Regulation of leaf organ size by the Arabidopsis RPT2a 19S proteasome subunit

Authors


*(fax +81 11 706 2737; e-mail jjyama@sci.hokudai.ac.jp).

Summary

The ubiquitin/26S proteasome pathway plays a central role in the degradation of short-lived regulatory proteins, to control many cellular events. To further understand this pathway, we focused on the RPT2 subunit of the 26S proteasome regulatory particle. The Arabidopsis genome contains two genes, AtRPT2a and AtRPT2b, which encode paralog molecules of the RPT2 subunit, with a difference of only three amino acids in the protein sequences. Both genes showed similar mRNA accumulation patterns. However, the rpt2a mutant showed a specific phenotype of enlarged leaves caused by increased cell size, in correlation with increased ploidy. Detailed analyses revealed that cell expansion is increased in the rpt2a mutant by extended endoreduplication early in leaf development. The transcription of genes encoding cell cycle-related components, for DNA replication licensing and the G2/M phase, was also promoted in the rpt2a mutant, suggesting that extended endoreduplication was caused by increased DNA replication, and disrupted regulation of the G2/M checkpoint, at the proliferation stage of leaf development.

Introduction

Multicellular organisms have an inherent organ size. In plants, the final leaf size is determined by the two indices of leaf cell number and average cell size. However, the mechanisms for regulation of cell size remain to be fully explained. Somatic polyploidy via the process of endoreduplication has been implicated in cell size control. Endoreduplication is a type of cell cycle where nuclear chromosomal DNA replication occurs without cell division, and is widespread among eukaryotes, although it is most common in plants (Edgar and Orr-Weaver, 2001; Nagl, 1976). In Arabidopsis thaliana, the size of mature leaf pavement cells is correlated with their ploidy levels, which vary from 2C to 32C as a result of differences in the number of endoreduplication cycles that they have undergone (Melaragno et al., 1993). The Arabidopsis trichome, a large branched cell on the surface of aerial organs, generally has a DNA content of 32C (Hülskamp et al., 1994). Other organs, such as hypocotyls and roots, also exhibit high ploidy levels (Barow, 2006).

Several mutants and transgenic plants that have aberrant levels of endoreduplication have been isolated, and have led to the identification of key regulators of endoreduplication (Sugimoto-Shirasu and Roberts, 2003). Examples of key regulators are the DNA replication licensing factors CDC6a and CDT1a, which cause increased ploidy levels when they are ectopically overexpressed (Castellano et al., 2001, 2004). The CDC6a, CDT1a and KRP2 proteins are known to be degraded by the 26S proteasome, indicating that regulation of their proteolysis plays an important role in the endoreduplication pathway.

The ubiquitin/26S proteasome pathway functions to degrade short-lived regulatory proteins that are involved in cell cycle regulation, signal transduction, apoptosis and metabolic regulation, as well as those involved in the elimination of damaged or misfolded proteins (Hershko and Ciechanover, 1998; Glickman and Ciechanover, 2002). Proteins destined for degradation in this pathway are first modified by the covalent attachment of polyubiquitin chains. The ubiquitinated proteins are then recognized and degraded by the 26S proteasome, a multisubunit ATP-dependent protease with broad substrate specificity (Hershko and Ciechanover, 1998).

The 26S proteasome is a 2-MD complex assembled from two particles: the 20S core particle (CP) and the 19S regulatory particle (RP) (Voges et al., 1999). The proteolytic activities reside within the central chamber of the CP, which is a hollow cylinder composed of four stacked rings (Groll et al., 1997; Löwe et al., 1995). The RP presumably helps identify appropriate substrates for breakdown, removes the polyubiquitin chain and directs the entry of unfolded proteins into the CP lumen for degradation.

The RP can be divided further into two subcomplexes: the base and the lid. The base consists of six AAA-ATPase subunits, RPT1–RPT6, and the non-ATPase subunits, RPN1, RPN2 and RPN10. The lid binds to the base and contains non-ATPase subunits: RPN3, RPN5–RPN9, RPN11 and RPN12 (Voges et al., 1999; Glickman, 2000; Fu et al., 2001). Each proteasome subunit is presumed to have a specific function, although only the roles of some of the non-ATPase subunits are known. RPN11 has a metalloprotease activity that appears to help release polyubiquitin chains from substrates for degradation (Verma et al., 2002; Yao and Cohen, 2002). In plants, RPN10 is involved in the regulation of ABA signaling by targeting ABA signaling proteins for degradation (Smalle et al., 2003), whereas RPN12a participates in the cytokinin response (Smalle et al., 2002). Recently, RPN9 has also been reported to participate in the regulation of auxin transport and brassinosteroid signaling (Jin et al., 2006), and RPN1 is essential for embryogenesis (Brukhin et al., 2005).

RPT2 is essential for the channel opening of the α-ring of the CP, and is thus necessary for proteasome activity in yeast (Groll et al., 2000). In plants, RPT2a has been shown to have an important role in meristem maintenance (Ueda et al., 2004), and has been linked to the oxidative stress response (Kurepa et al., 2008). Here, we demonstrate that RPT2a also has a specific role in the regulation of leaf organ size that involves control of endoreduplication and cell size, rather than alteration of leaf cell number.

Results

Arabidopsis contains duplicate RPT2 genes

The Arabidopsis genome contains duplicated RPT2 genes, AtRPT2a (At4g29040) and AtRPT2b (At2g20140), which each encode a paralog molecule of the 26S proteasome subunit RPT2. The open reading frame of both AtRPT2a and AtRPT2b genes encode a deduced 443-amino acid residue protein, with a predicted molecular mass of 49.3 kDa. Both AtRPT2 paralogs share a high amino acid sequence similarity that differs by only three amino acid residues (Figure 1a). A detailed motif analysis of the RPT2 protein sequence confirmed that well-known RPT motifs are conserved in both AtRPT2 paralogs, including the ATP/GTP-binding site P-loop (motif A), the AAA-protein family signature (motif B) and a putative nuclear localization signal (NLS) (Figure 1a; Beyer, 1997; Fu et al., 1999; Yang et al., 2004; Shibahara et al., 2002, 2004).

Figure 1.

 Multiple alignment and expression patterns of AtRPT2a and AtRPT2b.
(a) The alignment of the predicted amino acid sequences of AtRPT2a and AtRPT2b (At4g29040 and At2g20140, respectively). Black boxes indicate conserved amino acid residues. Black lines indicate the relative position of the ATP/GTP-binding site P loop (motif A) and the AAA-protein family signature (motif B). A dotted line indicates the relative position of the putative nuclear localization signal (NLS).
(b) RT-PCR analysis of the AtRPT2a, AtRPT2b and EF1α (control) genes in the wild-type. Total RNA was extracted from the different organs at 16 and 45 days after sowing (DAS). For the experimental details, see Experimental procedures.
(c) Histochemical localization of pAtRPT2a and pAtRPT2b::GUS expression. The seedlings of pAtRPT2a (i – iv) and pAtRPT2b::GUS (v–viii) transgenic plants were grown on germination-inducible media (GIM) at 10 DAS. Similar results were observed with three independent lines. Scale bars: 500 μm.

In order to establish the distribution of RPT2 gene expression in Arabidopsis, the accumulation of organ-specific AtRPT2 transcripts was analyzed by RT-PCR (Figure 1b). Transcripts for AtRPT2a and AtRPT2b were detectable in all organs tested at both 16 and 45 days after sowing (DAS), indicating that both paralog genes of Arabidopsis RPT2 are constitutively expressed. To evaluate the expression of the RPT2 paralog genes in further detail, we generated transgenic Arabidopsis plants containing the AtRPT2a and AtRPT2b promoter β-glucuronidase (GUS) reporter gene fusions, pAtRPT2a::GUS and pAtRPT2b::GUS, respectively. GUS activity showed a similar spatial distribution under the AtRPT2a and AtRPT2b promoters, and was detectable in various tissues, such as expanded cotyledons, vascular cells and trichomes (Figure 1c). The AtRPT2a promoter showed strong activity in the shoot meristem, whereas AtRPT2b showed strong activity in both the shoot and root meristems (Figure 1c).

AtRPT2a and AtRPT2b function in a specific manner

In order to investigate the function of the AtRPT2 protein, loss-of-function mutants were identified for both paralog genes. Two T-DNA insertion lines for AtRPT2a (SALK_130019, rpt2a-1; SALK_005596, rpt2a-2) and a T-DNA insertion line for AtRPT2b (SALK_043450, rpt2b-1) were obtained from the Arabidopsis Biological Resource Center (ABRC) (Figure 2a). The rpt2a-1 mutant allele harbored a T-DNA insertion in the first exon of the AtRPT2a gene, whereas the rpt2a-2 insert was located in the fourth exon. The rpt2b-1 mutant allele harbored a T-DNA insertion in the fourth exon of the AtRPT2b gene. The RT-PCR analysis of homozygous mutant plants revealed no transcript from respective genes containing a T-DNA insertion (Figure 2b), indicating that all the inserted lines used in this study are null mutants. As the RPT2 protein is a member of the 26S proteasome, the possibility existed that the null mutations disturbed proteasome activity. The accumulation of ubiquitinated-protein conjugates was determined by detection with a specific antibody (Fujimuro et al., 1994). No major change in the pattern of ubiquitinated proteins was observed in both the single rpt2a-2 and rpt2b-1 mutants, compared with that in the wild type (Figure 2c). A previous study has also shown that the polyubiquitinated protein levels in the rpt2a-2 mutant were not significantly different from wild type. The rpt2a-2 mutant reduced the rate of ubiquitination-dependent proteolysis, whereas the level of 20S CP activity was increased, and thus the apparent proteasomal activity was only slightly decreased in rpt2a-2 (Kurepa et al., 2008). In addition, multiple attempts to generate a double rpt2a rpt2b mutant in the current study were unsuccessful (data not shown). These results suggest that although AtRPT2a and AtRPT2b may function in a specific manner, they are also likely to share a common essential function in the 26S proteasome.

Figure 2.

 Isolation and characterization of T-DNA knock-out mutants of AtRPT2a and AtRPT2b.
(a) Exon structure of the AtRPT2a and AtRPT2a gene, and position of the T-DNA insertions in the respective mutants. Black boxes represent the exons, white boxes represent the intron and grey boxes represent 5′- and 3′-untranslated regions.
(b) Expression of the AtRPT2a and AtRPT2b genes in the wild type, and in the rpt2a-2 and rpt2b-1 mutants. Total RNA from 21-days after sowing (21-DAS; i.e. 3-week-old) vegetative shoot tissue was analyzed by RT-PCR to monitor the accumulation of AtRPT2a, AtRPT2b and EF1α transcripts. Each PCR amplification of the cDNA was done in parallel with primers specific for AtRPT2a, AtRPT2b and EF1α (see Table S2).
(c) Levels of ubiquitin–protein conjugates (Ubn) in the wild type, and in the rpt2a-2 and rpt2b-1 mutants. Equal quantities of total protein were subjected to SDS-PAGE and immunoblot analysis with anti-multiubiquitin chain protein antibodies.

The rpt2a mutant has enlarged rosette leaves, caused by increased cell size

The rpt2a-2 mutant as well as the rpt2a-1 mutant displayed a phenotype of enlarged rosette leaves, whereas the rpt2b-1 mutant did not show any significant morphological difference, compared with the wild type (Figures 3a and S1). The rpt2a-2 mutant contained greatly expanded first and seventh rosette leaves, and serrated leaves, compared with the wild type (Figure 3b,c). To evaluate the effect of the rpt2a-2 mutation on leaf development, a kinematic analysis was carried out on leaves 1 and 2, which are known as the first leaves that develop synchronously. The first leaves have well-characterized cell cycle and expansion parameters (De Veylder et al., 2001; Beemster et al., 2005), beginning with a proliferation phase period up to 11 DAS (cells divide and expand simultaneously), followed by an expansion phase up to 17 DAS (expansion in absence of division), and then a maturation phase up to 31 DAS (no more cell growth), respectively (Figure 4a). In the wild type, the area of the first leaves expanded linearly up to 23 DAS, where they reached mature size (Figures 4a and 5a). In contrast, the area of the rpt2a-2 mutant leaves was larger than that of the wild type at 17 DAS, and the linear expansion continued up to 31 DAS. The average area of the rpt2a-2 first leaf was only slightly greater than the wild type at 11 DAS (3.13 and 2.98 μm2, respectively), but this difference increased significantly to approximately 1.5-fold at 17 DAS, and to twofold at 31 DAS (Figure 5a).

Figure 3.

 Morphology of the rpt2a mutant.
(a) Morphology of the wild type, and the rpt2a-2 and rpt2b-1mutants at 49 days after sowing (DAS).
(b) The fifth rosette leaf of the wild type and the rpt2a-2 mutant at 35 DAS.
(c) The seventh rosette leaf of the wild type and rpt2a-2 mutant at 35 DAS. Scale bars: 1 cm.

Figure 4.

 Growth, cell area and ploidy level of the first leaf pair in an rpt2a mutant.
(a) First-leaf areas of the wild type (WT) and rpt2a-2 mutant at major developmental stages. Scale bars: 1 mm.
(b) Abaxial epidermis cell areas of the first leaves of the WT and the rpt2a-2 mutant at each development stage. Scale bars: 100 μm.
(c) DNA ploidy level distribution of the first leaves of the wild type and the rpt2a-2 mutant at each developmental stage.
(d) DAPI staining of the nuclei of the peeled adaxial epidermal layer of the fifth leaf in the wild type and the rpt2a-2 mutant at 42 days after sowing. Scale bars: 100 μm.

Figure 5.

 Kinematic analysis of the growth of first leaves of the rpt2a mutant.
(a) First-leaf areas of the wild type (WT) and the rpt2a-2 mutant at each developmental stage.
(b) Epidermal cell number on the abaxial side of first leaves of the WT and the rpt2a-2 mutant.
(c) Average cell area on the abaxial side of the first leaves of the WT and the rpt2a-2 mutant. The developmental stages for leaf development are: 11 days after sowing (DAS), proliferation (cells divide and expand simultaneously); 17 DAS, expansion (expansion in the absence of division); 23 and 31 DAS, mature (no more cell growth).

The epidermis has an important role in leaf morphology, and has been well characterized as a model for leaf cell development. Observation of epidermal cells in the first leaves revealed that rpt2a-2 mutant leaves contained larger cells compared with that of the wild type during all of the growth stages (Figures 4b and 5c). In the wild-type plant, the average cell size in the first leaves increased linearly up to 23 DAS, similar to that observed for leaf area (Figure 5c). In contrast, the average cell size in rpt2a-2 leaves increased up to 31 DAS, indicating continuous cell expansion, not only during the expansion stage but also in the mature stage. Indeed, extremely expanded cells were observed in the mature leaf of the rpt2a-2 mutant (Figure 6). The number of cells in the first leaves of the wild type and rpt2a-2 remained similar during development (Figure 5b). Therefore, the greater expansion of leaf area in rpt2a-2 was caused primarily by an increase in cell size, and not by an increase in cell number.

Figure 6.

 Area size distribution of the epidermal cells of first leaves of the wild type and of the rpt2a-2 mutant, at each developmental stage: (a) 11 days after sowing (DAS), (b) 17 DAS, (c) 23 DAS and (d) 31 DAS.

The expanded cell size in rpt2a mutant leaves correlates with extended endoreduplication

Endoreduplication is a type of cell cycle where nuclear chromosomal DNA replication occurs without cell division (Edgar and Orr-Weaver, 2001; Nagl, 1976), and there is a positive correlation between higher ploidy levels and larger cell sizes in Arabidopsis (Galbraith et al., 1991; Melaragno et al., 1993). We therefore investigated the ploidy level in the first leaves of wild-type and rpt2a-2 plants using flow cytometry (Figure 4c). The distribution of cell ploidy levels in wild-type and rpt2a-2 plants was similar at 11 DAS; however, theye differed remarkably at 17, 23 and 31 DAS. The rpt2a-2 mutant started to show relatively more 8C nuclei compared with the wild type at 17 DAS, and a reduction in 2C and 4C nuclei coincided with an increase in 8C and 16C nuclei at 23 DAS. During the 23–31-DAS phase, the fraction of 4C cells represented the highest peak in the wild type. In contrast, the fraction of 8C cells represented the major peak in rpt2a-2 over the equivalent stage. Moreover, only rpt2a-2 leaves showed a fraction of 32C cells at 31 DAS. These results illustrate that endoreduplication was extended in the first leaves of the rpt2a-2 mutant, and continued to a later phase compared with that in wild-type plants.

To evaluate the correlation between the enlarged leaves and cell nuclei size, the nuclei of epidermal cells from the fifth rosette leaves were observed by DAPI staining. Consistent with the increase in endoreduplication, larger nuclei were observed in the larger cells of rpt2a-2 compared with the wild type (Figure 4d).

Increased branch number and nuclear size of rpt2a-2 trichomes

Trichomes are polyploid cells with a well-established morphogenetic pattern, where branching is genetically defined and associated with the occurrence of endocycles (Hülskamp et al., 1998). To further evaluate the extended endoreduplication in the rpt2a-2 mutant, we examined the shape and nuclear size of trichomes on the fifth rosette leaf. Trichomes of the rpt2a-2 mutant were larger than the wild type (Figure 7a), and approximately 40% of rpt2a-2 trichomes developed four or five branches compared with the mostly triradiate trichomes of the wild type (Figure 7b). In the rpt2a-2 mutant, the proportion of three-branched trichomes was reduced from 95.77 to 66.50%, whereas the number of four-branched trichomes increased from 1.80 to 28.93%, compared with the wild type (Table S1). Trichome nuclear size was determined by measuring two-dimensional images of DAPI-stained trichomes, and revealed that the size of rpt2a mutant trichome nuclei are, on average, larger than those of the wild type (Figure 7c). The increased trichome branch numbers in the rpt2a-2 mutant correlated with the large increase in nuclear size, suggesting that rpt2a mutant trichomes also undergo extra endoreduplication. The kaktus mutant also displays increased branch numbers of trichomes as a result of increased ploidy levels. A double mutant generated by crossing rpt2a with kaktus (rpt2a kak) showed an additive branching phenotype (Figure S2). KAKUTUS encodes a putative HECT-domain E3 ligase (El Refy et al., 2003; Spitzer et al., 2006). However, these results suggest that RPT2a functions to regulate trichome endoreduplication and branching separately from KAKTUS.

Figure 7.

 Morphology and population of trichomes on the fifth leaves of the rpt2a mutant.
(a) Scanning electron microscopic image of a trichome (i, iv), and DAPI staining of the nuclei of the isolated trichome, from fifth leaves of the wild type and the rpt2a-2 mutant. Scale bars: (i) and (iv), 300 μm; (ii) and (v), 100 μm; (iii) and (v), 50 μm.
(b) Distribution of trichome populations on rosette leaves. Trichomes were counted on the fifth leaves of seven plants for each genotype.
(c) Nuclear size distribution of the isolated trichomes from the fifth leaves of each genotype.

Increased expression of genes encoding DNA replication licensing components in rpt2a-2

In order to study the temporal expression of AtRPT2 genes during the development of the first leaves, the accumulation of AtRPT2a transcripts was analyzed by RT-PCR. The accumulation of both AtRPT2a and AtRPT2b transcripts was similar, and had higher accumulation at 11 DAS, whereas the AtPBF gene (encoding the 20S CP subunit) was constitutively expressed during development (Figure 8a).

Figure 8.

 Expression of G1/S-related genes in the rpt2a-2 mutant.
(a) RT-PCR analysis of 26S proteasome subunit genes: AtRPT2a, AtRPT2b (19S RP subunit) and AtPBF (20S CP subunit). EF1α was used as a control.
(b) RT-PCR analysis of cell cycle-specific genes: CYCD3;1, CDC6a, CDC6b, CDT1a, CDT1b, HISH4, CYCA3;1 and EF1α (control).
(c) RT-PCR analysis of cell cycle-specific genes: KRP1, KRP2, CDKB1;1, CYCB1;1 and EF1α (control). Total RNA was prepared from shoots of the wild type (WT) and the rpt2a-2 mutant.
(d) Quantification of cell cycle-specific gene expression. Relative levels of CDC6b, CDT1b, CDKB1;1 and CYCB1;1 gene transcripts in rpt2a-2 mutants at 11 days after sowing. Values are the averages of three experiments, and the small bars represent the SEs.

Next, we examined the expression of cell cycle-related genes specific for the G1 phase (CYCD3;1; Riou-Khamlichi et al., 2000) and the S phases (CDC6a, CDC6b, CDT1a, CDT1b, HISH4 and CYCA3;1; Mariconti et al., 2002; Menges et al., 2005). Transcripts of CYCD3:1, CDC6b, CDT1a, CDT1b, HISH4 and CYCA3;1 were increased in the rpt2a-2 mutant compared with the wild type at 11 DAS, whereas no obvious difference was observed at 17 and 31 DAS (Figure 8b,d). It has been reported that the ectopic expression of CDC6a and CDT1a increases the endoreduplication level (Castellano et al., 2004). These results suggest that the extended endoreduplication in rpt2a-2 may result from the disrupted regulation of DNA replication licensing components at the proliferation stage.

The expression of cell cycle-related genes specific for the G2/M phase (CDKB1;1, CYCB1;1, KRP1, KRP2; Shaul et al., 1996) were also examined. At the proliferation stage (11 DAS), there was no difference in KRP2 transcript accumulation between the wild type and rpt2a-2. However, the expression of CDKB1;1, CYCB1;1 and KRP1 was increased in rpt2a-2 compared with the wild type (Figure 8c,d). The enhanced expression of CDKB1;1 and CYCB1;1 is likely to promote cell division, whereas the enhanced expression of KRP1 inhibits entry into mitosis (Weinl et al., 2005). The changes in expression of these genes were not consistent with the rpt2a-2 phenotype, which has a normal cell number, despite the extended endoreduplication, suggesting that a loss of AtRPT2a results in the disrupted regulationof the G2/M checkpoint.

RPT2b can not complement the rpt2a mutant phenotype

Our analyses showed that RPT2a and RPT2b had a similar expression pattern, but were not redundant with regards to their role in leaf development. In order to examine whether these paralog proteins are indeed functionally equivalent, we attempted to compliment the rpt2a mutant with constructs expressing RPT2b cDNA under the control of the RPT2a promoter (pRPT2a) (Figure 9a). Constructs containing the RPT2a cDNA under the control of pRPT2a generally complemented the enlarged-leaves phenotype of the rpt2a mutant (Figure 9a–c). On the other hand, the pRPT2a-RPT2b was unable to complement the rpt2a mutant phenotype, namely in the extended endoreduplication (Figure 9a–c). These results confirm that the RPT2a and RPT2b proteins are indeed functionally distinct with regards to a role in leaf development.

Figure 9.

 Complementation analysis.
(a) Morphology of the wild type (WT), the rpt2a-2 mutant carrying pRPT2a-RPT2a and the rpt2a-2 mutant carrying pRPT2a-RPT2b.
(b) Expression of the endogenous and exogenous AtRPT2a, endogenous and exogenous AtRPT2b, in the WT, and the expression of rpt2a-2 carrying pRPT2a-RPT2a and rpt2a-2 carrying pRPT2a-RPT2b in the mutant.
(c) DNA ploidy level distribution of the rosette leaves of the wild type, rpt2a-2 carrying pRPT2a-RPT2a and rpt2a-2 carrying pRPT2a-RPT2b.

Discussion

We show here that the loss of function of the AtRPT2a gene results in an enlargement of leaf organs, an increased epidermal cell size, with extended endoreduplication, and increased polyploidy, and also an increase in trichome branching and nuclear size. These observations indicate AtRPT2a, a 19S proteasome subunit protein, is required for the control of leaf organ size through the regulation of endoreduplication and the control of cell size in Arabidopsis leaf cells.

The development of the first leaf pair is a good model for analyzing the cell cycle and endoreduplication in plants (Beemster et al., 2005). In yeast and animals, the cell cycle is primarily regulated by proteolysis of the 26S proteasome (Reed, 2003). Paralog proteins of important cell-cycle regulators (cyclinD, cyclinA and cyclin dependent kinase inhibitor) have also been reported to be degraded by the 26S proteasome in higher plants (Criqui and Genschik, 2002; Planchais et al., 2004). The first leaf pair at 11 DAS contains mainly proliferating cells (2C as DNA content). In contrast with that in the wild type, the rpt2a mutant showed a higher proportion of cells with a DNA content of 4C at this stage. As cell numbers per leaf do not differ between the wild type and the rpt2a mutant, these 4C cells are likely to represent those that have already entered the endoreduplication cycle (endocycle). Control of DNA replication is exerted at the expression level of G1-specific factors, such as the CYCD and DNA replication licensing components CDT1 and CDC6 (Kelly and Brown, 2000; Bell and Dutta, 2002; Sugimoto et al., 2004). In Arabidopsis, plants overexpressing AtCDT1a and AtCDC6a display extra endocycles, leading to an increase in DNA content for the leaf and trichome cells that results in extra trichome branches (Castellano et al., 2004). The increased expression of AtCDT1b and AtCDC6b genes, as well as S-specific genes such as HISH4 and CYCA3;1, in the rpt2a-2 mutant is therefore consistent with the endoreduplication phenotype, and indicated that the DNA replication process is promoted at the early stage of leaf development.

During the mitotic cycle, inhibition of G2/M has been reported to promote entry into the endocycle (Boudolf et al., 2004; Verkest et al., 2005a,b). In the rpt2a-2 mutant, the CDKB1;1 gene was transcriptionally enhanced, although the CDKA;1 remained relatively constant. CDKB1;1 acts to promotes mitosis and cell division; however, the rpt2a-2 mutant did not show increased cell numbers per leaf, indicating that the G2/M checkpoint is deregulated to allow endoreduplication. Taken together, these results imply that AtRPT2a associates with the G2/M checkpoint to inhibit transition from mitosis to the endocycle at the early stage of leaf development.

Plant genomes generally contain duplicated gene pairs for CP and RP subunits of the 26S proteasome (Smalle et al., 2002; Fu et al., 1998, 1999; Shibahara et al., 2002, 2004). Site-directed mutagenesis analyses in yeast revealed that each RPT1–6 subunit is not functionally redundant, despite its multiplicity and sequence similarity (Rubin et al., 1998), and that opening of the CP channel is primarily mediated by the RPT2 subunit (Rubin et al., 1998; Köhler et al., 2001). The Arabidopsis RPT2 genes, AtRPT2a and AtRPT2b, share a highly conserved amino acid sequence, and have very similar expression patterns. We have not been able to generate a double rpt2a rpt2b mutant, suggesting that each gene shares a common essential function. On the other hand, only the AtRPT2a loss-of-function mutant had enlarged leaves, and RPT2b could not complement the rpt2a phenotype, implying that both AtRPT2 paralogs are not functionally equivalent for leaf-size control. It is possible that AtRPT2a and AtRPT2b undergo different post-transcriptional modifications.

Other Arabidopsis mutants showing increased endoreduplication also have larger leaf cells; however, this is at the expense of mitotic cycles, and the leaf organ size is reduced because of the dramatic reduction in cell number per leaf (De Veylder et al., 2001; Boudolf et al., 2004; Verkest et al., 2005a). In this study, the loss of function of AtRPT2a resulted in an increase in the size of leaf cells, without a reduction in cell number. These results imply that AtRPT2a-type 19S proteasome (19SAtRPT2a) activity is important for the normal specification of leaf cell ploidy and organ size after the normal completion of the mitotic cycle phase, possibly through the regulation of endoreduplication in developing leaves.

Experimental procedures

Plant materials and growth conditions

For the germination of the A. thaliana (ecotype Columbia-0) wild type and mutants, seeds were surface-sterilized and placed on MS medium supplemented with 2% sucrose (germination-inducible medium, GIM). After cold treatment for 2 days to synchronize germination, seeds were transfered to 22°C and 50% relative humidity, under a 16-h light/8-h dark cycle (this time point indicates 0 DAS). Seeds of the rpt2a-2, rpt2a-2 and rpt2b-1 mutants were obtained from ABRC (stock numbers SALK_005596, SALK_130019 and SALK_043450, respectively; http://www.biosci.ohio-state.edu/pcmb/Facilities/abrc/abrchome.htm). The sequences bordering the T-DNA insertion were determined using the primer pairs listed in Table S2.

Transcript level analysis

Total RNA was extracted by the guanidine thiocyanate method (Chomczynski and Sacchi, 1987). Total RNA (0.6 μg RNA) was used as a template for the first-strand cDNA synthesis using ReverTraAce-α-® reverse transcriptase (TOYOBO, http://www.toyobo.co.jp/e/index.htm). First-strand cDNA (0.7 μl) was then assayed with gene-specific DNA fragments, using the primer pairs listed in Table S2. PCR amplification was performed with optimum cycles for each gene using the Taq DNA polymerase (New England BioLabs, http://www.neb.com). The amplified fragments were electrophoresed on 1.2% (w/v) agarose gels, and were visualized by ethidium bromide staining. Real-time PCR was performed with the Power SYBR Green PCR Master Mix (Applied Biosystems, http://www.appliedbiosystems.com) on an Applied Biosystems 7300 Real-Time PCR system (Applied Biosystems). The relative quantitation of the gene expression is based on the comparative CT method (User Bulletin No. 2: ABI PRISM 7700 Sequence Detection System, 1997), using AtEF1α as the reference gene. The following PCR program was used: 2 min at 50°C; 10 min at 95°C; 40 cycles of 15 s at 95°C and 1 min at 60°C. Two biological and three technical replicates were performed. The sequences of the primers used are listed in Table S2.

Promoter::GUS fusion construct and GUS staining assays

Genomic fragments of the promoter region of AtRPT2a (2955-bp upstream of the ATG) and AtRPT2b (501-bp upstream of the ATG) were amplified from the genomic DNA of the wild type using the primer sets listed in Table S2. Construction of the promoter::GUS::NOS fusion protein was performed essentially according to the protocol supplied with the GATEWAY system (Invitrogen, http://www.invitrogen.com). In brief, promoter regions were first introduced into the TOPO cloning vector pENTRTM/D-TOPO using specifically designed primers, as described by the manufacturer. After verifying the orientation of the inserted sequence in the plasmid, the promoter region was further transformed into the destination vector pMDC164, a binary vector in which the GUS gene is fused at the C terminus, and in which there is no promoter, through an LR reaction. The fusion plasmid was transformed into A. tumefaciens GV3101 by electroporation, and then introduced into wild-type Arabidopsis plants using the Agrobacterium-mediated transformation method, as described by Clough and Bent (1998). To select for transgenic progeny, T1 seeds from primary transformants were planted on germination medium containing 50 μg L−1 hygromycin.

GUS assays were performed according to the standard protocols (Jefferson et al., 1987), with minor modifications. The young seedlings (T2) were incubated in 90% acetone for 15–30 min on ice, and then the acetone was removed. After that, the GUS staining buffer was added to the sample (100 mm of 5-bromo-4-chloro-3-indolyl β-d-glucuronide, 2 mm of potassium ferricyanide, 2 mm of potassium ferrocyanide, 10 mm EDTA, 0.1% Triton X-100, 50 mm Na phosphate buffer, pH 7.0). The reaction was performed at 37°C in the dark for 1 h (pAtRPT2a::GUS) or 5 h (pAtRPT2b::GUS), and the material was passed through a 70% ethanol, ethanol/acetic acid (6:1) and 70% ethanol wash to remove the chlorophyll. Samples were cleared using chloral hydrate/glycerol solution (56 g:7 ml in 50 ml deionized water) before observation under a light microscope (MZFLIII; Leica Geosystems, http://www.leica-geosystems.com).

Trichome analysis

Trichomes were isolated from the fifth rosette leaves, as described by Zhang and Oppenheimer (2004). The isolated trichomes were washed three times with wash buffer (0.02% Tween-20 and 50 mm Na phosphate buffer, pH 7.0), and DAPI staining buffer was added (1 μg μl−1; 4′,6-deamidinopphenylindole, 0.02% Tween-20, 0.1 mg ml−1p-phenylene diamine in Na phosphate buffer, pH 7.0) for 15 min. The DAPI-stained nuclei were photographed by fluorescence microscopy (DMR; Leica Geosystems). The nuclear size of trichomes was measured using ImageJ software (http://rsb.info.nih.gov/ij/) Trichomes were observed using low vacuum scanning electron microscopy (S-3000N; Hitachi, http://www.hitachi.com/), following the manufacturer’s protocol.

Kinematic analysis, DAPI staining and ploidy analysis of the first leaves

The first leaves were fixed with FAA solution for 1 h, and washed twice with Na phosphate buffer. The first leaves were then cleared and DAPI stained as described for the trichome analysis. Kinematic analysis was performed as described previously (De Veylder et al., 2001; Bemis and Torii, 2007) using ImageJ software. For flow cytometric analysis, the first-leaves nuclei were extracted and stained with CyStain UV precise P (Partec, http://www.partec.com), following the manufacturer’s protocol, and performed as described previously (Yoshizumi et al., 2006).

Complementation test

The complementation constructs were based on the plasmid pMDC32 (Curtis and Grossniklaus, 2003), where two copies of the 35S promoter were exchanged with the RPT2a promoter (2955-bp upstream of ATG; identical promoter size to pAtRPT2a::GUS) using HindII and KpnI restriction sites. RPT2a or RPT2b cDNA was inserted downstream of the selected promoter by recombination cloning. rpt2a mutants were transformed as described above. Transformed plants were evaluated for ploidy analysis.

Acknowledgements

We are grateful to Drs Takeshi Yoshizumi and Minami Matsui for flowcytometery analysis, Drs Minako Ueda and Kiyotaka Okada for providing the hlr mutant, and to Ms Yoko Osaka for technical assistance. We are also grateful to Dr Derek B. Goto (Hokkaido University) for his critical reading of the manuscript. This work was supported by a Grant-in-Aid for Scientific Research (nos 19657013 and 19039001) to JY, in part by the Program for Basic Research Activities for Innovative Bioscience (PROBRAIN), by grants from the 21st century COE Hokkaido University (YS and JY) and The Akiyama Foundation to JY. KS acknowledges Research Fellowships from the Japan Society for the Promotion of Science for Young Scientists (2008–2010).

Ancillary