK1- and ASK1-overexpressing lines. Gene duplication, followed by functional diversification, is a mechanism of gene evolution. Homologous sequences might share redundancy but they could diverge in functions through changes in expression patterns and accumulating variations in the coding regions. Analysis of evolution in SKP1 gene family demonstrated that the low-evolving genes may possess the conserved function and express widely at high levels, whereas the rapid evolving members may acquire new functions and express at low levels or in specific tissue (Kong et al. 2004). Our previous report revealed the TSK1 gene was mainly expressed in young roots and spikes and at a very low level in leaves (Li et al. 2006). ASK1 and ASK2 were expressed ubiquitously and at a very high level in leaves (Zhao et al. 2003b). The tissue specific expression pattern suggested TSK1 may evolve rapidly and obtain new function compared with ASK1 and ASK2. Taking together, our results provide evidence for the conservation of the ancestral function of SKP1 genes in regulating ABA signalling pathway since the separation of eudicots and monocots about 150 million years ago. However, they also indicated the monocot-specific neo- or sub-functionalization during evolution.

In conclusion, constitutive overexpression of TSK1 in Arabidopsis results in an ABA-hypersensitive phenotype. The double mutant ask1/ask1 ASK2/ask2 exhibits reduced ABA sensitivity. Additionally, ASK1/2 overexpression rescued ABA sensitivity in the abi5-1 mutant. The expression level of ASK1/2 was reduced in the abi5-1 mutant. These data suggest that SKP1 is involved in ABA signalling. This study supports the hypothesis that ABA signalling is co-regulated by SKP1.

ACKNOWLEDGMENTS

We thank Professors Zheng Meng and Hongzhi Kong (Institute of Botany, CAS) for their useful comments. This work was supported by the Ministry of Science and Technology of China (2007CB948202) and NSFC for Innovative Research Groups (No. 30821007).

Ancillary