HAB1 was originally cloned on the basis of sequence homology to ABI1 and ABI2, and indeed, a multiple sequence alignment of 32 Arabidopsis protein phosphatases type-2C (PP2Cs) reveals a cluster composed by the four closely related proteins, ABI1, ABI2, HAB1 and At1g17550 (here named HAB2). Characterisation of transgenic plants harbouring a transcriptional fusion ProHAB1: green fluorescent protein (GFP) indicates that HAB1 is broadly expressed within the plant, including key target sites of abscisic acid (ABA) action as guard cells or seeds. The expression of the HAB1 mRNA in vegetative tissues is strongly upregulated in response to exogenous ABA. In this work, we show that constitutive expression of HAB1 in Arabidopsis under a cauliflower mosaic virus (CaMV) 35S promoter led to reduced ABA sensitivity both in seeds and vegetative tissues, compared to wild-type plants. Thus, in the field of ABA signalling, this work represents an example of a stable phenotype in planta after sustained overexpression of a PP2C genes. Additionally, a recessive T-DNA insertion mutant of HAB1 was analysed in this work, whereas previous studies of recessive alleles of PP2C genes were carried out with intragenic revertants of the abi1-1 and abi2-1 mutants that carry missense mutations in conserved regions of the PP2C domain. In the presence of exogenous ABA, hab1-1 mutant shows ABA-hypersensitive inhibition of seed germination; however, its transpiration rate was similar to that of wild-type plants. The ABA-hypersensitive phenotype of hab1-1 seeds together with the reduced ABA sensitivity of 35S:HAB1 plants are consistent with a role of HAB1 as a negative regulator of ABA signalling. Finally, these results provide new genetic evidence on the function of a PP2C in ABA signalling.
Protein phosphorylation/de-phosphorylation events in ABA signalling involve several known protein kinases and phosphatases (Finkelstein et al., 2002; Leung and Giraudat, 1998). For instance, the guard-cell-specific protein kinase ABA-activated protein kinase (AAPK) from Vicia faba regulates ABA-induced stomatal closure (Li et al., 2000). Expression of a dominant negative AAPK allele renders stomata insensitive to ABA-induced closure through inhibition of plasma membrane anion channels (Li et al., 2000). Recently, the Arabidopsis open stromata 1 (OST1) protein kinase, which is related to AAPK, was shown to mediate the regulation of stomatal aperture by ABA (Mustilli et al., 2002). ABA-induced ROS production was disrupted in ost1 guard cells, indicating that OST1 acts upstream of ROS production (Mustilli et al., 2002). Another protein kinase involved in ABA signalling is ABA-induced protein kinase 1 (PKABA1), which is induced by ABA and which suppresses gibberellin (GA)-inducible gene expression in barley aleurone layers (Gomez-Cadenas et al., 1999). In addition to calcium-independent protein kinases, the calcium-dependent protein kinases (CDPK)1 and CDPK1a also mediate ABA signalling, as expression of constitutively active CDPK1 and CDPK1a leads to activation of the ABA-inducible Hordeum vulgare ABA-responsive 1 (HVA1) promoter (Sheen, 1996).
Conversely, ABA signalling is also regulated by several protein serine/threonine phosphatases (PP). PP are classified into four major classes, PP1, PP2A, PP2B and protein phosphatases type-2C (PP2C), on the basis of their substrate specificities, divalent cation requirements and sensitivity to inhibitors (Cohen, 1989). The use of okadaic acid (an inhibitor of PP1 and PP2A) enhances ABA-induced stomatal closing in fava bean (Schmidt et al., 1995) and activates ABA-responsive promoters in tomato hypocotyls (Wu et al., 1997). However, the same drug reduces ABA-induced stomatal closure in Arabidopsis (Pei et al., 1997). Interestingly, it has been recently found that disruption of the PP2A regulatory subunit RCN1 confers ABA insensitivity in Arabidopsis, which suggests that RCN1 functions as a positive transducer of ABA signalling (Kwak et al., 2002). Finally, inactivation of Ca2+-sensitive inward K+ channels in fava bean guard cells was prevented by inhibitors of PP2B (Luan et al., 1993), and ABA responses in pea epidermal peels were reduced by an inhibitor of PP2B (Hey et al., 1997). However, these results have to be interpreted with caution as no catalytic PP2B subunit is found in the Arabidopsis genome (Kerk et al., 2002).
In this work, we present evidence for a novel PP2C gene, named HAB1, functioning as a negative regulator of ABA signalling. HAB1 was originally cloned on the basis of sequence homology to ABI1 and ABI2, and it was tentatively named AtP2C-HA for homology to ABI1/ABI2 (Rodriguez et al., 1998b). Following standards for nomenclature (Meinke and Koornneef, 1997), we have re-named the gene as AtP2C-HAB1 and abbreviated it as HAB1. HAB1 encodes a functional PP2C (Meskiene et al., 2003), whose expression was previously detected in root, leaf, stem, flower and silique (Rodriguez et al., 1998b). In this work, a detailed analysis of transgenic plants harbouring a transcriptional fusion between the green fluorescent protein (GFP) reporter gene and the HAB1 promoter was performed. As a result, we could confirm that HAB1 is expressed ubiquitously in the plant, including key target sites of ABA action, such as guard cells and seeds. HAB1 expression in vegetative tissues was found to be low in the absence of ABA and was strongly upregulated by ABA (Rodriguez et al., 1998b); however, genetic evidence involving this gene in ABA signalling was lacking. In this work, we show that transgenic Arabidopsis plants, expressing a 35S:HAB1 transgene, show a remarkable ABA insensitivity. Blockage of ABA signalling by other PP2Cs (ABI1 and PP2CA) in transient overexpression experiments was previously reported by Sheen (1998). Our work on HAB1 and the one recently published by Gonzalez-Garcia et al. (2003) on FsPP2C1 further show reduced ABA signalling in planta upon sustained overexpression of PP2C genes. Conversely, a T-DNA disrupted allele of HAB1 (here named hab1-1) displays enhanced response to ABA in seeds. Both phenotypes support the role of HAB1 as a negative regulator of ABA signalling.
HAB1 in the context of the complex family of plant PP2Cs
At least 69 PP2C candidate gene products can be found in the genome of Arabidopsis (Kerk et al., 2002) and additional PP2Cs have been described in other plant species (Lorenzo et al., 2001, 2002; Meskiene et al., 1998; Miyazaki et al., 1999). A full alignment including 169 PP2Cs has recently been reported by Kerk et al. (2002). This complex alignment, in addition to plant PP2Cs, includes bacterial, fungal and animal PP2Cs. Therefore, in order to provide a comprehensive overview on the sequence similarity among HAB1 and a representative subset of plant PP2Cs, we compiled a partial alignment including 32 Arabidopsis PP2Cs, as well as 7 PP2Cs from other plant species (Figure 1).
The alignment of the catalytic cores of these 39 PP2Cs defines six distinct clusters. Physiological characterisation of PP2Cs from cluster #1 has not yet been reported. Cluster #2 includes the recently characterised POLTERGEIST PP2C (Yu et al., 2003). In the case of cluster #3, the single distinctive feature shared by the members of the group is the presence of two bipartite nuclear targeting sequences next to the C-terminal region of the gene product (Lorenzo et al., 2002). However, experimental evidence on the subcellular localisation of this group of PP2Cs is lacking. Cluster #4 includes the Arabidopsis KAPP and KAPP-like proteins from rice and maize. Cluster #6 includes, as a representative member, the well-characterised alfalfa MP2C enzyme (Meskiene et al., 1998, 2003).
Finally, cluster #5 contains PP2Cs involved in ABA signalling, such as ABI1, ABI2 and PP2CA. Two sub-branches can be distinguished, one including several PP2CA-like proteins and the other including ABI1, ABI2, HAB1 and At1g17550. Pair-wise comparisons further subdivide this latter sub-branch into two pairs of PP2Cs, i.e. ABI1/ABI2 and HAB1/At1g17550 (Figure 1). Indeed, overall amino acid sequence comparison between HAB1 and At1g17550 reveals 75% identity. Moreover, the N-terminal extension of HAB1 exhibits 60% identity to the corresponding one from At1g17550, and some stretches showing amino acid sequence similarity at the N-terminus are apparent in the four PP2Cs (Figure 2a). From these data, a close functional relationship between HAB1 and At1g17550 can be predicted. Therefore, we propose to name At1g17550 as HAB2. ABI1, ABI2 and HAB1 were reported to be upregulated by ABA (Leung et al., 1997; Rodriguez et al., 1998b). HAB2 is also upregulated in response to ABA; however, its level of expression is lower than the one observed for HAB1 (Figure 2b). An analogous situation occurs in the case of ABI1/ABI2, where the level of expression in response to ABA is notably higher for ABI1 than ABI2 (Figure 2b).
Tissue and cellular expression pattern of ProHAB1:GFP
Using Northern blot analysis, the HAB1 mRNA was detected previously in root, leaf, stem, flower and silique (Rodriguez et al., 1998b). To further extend this analysis, a characterisation of transgenic plants harbouring a transcriptional fusion between the GFP reporter gene and the HAB1 promoter was carried out. Ten transgenic lines were analysed and they showed similar patterns of GFP expression, which was upregulated by ABA in all the cases (Figure 3i). Expression of ProHAB1:GFP was observed in all plant tissues examined, particularly in key target sites of ABA action, such as guard cells (Figure 3a), seeds (Figure 3b,c), embryo (Figure 3d) and silique (Figure 3e). GFP expression driven by the HAB1 promoter was also observed in meristematic tissues, including the root meristem (Figure 3g) and shoot apical meristem (Figure 3h). Roots (Figure 3f,g), stem and leaves (data not shown) of ProHAB1:GFP transgenic plants also showed GFP expression. These results, together with previous Northern blot analysis, indicate that HAB1 is expressed ubiquitously in all plant organs.
Construction and characterisation of 35S:HAB1 lines
Sequence homology of HAB1 to ABI1/ABI2 suggests that HAB1 might be involved in ABA signalling. To investigate HAB1 function, we generated transgenic Arabidopsis plants expressing a 35S:HAB1 transgene. Initial characterisation of the recovered 64 T1 lines revealed that approximately 70% showed a wilty phenotype under low-humidity conditions (data not shown). Finally, 10 T3 homozygous lines with high mRNA expression levels of the 35S:HAB1 transgene were recovered. Southern blot analysis of five T3 homozygous lines revealed the presence of a single insertion of the 35S:HAB1 transgene in addition to the endogenous HAB1 gene (Figure 4a). The 3 kb BglII fragment, corresponding to the endogenous HAB1 gene, showed a higher intensity in the wild-type track because threefold more DNA was loaded in the gel (data not shown). Northern blot analysis showed substantially higher (5–10-fold) expression of the 35S:HAB1 transgene compared to endogenous HAB1 (Figure 4b). Furthermore, the steady-state levels of the endogenous HAB1 mRNA were not appreciably altered in the transgenic lines as compared to the wild-type control (Figure 4b).
To determine whether sustained transcriptional upregulation of HAB1 affects ABA sensitivity, we compared germination and early seedling development of 35S:HAB1 and wild-type seeds in media supplemented with ABA (Figure 4c,d). At 5 days post-stratification, 35S:HAB1 seeds developed green cotyledons on 1 µm ABA medium (Figure 4d), and these seedlings further developed green leaves (Figure 4c) in contrast to wild-type plants. These data indicate a reduced sensitivity of 35S:HAB1 seeds to ABA with respect to wild type. Finally, at 12 days post-stratification, approximately 40% of 35S:HAB1 seeds were able to germinate and develop green cotyledons on medium supplemented with 3 µm ABA, whereas wild-type seeds did not (Figure 4d). All five transgenic lines showed similar germination phenotypes in the presence of ABA (data not shown).
We wished to further examine the response of 35S:HAB1 seeds in other germination assays. Osmotic stress delays seed germination and arrests early seedling development mainly through ABA action. Indeed, both ABA-insensitive as well as ABA-deficient mutants are able to bypass the delay in germination and growth arrest induced by osmotic stress (Gonzalez-Guzman et al., 2002; Leon-Kloosterziel et al., 1996; Lopez-Molina et al., 2001; Werner and Finkelstein, 1995). Seed germination under 400 mm mannitol led to a severe delay in radicle emergence and growth arrest in wild-type individuals; in contrast, 35S:HAB1 seeds were able to germinate and they developed green cotyledons under such conditions (Figure 4e). An additional seed germination assay was carried out in the presence of paclobutrazol, an inhibitor of GA biosynthesis. GAs and ABA play antagonistic roles in seed germination, and ABA-insensitive (or ABA-deficient) mutants have a lower need for GAs during germination (Koornneef and Karssen, 1994). This requirement was compared in wild-type, 35S:HAB1 and abi2-1 seeds germinated in a medium containing paclobutrazol. In contrast to wild type, both 35S:HAB1 and abi2-1 seeds germinated and developed green cotyledons in medium supplemented with 10 µm paclobutrazol, indicating a reduced requirement for GAs at this developmental stage (Figure 4e).
35S:HAB1 plants show enhanced sensitivity to drought, higher transpiration rate, ABA-resistant growth and diminished expression of ABA-responsive genes
Stomatal closure and consequent reduction in water loss is a key ABA-controlled process that preserves water under drought conditions. Therefore, the ABA-insensitive abi1-1 and abi2-1 mutants are very sensitive to water-stress conditions because of impaired regulation of stomatal closure. We mimicked drought conditions by exposing plants to the drying atmosphere of a flow laminar hood. Figure 5(a) shows that, compared to wild type, 35S:HAB1 and abi2-1 plants developed a severe wilty phenotype 3 h after exposure to such conditions. This result suggests that 35S:HAB1 plants show a higher transpiration rate than wild-type plants. When measured by the loss of FW of detached rosette leaves, the water loss in wild-type plants was approximately twofold lower than that in 35S:HAB1 and abi2-1 plants (Figure 5b). Thus, constitutive overexpression of HAB1 resulted in increased transpiration and reduced tolerance to drought.
ABA has an inhibitory effect on root growth, and accordingly, ABA-insensitive mutants show higher ABA-resistant root growth than wild-type plants (Himmelbach et al., 1998). Figure 5(c) shows that 35S:HAB1 plants had a reduced sensitivity to the ABA-promoted inhibition of root growth as compared to wild-type plants. In presence of 10 µm ABA in the medium, root growth of 35S:HAB1 plants was approximately twofold higher than that of wild-type plants. Additionally, prolonged culture of wild-type plants on 30 µm ABA-containing medium led to growth arrest of the aerial part of the plant and yellowing of the leaves (Figure 5d). In contrast, after 2 weeks in a medium supplemented with 30 µm ABA, 35S:HAB1 plants showed little inhibition of leaf growth and green leaves, and eventually flowered under these conditions (Figure 5d).
Finally, to further establish the role of HAB1 in ABA signalling, we examined whether the reduction in ABA sensitivity in transgenic plants was accompanied by altered expression of ABA-responsive genes (Figure 5e). RAB18 is an ABA-inducible gene, whose expression is drastically inhibited both in abi1-1 and abi2-1 mutants (Leung et al., 1997). 35S:HAB1 plants also showed a severe reduction in the expression of RAB18 upon ABA treatment (Figure 5e). Accumulation of delta 1-pyrroline-5-carboxylate synthase (P5CS1) mRNA is induced by drought, salinity and ABA, and it is reduced by approximately 50% in the abi1-1 mutant compared to wild-type control (Strizhov et al., 1997). Likewise, 35S:HAB1 plants reduced P5CS1 transcript levels to about half of those detected in wild-type plants. RD29A is a cold-, drought- and ABA-inducible gene that, in addition to the dehydration-responsive element (DRE), contains ABA-responsive elements in its promoter (Shinozaki and Yamaguchi-Shinozaki, 1997). Impaired induction of RD29A by ABA was observed in 35S:HAB1 plants compared to wild type. Therefore, the reduced physiological responses to ABA of 35S:HAB1 plants were also correlated with a reduced expression of three ABA-responsive genes.
Identification of a T-DNA insertion mutant of HAB1
A T-DNA insertion mutant of HAB1 was identified in the Salk collection of T-DNA lines, which corresponds to donor stock number SALK_002104 (http://signal.salk.edu/cgi-bin/tdnaexpress). In order to identify individuals homozygous for the T-DNA insertion, genomic DNA was obtained from kanamycin-resistant seedlings of the SALK_002104 line and was submitted to Southern blot analysis (Figure 6b). Homozygous and hemizygous individuals were thus distinguished (Figure 6b), and plants homozygous for the T-DNA insertion in HAB1 were selected for further studies. The DNA sequence of the T-DNA flanking region of the SALK_002104 line indicated that the T-DNA insertion lied at the end of the third intron of HAB1 (http://signal.salk.edu/cgi-bin/tdnaexpress). In order to confirm these data, a genomic fragment adjacent to the left border of the T-DNA insertion was isolated by PCR and was sequenced. Sequence analysis confirmed that the T-DNA insertion was localised at nucleotide 1513 of the HAB1 gene (numbering refers to the ATG start codon). Therefore, the T-DNA insertion lies at the third intron of HAB1, only eight nucleotides upstream of the beginning of the fourth exon of HAB1 (Figure 6a). The 5-kb T-DNA insertion might disrupt splicing or affect the stability of the HAB1 transcript. In any case, this T-DNA-disrupted allele of HAB1 is predicted to result at least in the loss of the last 111 amino acids of the HAB1 protein, which includes essential residues for PP2C function (Das et al., 1996; Rodriguez, 1998). Therefore, the SALK_002104 line likely contains a null allele of HAB1. Indeed, Northern blot analysis failed to detect a full-length HAB1 transcript in the hab1-1 mutant (Figure 6c).
hab1-1 mutant seeds are ABA-hypersensitive
Progeny of hab1-1 homozygous individuals was harvested and, subsequently, seed germination assays were performed (Figure 7). In the absence of exogenous ABA, hab1-1 mutant seeds showed wild-type germination ratios after 3 days stratification at 4°C. However, in the presence of exogenous ABA, the hab1-1 mutant showed ABA hypersensitive inhibition of seed germination (Figure 7b; IC50 of 0.37 µm ABA for hab1-1 versus IC50 of 0.67 µm ABA for Columbia wild type). Wild-type and hab1-1 mutant plants were crossed and the resulting F1 seeds showed wild-type germination ratios on 0.5 µm ABA (data not shown). F2 seeds showed a segregation of the hab1-1 phenotype of 115–321 corresponding to a ratio of about 1–3 (χ2 = 0.44; P > 0.5). F2 ABA-hypersensitive seeds showed kanamycin-resistance and were homozygous for the T-DNA insertion as determined by PCR analysis (n = 40). These data suggest that the hab1-1 mutation is recessive and segregates as a single nuclear locus linked to the T-DNA insertion.
Additional dose–response analyses of germination in media supplemented with increasing concentrations of NaCl (Figure 7c) or mannitol (Figure 7d) were also performed. Compared to wild-type seeds, hab1-1 mutant shows enhanced inhibition of germination under these conditions. Osmotic stress blocks germination through ABA action; therefore, these results are consistent with the ABA-hypersensitive inhibition of germination observed above. Finally, an enhanced sensitivity to inhibition of germination by paclobutrazol was observed in hab1-1 mutant compared to wild-type seeds (Figure 7e). This result indicates a higher requirement for GAs during germination in hab1-1 compared to wild type, which is in agreement with both the antagonistic role of GAs to ABA during germination and the hypersensitive response of hab1-1 to ABA.
Finally, we transformed hab1-1 plants with the HAB1 cDNA under the control of its own promoter. Seeds from three independent hab1-1::HAB1 lines showed wild-type germination in the presence of either 0.5 µm ABA or 250 mm mannitol (Figure 7g). These results confirm that hab1-1 mutation is responsible for the enhanced response to ABA observed in the mutant seeds.
Transpiration rate in hab1-1
Water loss in detached leaves of hab1-1 mutant was similar to that of wild-type plants (data not shown), and whole plant transpiration rate was quite similar in wild-type and hab1-1 plants (Figure 7f). This result is in accordance to previous analyses of water loss in loss-of-function alleles of ABI1 and ABI2, or PP2CA antisense plants, which show kinetics of water loss similar to that of wild-type plants (Gosti et al., 1999; Merlot et al., 2001; Tahtiharju and Palva, 2001). Taking into account the complexity of the PP2C subfamily involved in ABA signalling, it is conceivable that loss-of-function of HAB1 in stomata is masked by the activity of other PP2Cs.
The importance of the PP2C class of protein serine/threonine phosphatases in plants is highlighted by the high number of PP2C genes (at least 69) found in the Arabidopsis genome (Kerk et al., 2002). In contrast, no more than 15 PP2Cs are found in the human genome (Cheng et al., 2000) and only 6 are found in the yeast genome (Stark, 1996). Therefore, with respect to other eukaryotic organisms, plants seem to have greatly expanded the regulatory mechanisms based on PP2C de-phosphorylation. In fungi, mammals and plants, examples are found where PP2Cs act as negative regulators of protein kinase cascades activated as a result of stress (Rodriguez, 1998; Takekawa et al., 1998; Warmka et al., 2001). Particularly, in plants, the stress-induced MP2C acts as a negative regulator of a MAPK pathway activated by wounding or salt stress (Meskiene et al., 1998, 2003). However, the regulatory role of eukaryotic PP2Cs clearly extends to other processes. For instance, both yeast and human PP2Cs de-phosphorylate cyclin-dependent kinases and, therefore, participate in cell cycle regulation (Cheng et al., 2000). Additional examples of singular functions of PP2Cs are well documented in plants. Thus, Arabidopsis KAPP and POLTERGEIST are negative regulators of the CLAVATA pathway, which is involved in regulation of shoot and floral meristem size (Williams et al., 1997; Yu et al., 2003). Finally, the Arabidopsis PP2CA regulates the K+ channel AKT2 and therefore extends the role of PP2C to control K+ transport and membrane polarisation (Cherel et al., 2002).
The field of ABA signalling also offers a significant contribution to our understanding of a plant process regulated by PP2Cs. Studies on abi1 and abi2 mutants, transient expression experiments, as well as analysis of transgenic PP2CA antisense plants, indicate that PP2Cs act as negative regulators of ABA signalling (Gosti et al., 1999; Merlot et al., 2001; Sheen, 1998; Tahtiharju and Palva, 2001). Transient expression studies in maize protoplasts demonstrated that overexpression of ABI1 and PP2CA blocked the induction of a reporter gene driven by an ABA-inducible promoter (Sheen, 1998). The identification and physiological characterisation of loss-of-function alleles (abi1-1R1–abi1-1R7) of the ABI1 gene were crucial to provide in planta evidence on the role of ABI1 as a negative regulator of ABA signalling (Gosti et al., 1999). This notion was further supported for ABI2 by the isolation of the loss-of-function abi2-1R1 allele and analysis of ABA responses in double mutants abi1-1R4 abi2-1R1 or abi1-1R5 abi2-1R1 (Merlot et al., 2001). Recent controversy on the role of ABI1 has arisen from the work of Wu et al. (2003), which questions the role of ABI1 as a negative regulator of ABA signalling on the basis that ABI1 overexpression in Arabidopsis does not affect the ABA signalling pathway. However, it seems logical to think that in case ABI1 was a positive regulator of ABA signalling instead of a negative one, 35S:ABI1 lines should show enhanced response to ABA, which was not reported by Wu et al. (2003). Additionally, the results of Wu et al. (2003) are opposed to those of Sheen (1998), which showed that overexpression of ABI1 in maize protoplasts led to a blockade of ABA-inducible gene expression. Additional experiments will be required to resolve this controversy.
Analysis of PP2C activity in abi1-1R5 abi2-1R1 plants revealed that ABI1 and ABI2 contribute to approximately 50% of the ABA-induced PP2C activity (Merlot et al., 2001). These data indicated that additional PP2Cs participate in ABA signalling. Indeed, it has been shown that antisense inhibition of PP2CA expression leads to increased sensitivity to ABA during development of frost tolerance and seed germination (Tahtiharju and Palva, 2001). In this context, the features of HAB1, namely ABI1/ABI2 sequence similarity and ABA-induced upregulation, made it a likely additional candidate to regulate ABA signalling. Considering the complexity of the plant subfamily of PP2Cs involved in ABA signalling (see cluster #5 in Figure 1) and the potential functional redundancy of these proteins, an overexpression approach was initially chosen. We reasoned that in case HAB1 was a positive regulator of ABA signalling, introduction of a 35S:HAB1 transgene might lead to enhanced or even constitutive response to ABA. Alternatively, sustained upregulation of HAB1 might lead to an ABA-insensitive phenotype in case it was a negative regulator of ABA signalling. The analysis of ABA response in 35S:HAB1 lines supports the latter hypothesis. Thus, transpiration assays in 35S:HAB1 T1 lines indicated enhanced water loss as compared to control plants, as expected for an ABA-insensitive phenotype. This result was confirmed in T3 plants, which showed an approximately twofold higher transpiration rate than wild-type plants, and consequently an increased sensitivity to drought stress (Figure 5a,b).
Further analysis of ABA-mediated responses both in seeds as well as in vegetative tissues indicated that sustained upregulation of HAB1 led to a reduced ABA sensitivity (Figures 4 and 5). For instance, root growth assays also revealed a reduced sensitivity to inhibition of growth by ABA in 35S:HAB1 plants compared to wild type (Figure 5c,d). The inhibitory effect of high ABA concentrations on root growth has been attributed to activation of the ethylene response pathway (Beaudoin et al., 2000; Ghassemian et al., 2000). Additionally, a possible link between the ABA growth-control pathway and cell cycle control has been provided by the discovery of cyclin-dependent protein kinase inhibitor (ICK1), which is induced by ABA (Wang et al., 1998). Both root and shoot meristematic activities of 35S:HAB1 plants were remarkably resistant to the inhibitory effect of 30 µm ABA compared to wild-type plants (Figure 5d). Indeed, under such high ABA concentration, the shoot apical meristem of 35S:HAB1 plant is able to switch from vegetative to reproductive growth and to produce flowers (Figure 5d). As enhanced HAB1 expression attenuates inhibition of growth by ABA, the ABA-mediated upregulation of HAB1 might contribute to regulation of the cell cycle arrest induced by ABA in meristematic tissues. Finally, the changes in the mRNA levels of three genes responsive to exogenous ABA –RAB18, P5CS1 and RD29A– also reflect a reduced ABA signalling in 35S:HAB1 plants compared to wild-type control (Figure 5e).
In order to complete our study on HAB1 function, we also analysed a T-DNA insertion mutant of HAB1. No gene deletion mutant or T-DNA-disrupted allele of PP2C genes involved in ABA signalling had been characterised previously. The T-DNA insertion present in the hab1-1 allele analysed in this study is predicted to result at least in the loss of the C-terminal 111 amino acid residues of HAB1, which include 5 out of the 11 conserved motifs of the PP2C family (Rodriguez, 1998). Indeed, Northern blot analysis failed to detect a full-length HAB1 transcript. Therefore, the T-DNA-disrupted hab1-1 allele is likely to be null and results in the loss of HAB1 function. ABA dose–response analyses indicated that hab1-1 seeds were hypersensitive to the inhibition of germination by ABA, as compared to wild type (Figure 7). Accordingly, an increased requirement for GAs during germination was observed in hab1-1 seeds compared to wild type (Figure 7) and, consequently, HAB1 function could play an important physiological role to promote germination. Thus, we suggest that promotion of germination would imply not only a positive action of GAs, but also a negative regulation of ABA signalling by HAB1 activity.
Whereas an enhanced response to ABA in germination was observed for hab1-1 mutant, its transpiration rate was similar to that of wild-type plants. A similar result was obtained for abi1-1R1 to abi1-1R7 revertants, which were ABA-supersensitive in germination, but showed kinetics of water loss similar to wild type (Gosti et al., 1999). Additionally, the loss-of-function aba2-1R1 mutant displayed wild-type ABA-induced stomatal closing (Merlot et al., 2001), and ABA-mediated drought responses were not affected by inhibition of PP2CA expression (Tahtiharju and Palva, 2001). Taking into account the fact that at least four PP2Cs are negative regulators of ABA signalling, a partial functional redundancy might explain the lack of phenotype in transpiration assays for single loss-of-function alleles. Moreover, analysis of the Arabidopsis genome reveals a putative paralog of HAB1, i.e. HAB2. Although these two homologous PP are unlikely to have completely redundant roles and HAB2 transcriptional upregulation in response to ABA is significantly lower than HAB1 (Figure 2b), it is possible that HAB2 (or ABI1/ABI2) activity partially masks the loss-of-function of HAB1. It will be necessary to test whether simultaneous inactivation of these PP2Cs affects transpiration rate.
The abi1-1 and abi2-1 mutants, as well as 35S:HAB1 plants, show a reduced ABA sensitivity. It has been previously suggested (Gosti et al., 1999) that abi1-1 and abi2-1 proteins have dominant negative effects and might act by trapping their endogenous substrates (positive regulators of ABA signalling) in a dead complex. To explain the molecular basis of HAB1 action, it will be required to identify its endogenous substrate. However, according to the phenotype of 35S:HAB1 plants, a simple model can be envisaged, where enhanced HAB1 phosphatase activity attenuates the ABA transduction cascade by de-phosphorylating a positive regulator of ABA signalling. Thus, positive regulators of ABA signalling could be inactivated either through formation of poison complexes (abi1-1, abi2-1) or by sustained de-phosphorylating activity (35S:HAB1). The target of HAB1 is likely different from those of ABI1 and ABI2, as loss of HAB1 function leads to enhanced response to ABA. HAB1 might function either in the same or in a different branch of ABA signalling than ABI1 and ABI2. Characterisation of new loss-of-function alleles of ABI1 and ABI2 and subsequent epistasis analyses with hab1-1 are required to address this question. Additionally, it will help to understand the relative contribution of each PP2C in ABA signalling. Finally, ABA promotes transcriptional upregulation of ABI1, ABI2 and HAB1, therefore, leading to enhanced de-phosphorylating activity and attenuation of the signal. As ABI1, ABI2 and HAB1 expression is itself upregulated by ABA, attenuation of the signal would lead to diminished de-phosphorylating activity, restoring the capacity of ABA response.
Negative regulation of signal transduction is required for a proper control of the complex signalling pathways that operate in a cell. Mechanisms of negative feedback, blockage of downstream signalling or transcriptional repression are a common issue in signalling pathways, particularly in hormone action (McCourt, 1999). With respect to ABA signalling, in addition to ABI1, ABI2, PP2CA and HAB1, other genes have been identified as negative regulators of the pathway. Thus, constitutive expression of the A. thaliana homeodomain protein 6 (ATHB6) leads to ABA insensitivity in a subset of ABA responses, which suggests that ATHB6 represents a negative transcriptional regulator of the ABA signal (Himmelbach et al., 2002). Interestingly, ATHB6 is the first target described of ABI1 (Himmelbach et al., 2002). The farnesyl transferase β-subunit enhanced response to ABA1 (ERA1) also plays a crucial role as a negative regulator of ABA signalling (Cutler et al., 1996; Pei et al., 1998). The recessive era1 mutant shows enhanced response to ABA and, therefore, protein farnesylation of certain signalling proteins appears crucial for negative regulation of ABA signalling. The mRNA cap-binding protein ABH1 is a novel modulator of ABA signalling (Hugouvieux et al., 2001). A recessive loss-of-function abh1 allele shows ABA hypersensitivity, indicating that ABH1 negatively modulates ABA signalling (Hugouvieux et al., 2001). Finally, the mutant ade1 exhibits sustained and enhanced levels of an ABA-inducible gene, suggesting a negative regulatory function for this locus (Foster and Chua, 1999). Therefore, a complex negative regulatory mechanism seems to have evolved to reset ABA signalling and to avoid undesirable effects because of sustained activation of the ABA pathway. It remains as a major challenge for the future to identify additional signalling elements linking the known intermediates, particularly the targets of the negative regulators of the pathway.
Arabidopsis thaliana plants were routinely grown under greenhouse conditions in pots containing a 1 : 3 vermiculite:soil mixture. For in vitro culture, seeds were surface-sterilised by treatment with 70% ethanol for 20 min, followed by commercial bleach (2.5%) containing 0.05% Triton X-100 for 10 min and, finally, four washes with sterile distilled water. Stratification of the seeds was conducted for 3 days at 4°C. Afterwards, seeds were sowed on Murashige–Skoog (MS) plates (Murashige and Skoog, 1962) containing solid medium composed of MS basal salts and 1% sucrose, solidified with 1% agar, and the pH was adjusted to 5.7 with KOH before autoclaving. Plates were sealed and incubated in a growth chamber having a controlled environment at 22°C under a 16-h light/8-h dark photoperiod at 80–100 µE m−2 sec−1.
Recombinant constructs and generation of transgenic plants
The coding region of the HAB1 cDNA was excised from the pSKAtP2C-HA construct (Rodriguez et al., 1998b) using an Ecl136II–EcoRI double digestion and subcloned into SmaI–EcoRI doubly digested pBIN121 (Clontech, Palo Alto, USA). The pBIN121-HAB1 construct was transferred to Agrobacterium tumefaciens C58C1 (pGV2260; Deblaere et al., 1985) by electroporation, and Arabidopsis plants (La-er ecotype) were transformed by the floral dip method (Clough and Bent, 1998). Seeds of plants transformed with pBIN121-HAB1 were harvested and plated on kanamycin selection medium to identify T1 transgenic plants. T2 seeds, plated on selection medium to assay the segregation ratio, and transgenic lines with a 3 : 1 (resistant/sensitive) ratio were selected. Southern blot analysis was performed to select lines carrying a single T-DNA copy. T3 progenies, homozygous for the selection marker, were used for further studies.
The 2-kb fragment of the HAB1 promoter region used in this work was obtained by PCR-mediated amplification from Columbia plants using oligonucleotides FpHAB1: 5′-CAACAGCAATATATGTATCTACG and RpHAB1: 5′-CCTCCATGGATCCTCCAAAATCAGAGATTTCC. This latter primer introduces a unique BamHI site around the ATG start codon of the HAB1 coding sequence. The amplified DNA was cloned into the BamHI site of a pBluescript SK-GFP vector. pBluescript SK-GFP bears unique BamHI and NcoI sites in front of the start codon of a promoterless GFP coding sequence located upstream of the nopaline synthase (NOS) terminator. Thus, the recombinant clone (pBluescript SK-ProHAB1:GFP) harboured a transcriptional fusion between the HAB1 mRNA 5′ untranslated sequence and the GFP coding sequence. The complete expression cassette comprising the HAB1 promoter, the GFP coding sequence and the NOS terminator was subcloned into SacI–SalI doubly digested pCAMBIA 2300. The resulting construct was named pCAMBIA2300-ProHAB1:GFP and was used to transform Arabidopsis plants as described above. The GFP reporter gene used in this study (Chiu et al., 1996) was kindly provided by J. Sheen (Boston, USA).
For functional complementation of the hab1-1 mutant, the GFP reporter gene of the pBluescript SK-ProHAB1:GFP construct was replaced with the coding sequence of HAB1, generating the pBluescript SK-ProHAB1:HAB1 construct. The complete expression cassette comprising the HAB1 promoter, the HAB1 coding sequence and the NOS terminator was subcloned into SacI–SalI doubly digested pCAMBIA 1300 (HYGR). The resulting construct was named pCAMBIA1300-ProHAB1:HAB1 and used to transform hab1-1 (KANR) plants as described above. Transgenic plants were screened in vitro on a MS medium (Sigma M5524) with 20 mg l−1 hygromycin B (Sigma H9773, Sigma-Aldrich, St Louis, MO, USA).
To measure ABA sensitivity, seeds were plated on solid medium, composed of MS basal salts, 1% sucrose and increasing concentrations of ABA. To determine sensitivity to inhibition of germination by high osmoticum or paclobutrazol, the medium was supplemented with increasing concentrations of either sodium chloride and mannitol, or paclobutrazol, respectively. In order to score seed germination, the percentage of seeds that had germinated and developed fully green expanded cotyledons was determined.
Root growth and transpiration assays
The root growth assay for scoring ABA sensitivity was performed by measuring root growth after 5 days of the transfer of 5-day-old seedlings onto MS plates containing 10 µm ABA. Kinetics analysis of water loss was performed in detached leaves at the same developmental stage and size from single 3-week-old plants. Five leaves per individual were excised and FW was determined at ambient conditions (25°C and approximately 40% relative humidity (RH)) after the indicated periods of time.
Whole plant transpiration was measured basically as described by Pei et al. (1998). Both wild-type and hab1-1 plants (five individuals per experiment, three independent experiments) were grown under normal watering conditions for 21 days and were then subjected to drought stress by completely terminating irrigation and minimising soil evaporation by covering pots with Saran Wrap. Pots were weighed every day at the same time. Pots containing no plants were subjected to the same treatment to determine the background rate of water loss.
About 10–12 seven-day-old seedlings were transferred from MS plates to 125-ml flasks containing 25 ml of MS solution and 1% sucrose. The flasks were shaken under cool fluorescent light. After 10 days, seedlings were mock-treated or treated with 50 µm ABA. Plant material was collected and frozen in liquid nitrogen. Total RNA was extracted as described by Gonzalez-Guzman et al. (2002), separated on formaldehyde–agarose gels and blotted to a nylon membrane. Blots were hybridised with random-priming 32P-labelled probes. mRNA levels were quantified by phosphorimage analysis of Northern blots using a Bioimaging analyser BAS1500 (Fujifilm España). A full-length cDNA probe for HAB1 was prepared as described previously by Rodriguez et al. (1998b). The P5CS1 probe was kindly provided by L. Szabados (Institute of Plant Biology, Hungary). The RAB18, RD29A, tubulin (TUB) and HAB2 probes were prepared by PCR amplification from genomic DNA of wild-type Columbia plants. Primers RAB18, RD29A and TUB have been described previously by Gonzalez-Guzman et al. (2002). Primer HAB2: 5′-GTGTAATCAGAAAAGACAAAG and 5′-GCCACCTCCATATTCACATCG.
Molecular characterisation of hab1-1 allele
A 158-bp genomic fragment adjacent to the left border of the T-DNA insertion was isolated from hab1-1 plants by PCR using primers LBb1: 5′-GCGTGGACCGCTTGCTGCAACT and R380: 5′-TCCGGTTCTGGGATCACAT. The amplified product was sequenced on both strands.
The fluorescence photographs of plants expressing the GFP reporter gene under control of the HAB1 promoter were taken using a Leica TCS-SL confocal microscope and laser scanning confocal imaging system. For GFP detection, the excitation source was an argon ion laser at 488 nm and emission was observed between 510 and 530 nm. Chloroplast auto-fluorescence was detected between 660 and 700 nm.
We thank Miguel A. Andrade for assistance in database analysis, and Roberto Solano and Christopher Rock for critical reading of the manuscript. We acknowledge Maria D. Gomez for her assistance with confocal microscopy. We thank Joseph Ecker and the Salk Institute Genomic Analysis Laboratory for providing the sequence-indexed Arabidopsis T-DNA insertion mutants, and ABRC/NASC for distributing these seeds. M.G.G. was supported by a Ministerio de Educacion y Cultura fellowship. P.L.R. was supported by a Ramon y Cajal research contract. This work was supported by Grants BFI2000-1361 and BIO2002-03090 from the Ministerio de Ciencia y Tecnologia and FEDER.