A protein phosphatase 2A catalytic subunit is a negative regulator of abscisic acid signalling1

Authors


(fax +34 915 85 4506; e-mail jjss@cnb.uam.es).

Summary

The key regulatory role of abscisic acid (ABA) in many physiological processes in plants is well established. However, compared with other plant hormones, the molecular mechanisms underlying ABA signalling are poorly characterized. In this work, a specific catalytic subunit of protein phosphatase 2A (PP2Ac-2) has been identified as a component of the signalling pathway that represses responses to ABA. A loss-of-function pp2ac-2 mutant is hypersensitive to ABA. Moreover, pp2ac-2 plants have altered responses in developmental and environmental processes that are mediated by ABA, such as primary and lateral root development, seed germination and responses to drought and high salt and sugar stresses. Conversely, transgenic plants overexpressing PP2Ac-2 are less sensitive to ABA than wild type, a phenotype that is manifested in all the above-mentioned physiological processes. DNA microarray hybridization experiments reveal that PP2Ac-2 is negatively involved in ABA responses through regulation of ABA-dependent gene expression. Moreover, the results obtained indicate that ABA antagonistically regulates PP2Ac-2 expression and PP2Ac-2 activity thus allowing plant sensitivity to the hormone to be reset after induction. Phenotypic, genetic and gene expression data strongly suggest that PP2Ac-2 is a negative regulator of the ABA pathway. Activity of protein phosphatase 2A thus emerges as a key element in the control of ABA signalling.

Introduction

Reversible protein phosphorylation is a regulatory mechanism essential to many cellular processes (Hunter, 1995; Millward et al., 1999). The response of the cell to environmental and nutritional stress as well as to other extracellular signals often depends on the dynamic phosphorylation and dephosphorylation of key regulatory proteins controlling the corresponding signalling pathways. The phosphorylation status of these proteins reflects the balance between the opposing activities of specific protein kinases and phosphatases. Over the recent years, increasing effort has been devoted to the identification and characterization of protein kinases and phosphatases that modulate specific signal transduction pathways in plants. Although substantial progress has been with the protein kinases, work on protein phosphatases (PPs) lags behind in comparison.

Protein phosphatases are divided in two major types: serine/threonine phosphatases and tyrosine phosphatases. Depending on their biochemical and structural characteristics, serine/threonine phosphatases are further divided into PP1, PP2A, PP2B and PP2C (Shenolikar, 1994; Wera and Hemmings, 1995). Protein phosphatase 2A (PP2A) is one of the most abundant types of serine/threonine phosphatase in all eukaryotic cells, comprising 1% of the total protein content in a mammalian cell (Cohen, 1997; Kremmer et al., 1997). Protein phosphatase 2A is a holoenzyme consisting of three subunits: a 36-kDa catalytic subunit C complexes with a 65-kDa scaffolding subunit A to form the AC core enzyme. This core enzyme binds a third subunit, the regulatory subunit B, to form a variety of heterotrimeric complexes (Mayer-Jaekel and Hemmings, 1994). This latter regulatory subunit determines the substrate specificity, subcellular localization and catalytic activity of the PP2A holoenzyme (Virshup, 2000; Janssens and Goris, 2001). This heterotrimeric assembly is conserved throughout eukaryotes, including plants (Luan, 2003). In the genome of the model plant Arabidopsis thaliana, five genes have been identified that code for isoforms of catalytic subunits (Ariño et al., 1993; Casamayor et al., 1994; Pérez-Callejón et al., 1998;Kerk et al., 2002), three for scaffolding subunit A (Slabas et al., 1994; Garbers et al., 1996) and at least 16 for regulatory B subunits (Haynes et al., 1999; Terol et al., 2002). The large number of isoforms of each subunit that can combine to obtain functional heterotrimers underlies the versatility of these PPs in the regulation of different processes (Janssens et al., 2005). These manifold possible combinations may have limited the experimental evidence concerning processes regulated by PP2A in A. thaliana. Pharmacological approaches have been useful for assessing the involvement of PP2A in the regulation of diverse cellular functions. Okadaic acid, a natural PP2A inhibitor, has been extensively used to reveal roles for PP2A in metabolism (Siegl et al., 1990), seed germination (Chang et al., 1999), cold acclimation (Monroy et al., 1998), cytoskeleton control (Camilleri et al., 2002), hormonal signalling (Kuo et al., 1996; Wu et al., 1997) and wound responses (Rojo et al., 1998). Only one mutant with reduced levels of PP2A activity, rcn1, has so far been reported. RCN1 encodes a scaffold subunit of PP2A (PP2A-Aα) (Deruère et al., 1999). rcn1 plants were isolated as auxin transport mutants (Garbers et al., 1996; Rashotte et al., 2001), but two alleles of rcn1 were identified as abscisic acid (ABA)-insensitive (Kwak et al., 2002) and ethylene-hypersensitive mutants, respectively (Larsen and Cancel, 2003).

In this work, a reverse genetic approach has been used to identify loss-of-function mutants in PP2A catalytic subunits (PP2Ac) that may serve as tools to elucidate the role of specific PP2As in the regulation of signal transduction networks in A. thaliana. A recessive mutant was identified with a T-DNA-mediated disruption of the gene encoding the catalytic subunit PP2Ac-2. This mutation leads to a reduction in the total PP2A activity in the plant. The characterization of this mutant and of transgenic A. thaliana plants overexpressing the same PP2Ac-2 gene is reported here. The pp2ac-2 mutant shows hypersensitivity to ABA in different processes regulated by this hormone: lateral and primary root growth, seedling development, germination, dormancy and ABA-dependent gene expression. Conversely, plants ectopically overexpressing PP2Ac-2 have reduced sensitivity to ABA. The genetic interaction between pp2ac-2 and abi1-1, an ABA-insensitive mutant altered in a PP2C gene is also described. Additionally, PP2Ac-2 is subject to differential ABA regulation at the transcriptional and post-transcriptional level that provides a desensitization/activation mechanism for ABA signalling. These results reveal PP2Ac-2 to be a specific component of the ABA signalling pathway and highlight the significance of PP activity as a negative regulator of ABA signal transduction in A. thaliana.

Results

Identification and characterization of pp2ac-2 and PP2Ac-OE transgenic plants

A collection of T-DNA insertion lines generated in A. thaliana (see Experimental procedures) was screened for mutants in PP2Ac genes. The PCR analysis of this collection identified a line with a T-DNA insertion in the PP2Ac-2 gene (At1g10430, GenBank accession number NM100918; for nomenclature of PP2Ac genes see Casamayor et al., 1994) Southern analysis of pp2ac-2 homozygous plants revealed that this line contains a single T-DNA insertion (data not shown). Sequencing of the flanking DNA localized this T-DNA within the PP2Ac-2 coding region to serine 168 at the beginning of the fifth exon. Moreover, sequencing revealed that the uidA gene that is part of the inserted T-DNA is in frame with the PP2Ac-2 coding region. This arrangement would thus give rise to a fusion protein consisting of a truncated PP2Ac-2 lacking its C-terminal regulatory region essential for its function (Favre et al., 1994; Wu et al., 2000) fused to the N-terminus of β-glucuronidase (GUS).

Ribonucleic acid gel blot hybridization was used to determine the accumulation of PP2Ac-2 transcript in pp2ac-2 mutants. A faint band hybridizing with the PP2Ac probe is detected in wild-type plants that is absent in the pp2ac-2 mutant (Figure 1a). However, the accumulation of a larger transcript is observed in the mutant. This transcript also hybridizes with a probe derived from the uidA gene (not shown) and is thus likely to be derived from the disrupted pp2ac-2 gene.

Figure 1.

 Characterization of pp2ac-2 and PP2Ac-OE plants.
(a) Northern blot analysis of PP2Ac-2 (catalytic subunit of protein phosphatase 2A) transcript levels in wild-type (wt, Wassilewskija ecotype), mutant pp2ac-2 and two transgenic PP2Ac-OE lines (OE1 and OE2). The bottom panel shows ethidium bromide staining of the RNA gel. Each lane contained 15 μg of total RNA from leaves.
(b) Western blot analysis of PP2Ac in wild type (wt), mutant (pp2ac-2) and two transgenic PP2Ac-OE lines (OE1 and OE2). Protein extracts from leaves (each lane contained 30 μg of total protein) were separated by sodium dodecyl sulphate–polyacrylamide-gel electrophoresis and transferred onto nylon membranes. The membranes were immunoblotted with an antibody specific for the C-terminal peptide of the PP2A catalytic subunit.
(c) The PP2A activity in wild type (wt), mutant (pp2ac-2) and two transgenic PP2Ac-OE lines (OE1 and OE2). The PP2A activity in the corresponding leaf extracts is expressed in pmol of inorganic phosphate (Pi) released per min and per μg of total protein. Values shown are mean ± SD of three replica samples. Similar results were obtained in three independent experiments.

Western analyses were performed with an antibody that recognizes the C-terminal part that is conserved in many of the PP2Ac isoforms in A. thaliana. The PP2Ac-2/GUS fusion protein lacks this epitope, and thus should not be recognized by the antibody. However, no differences in protein accumulation are observed between wild-type and pp2ac-2 plants (Figure 1b), suggesting that PP2Ac-2 constitutes only a minor fraction of the total PP2Ac content in A. thaliana.

An overexpression approach was also taken to further analyse the in vivo function of PP2Ac-2. The coding region of the PP2Ac-2 was fused to the 35S promoter of cauliflower mosaic virus and transferred into A. thaliana plants. After preliminary screening of several transgenic lines (not shown), those with higher PP2Ac-2 expression levels (PP2Ac-OE1 and -OE2) were selected for a more detailed analysis. The PP2Ac-2 transcripts accumulate to a much higher level in the overexpression lines than in wild-type plants (Figure 1a). This increase in PP2Ac-2 transcript is mirrored by a higher content in the corresponding protein (Figure 1b).

Total protein extracts from wild-type, pp2ac-2 and PP2Ac-OE plants were analysed for PP2A activity, considered as the fraction of PP activity that is sensitive to inhibition by a nanomolar concentration of okadaic acid (Hardie, 1989). A significant reduction was detected in the pp2ac-2 mutant, with values 20% lower than in wild-type plants (Figure 1c). In contrast, PP2A activity increased by 40–50% in the overexpression lines analysed. These results show that manipulation of PP2Ac-2 transcript levels leads to significant alterations of PP2A activity, and thus emerges as a good strategy for the study of the specific role(s) that this particular catalytic subunit may play in A. thaliana.

Histochemical analysis of PP2Ac-2 expression

The in-frame insertion of the uidA gene in the PP2Ac-2 coding sequence allows a detailed study of PP2Ac-2 expression. With this type of analysis, the disadvantages derived from cross-reactivity with the other, highly similar, PP2Ac genes can be overcome.

A histochemical analysis of the pp2ac-2 mutant reveals weak GUS staining in the aerial part of the seedlings (cotyledons and hypocotyls) that becomes more intense at the root/shoot junction, vascular tissue and stipules (Figure 2a). In adult plants, weak GUS staining is observed in mesophyll and guard cells (Figure 2b,c). GUS activity is also detected in anthers and pollen of inflorescences but is absent in young pods and siliques (Figure 2d). In the roots, GUS staining is observed in the apical part and vascular tissues. Strong staining is detected in root meristems (Figure 2e). Interestingly, a speckled staining pattern is observed along the root axis that may correspond to locations of lateral root formation (Figure 3a).

Figure 2.

 Histochemical analysis of PP2Ac-2 (catalytic subunit of protein phosphatase 2A) promoter activity.
GUS staining of pp2ac-2 plants reveals the activity of the PP2Ac-2 promoter.
(a) Aerial part of a 10-day-old plantlet, showing strong GUS staining at the root–shoot junction, and the vascular tissue.
(b) Close up of the tip of an adult leaf, showing GUS staining in the vasculature while it is absent in the trichomes.
(c) Close-up showing GUS staining in the guard cells.
(d) Inflorescence of adult plants, with strong staining at the anthers.
(e) Strong GUS staining is observed at the primary root tip of a ten-day old plantlet.

Figure 3.

 Histochemical analysis of PP2Ac-2 promoter activity in roots.
(a) The PP2Ac-2 promoter activity in the primary root of a 10-day-old pp2ac-2 plantlet. GUS staining marks the position of lateral root primordia.
(b, c and d) Close ups with a longitudinal view at different stages of lateral root development: (b) GUS activity at the establishment of a lateral root primordium after the asymmetrical division has occurred; (c) GUS activity stains the forming lateral root dome; and (d) is maintained in the emerging lateral root. Nomarski optics are used to reveal structural details.
(e) Once the lateral root has emerged, GUS staining localizes to the meristematic tissue of the root.

PP2Ac-2 expression correlates with lateral root development

The correlation between the GUS staining pattern observed in pp2ac-2 roots and the formation of lateral roots was investigated in more detail using histochemical techniques. Lateral roots in A. thaliana develop from a subset of pericycle cells termed pericycle founder cells, which are adjacent to xylem vessels. These cells undergo simultaneous asymmetrical transverse divisions giving rise to a dome-shaped primordium. Growth of this lateral primordium continues through the outer cell layers to ultimately emerge from the parental root. After emergence, lateral root primordium behaves as an apical meristem and leads the elongation of the lateral root (for a review see Casimiro et al., 2003). Mutant pp2ac-2 plants show intense staining in the meristematic cells in the apex of the primary root (Figure 2e). A faint GUS staining can also be observed in the pericycle throughout the length of the primary root. In addition, several spots of a more intense staining mark the positions of lateral root formation (Figure 3a). The activity profile of PP2Ac-2/GUS follows the development of lateral roots. Intense GUS staining is observed in the pericycle founder cells, at the early steps of establishment of lateral root primordium, and is maintained in the proliferating cells until emergence (Figure 3b–d). At later stages, subsequent elongation of fully developed lateral root leads to a localization of GUS staining to the apical position (Figure 3e) with a pattern similar to that exhibited in the primary root. This specific expression pattern in lateral root development suggests a role for PP2Ac-2 in this developmental event.

PP2Ac-2 influences the response of lateral and primary root growth to ABA

Lateral root formation is stimulated by auxins at several stages of their development (Xie et al., 2000; Casimiro et al., 2001; Bhalerao et al., 2002; Marchant et al., 2002). In addition, ABA may negatively regulate lateral root formation (De Smet et al., 2003). It was therefore interesting to elucidate whether regulation of lateral root formation by auxin and/or ABA was altered in pp2ac-2 plants. Seedlings were grown on hormone-free medium and subsequently transferred to vertical plates with either 100 nm NAA or 0.25 μm ABA, respectively. Under standard culture conditions, both pp2ac-2 and PP2Ac-OE plants developed a root system that could not be distinguished from that of wild-type plants. Moreover, auxin treatment did not result in any significant difference between wild-type, pp2ac-2 mutants and PP2Ac-OE plants (data not shown). In contrast, pp2ac-2 seedlings transferred to plates with ABA clearly showed a more severe inhibition of lateral root development than wild-type or PP2Ac-OE plants (Figure 4a).

Figure 4.

 Inhibition of lateral root growth by abscisic acid (ABA).
(a) Nine-day-old seedlings of wild type, pp2ac-2 and transgenic PP2Ac-OE1 line (OE1), 6 days after transfer onto agar plates supplemented with 0.25 μm ABA.
(b) Number of lateral root primordia in 9-day-old seedlings grown on medium without ABA or 6 days after transfer to 0.25 μm ABA-containing medium. The values are the mean of six plants ± SD.
(c) Length of lateral roots per centimetre of primary root in 9-day-old seedlings grown on medium without ABA or 6 days after transfer to 0.25 μm ABA-containing medium. The values are the mean of six plants ± SD. Wild type (open bars); pp2ac-2 (grey bars); transgenic PP2Ac-OE1 (black bars).The experiments were performed twice with similar results.

Abscisic acid is reported to inhibit lateral root development at various stages, both prior to and after emergence of lateral roots (De Smet et al., 2003). To investigate which step is affected in pp2ac-2 plants, we estimated the number of lateral primordia (including lateral roots) per centimetre of primary root (thus normalizing for effects of the treatment on root length). As in wild-type plants, ABA treatment reduced the number of lateral primordia in both pp2ac-2 and PP2Ac-OE plants to a similar extent (a 40% reduction, Figure 4b). However, significant differences were detected in the length of visible lateral roots. Whereas ABA-treated wild-type plants had a 78% reduction in the length of lateral roots, the length of lateral roots in pp2ac-2 plants show a 96% reduction (Figure 4c). Moreover, plants overexpressing PP2Ac-2 were less sensitive to inhibition (root length was reduced by 58%) than wild-type or pp2ac-2 plants. These results show that rather than initiation, a null mutation in PP2Ac-2 affects elongation of lateral roots, and suggest a role for PP2Ac-2 in lateral root development via ABA, and not auxin, signalling.

It was interesting to determine whether the higher responsiveness to ABA observed in the lateral roots of pp2ac-2 plants also affects the primary roots. To this end, seedlings grown on hormone-free medium were transferred to vertical plates supplemented with 10 and 30 μm ABA. Root length was scored 8 days after transfer and compared with control plants cultured without ABA. No differences in root growth were visible between wild-type, pp2ac-2 and PP2Ac-OE plants grown in hormone-free medium (not shown). However, similar to the effects determined for lateral roots, pp2ac-2 seedlings showed an increased inhibition of primary root growth compared with the wild-type plants. Moreover, PP2Ac-OE seedlings were significantly less sensitive to ABA in their response to growth inhibition of the primary root (Figure 5a). The differences in the sensitivity to ABA inhibition were more striking for the lower concentration tested. Compared with wild-type plants, roots that were 33% shorter were scored in pp2ac-2 plants grown on 10 μm ABA while, under the same conditions, the roots of PP2Ac-OE plants were on average 70% longer. Although the relative sensitivities of wild-type, pp2ac-2 and PP2Ac-OE plants are essentially maintained at 30 μm ABA, the absolute differences are smaller because of the shorter roots formed on average in this ABA concentration (Figure 5b).

Figure 5.

 Inhibition of root-growth by abscisic acid (ABA).
Four-day-old wild-type, pp2ac-2 and PP2Ac-OE1 seedlings (OE1) grown without ABA were transferred to plates containing 10 and 30 μm ABA. After a further 8 days the pictures were taken (a), and elongation of new root was measured and compared with the value determined for the same plant type grown without ABA (b). The values are the mean of six plants ± SD. Wild type (open bars); pp2ac-2 (grey bars); transgenic PP2Ac-OE1 (black bars).The experiment was performed four times with similar results.

Increased ABA sensitivity in vegetative tissues and seeds of pp2ac-2 plants

The above-mentioned results show that PP2Ac-2 modulates the responses to ABA in root tissues. The effect of an alteration in PP2A activity on the sensitivity of other vegetative tissues to ABA was also tested. Some pp2ac-2 and PP2Ac-OE plants were germinated and grown on plates containing various concentrations of ABA. On 0.5 μm ABA, development of wild-type plants was arrested just after expansion of the cotyledons, whereas pp2ac-2 mutant plants could hardly expand them. On ABA concentrations of 1 μm or higher, growth of both pp2ac-2 and wild-type seedlings was halted just after the emergence of the radicle (Figure 6). In contrast, PP2Ac-OE plants developed normally on 0.5 μm ABA, without any visible inhibition of root growth or cotyledon greening and expansion, at a rate similar to abi1-1, a classical ABA-insensitive mutant. On 1 μm ABA, the development of PP2Ac-OE seedlings was delayed in comparison with abi1-1 plants, but they grew noticeably better than wild-type and pp2ac-2 plants. The hypersensitive phenotype of pp2ac-2 mutant plants to ABA and the opposite effect observed in the PP2Ac-OE plants, suggests that PP2A activity represses ABA signalling in vegetative tissues of A. thaliana.

Figure 6.

 Abscisic acid (ABA) sensitivity of pp2ac-2 and PP2Ac-OE1 seedlings.
Ten-day-old Wassilewskija wild type (wt), pp2ac-2 and PP2Ac-OE1 seedlings (OE1) grown on agar plates supplemented with 0, 0.5 and 1 μm ABA. For comparison, the classical ABA-insensitive abi1-1 mutant and its corresponding wild-type background (Ler), are also included.

Abscisic acid plays a key regulatory role during seed germination (Leung and Giraudat, 1998; Rock, 2000; Finkelstein et al., 2002). Therefore, the rates of germination of pp2ac-2 and PP2Ac-OE seeds were scored in the absence or presence of exogenous ABA. In a hormone-free medium, 45% of the pp2ac-2 seeds germinated after 2 days, 19% fewer than wild-type seeds. In contrast, germination of PP2Ac-OE seeds was 26% higher than the wild-type (Figure 7a). In the presence of ABA, the germination of pp2ac-2 seeds was further inhibited (down to 15% at 0.5 μm ABA, nearly the half of the rate of germination shown by wild-type seeds). In medium with ABA, PP2Ac-OE seeds germinated earlier than wild type, and the differences increased with the higher concentration of ABA. Thus, the maximum relative differences between PP2Ac-OE and wild type were scored with seeds germinating on 0.5 and 1 μm ABA (at these concentrations the number of wild-type seeds germinated was one-third of the number of germinated PP2Ac-OE seeds).

Figure 7.

 Germination and dormancy of pp2ac-2 and PP2Ac-OE seeds.
(a) The abscisic acid (ABA) dose response of germination. After 4 days’ vernalization at 4°C, wild-type (open bars), pp2ac-2 (grey bars) and transgenic PP2Ac-OE1 (black bars) seeds were germinated on plates containing different concentrations of ABA. Seeds with emerged radicles were scored after 2 days.
(b) Seed dormancy of wild-type, pp2ac-2 and PP2Ac-OE seeds. Freshly harvested seeds were plated on ABA-free medium and maintained at 22°C. The time course of germination of wild-type (closed squares), pp2ac-2 (closed circles) and PP2Ac-OE1 seeds (closed triangles).
(c) Freshly harvested wild-type (wt), pp2ac-2 and PP2Ac-OE1 seeds (OE1) germinating for 48 and 60 h on ABA-free medium. All experiments were performed three times (= 50 seeds each) and the values given are the mean ± SD.

The different rates of germination of pp2ac-2 and PP2Ac-OE seeds compared with wild-type seeds, in the absence of ABA, suggests that these seeds may also have altered dormancy, a process maintained by endogenous ABA (Gubler et al., 2005). In order to explore this possibility, freshly harvested wild-type, pp2ac-2 and PP2Ac-OE seeds were directly sown on plates without previous vernalization. As shown in Figure 7c, the radicle of wild-type seeds was fully developed after 2 days while it was hardly emerging in pp2ac-2 seeds. Consistent with a negative effect of PP2Ac-2 on seed dormancy, PP2Ac-OE seeds not only had a developed radicle but the cotyledons were already expanding and greening. Figure 7b summarizes the percentage of germinated seeds determined at different time points. After 3 days, only 16% of pp2ac-2 seeds were germinated compared with 36% of the wild-type seeds and 74% of the PP2Ac-OE seeds. Indeed, the latter were all germinated after 5 days whilst for wild type and pp2ac-2 the same levels of germination were only attained after 7 days. Taken together, these results indicate that at the germination stage pp2ac-2 seeds are hypersensitive to both endogenous and exogenously applied ABA, showing retarded germination and increased dormancy. Conversely, PP2Ac-OE seeds are insensitive to ABA with earlier and higher germination rates and decreased dormancy.

Compared with wild type, no significant differences were observed in pp2ac-2 and PP2Ac-OE plants grown on auxins, jasmonic acid, ethylene, cytokinins or gibberellins, suggesting that these plants have specifically altered their response to ABA (data not shown).

ABA-related stress responses of pp2ac-2 and PP2Ac-OE plants

In addition to its role in plant growth and development, ABA modulates adaptive responses to several environmental stresses such as salinity, extreme temperatures and drought (Shinozaki and Yamaguchi-Shinozaki, 2000; Xiong and Zhu, 2001; Xiong et al., 2002a,b). Indeed, these environmental conditions increase the level of endogenous ABA, which in turn triggers the expression of genes related to tolerance to water deficit that underlies these stress situations (Zhu, 2002). Furthermore, ABA-insensitive (abi) and ABA-deficient (aba) mutants identified in Arabidopsis are impaired in their responses to these stresses (Gosti et al., 1999; Xiong et al., 2002a). Because pp2ac-2 and PP2A-OE plants are altered in their ABA-mediated developmental responses, we tested whether they were also disturbed in their ABA-related stress responses.

Growth of pp2ac-2 and PP2Ac-OE plants was examined in a high-salt medium. To this end, seeds were germinated in a medium without salt, and transferred after 6 days to a medium containing 150 mm NaCl. Compared with wild-type plants, pp2ac-2 plants exhibited a more severe inhibition of root growth, less greening of cotyledons and slightly more branched roots (Figure 8a). In contrast, this high salt concentration inhibited to a similar extent the growth of wild-type and PP2Ac-OE plantlets. However, a better performance of PP2Ac-OE plants at other salt concentrations cannot be ruled out. Control experiments with equivalent concentrations of mannitol were performed and no significant differences in growth were noticed (data not shown), suggesting that the effects observed in germination were not due to the osmotic pressure induced by the high NaCl concentration. The inhibition of germination by a high salt concentration was also tested in pp2ac-2 and PP2Ac-OE seeds. Germination of pp2ac-2 was more sensitive to NaCl than that of wild-type seeds. Indeed, only one-quarter of pp2ac-2 seeds germinated in 150 mm NaCl compared with the half of the wild-type seeds. Even though no differences were observed between wild-type and PP2Ac-OE seedlings exposed to 150 mm NaCl, PP2Ac-OE had a significantly higher percentage of seeds germinated on this salt concentration than wild type (80%; Figure 8b). Results from these two experiments suggest that a deficiency in the activity of a specific PP2A in Arabidopsis provokes changes in salt tolerance at both stages, seedling development and germination.

Figure 8.

 Stress sensitivity of pp2ac-2 and PP2Ac-OE plants.
(a) Effect of high salt on the growth of wild-type, pp2ac-2 and PP2Ac-OE1 seedlings (OE1). Six-day-old seedlings were transferred to a medium containing 150 mm NaCl. After 10 days, a picture of representative seedlings was taken.
(b) Germination of pp2ac-2 and PP2Ac-OE in high-salt medium. Wild-type (wt, open bar), pp2ac-2 (grey bar) and PP2Ac-OE1 seeds (OE1, black bar) were germinated on plates containing 150 mm NaCl, after 4 days’ vernalization at 4°C. Seeds with emerged radicles were scored after 2 days. The values are the mean of 30 seeds ± SD.
(c) Rate of transpirational water loss from pp2ac-2 and PP2Ac-OE plants. Rosette leaves of wild-type (circles), pp2ac-2 (squares) and PP2Ac-OE1 plants (triangles) were detached and weighed. Leaves were kept on filter paper, and weighed again at the indicated time points. The percentage water loss was calculated from the difference between initial and final weights. The values are the mean of three experiments (six plants each) ± SD.
(d) Response to high glucose treatment of pp2ac-2 and PP2Ac-OE plants. Wild-type (wt), pp2ac-2 and PP2Ac-OE1 seeds (OE1) were germinated and grown on plates with 3% glucose or 3% mannitol, and maintained in a vertical position. Pictures were taken after 2 weeks.

Upon water deficit, ABA mediates stomatal closure, thus reducing water loss through transpiration (Schroeder et al., 2001; Hetherington, 2001). As GUS staining was also detected in the guard cells of plants carrying the PP2Ac-2/GUS fusion (Figure 2e), and increased ABA sensitivity is frequently accompanied by a reduced water loss from transpiration, the drought tolerance of pp2ac-2 and PP2Ac-OE plants was tested. Comparing rates of water loss of detached rosette leaves, no significant differences were detected between wild-type and PP2Ac-OE plants. A slight delay in loss of water from pp2ac-2 plants was observed (Figure 8c), but this was not sufficient to confer higher survival rates under drought treatment (data not shown).

Abscisic acid also affects glucose and sucrose signal transduction pathways (Finkelstein and Gibson, 2002; Gazzarrini and McCourt, 2001). In fact, several glucose-insensitive mutants are allelic to ABA-insensitive or ABA-defective mutants. The sensitivity of pp2ac-2 and PP2Ac-OE plants to glucose was therefore tested. Mutant and overexpression seeds were grown in vertical plates with 3% glucose or mannitol and allowed to develop for 2 weeks. After that time, the pp2ac-2 seedlings showed a stronger inhibition of cotyledon greening and leaf development compared with the wild-type (Figure 8d). This enhanced response was not observed when similar experiments were performed with mannitol, which inhibited the development of all plant types to the same extent. Growth of PP2Ac-OE plants in the presence of this concentration of glucose was essentially identical to that of wild-type plants.

PP2Ac-2-mediated regulation of gene expression

To further investigate the role of PP2Ac-2 in the regulation of ABA signalling, transcriptomic profiles of pp2ac-2 and wild-type plants treated with ABA were obtained and compared using whole-genome Arabidopsis long-oligonucleotide microarrays (see Experimental procedures). As shown in Supplementary Table S1, 57 genes (20 upregulated and 37 downregulated in ABA-treated pp2ac-2 plants) were identified as differentially expressed using two criteria: a FDR (false discovery rate) below 10% and a fold change over 1.5. Independent confirmation of differential expression was obtained for two of the genes in Table S1 by quantitative PCR (QPCR; Supplementary Figure S1). Meta-analysis of the upregulated genes using Genevestigator db (http://www.genevestigator.ethz.ch) showed that more than half of them are also upregulated by ABA, and 90% are downregulated in the presence of Norfluorazon, an inhibitor of ABA biosynthesis (Figure 9). These results indicate that pp2ac-2 mutants have an enhanced expression of ABA-regulated genes. Moreover, analysis of putative cis-regulatory sequences over-represented in the promoter of differentially expressed genes using Promomer (http://bbc.botany.utoronto.ca) and Motif Analysis (TAIR) revealed the ABA-response element (ABRE; ACGTG) as being the most overrepresented sequence in these promoters [Supplementary Table S2; statistical significance < 0.001 (Promomer) and 1.2 × 10−4 (Motif Analysis)]. Taken together, these data lend strong support at the molecular level to the previous conclusion that PP2Ac-2 is a repressor of ABA-dependent gene expression.

Figure 9.

 Induction of abscisic acid (ABA)-regulated genes in protein phosphatase 2A (pp2ac-2) mutants.
Meta-analysis of differentially regulated genes in pp2ac-2 mutants compared with wild-type plants after ABA treatment. Data from upregulated genes (pp2ac-2) are clustered together with data from ABA and Norfluorazon (NF) treatments obtained from Genevestigator db for the same genes. A hierarchical cluster was obtained using Pearson’s uncentred distance in the Multiarray Experiment Viewer Clustering tool from TIGR. Differentially expressed genes are listed on the right.

The pp2ac-2 mutation partially suppresses the ABA-insensitive phenotype of abi1-1

The ABA-related phenotypes of pp2ac-2 and PP2Ac-OE plants suggest that PP2Ac-2 is a component of ABA signalling. In order to place PP2Ac-2 in the ABA signal transduction pathway, plants carrying the dominant insensitive abi1-1 mutation were crossed with the recessive hypersensitive pp2ac-2 plants. The sensitivity of root growth to ABA was tested in the F2 segregating population. In addition, double mutants abi1-1/pp2ac-2 were identified by PCR (see Experimental procedures) among the segregating seedlings, and their phenotypes in the presence of ABA retested in the F3 generation. As shown in Figure 10, the double abi1-1/pp2ac-2 mutant was more sensitive to ABA than the parental abi1-1 plants, showing reduced root growth in 1 μm ABA. In this concentration of the hormone, development of pp2ac-2 plantlets is arrested after germination. In spite of the dominant character of the abi1-1 mutation, the enhanced sensitivity conferred by the homozygous recessive pp2ac-2 mutation was more apparent in heterozygous seedlings ABI1/abi1-1 than in the homozygous abi1-1 mutant (not shown). These results show that pp2ac-2 mutation partially suppresses ABA insensitivity in the abi1-1 background. They are therefore consistent with PP2Ac-2 acting downstream of, or at the same level as, ABI1 in the ABA signalling pathway, and thus regulating a subset of the plant responses to ABA. However, these results do not exclude the possibility that PP2Ac-2 and ABI1 act in two independent parallel pathways both contributing to ABA responses.

Figure 10.

 Partial suppression of abi1-1 phenotype by a null protein phosphatase 2Ac (pp2ac-2) mutation.
An abscisic acid (ABA) sensitivity assay of abi1-1/pp2ac-2 plants. Seeds of double mutant abi1-1/pp2ac-2, abi1-1 and pp2ac-2 plants (in a ABI1/abi1-1 background) were germinated and grown for 10 days on medium containing 1 μm ABA before the pictures were taken.

Regulation of PP2Ac-2 activity by ABA

Considering the involvement of PP2Ac-2 in the regulation of responses to ABA in A. thaliana, it was therefore relevant to determine if, in turn, ABA controls PP2A activity. To this end, PP2Ac-2 transcription was analysed after treatment with ABA. An increase in PP2Ac-2 mRNA was detected 20 min after the addition of ABA to the medium, and was maximal at 10 h (Figure 11a). Hybridization with the ABA-responsive KIN2 gene showed that induction of PP2Ac-2 transcript by ABA was delayed compared with KIN2 that was already detected 5 min after treatment. Histochemical GUS staining provided indirect evidence that ABA-induced PP2Ac-2 transcription may result in increased protein accumulation. Upon treatment of pp2ac-2 plants with ABA, GUS staining was detected in essentially all tissues. As soon as 20 min after ABA treatment, increased staining was observed that became more intense at the longer times analysed (Figure 11b). This ABA-dependent increase of PP2Ac-2 may seem to conflict with its role as repressor of ABA signalling deduced from all the above-mentioned results. To gain further insight into this apparent contradiction, total PP2A activity was measured in wild-type, pp2ac-2 and PP2Ac-OE plants at different times after ABA treatment. In wild-type plants, PP2A activity rapidly decreased upon ABA treatment, with activity levels nearly cut by half after 10 min in the presence of the hormone. Consistent with the observed induction of PP2Ac-2 transcription and protein accumulation, a slow increase in activity was already detected 45 min after ABA treatment, followed by a steady increase to return to basal levels of activity after 10 h. Variation in PP2A activity followed a similar pattern in both pp2ac-2 and PP2A-OE plants, the latter showing a more severe reduction and increase of activity after ABA treatment. The fluctuations in PP2A activity monitored in ABA-treated pp2ac-2 plants suggest that the activity of other PP2Acs may also be subject to ABA regulation.

Figure 11.

 The abscisic acid (ABA) regulation of protein phosphatase 2A (PP2Ac-2).
(a) Northern blot analysis of PP2Ac-2 and KIN2 transcript levels in wild-type plants after ABA treatment. The bottom panel shows ethidium bromide staining of the RNA gel. Fifteen micrograms of total leaf RNA was loaded on each lane.
(b) Histochemical GUS staining of pp2ac-2 seedlings after ABA treatment. Seedlings of the pp2ac-2 mutant were treated with 100 μm ABA, and samples were taken and stained at indicated times.
(c) The PP2A activity in wild-type, pp2ac-2 and PP2Ac-OE plants after ABA treatment. Leaf protein extracts were prepared from wild-type (open bars), pp2ac-2 (grey bars) and PP2Ac-OE1 (black bars) plants treated with 100 μm ABA for the indicated times. The PP2A activity is expressed in pmol of inorganic phosphate (Pi) released per min and per μg of total protein. Values shown are means ± SD of three replicate samples. Similar results were obtained with three independent experiments.

In summary, following repression of PP2A activity at early times after ABA treatment, the increase in RNA and protein levels leads to a return of the activity to the initial levels. These results suggest that ABA signal transduction requires the early release of PP2A repression. This regulation is exerted at levels of both gene transcription and enzymatic activity. The ABA-induced transcription of (at least) PP2Ac-2 probably restores basal levels of PP2A protein and activity, and may thus reset the plant’s sensitivity to ABA.

Discussion

Abscisic acid plays a key regulatory role in several physiological processes during the life of plants, such as seed dormancy and germination, stomatal closure, root growth and responses to environmental stress (Schroeder et al., 2001; Finkelstein et al., 2002; Zhu, 2002). The characterization of a T-DNA insertion mutant in a gene encoding a catalytic subunit of PP2A (PP2Ac-2) in A. thaliana, and of the corresponding transgenic plants ectopically overexpressing PP2Ac-2, has revealed that this specific catalytic subunit is a component of the ABA signal transduction pathway regulating developmental and physiological programmes. Our results show that the reduced PP2A activity in the loss-of-function pp2ac-2 mutant confers a hypersensitive response to ABA. In contrast, overexpression of PP2Ac-2 and subsequent increase in PP2A activity reduces the sensitivity of the transgenic plants to this hormone. The opposite sensitivities correlate well with the alterations observed in different ABA-regulated processes, supporting a negative role for PP2Ac-2 in ABA signalling.

The genome of A. thaliana contains five highly similar PP2Ac genes. Indeed, these genes exhibit sequence identities above 90% (Casamayor et al., 1994). Moreover, as these genes are apparently expressed in a nearly constitutive fashion T. Domínguez and JJSS, CNB-CSIC, Madrid, Spain, unpublished), it is surprising that disruption of a single one results in an observable phenotype, suggesting that in the case of ABA signalling the function of the other PP2Ac genes is not redundant. This absence of compensatory functions is especially remarkable in the case of PP2Ac-1 that shares 97% identity with the amino acid sequence of PP2Ac-2. However, mutant pp2ac-1 plants exhibit wild-type responses to ABA T. Domínguez, A. Duprat and JJSS, CNB-CSIC, Madrid, Spain, unpublished) highlighting the exquisite specificity of the interaction of catalytic and regulatory subunits to form functionally distinct complexes.

Protein phosphatase 2Ac-2 is essential for the orchestration of adequate responses in different ABA-regulated processes in both seeds and vegetative tissues of the plant. Abscisic acid plays a major role in the development and germination of seeds. Accordingly, pp2ac-2 plants have delayed seed germination and increased dormancy, whereas PP2Ac-OE seeds germinate rapidly without vernalization even in the presence of a low concentration of ABA. These results indicate that PP2Ac-2 is required for the correct response of the plant to both endogenous and exogenously applied ABA.

The in-frame insertion in the PP2Ac-2 coding region of the uidA gene (carried in the inserted T-DNA) has allowed promoter activity in its bona fide genomic environment to be monitored. Histochemical analyses of pp2ac-2 plants detect GUS activity in root meristematic tissues, both in primary roots and at early stages of lateral root development. Weak GUS staining is seen in the pericycle cells that run along the primary root, and becomes stronger once asymmetrical cell division that initiates lateral root formation occurs. At later stages, GUS staining concentrates in the meristems of the emerging lateral root, very much resembling the pattern observed in primary roots, and suggesting that PP2Ac-2 expression is related to the cell proliferation that accompanies root growth.

Treatment with ABA arrests lateral root development at different steps, both before and after emergence of the lateral root primordium (Casimiro et al., 2003; De Smet et al., 2006). Compared with the wild type, lateral and primary root growth is severely inhibited in pp2ac-2 seedlings upon mild ABA treatment, a phenotype that is consistent with the observed histological pattern of promoter activity. In contrast, PP2Ac-OE plants exhibit a noticeable insensitive phenotype to the hormone in this developmental programme. Remarkably, ABA treatment of pp2ac-2 plants results in a reduction in the length of primary and lateral roots but not in the number of primordia developing from the main parental root. This result suggests that although PP2Ac-2 is expressed in all stages of lateral root formation, its ABA-repressing function is exerted at the last stages of lateral root development, after the emergence of lateral roots.

Treatment with ABA also inhibits the growth of primary roots. Whilst inhibition of lateral root formation is already observed at 0.1 μm ABA, a concentration 100 times higher is required to inhibit primary root growth. Protein phosphatase 2Ac-2 is already expressed at the tip of the main root by the time of the emergence of the radicle in the germinating seed, and this pattern of expression is maintained throughout root development. Consistent with a role in repressing ABA signalling, pp2ac-2 primary roots are more sensitive and PP2Ac-OE roots more resistant to ABA treatment than wild-type roots. Although sensitivity to 0.1 μm ABA may identify signalling components and pathways that are different from those targeted at 10 μm, the results reported here indicate that PP2Ac-2 is none the less a common component in both signalling pathways.

Weak constitutive GUS staining is also observed in the aerial parts of pp2ac-2 plants. In the leaves, staining concentrates in the veins, and extends to all leaf tissues upon ABA treatment. In the flowers of non-treated plants, anthers and pollen grains are heavily stained, suggesting again the involvement of PP2Ac-2 in cell division. An essentially identical pattern of staining has been observed in transgenic A. thaliana lines transformed with a 561-bp PP2Ac-2 promoter/GUS fusion (Thakore et al., 1999). The pattern of GUS staining in the loss-of-function pp2ac-2 plants and in plants hemizygous for the T-DNA insertion is nearly identical, suggesting that PP2Ac-2 activity has little, if any, influence on its own gene expression.

Abscisic acid not only regulates developmental processes but also the adaptation of plants to different environmental stresses (Xiong et al., 2002a; Zhu, 2002). It has been reported that ABA-insensitive and ABA-defective mutants are resistant to high salt and/or high sugar concentrations (Huijser et al., 2000; Laby et al., 2000; Quesada et al., 2002; Xiong et al., 2001b; González-Guzmán et al., 2002). It should thus be expected that ABA-hypersensitive mutants, such as pp2ac-2, are more sensitive to these stresses. Indeed, pp2ac-2 plants are more sensitive than wild-type ones to high salt and high sugar stresses, both at germination and during vegetative growth. Remarkably, the hypersensitive mutant era1 (enhanced response to ABA) shows increased tolerance to these stresses (Xiong et al., 2001a; Wang et al., 2005), highlighting the complex interactions between ABA and stress responses. None the less, the involvement of PP2A2c-2 in all the above-mentioned developmental and environmental processes points to a central role for this PP in ABA signalling.

Transcriptomic analyses have revealed that PP2Ac-2 exerts its role in ABA signalling through regulation of gene expression. A loss-of-function mutation in PP2Ac-2 enhances the expression of ABA-regulated genes, such as RD26 (responsive to desiccation gene 26), DREB-1B and DREB-1C (dehydration responsive element-binding factor 1B and 1C) (Fujita et al., 2004; Knight et al., 2004; Liu et al., 1998). DREB-1B and DREB-1C are members of the DREB/CBF family of transcription factors involved in drought, osmotic and cold response, whereas RD26 is another transcription factor that regulates ABA-mediated responses to dehydration and salt stress. Protein phosphatase 2Ac-2 may thus act on ABA signalling at least in part through downregulation of these transcriptional activators. Moreover, an increase in the representation of the ABRE (Shinozaki et al., 2003) is observed in the promoters of this group of genes. Given the ABA hypersensitivity of pp2ac-2, the presence of ABA-regulated genes in the cluster of overexpressed genes in this mutant confirms that PP2Ac-2 is a suppressor of ABA signalling in A. thaliana plants.

Research on hormone signalling has largely focused on the activation of the corresponding signalling pathways (Himmelbach et al., 2003). However, understanding how cells are desensitized and return to basal activity once the response to the hormone has been triggered is lagging behind. Regulation of PP2Ac-2 expression and PP2Ac-2 activity offers a mechanistic explanation on how the sensitivity to ABA is reset after initial activation. According to the results obtained, constitutive PP2A activity limits perception of ABA in non-challenged plants. Upon exogenous application of ABA or after ABA synthesis in stressed plants, PP2A activity decreases, ABA is thus perceived, and the corresponding responses are triggered. Recently, it has been reported that the scaffolding subunits PP2A-A2 and PP2A-A3 can be ubiquitinated in vitro by an E3 ubiquitin ligase (AtCHIP) (Luo et al., 2006). One possibility is that PP2Ac-2 may also undergo this post-transcriptional regulation, resulting in a fast proteolytic degradation that would account for the observed ABA-mediated repression of PP2A activity. After ABA treatment or stress release, increased ABA-dependent transcription/translation of PP2Ac-2 restores activity levels of PP2A. The newly synthesized PP2Ac-2 represses ABA signalling, thus allowing a full response of the plant to subsequent stress.

The phenotypic and molecular evidence presented here lends strong experimental support for a fundamental role of PP2Ac-2 in ABA signalling. Reversible protein phosphorylation has previously been involved in ABA signal transduction. In particular, protein phosphatase 2C has consistently been involved in the regulation of ABA-mediated developmental and environmental processes. Protein phosphatase 2C is encoded by a multigene family in A. thaliana, and some gene members appear to have distinct, essentially non-overlapping roles in the ABA signalling network (Kuhn et al., 2006; Yoshida et al., 2006). Loss-of-function pp2c mutants exhibit an ABA-hypersensitive phenotype in several developmental and environmental responses of the plant, indicating that, similar to PP2Ac-2, PP2C acts as a negative regulator of ABA signalling (Gosti et al., 1999; Merlot et al., 2001; Leonhardt et al., 2004; Saez et al., 2004; Schweighofer et al., 2004). However, the gain-of-function PP2C mutant abi1-1 confers ABA insensitivity in a dominant manner (Meyer et al., 1994). Abscisic acid sensitivity is partially restored in the abi1-1/pp2ac-2 double mutant, suggesting that both PPs act in the same ABA signalling pathway. To date only another member of the PP2A family of phosphatases, RCN1, has been associated with this hormonal pathway (Kwak et al., 2002). However, RCN1 is a scaffolding subunit (PP2A-A1) that is also involved in auxin transport and ethylene responses whose mutants therefore exhibit pleiotropic phenotypes (Rashotte et al., 2001; Larsen and Cancel, 2003). In contrast, PP2Ac-2 is a catalytic subunit with a highly specific role in ABA signalling. With the identification of PP2Ac-2 as a negative regulator of ABA responses, PP activity emerges as a key negative element in the regulation of this hormonal pathway.

Experimental procedures

Plant materials, mutant screening and transformation

Arabidopsis thaliana pp2ac-2 and PP2Ac-OE plants are in the Wassilewskija (ws) ecotype, whereas abi1-1 mutants are in Landsberg erecta (Ler) background. The CSIC-INIA T-DNA insertion mutant collection was generated by Agrobacterium tumefaciens-mediated random transformation with the pGKB5 vector (Bouchez et al., 1993). To obtain the overexpression PP2Ac-OE plants, the coding region of PP2A2c-2 (Ariño et al., 1993) was cloned in the sense orientation in the BIN19 vector under the control of the 35S promoter region of the cauliflower mosaic virus and the 3′ terminator sequence of the octopine synthase gene. The C58C1 strain of A. tumefaciens containing the above construct was used to transform A. thaliana ecotype Wassilewskija by in planta vacuum infiltration (Bechtold et al., 1993). Kanamycin-resistant T1 plants were selected by plating seeds on Murashige and Skoog medium (Sigma; http://www.sigmaaldrich.com/) supplemented with 100 μg ml–1 kanamycin and transferring kanamycin-resistant seedlings to soil. For analysis of epistasis, pp2ac-2 plants were crossed with abi1-1 and the homozygous double mutants were isolated from F2 progeny by PCR analysis (Leung et al., 1994).

Plant growth, germination and stress conditions

All seeds were germinated on Johnson medium (Sigma) supplemented with ABA, glucose and NaCl as indicated. Before plating, seeds were surface-sterilized for 1 min in 75% bleach and 0.5% Tween-20, washed three times in sterile water and incubated at 4°C for 3 days. Growth conditions were a 16-h light/8-h dark photoperiod at 23°C. For germination assays, seeds collected on the same date and stored at 4°C for a month or longer were used unless indicated otherwise, and germination (emergence of the radicle through the seed coat) was scored at indicated times. For the salt sensitivity assay, seeds were germinated on Johnson medium plates and transferred to plates supplemented with different concentrations of NaCl and mannitol.

For root growth measurements, plants were germinated on Johnson medium plates and transferred after 3 days to fresh plates containing the additional compounds at the final concentrations indicated in the text. Lengths of primary and lateral roots were measured from images captured using a Leica MPS-60 camera (http://www.leica.com/), with the help of the NIH image programme. Lateral root primordia were identified and scored with a Leica DMR microscope.

Histochemical staining of GUS activity and microscopy

Tissues were fixed in 0.3%para-formaldehyde, 100 mm sodium phosphate pH 7, 1 mm EDTA, and soaked in a staining solution of 1 mm 5-bromo-4-chloro-3-indolyl-β-glucuronic acid (Biosynth, Switzerland), 100 mm sodium phosphate pH 7, 1 mm EDTA, and 0.5% Triton X-100. After applying a vacuum for 5 min, plant tissue was incubated for 6 h at 37°C in the dark, and reactions stopped by replacing the staining solution with 70% ethanol. Repeated washing with 70% ethanol was used for clearing the green tissues. The samples were mounted in dePex (BDH, http://uk.vwr.com) and photographed using either a Leica MZFLIII dissecting microscope or Leica DMR microscope with Nomarski optics, equipped with a Leica MPS-60 camera. The images were processed using photoshop 6.0 (Adobe Systems Inc.; http://www.adobe.com/).

Protein analysis: extraction, Western blot analysis and PP assay

Plant tissue was ground in liquid nitrogen and resuspended in 50 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (Tris)–HCl pH 7.4, 1 mm EDTA, 10% glycerol, 0.1%β-mercaptoethanol and protease inhibitor cocktail (Sigma). After 10 min centrifugation at 15 7000 g, the supernatant was removed and the protein concentration was determined by the Bradford method (Bio-Rad; http://www.bio-rad.com/). Extracts from each sample were separated by SDS-PAGE, and electroblotted onto nitrocellulose membranes. Proteins were detected using a rabbit anti-protein phosphatase 2A/C antibody (directed to a synthetic peptide encompassing amino acids 298–309 in the mammalian PP2Ac sequence; Calbiochem, http://www.merckbiosciences.co.uk), peroxidase-conjugated donkey anti-rabbit antibody (Amersham Pharmacia Biotech; http://www5.amershambiosciences.com/), and ECL (enhanced chemiluminescence) Western blotting detection reagents (Amersham Pharmacia Biotech).

Protein phosphatase 2A activity was measured using serine/threonine phosphatase assay system (Promega; http://www.promega.com/) according to the manufacturer’s instructions with minor modifications. The PP2A activity in protein extracts was assayed with and without okadaic acid (Sigma) using the phosphopeptide RRA (pT)VA as substrate. The PP2A activity is considered as the fraction of PP activity inhibited by 10 nm okadaic acid.

RNA gel blot analysis

Total RNA was extracted from leaves using RNA Wiz (Ambion; http://www.ambion.com/) as described by the manufacturer. Extracted RNAs were subjected to electrophoresis on 1.5% formaldehyde/agarose gels and blotted onto Hybond N+ membranes (Amersham). For ABA-regulated gene expression, 10-day-old seedlings were treated with 100 μm ABA and harvested at indicated times. The PP2Ac-2 (Ariño et al., 1993) and kin2 (Kurkela and Borg-Franck, 1992) probes were labelled with 50 μCi of 32P-dCTP. Blots were exposed up to 24 h on a PhosphorImager screen (Molecular Dynamics, http://www.mdyn.com).

RNA amplification and labelling for microarrays analysis

Total RNA (1 μg) was amplified and aminoallyl-labelled using MessageAmp® II aRNA kit (Ambion, http://www.ambion.com) and 5-(3-aminoallyl)-2′-deoxyuridine-5′-triphosphate (aa-dUTP, Ambion), following the manufacturer’s instructions. Approximately 40–50 μg of aRNA was obtained. For each sample, 7.5 μg of aminoallyl-labelled aRNA was resuspended in 0.1 m Na2CO3 (pH 9.0) and labelled with either Cy3 or Cy5 Mono NHS Ester (CyTM Dye Post-labelling Reactive Dye Pack, Amersham). The samples were purified with MegaclearTM (Ambion) following the manufacturer’s instructions. Incorporation of Cy3 and Cy5 was measured using 1 μl of the probe in a Nanodrop spectrophotometer (Nanodrop Technologies Inc.; http://www.nanodrop.com/). For each hybridization 200 mol of Cy3 and Cy5 probes was mixed, dried in a speed-vac, and resuspended in 9 μl of RNase-free water. Labelled aRNA was fragmented by adding 1 μl of 10× fragmentation buffer (Ambion) and incubating at 70°C for 15 min. The reaction was stopped with 1 μl of stop solution (Ambion). The integrity and average size of total RNA, aRNA and fragmented aRNA were evaluated using a Bioanalyzer 2100 (Agilent; http://www.agilent.com/). The average size of aRNAs was about 1000 nucleotides and that of fragmented aRNAs 100 nucleotides. The final volume of the probe was diluted to 100 μl in hybridization solution.

Microarray hybridization

Four biological replicates were independently hybridized for each transcriptomic comparison. Microarray slides were composed of synthetic 70mer oligonucleotides from the Operon Arabidopsis Genome Oligo Set Version 1.0 (Qiagen; http://www.qiagen.com/) spotted on aminosilane-coated slides (Telechem, http://www.arrayit.com) by the University of Arizona. Slides were rehydrated and UV-crosslinked according to the details on the supplier’s web page http://ag.arizona.edu/microarray/methods.html. The slides were then washed twice for 2 min in 0.1% SDS and in ethanol for 30 sec. Arrays were drained with a 528 g spin for 2 min. Slides were pre-hybridized in 6× SSC, 0.5% SDS (w/v), and 1% BSA (w/v) at 42°C for 1 h., followed by five rinses with milliQ water (Millipore; http://www.millipore.com/). Excess water was drained with a 528 g spin for 2 min.

For the hybridization, equal amounts of dye of each aRNA labelled with either cy3 or cy5, ranging from 200 to 300 pmol, were mixed with 20 μg of polyA and 20 μg of yeast tRNA (Sigma-Aldrich) in a volume of 9 μl. To this volume 1 μl of RNA fragmentation buffer was added, (RNA Fragmentation Reagents, Ambion) and, after 15 min at 70°C, 1 μl of stop solution. Formamide, 20× SSC, 50× Denhart’s and 20% SDS were added to a final concentration of 50% formamide, 6× SSC, 5× Denhart’s and 0.5% SDS. This mix was boiled for 3 min at 95°C and then added to the pre-hybridized slide. Hybridization took place overnight at 37°C in a hybridization chamber. Arrays were then washed for 5 min at 37°C in 0.5× SSC, 0.1% SDS, twice for 5 min at room temperature (RT; 21°C) with 0.5× SSC, 0.1% SDS, three times with 0.5× SSC at RT, and 5 min with 0.1× SSC. The slides were then drained with a 528 g spin for 2 min. The slides were stored in darkness until they were scanned.

Images from Cy3 and Cy5 channels were equilibrated and captured with a GenePix 4000B (Axon; http://www.axon.com/) and spots quantified using genpix pro 5.1 software (Axon).

Background correction and normalization of expression data were performed using limma (Smyth and Speed, 2003; Smyth, 2004). To avoid exaggerated variability of log-ratios for low-intensity spots during local background correction we use the ‘normexp’ method in limma to adjust the local median background estimates. The resulting log-ratios were print-tip loess normalized for each array (Smyth and Speed, 2003). To have similar distribution across arrays and to achieve consistency among arrays, log-ratio values were scaled using as the scale estimator the median-absolute-value (Smyth and Speed, 2003).

The mean of the three replicate log-ratio intensities and their SD were generated.

Identification of differentially expressed genes

The rank products (RPs) method was used to determine differentially expressed genes. Probes were sorted by their normalized expression ratio for each chip. The RPs were calculated for each gene according to Breitling et al. (2004). The RPs were compared with the RPs of 5000 random permutations of the same data to assign E-values. To correct for the multiple testing problem inherent in microarray experiments we employed the FDR (Storey, 2003), i.e. we divided the E-value of each gene by its position in the list of changed transcripts. An FDR of 5% means that only 5% or fewer of the genes up to this position are expected to be observed by chance (false positives), the remaining 95% being genes that are indeed significantly affected (true positives).

Data for ABA or Norfluorazon from Affymetrix microarrays were downloaded from the meta-analysis program of the ‘Genevestigator’ database. The hierarchical cluster was calculated and drawn using the TIGR MeV (multiarray experiment viewer; Saeed et al., 2003) software provided by the TIGR Institute.

Acknowledgements

We are grateful to Dr Albert Ferrer for the kind donation of PP2Ac cDNAs, and encouragement during the initial stages of this work. We thank Drs Antonio Fuertes and Javier Paz-Ares for making the CSIC-INIA seed collection available to us. We also thank Pilar Paredes for her excellent technical help. Financial support for this work was obtained from the Spanish Ministerio de Educación y Ciencia (grants BIO2005-08528 to JJSS, and BIO2004-02502 and GEN2003-20218-CO2-02 to RS), the Comunidad de Madrid (grant R/SAL/0674/2004 to RS) and the European Commission TMR-CRISP Network (HPRN-CT-2000-00093 to JJSS).

Ancillary