Arabidopsis ULTRAVIOLET-B-INSENSITIVE4 Maintains Cell Division Activity by Temporal Inhibition of the Anaphase-Promoting Complex/Cyclosome^,

This work identifies ULTRAVIOLET-B-INSENSITIVE4 as a plant-specific inhibitor of the Anaphase-Promoting Complex/Cyclosome E3-ubiquitine ligase. Its activity allows the temporal accumulation of mitotic cyclins during the DNA replication phase, thereby linking DNA replication with mitosis.

Abstract

The anaphase-promoting complex/cyclosome (APC/C) is a multisubunit ubiquitin ligase that regulates progression through the cell cycle by marking key cell division proteins for destruction. To ensure correct cell cycle progression, accurate timing of APC/C activity is important, which is obtained through its association with both activating and inhibitory subunits. However, although the APC/C is highly conserved among eukaryotes, no APC/C inhibitors are known in plants. Recently, we have identified ULTRAVIOLET-B-INSENSITIVE4 (UVI4) as a plant-specific component of the APC/C. Here, we demonstrate that UVI4 uses conserved APC/C interaction motifs to counteract the activity of the CELL CYCLE SWITCH52 A1 (CCS52A1) activator subunit, inhibiting the turnover of the A-type cyclin CYCA2;3. UVI4 is expressed in an S phase-dependent fashion, likely through the action of E2F transcription factors. Correspondingly, uvi4 mutant plants failed to accumulate CYCA2;3 during the S phase and prematurely exited the cell cycle, triggering the onset of the endocycle. We conclude that UVI4 regulates the temporal inactivation of APC/C during DNA replication, allowing CYCA2;3 to accumulate above the level required for entering mitosis, and thereby regulates the meristem size and plant growth rate.

INTRODUCTION

Unidirectional and irreversible progression through the cell cycle is mediated through the selective proteolytic regulation of both positive and negative regulators of cyclin-dependent kinases (CDKs). Cyclins, cell cycle inhibitors, and many other proteins are marked with ubiquitin by E3 ubiquitin ligases and, subsequently, recognized and destroyed by the 26S proteasome (Hershko, 1997; Peters, 2002). Two related E3 ubiquitin-ligase complexes are most intimately dedicated to basic cell cycle regulation, namely Skp1-Cullin-F-box (SCF)-related complexes and the anaphase-promoting complex/cyclosome (APC/C), which operate at the G1-to-S transition and in the M-G1 phases, respectively (Marrocco et al., 2010).

The APC/C is a multisubunit complex that is highly conserved among eukaryotes, consisting of at least 11 different subunits, including a catalytic core composed of APC2 and APC11 (Page and Hieter, 1999; Tang et al., 2001). Besides this core, the APC/C requires essential subunits for its activation, including the docking factor APC10/Doc1 (Destruction of D-box) and the CDC20/Fizzy or CDH1/Fizzy-related (FZR) activator subunits. Both CDC20 and CDH1 activators confer substrate specificity, in part by their different activation timing and by the recognition of distinct destruction degrons, including the D-box (RxxLxxxxN), KEN-box (KENxxxE/D/N), and GxEN-box (Pfleger and Kirschner, 2000; Castro et al., 2003; Lénárt and Peters, 2006).

In addition to activators, the APC/C is controlled by inhibitory proteins. Such inhibitors have been described for different species. In Homo sapiens, Mus musculus, and Xenopus laevis, the Early mitotic inhibitor1 (Emi1), known as Regulator of cyclin A1 (Rca1) in Drosophila melanogaster, inhibits APC/C^CDH1 activity by operating as a pseudosubstrate (Di Fiore and Pines, 2007). Expression of Emi1 is regulated by E2F transcription factors at the G1-to-S transition, restricting the APC/C^CDH1 activity to the interphase and allowing the accumulation of the APC/C^CDH1 targets, such as geminin and A- and B-type cyclins, during the late G1 and early S phases (Hsu et al., 2002; Verschuren et al., 2007; Tategu et al., 2008). Xenopus extracts depleted from Emi1 refrain from entering mitosis, because they do not accumulate mitotic cyclins (Reimann et al., 2001). Likewise, Drosophila rca1 mutants fail to accumulate A-type cyclins and arrest in G2 (Grosskortenhaus and Sprenger, 2002). Similarly, in budding yeast (Saccharomyces cerevisiae), APC/C^CDH1 modulator1 (Acm1) suppresses degradation of mitotic cyclins by acting as a pseudosubstrate of APC/C^CDH1, although it is unrelated in sequence to Emi1 and Rca1 (Burton et al., 2011). Based on sequence identity, no proteins orthologous to Emi1/Rca1 or Acm1 have been recognized in plants.

The genome of Arabidopsis thaliana encodes three CDH1/FZR-related proteins, annotated as the CELL CYCLE SWITCH 52 genes (CCS52A1, CCS52A2, and CCS52B) (Cebolla et al., 1999). Whereas the function of CCS52B is unclear, both CCS52A1 and CCS52A2 have been implicated in the control of endocycle onset in different species (Cebolla et al., 1999; Lammens et al., 2008; Narbonne-Reveau et al., 2008; Mathieu-Rivet et al., 2010). The endoreduplication cycle represents an alternative to the mitotic cell cycle, during which mitosis is completely repressed and DNA re-replication is stimulated, resulting in a stepwise doubling of the DNA content. Whereas this phenomenon occurs only in certain cell types in mammals, endoreduplication is widespread among plants, particularly in angiosperms, in which it might occur in most cell types except gametes, meristematic tissues, and guard cells (Ullah et al., 2009). In plants, the onset of endoreduplication is often coupled to cell differentiation, because it coincides with the switch from the proliferation to the differentiation program (Beemster et al., 2005; Inzé and De Veylder, 2006). Furthermore, in Arabidopsis, it plays a decisive role in determining differentiation of leaf trichome precursor cells (Bramsiepe et al., 2010).

Major efforts have advanced our understanding of how plants exit the mitotic cell cycle and switch to the endocycle. Among others, cell cycle proteins important for the timing of endocycle onset in Arabidopsis include the A-type cyclin CYCA2;3 and the plant-specific B-type CDK CDKB1;1, which, as a complex, repress the exit from the mitotic cell cycle (Imai et al., 2006; Boudolf et al., 2009). Recently, the cyclin CYCA2;3 has been identified as one of the targets of the APC/C^CCS52A1 in the root cell elongation zone, hinting at a role in the regulation of the endocycle by preventing an increase in CYCA2;3-CDKB1;1 activity at the G2-to-M transition (Boudolf et al., 2009).

Through tandem affinity purification (TAP), we purified the subunits of the Arabidopsis APC/C, revealing the presence of plant-specific binding partners (Van Leene et al., 2010), including ULTRAVIOLET-B INSENSITIVE4 (UVI4), which has been demonstrated previously to affect endoreduplication in leaves (Hase et al., 2006). Similarly, the UVI4 homolog OMISSION OF SECOND DIVISION (OSD1) was found to be an APC/C subunit (Van Leene et al., 2010) and has been proposed to regulate the oscillations of CDK activity that are required to progress through the meiosis I-to-meiosis II transition during pollen gamete formation (d’Erfurth et al., 2009, 2010; Bulankova et al., 2010). Here, we demonstrate that UVI4 regulates the APC/C activity through binding to the CCS52A activator subunits. UVI4 is likely under the direct regulation of the E2F pathway and suppresses endocycle onset by inhibiting CYCA2;3 destruction in a timely fashion and hence regulating the meristem size. Based on structural, biochemical, and genetic data, we conclude that UVI4 is a functional homolog of Emi1, representing a plant APC/C inhibitor.

RESULTS

UVI4 Is a Part of the APC/C

By means of TAP, UVI4 had been identified as an APC/C copurifying protein (Van Leene et al., 2010). Although TAP allowed the detection of the entire composition of APC/C complexes, it did not convey any information on the direct interaction between core subunits and associated factors. Therefore, to study the interaction of UVI4 with the different APC/C subunits in more detail, we set up a yeast two-hybrid (Y2H) interaction matrix, testing all possible interactions between 18 different APC/C subunits. Y2H interactions were analyzed with a yeast mating protocol in a microtiter plate format (Boruc et al., 2010) and by scoring for His prototrophy as output signal. Three clones could not be used as bait, because of autoactivation (APC2, APC10, and APC11), but could still be used as prey. The interaction between two proteins was considered positive when growth was more than threefold enhanced on His-lacking medium relative to the background signal (see Methods). In comparison with the robust binding among the APC/C subunits observed in the TAP assay, the number of interacting partners identified by Y2H was relatively low (for raw data of all tested interactions, see Supplemental Data Set 1 online), indicating that most APC/C subunits associate through cooperative binding. Out of the 288 tested interactions, only 14 protein pairs scored positive, including the previously identified APC2-APC11, APC10-HOBBIT (HBT), and APC7-HBT interactions (Capron et al., 2003; Eloy et al., 2006; Pérez-Pérez et al., 2008). Interestingly, in addition to the association with the APC5 scaffold protein, UVI4 also bound directly to the APC/C-activating subunit CCS52A1, implying that UVI4 binds near the APC/C catalytic site (Figure 1).

Figure 1.

Y2H Interactome of APC/C Proteins.

All known APC/C subunits, including several CDC20 and CCS52 activators and recently found UVI4, OSD1, and At1G32310, were tested for binding to each other in an 18×18 matrix. Interactions were found to be positive when the OD₆₀₀ value exceeded threefold the background value of the negative control of yeast grown in medium without His (−HIS). Arrowheads point to the interacting prey; double arrowheads indicate both interact as prey.

UVI4 Contains Highly Conserved Motifs Required for CCS52A1 Binding

UVI4 seems to be a plant-specific protein without sequence homology with any annotated protein. However, alignment of UVI4 with its dicot homologs identified by the PLAZA online Web tool (Proost et al., 2009) revealed the presence of many highly conserved putative APC/C interaction domains, including a D (RxxL)- and a GxEN-box. In addition, nearly all UVI4-like proteins share a C-terminal Met-Arg (MR) tail, which resembles the known Ile-Arg tail required for binding of the APC/C activator proteins to the tetratricopeptide repeat-containing subunits (Vodermaier et al., 2003) (Figure 2A). To verify whether these domains represent true APC/C interaction motifs, a mutagenesis analysis was undertaken. For both the D-box and GxEN-box, the amino acid residues essential for APC/C binding were substituted to Ala (mD2, RxxL→AxxA; mG, GxEN→AxAA) (Figure 2B), whereas the MR tail was deleted (−MR). We also mutated a putative second D-box motif (mD1) in UVI4 that was seemingly not conserved among the different plant proteins. Subsequently, we tested the interaction of the mutant UVI4 proteins with CCS52A1 by means of the Y2H approach. In contrast with the wild-type and mD1 UVI4 proteins, the mD2, mG, and −MR mutants were unable to interact with CCS52A1, illustrating that these three mutated domains are essential for the APC/C interaction (Figure 2C).

Figure 2.

Structural-Functional Analysis of UVI4.

(A) Alignment of protein sequences homologous to UVI4 obtained from the PLAZA online Web tool. The conserved GxEN-box, D-box, and MR tail motifs are indicated.

(B) Schematic representation of UVI4, showing the typical APC interaction motifs and their respective amino acid substitution in the mutant alleles.

(C) Y2H interactions between wild-type (WT) and mutant UVI4 proteins and the CCS52A1 activator subunit. Interactions are positive upon growth in the absence of His (−HIS). The GUS protein was used as a negative control.

(D) and (E) DNA ploidy and leaf trichome complementation analysis of 21-d-old wild-type Columbia-0 (Col-0) (D) versus representative uvi4 lines expressing mD1, mD2, mG, and −MR mutant alleles (E).

Bars in (D) = 300 μm; bars in (E) = 500 μm.

To investigate whether different conserved domains are needed for UVI4 functionality, we tested the capacity of genes encoding the mutant alleles to complement the uvi4 mutant phenotype. Previously, uvi4 loss-of-function plants had been shown to display an increased DNA ploidy level because of a faster endocycle onset, correlating with an overbranched leaf trichome phenotype (Hase et al., 2006) (Figure 2D). Wild-type Arabidopsis trichomes have on average three branches, whereas uvi4 leaf hairs mostly display four to five branches. The wild-type and mutant alleles of UVI4 were cloned under the control of the UVI4 promoter (1046 bp upstream of the start codon) and transformed into a uvi4 loss-of-function mutant background. Wild-type UVI4 and mutated mD1 and mD2 uvi4 complemented the trichome phenotype, whereas plants expressing the mG or −MR alleles were not rescued (Figure 2E, Table 1). In addition to trichome branching, the DNA ploidy levels of the mature first leaves were analyzed. Complementation of the endoreduplication phenotype could be observed in uvi4 mutants harboring the wild-type, mD1, and mD2 alleles. By contrast, no complementation could be seen in mutants transformed with the mG and −MR uvi4 alleles (Figure 2D, Table 1). These data show that loss of the conserved GxEN-box or MR tail renders UVI4 unable to associate with the APC/C holocomplex and to exert its function.

Table 1.

Complementation Analysis of uvi4 Mutants

UVI4 Allele	Independent Lines Screened	Lines Showing Complementation
Wild type	4	4
mD1	2	2
mD2	3	3
mG	3	0
−MR	3	0

CCS52A1 Is Required for UVI4 Activity

Our data indicated that UVI4 might regulate APC/C activity by interacting with at least the APC/C-activating subunit CCS52A1. To test this hypothesis, we generated a uvi4 ccs52a1-1 double mutant. Mutation of uvi4 leads to increased DNA ploidy levels, whereas ccs52a1 mutants show decreased levels (Lammens et al., 2008). Investigation of the DNA ploidy levels of the mature first leaves of the uvi4 ccs52a1-1 double mutant demonstrated that its ploidy distribution was more similar to that of the ccs52a1-1 single mutant (Figure 3A), indicating that ccs52a1-1 might be epistatic over uvi4. Nevertheless, endoreduplication levels in the uvi4 ccs52a1-1 double mutant were slightly higher than those of the ccs52a1-1 mutant (see Supplemental Table 1 online). This observation might be attributable to the regulation of endoreduplication in leaves by CCS52A2, in addition to CCS52A1 (Lammens et al., 2008). By contrast, CCS52A1 is the predominant APC/C activator involved in endocycle onset in trichomes, as illustrated by a decrease in trichome branch number and DNA content in ccs52a1 but not in ccs52a2 mutants (Kasili et al., 2010). When trichomes of the uvi4 ccs52a1-1 double mutant were examined, mainly trichomes with only two branches were detected, similar to the ccs52a1-1 single mutant trichome phenotype (Figure 3B; see Supplemental Figure 1 online). To investigate whether the trichome branch number phenotype correlated with the DNA ploidy level of the hair cells, trichome nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI) and analyzed by epifluorescence microscopy. Mutant uvi4 trichome nuclei were significantly larger compared with wild-type trichome nuclei, corresponding to an increased DNA content. By contrast, ccs52a1-1 mutant trichomes contained significantly smaller nuclei compared with the wild type. In the uvi4 ccs52a1-1 double mutant, the size of the trichome nucleus was identical to that of the ccs52a1-1 single mutant (Figure 3C; see Supplemental Figure 2 online), indicating that ccs52a1-1 is epistatic over uvi4. These data imply that the presence of the CCS52A1 activator subunit is required for UVI4 to exert its function on the APC/C activity.

Figure 3.

CCS52A1 Is Required for UVI4 Activity.

(A) Flow cytometric analysis of wild-type (Col-0), uvi4, ccs52a1-1, and uvi4 ccs52a1-1 3-week-old first leaves. Data are representative for the mean (n = 3).

(B) Scanning electron microscope images of wild-type (Col-0), uvi4, ccs52a1-1, and uvi4 ccs52a1-1 trichomes.

(C) Epifluorescence images of DAPI-stained wild-type (Col-0), uvi4, ccs52a1-1, and uvi4 ccs52a1-1 trichome nuclei. Images are representative for the mean (n ≥ 13).

Bars in (B) = 300 μm; bars in (C) = 25 μm.

[See online article for color version of this figure.]

UVI4 Regulates CYCA2;3 Stability

Previously, CCS52A1 had been shown to regulate CYCA2;3 abundance (Boudolf et al., 2009). Therefore, because UVI4 associates with CCS52A1, we tested whether UVI4 deficiency affected cyclin abundance by introducing a β-estradiol-inducible CYCA2;3-green fluorescent protein (GFP) construct under control of the minimal CaMV35S promoter into the uvi4 mutant background (Zuo et al., 2000; Imai et al., 2006). Two days after CYCA2;3-GFP induction, the GFP signal within the Arabidopsis root tip was assessed by confocal microscopy. A CYCA2;3-GFP signal was visible in the columella cells of control plants, but was barely detectable in the uvi4 mutant background (Figure 4A). A decrease in CYCA2;3-GFP levels was confirmed by protein immunoblots (Figure 4B). The GFP abundance did not decrease in plants expressing a mutant allele of CYCA2;3 harboring a mutant D-box (Figures 4A and 4B), illustrating that the increased turnover of CYCA2;3 in the uvi4 background depended on the cyclin destruction motif.

Figure 4.

Control of CYCA2;3 Abundance by UVI4.

(A) Confocal images of induced CYCA2;3-GFP and D-box-mutated mDBCYCA2;3-GFP in wild-type or uvi4 mutant root meristems. The top row of images shows the CYCA2;3-GFP signal, and the bottom row shows the CYCA2;3-GFP signal in combination with propidium iodide (PI) staining.

(B) Protein gel blot analysis with an anti-GFP antibody. LC, loading control corresponding to a nonspecific cross-reacting protein.

(C) Relative CYCA2;3 expression levels of the lines presented in (B). Expression levels in the wild type (Col-0) were arbitrarily set to one. Data represent mean ± se (n = 3).

(D) CYCA2;3-HA protein levels in control (PSB-D), control transformed (ROLD:CYCA2;3 and CaMV35S:GUS), and UVI4-overexpressing (ROLD:CYCA2;3 and CaMV35S:UVI4) cell cultures. CDKA;1 proteins levels serve as loading control.

(E) and (F) Relative UVI4 (E) and CYCA2;3 (F) expression levels of the lines presented in (D). Expression levels in wild type (PSB-D) were arbitrarily set to one. Data represent mean ± se (n = 3).

Bars in (A) = 20 μm.

Although no UVI4-overexpressing plants could be obtained, the observations made for the knockout plants could be confirmed using Arabidopsis cell suspension culture cells as a model system. Cultures stably co-overexpressing both UVI4 and hemagglutinin (HA)-tagged CYCA2;3 were generated by Agrobacterium-mediated transformation. As a control, cultures co-overexpressing both CYCA2;3-HA and the β-glucuronidase (GUS)-encoding gene were produced. Protein extracts were assessed for CYCA2;3 abundance by protein gel blotting with an anti-HA antibody. The CYCA2;3-HA protein was stabilized in cultures co-overexpressing UVI4 (Figure 4D), despite the fact that the cultures accumulated only a 1.6-fold increase in UVI4 transcripts (Figure 4E) and exhibited a 58% lower level of CYCA2;3-HA transcripts than in the GUS CYCA2;3-HA co-overexpressing culture (Figure 4F). Thus, even a minor increase in the UVI4 transcript level is sufficient to accumulate CYCA2;3, probably because of an enhanced APC/C inhibition.

UVI4 and CDKB1;1 Synergistically Inhibit Endocycle Onset

Previously, we demonstrated that expression of a dominant negative allele of CDKB1;1 (CDKB1;1.N161) induced endoreduplication, illustrating that CDKB1;1 activity drives the mitotic cell cycle and suppresses endocycle onset (Boudolf et al., 2004). Ectopic expression of the CDKB1;1 binding partner CYCA2;3 partially complemented the CDKB1;1.N161 phenotype, suggesting that the outtitration of CYCA2;3 is at least one of the mechanisms by which CDKB1;1.N161 induces endoreduplication (Boudolf et al., 2009). Because of the negative effect of uvi4 on CYCA2;3 abundance, the UVI4 deficiency would be expected to aggravate the dominant negative effect of CDKB1;1.N161 on endoreduplication. Indeed, an increase in the DNA ploidy level of the mature first leaves was observed in the uvi4 CDKB1;1.N161 double mutant (Figure 5; see Supplemental Table 2 online), suggesting that UVI4 and CDKB1;1 work synergistically to regulate endocycle onset.

Figure 5.

Synergistic Regulation of Endocycle Onset by UVI4 and CDKB1;1.

Flow cytometric analysis of wild-type (Col-0), uvi4, CDKB1;1.N161, and uvi4 CDKB1;1.N161 3-week-old first leaves. Data represent mean ± sd (n = 3).

UVI4 Inhibits Premature Cell Differentiation

CCS52A1 and CYCA2;3 antagonistically regulate the timing of cell cycle exit and endoreduplication onset (Lammens et al., 2008; Boudolf et al., 2009). As UVI4 interacts with CCS52A1 and regulates the abundance of CYCA2;3, the effects of UVI4 on the timing of cell differentiation were examined. First, we focused on the root meristem, because the cell cycle exit can be easily monitored along the proximal-distal axis, where stem cell daughters divide repeatedly before eventually starting to elongate and differentiate (Dolan et al., 1993). Root meristem size was defined as the number of cortex cells in a file extending from the quiescent center to the first elongated cortex cell (Perilli and Sabatini, 2010). A significant decrease in the number of meristematic cortex cells was observed in uvi4 compared with wild-type 1-week-old seedlings (26 ± 4 versus 33 ± 3 cells, respectively; P < 0.05, Student’s t test) (Figures 6A and 6B). By contrast, the mature cortex cell length of the two genotypes did not differ significantly (Figure 6C), with a reduced meristem size in the uvi4 mutant as a consequence (Figure 6A). Correspondingly, measurements of root growth revealed a significant (P < 0.05, Student’s t test) reduction in the uvi4 mutant (Figure 6D). In the uvi4 CDKB1;1.N161 double mutant, the number of meristematic cortex cells was reduced significantly (P < 0.05, Student’s t test) in comparison with the single mutants (Figures 6A and 6B). Despite an increase in the cortex cell length (Figure 6C), root growth of the uvi4 CDKB1;1.N161 double mutant was reduced significantly (P < 0.05, Student’s t test) (Figure 6D).

Figure 6.

UVI4 Regulates Root Meristem Size.

(A) Representative confocal microscopy images of wild-type (Col-0), uvi4, CDKB1;1.N161, and uvi4 CDKB1;1.N161 1-week-old root meristems stained with PI. Arrowheads indicate the meristem size based on the cortical cell length.

(B) and (C) Number of meristematic cortex cells (B) and their average cell length (C) for lines presented in (A). Data represent mean ± sd (n > 12). The uvi4 mutant has a significant decrease in the number of meristematic cortex cells compared with the wild type (Col-0) (*P value < 0.05, Student’s t test). The CDKB1;1.N161 mutant has a significant decrease in the number of meristematic cortex cells and an increase in their average cell length compared with the uvi4 mutant (**P value < 0.05, Student’s t test). The uvi4 CDKB1;1.N161 double mutant has a significant decrease in the number of meristematic cortex cells and an increase in their average cell length compared with both uvi4 and CDKB1;1.N161 single mutants (***P value < 0.05, Student’s t test).

(D) Root length of lines presented in (A) plotted from 5 to 9 d after sowing. Data represent mean ± sd (n > 27). Both uvi4 and CDKB1;1.N161 mutants have a significantly decreased root length compared with the wild type (Col-0) (*P value < 0.05, Student’s t test). The uvi4 CDKB1;1.N161 double mutant has a significantly decreased root growth compared with both uvi4 and CDKB1;1.N161 single mutants (**P value < 0.05, Student’s t test).

(E) to (G) Leaf size (E), average abaxial epidermal cell size (F), and cell number (G) of 3-week-old first leaves of the lines presented in (A). Data represent mean ± sd (n ≥ 9). The uvi4 mutant has a significant decrease in leaf size and cell number and an increased average abaxial epidermal cell size compared with the wild type (Col-0) (*P value < 0.05, Student’s t test). The CDKB1;1.N161 mutant has a significant decrease in leaf size and cell number and an increased average abaxial epidermal cell size compared with the uvi4 mutant (**P value < 0.05, Student’s t test). The uvi4 CDKB1;1.N161 double mutant has a significant decrease in leaf size and cell number, an increased average abaxial epidermal cell size compared with the uvi4 single mutant, and a significant decrease in leaf size and cell number compared with the CDKB1;1.N161 single mutant (***P value < 0.05, Student’s t test).

Bars in (A) =50 μm.

[See online article for color version of this figure.]

Premature cell differentiation was observed for leaves of both uvi4 and CDKB1;1.N161 lines, as suggested by the reduced mature leaf cell number, resulting in a decrease in leaf size, despite a slight increase in cell size (Figures 6E to 6G). In the uvi4 CDKB1;1.N161 double mutant, the leaf phenotype was more pronounced, again illustrating that UVI4 and CDKB1;1 synergistically inhibit cell differentiation (Figures 6E to 6G).

OSD1 Has No Direct Role in Meristem Maintenance

Because of the high level of similarity of OSD1 with UVI4 and its association with the APC/C (Van Leene et al., 2010), the osd1-3 mutants were tested for their effect on meristem maintenance. Both diploid homozygous knockouts and their tetraploid offspring were examined, because a lack of OSD1 provokes polyploidization in the next generation (d’Erfurth et al., 2009). Whereas the diploid osd1-3 mutant did not differ significantly in root length, the tetraploid generation displayed a longer root (3.81 ± 0.40 cm, 3.99 ± 0.43 cm, and 4.28 ± 0.52 cm 9 d after sowing for the wild type, diploid osd1-3, and tetraploid osd1-3, respectively). At the cellular level, the diploid osd1-3 mutants were quite similar to control plants in the number of meristematic cortex cells (33 ± 3 cells), but the tetraploid plants showed a decrease in cell number (28 ± 2 cells), which was compensated by an increase in cell size, eventually resulting in a meristem similar in size to that of control plants (Figures 7A and 7B). However, similar effects on meristem cell number (30 ± 2 cells) and size were observed in tetraploid control plants, illustrating that the observed defects were caused by the tetraploid genome rather than the absence of OSD1.

Figure 7.

OSD1 Does Not Regulate Meristem Size.

(A) Representative confocal microscopy images of wild-type (Col-0), osd1-3 diploid, osd1-3 tetraploid, and Col-0 tetraploid 1-week-old root meristems stained with PI. Arrowheads indicate the total meristem length based on the cortical cell length.

(B) Number of meristematic root cortex cells in control (Col-0), osd1-3 diploid, osd1-3 tetraploid, and Col-0 tetraploid plants. Data represent mean ± sd (n ≥ 19). The osd1-3 diploid mutant shows no significant difference compared with wild-type (Col-0) but has a significant increase in cortex cell number compared with the osd1-3 tetraploid mutant and Col-0 tetraploid plants (*P value < 0.05, Student’s t test). The osd1-3 tetraploid mutant shows no significant difference with Col-0 tetraploid plants.

(C) to (E) Leaf size (C), average abaxial epidermal cell size (D), and cell number (E) of 3-week-old first leaves of lines presented in (A). Data represent mean ± sd (n ≥ 7).The osd1-3 diploid mutant shows no significant difference compared with the wild type (Col-0) but has a significantly decreased leaf size and an average abaxial epidermal cell size and cell number compared with the osd1-3 tetraploid mutant and Col-0 tetraploid plants (*P value < 0.05). The osd1-3 tetraploid mutant shows no significant difference with Col-0 tetraploid plants.

Bars in (A) = 50 μm.

[See online article for color version of this figure.]

A similar conclusion could be drawn for leaf development. When examining mature leaves, diploid osd1-3 mutants did not significantly differ in leaf size and abaxial cell number and cell size compared with the wild type (Figures 7C to 7E). By contrast, the tetraploid osd1-3 mutant showed a significantly increased leaf size (24.6 ± 1.4 mm² and 31.1 ± 1.8 mm² for the diploid wild type and tetraploid osd1-3, respectively; P < 0.001, Student’s t test), because of an increase in cell size (0.0012 ± 0.0001 mm² versus 0.0020 ± 0.0004 mm²) and despite a decrease in cell number (16.313 ± 1.976 versus 10.033 ± 2.153). However, because the phenotypes of the tetraploid osd1-3 mutant resembled that of a tetraploid wild-type plant, it can be concluded that the observed cell number phenotypes only indirectly relate to OSD1 and that UVI4 plays the primary role in maintaining meristem size.

UVI4 Expression Is Regulated by E2F Transcription Factors

Expression of UVI4-GFP under the control of the UVI4 promoter revealed a patchy nuclear accumulation pattern in the root tissues, indicating a cell cycle phase-specific transcription (Figure 8A). To specify the cell cycle phase in which UVI4 was expressed predominantly, root tips were synchronized by hydroxyurea (HU) treatment (Cools et al., 2010). In this synchronized system, root tips are strongly enriched for S- and M-phase cells at the 6 h and 16 h time points, respectively. Synchronized root tips were harvested at 2-h intervals over a total time of 22 h, and the UVI4 expression levels were determined by reverse-transcription quantitative polymerase chain reaction (RT-qPCR). The UVI4 transcript levels increased 4 h after the onset of synchrony, which coincides with the G1-to-S transition, peaked during the S phase at the 6 h time point, and decreased thereafter. A second peak in expression occurred at the 22 h time point, corresponding with the synchronous onset of a second S phase (Figure 8B). A similar accumulation pattern of the UVI4 protein was observed (see Supplemental Figure 3 online). These data suggest that UVI4 exerts its role as an APC/C inhibitor during the S phase.

Figure 8.

Expression of UVI4 Is Cell Cycle Phase Dependently Regulated by E2F Transcription Factors.

(A) Confocal microscopy image of a UVI4:UVI4-GFP root tip counterstained with PI. The patchy nuclear GFP signal suggests a cell cycle phase-specific expression.

(B) Relative UVI4 expression levels in HU-synchronized root tips. The transcript level of nonsynchronized root tips (0 h) was arbitrarily set to 1. Data represent mean ± se (n = 2).

(C) and (D) Relative UVI4 expression levels in control (Col-0), E2Fa, DPa, and E2Fa-DPa overexpression plants (C) and e2fa-2, e2fb-1, and e2fa-2 e2fb-1 mutants (D). The transcript level of Col-0 was arbitrarily set to 1. Data represent mean ± se (*P value < 0.05, Student’s t test, n = 3).

(E) and (F) Promoter element enrichments after ChIP from 1-week-old Arabidopsis seedlings with the E2Fa (E) and E2Fb (F) antibodies and with actin and ETG1 as a negative and positive control, respectively. For E2Fa, the no-antibody control values were extremely low compared with those of the E2Fa antibody and IP and were hence not visible on the graph. Data represent mean ± se (n = 2).

Bar in (A) = 100 μm.

Expression of most S phase-specific genes in Arabidopsis is controlled by two transcriptional activators, E2Fa and E2Fb, whose overexpression triggers the transcription of directly and indirectly regulated E2F target genes (Berckmans and De Veylder, 2009). We tested whether UVI4 expression is regulated by E2F transcription factors, despite the apparent lack of a consensus E2F binding cis-acting element in its promoter. UVI4 transcript levels were found to increase 1.2-fold in E2Fa-overexpressing plants, and up to 4.9-fold after co-overexpression with its dimerization partner DPa (Figure 8C). By contrast, when UVI4 transcript levels were assessed in e2fa-2 loss-of-function mutants, no significant decrease could be observed. In e2fb-1 loss-of-function plants, a decrease in UVI4 expression levels of only 0.24-fold was observed, in contrast with the 0.54-fold reduction observed in an e2fa-2 e2fb-1 double mutant, indicating that both E2Fa and E2Fb regulate UVI4 expression (Figure 8D). To confirm this hypothesis, we performed a chromatin immunoprecipitation (ChIP) assay to analyze whether these transcription factors bind the UVI4 promoter. ETG1, which has been shown previously to be a target of both E2Fa and E2Fb, was included as a positive control, and actin as a negative control (Takahashi et al., 2008). After precipitation using the anti-E2Fa and anti-E2Fb antibodies, an enrichment of both the ETG1 and UVI4 promoters could be observed compared with the actin-negative control (Figures 8E and 8F). Together, these data show that both E2Fa and E2Fb transcription factors can bind to the UVI4 promoter and regulate its expression.

UVI4 Allows CYCA2;3 Accumulation during the S Phase

The induction of UVI4 transcripts at the G1-to-S transition and the control of CYCA2;3 abundance by UVI4 prompted us to investigate whether UVI4 might play a role in the temporal accumulation pattern of CYCA2;3. To this end, root tips from CYCA2;3-GFP-inducible overexpression plants were synchronized on medium containing both HU and β-estradiol. After 4, 6, and 8 h of synchronization, corresponding to maximum UVI4 transcript levels, root tips were harvested from plants harboring CYCA2;3-GFP either in a wild-type or uvi4 mutant background. CYCA2;3-GFP accumulation stabilized after 4 h of HU treatment in wild-type root tips and decreased after 6 h and 8 h of synchronization, but remained higher than initial levels. By contrast, in the uvi4 mutant background, only a minor accumulation of CYCA2;3-GFP could be detected after 4 h (Figure 9). These data indicate that plants lacking UVI4 fail to accumulate CYCA2;3-GFP during DNA replication.

Figure 9.

UVI4 Regulates CYCA2;3 Accumulation during the S Phase.

CYCA2;3-GFP levels in wild-type (Col-0) and uvi4 HU-synchronized root tips measured by protein gel blotting with an anti-GFP antibody.

DISCUSSION

UVI4 Modulates Meristem Size by Regulating the Abundance of CYCA2;3

Here, we identified UVI4 as a negative regulator of APC/C^CCS52A1 activity. By using the Y2H approach, we confirmed the previously reported TAP interaction between UVI4 and CCS52A1, but not with other APC/C subunits, including APC2, HBT, APC7, APC8, APC10, CDC20-1, and CCS52B (Van Leene et al., 2010). Failure to detect these interactions might be caused by the cooperative binding of UVI4 to different APC/C subunits. Nevertheless, CCS52A1 seems to be one of the most important APC/C contact sites, because the uvi4 mutant trichome and leaf phenotypes were absent or strongly attenuated in ccs52a1-1 knockout lines, which indicates that CCS52A1 is essential for UVI4 action. In leaf tissues, the endoreduplication phenotype of uvi4 was not completely repressed by the ccs52a1-1 mutation. Because both CCS52A1 and CCS52A2 regulate endoreduplication in leaves (Lammens et al., 2008), UVI4 likely targets both CCS52A proteins for inhibition. Accordingly, although UVI4 interacted only with CCS52A1 in the Y2H screen, both CCS52A proteins copurified with UVI4 by TAP, suggesting that UVI4 also regulates APC/C^CCS52A2. However, protein immunoblots revealed that UVI4 preferentially interacted with CCS52A1 (see Supplemental Figure 4 online), at least in Arabidopsis cell cultures. Previously, we demonstrated that CCS52A2 is regulated at the transcriptional level by the E2F/DP-E2F-Like1 (DEL1) transcription factor (Lammens et al., 2008). These data hint at the possibility that APC/C^CCS52A2 activity is regulated at two different levels: transcriptionally via DEL1 and posttranscriptionally by UVI4. Because the expression of both DEL1 and UVI4 is dependent on the phase of the cell cycle (see below), the combination of both mechanisms might fine-tune the timing of APC/C activity.

Arabidopsis encodes three E2F transcription factors, two transcriptional activators (E2Fa and E2Fb) and one repressor (E2Fc) (Berckmans and De Veylder, 2009). Analyses of e2f mutants and ChIP revealed that both the E2Fa and E2Fb transcription factors regulate UVI4 expression, and this is probably the reason why transcript levels peak during the S phase. Although no consensus E2F binding motifs are present within 1000 nucleotides upstream of the UVI4 start codon with the sequence matrix WTTSSCSS (where W = A or T, and S = C of G) that matches the most common E2F binding sites (Vandepoele et al., 2005), the UVI4 promoter harbors two closely related sequences (see Supplemental Figure 5A online). A ChIP promoter scanning experiment revealed enrichment for E2Fa binding around these binding sites, suggesting that the putative E2F cis-regulatory elements represent functional ones (see Supplemental Figure 5B online).

Cell cycle phase-dependent expression of UVI4 suggests that its gene product might play a role in the temporal control of APC/C^CCS52A1 activity. Previously, we have identified the S phase-specific cyclin CYCA2;3 as one of the critical substrates of the APC/C^CCS52A1 regulating endocycle onset, as illustrated by the CYCA2;3-GFP protein stabilization in the root elongation zone in a ccs52a1-1 mutant background (Boudolf et al., 2009), corresponding with the spatial expression pattern of CCS52A1 (Vanstraelen et al., 2009). Here, we showed that CYCA2;3 destruction through the APC/C is regulated by UVI4. Knockout plants for UVI4 display reduced CYCA2;3 levels, whereas overexpression of UVI4 results in cyclin stabilization. When comparing synchronized root tips of wild-type and uvi4 mutant plants, the UVI4 activity was found to be required for the accumulation of CYCA2;3 during the S phase progression. We hypothesize that this accumulation is a prerequisite for cells to progress through the G2-to-M transition (Figure 10A). CYCA2;3-regulated CDK activity might be needed to activate the three-repeat MYB transcription factors that regulate the expression of G2/M-expressed genes (Ito et al., 2001). In the absence of UVI4, cells can finalize the DNA replication but not enter into a mitotic event, eventually entering an endocycle instead, with polyploidy as a consequence (Figure 10B). Nevertheless, despite the smaller root meristem size, mitotic divisions are not totally suppressed in UVI4-deficient roots, suggesting that redundant mechanisms regulate CYCA2;3 abundance. In the young meristematic cells, strong CYCA2;3 transcription or CCS52A inactivation through CDK phosphorylation might represent alternative mechanisms to inhibit APC/C activity, allowing CYCA2;3 to accumulate. As cells age, the importance of these alternative mechanisms might decrease, at which point UVI4 action might become the most important factor to regulate CYCA2;3 abundance.

Figure 10.

Model for UVI4 Action.

In wild-type plants, UVI4 levels increase at the G1-to-S transition because of the E2F activity. UVI4 inhibits APC/C^CCS52A1 activity and allows the increase of CYCA2;3 levels and the occurrence of the G2-to-M transition (A). In the absence of UVI4, the APC/C^CCS52A1 activity probably drops only marginally, possibly through the negative regulation by CDK phosphorylation that allows a partial, but insufficient, increase in the CYCA2;3 abundance to trigger the G2-to-M transition, with endoreduplication as a consequence (B). The dashed black horizontal line represents a hypothetical CYCA2;3 accumulation threshold needed to proceed into mitosis. Full and dashed colored lines represent experimentally demonstrated and speculated activities derived from homology with other eukaryotic systems, respectively.

By mediating mitotic entry, UVI4 indirectly regulates the meristem size and the plant growth rate. CYCA2;3 has previously been found to be an important factor in the suppression of cell cycle exit. Ectopic accumulation of the CYCA2;3-GFP fusion protein activates cell proliferation, whereas the loss of function mutation stimulates cell cycle exit and onset of endoreduplication (Imai et al., 2006). CYCA2;3 forms a functional complex with CDKB1;1 (Boudolf et al., 2009), of which the kinase activity is essential in the maintenance of cell division (Boudolf et al., 2004). Together with the meristematic expression pattern of UVI4 (Hase et al., 2006), this suggests that UVI4 might allow the accumulation of CYCA2;3-CDKB1;1 activity in dividing cells, in that way negating the possible premature action of CCS52A1, and thus preventing premature cell cycle exit. In agreement with this hypothesis, uvi4 mutant plants display a reduced growth and meristem size. This effect is even more pronounced in a CDKB1;1.N161 overexpression background, probably because of an even stronger decrease in CYCA2;3 levels.

UVI4, a Functional Homolog of Emi1

UVI4 shares structural and functional features with animal APC/C inhibitors, such as Emi1. From both our protein interaction studies and genetic complementation tests with uvi4 mutant alleles, we conclude that the deletion of the conserved MR tail results in the loss of the APC/C interaction and UVI4 function. Likewise, the C-terminal RL tail in the Emi1 and Emi2/Erp1 sequence has been shown to be essential for the APC/C inhibitory activity (Miller et al., 2006). However, there are also structural differences. Besides the RL tail, detailed studies of the Emi1- and Erp1/Emi2-conserved D-boxes have revealed that this degron is needed for APC/C interaction as well, because its mutation reduces APC/C binding and activity (Miller et al., 2006; Ohe et al., 2010). By contrast, the D-box motif is seemingly not essential for UVI4 activity in Arabidopsis; however, its mutation abrogates the interaction with CCS52A1 in the Y2H system. Inhibitory activity of the D-box mutant can be explained by the cooperative binding of UVI4 to the APC/C holocomplex in vivo. The D-box might contribute to the CCS52A1 interaction, as suggested by the Y2H experiment, but interactions using the GxEN motif and MR tail might be sufficient to achieve APC/C inhibition. Thus, our results suggest that, as observed for Emi1 and Erp1/Emi2, the D-box might interact directly with CCS52A1, whereas the MR tail, together with the GxEN motif, might mediate the interaction of UVI4 with the APC/C holocomplex.

At the functional level, just like UVI4, the mammalian Emi1 is needed for cells to enter mitosis. Xenopus extracts depleted of Emi1 do not enter mitosis, because they cannot accumulate cyclin B (Reimann et al. 2001). Likewise, depletion of Emi1 in mammalian cells induces DNA re-replication, because APC/C^CDH1 cannot be turned off, resulting in the continuous degradation of several APC/C substrates, including cyclin A (Di Fiore and Pines 2007; Machida and Dutta 2007). Moreover, as observed in Arabidopsis, phenotypes of Emi1 mutant cells can be suppressed by mutations in CDH1. Finally, both UVI4 and Emi1 seem to be required for the repression of APC/C activity during the S-to-G2 transition, coupling DNA replication with mitosis, whereby the E2F transcription factors seemingly regulate the timing of activity of both genes (Hsu et al., 2002).

UVI4 and Emi1 are not only structurally and functionally related, but also both possess a homolog that operates specifically in the meiotic cell cycle, namely OSD1 and Erp1/Emi2, respectively (Tung et al., 2005; Ohe et al., 2007; d’Erfurth et al., 2009), that are both essential for the transition from meiosis I to meiosis II (Ohe et al., 2007; d’Erfurth et al., 2010). In Xenopus, this transition occurs through a stabilization of CYCB by Erp1/Emi2 (Ohe et al., 2007), whereas the targets of OSD1 remain to be identified. Taken together, we postulate that UVI4 and OSD1 correspond to functional homologs of Emi1 and Erp1/Emi2 that encode APC/C inhibitors, which play an important role in sustaining the mitotic and meiotic cell cycle, respectively. The Emi1 and Erp1/Emi2 APC/C inhibitors have been shown to operate by acting as pseudosubstrates by D-box recognition and by blocking the catalytic cleft by means of their conserved C-terminal zinc-binding region (Miller et al., 2006). Although no zinc binding or other putative APC/C inhibitory domains occurred in the sequence alignment of UVI4, we cannot exclude the possibility that an unidentified plant-specific domain also exerts the inhibitory function. Further functional mapping and three-dimensional structure information on UVI4 can shed light on how UVI4 plays its role as APC/C inhibitor.

METHODS

Antibody Generation

An anti-UVI4 antibody was raised against the N-terminally located GQIQFRSPQGRENMSLG(C) (83 to 99) peptide in rabbits. The C-terminal Cys was added to allow KLH coupling of the peptide. Peptides were synthesized and rabbits immunized by Agro-Bio.

Gene Constructs and Site-Specific Mutagenesis

The full-length cDNA of UVI4 was cloned, and the D1, D2, and G degrons were mutated with a two-step adaptor ligation PCR approach. Adaptor primers containing the mutated sequence were used to amplify the open reading frame (ORF) 5′ and 3′ of the respective site. Ligation of the two fragments was achieved in a second PCR step. The UVI4 −MR mutant was designed by deletion of the six corresponding 3′ terminal nucleotides. The UVI4 alleles were inserted into the pK7m24GW plasmid under control of the UVI4 promoter, the region of 1046 nucleotides upstream of the start codon. The GFP-UVI4 fusion construct was generated by recombining UVI4 and GFP into the pH7m34GW plasmid, under control of the UVI4 promoter. The PCR primer sequences used for cloning are given in Supplemental Table 3 online. Constructs were generated by Gateway cloning (Invitrogen).

Plant Medium and Growth Conditions

Plants were grown under a long-day/short-night regime (16 h light/8 h darkness) at 21°C on agar-solidified culture medium [Murashige and Skoog (MS) medium, 10 g/L saccharose, 0.43 g/L 2-(N-morpholino)-ethane-sulfonic acid, and 0.8% plant tissue culture agar]. UVI4 complementation plants were screened using 45 mg/L kanamycin selection; plants harboring the GFP-UVI4 fusion construct were selected using 15 mg/L hygromycin. The root meristem was synchronized as described (Cools et al., 2010). For CYCA2;3-GFP induction, plants were grown on vertical MS medium for 5 d and then transferred with nylon meshes (Prosep) to MS medium containing 10 μM β-estradiol (Sigma-Aldrich) for an additional 2 d. For synchronization of CYCA2;3-GFP plants, 2 d after induction, seedlings were transferred to MS plates containing both 10 μM β-estradiol and 2 mM HU (Sigma-Aldrich).

Arabidopsis Cell Suspension Culture Transformation and Cultivation

The UVI4 and CYCA2;3-HA ORF were inserted into the pK7m34GW2-8m21GW3D vector, under control of the CaMV35S and ROLD constitutive promoters, respectively, through recombination using the Gateway LR II Plus enzyme mix. As a control, the GUS-encoding gene was recombined instead of the UVI4 ORF. Transformation and cultivation of Arabidopsis thaliana cell suspension cultures was performed as previously described by Van Leene et al. (2007).

Mutant Lines

The osd1-3 transfer DNA (T-DNA) insertion mutant was obtained from the Koncz collection in Cologne. The T-DNA was inserted 58 nucleotides after the start codon of the second exon, resulting in the removal of the last 109 amino acids of the 243-amino-acid-long OSD1 (see Supplemental Figure 6 online). Primers used for genotyping are given in Supplemental Table 3 online. The ccs52a1-1, ccs52a2-1 (Lammens et al., 2008), CDKB1;1.N161 (Boudolf et al., 2004), e2fa-2 and e2fb-1 (Berckmans et al., 2011b), inducible CYCA2;3-GFP and mDBCYCA2;3-GFP (Imai et al., 2006), and tetraploid Col-0 (Tsukaya, 2008) lines had been described previously. The uvi4 mutant was obtained from Yoshihiro Hase (RIKEN Institute, Yokohama, Japan). Double mutants were generated by crossing.

Y2H Interactions

Plasmids encoding the bait (pDEST32) and prey (pDEST22) were transformed into the yeast strain PJ69-4α (MATα; trp1-901, leu2-3,112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2::GAL1-HIS3, GAL2-ADE2, met2GAL7-lacZ) and PJ69-4a (MATa; trp1-901, leu2-3,112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2TGAL1-HIS3, GAL2-ADE2, met2TGAL7-lacZ) by the LiAc method (Gietz et al., 1992). Transformed yeast cells were selected on synthetic dextrose plates without Leu (pDEST 32) or without Trp (pDEST22). Interactions between proteins were assayed by the mating method. All pDEST32 yeast cultures were inoculated in 200 μL synthetic dextrose without Trp in a 96-well microtiter plate (Falcon; BD Biosciences), whereas one pDEST22 yeast culture was inoculated in 50 mL synthetic dextrose medium without Leu. To scale up the yeast cultures, 20 μL of each culture grown for 2 d at 30°C was added to a microtiter plate containing 125 μL of Glc-containing rich medium (10 g/L bacto-yeast extract, 10 g/L bacto-peptone, and 20% dextrose) and again grown for 24 h at 30°C. The Glc-containing rich medium was replaced by synthetic dextrose medium without Leu and Trp. Diploid strains grown in a 96-well microtiter plate (NUNC) for 2 d at 30°C were diluted until OD₆₀₀ = 0.2 and then added to a 96-well microtiter plate (Falcon; BD Biosciences) containing either 190 μL synthetic dextrose medium without Leu and Trp but with His (as control), or synthetic dextrose medium without Leu, Trp, or His. The OD was measured after 2 d of incubation at 30°C. Interactions were scored positive when the ratio of the OD₆₀₀ of the synthetic dextrose medium without Leu, Trp, or His or that of the synthetic dextrose medium without Leu and Trp exceeded the threefold average background value of the negative controls.

Protein Extraction and Purification

Whole plant tissue and cell suspension cultures were ground with 4-mm metal beads, and proteins were extracted with homogenization buffer (Boudolf et al., 2009). CYCA2;3-GFP was purified with agarose-conjugated anti-GFP from MBL (D153-8) by incubating the agarose beads with 100 μg of total protein extract for 1 h at 4°C. After three washes with cold lysis buffer (Boudolf et al., 2009), beads were boiled in 5× sample buffer (350 mM Tris-HCL, pH 6.8, 10% SDS, 600 mM dithiothreitol, 0.012% bromophenol blue, 30% glycerol), and the supernatant was separated by SDS-PAGE. After protein immunoblotting, CYCA2;3-GFP was visualized with the Living Colors Monoclonal Antibody (JL-8) (Clontech). For the inducible CYCA2;3-GFP and UVI4-GFP root synchronization experiments, 35 μg and 250 μg total protein extract were incubated with anti-GFP agarose for 1 h at 4°C, respectively. Total cell suspension protein extract was made using the same method, and 300 μg was separated by SDS-PAGE. CYCA2;3-HA was visualized after protein gel blotting with the Anti-HA High-Affinity Rat monoclonal antibody (clone 3F10) (Roche).

ChIP

ChIP was performed with the anti-E2Fa and anti-E2Fb antibodies according to the protocol of Berckmans et al. (2011a). As a negative control, immunoprecipitation was performed without any antibody. The promoter-scanning ChIP experiment was performed using a Pro35S:E2Fa-overexpressing Arabidopsis cell suspension culture. Maintenance and stable transformation of Arabidopsis PSB-D cell suspension cultures was done according to Van Leene et al. (2007). qPCR experiments were performed on equal concentrations of total, mock ChIP, and E2Fa ChIP DNA, as measured with the Quant-iT dsDNA High Sensitivity Kit (Invitrogen). Primers used for PCR amplification were designed to span the complete upstream region of UVI4, up to the next gene, using Primer-BLAST (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). PCR primer sequences are given in Supplemental Table 3 online. Normalization of enrichment relative to total input DNA was calculated as 2^{(Ct(input)−Ct(purified))}.

Confocal and Scanning Electron Microscopy

Root meristems were analyzed using a Zeiss Axiovert 100M confocal laser scanning microscope. Plant material was incubated for 2 min in a 10-μm PI solution to stain the cell walls and was observed under an epifluorescence light with the appropriate filter set (excitation filter, 535 to 550 nm; dichroic mirror, 565 nm; barrier filter, 590 nm). GFP fluorescence was detected with a 500- to 550-nm band-pass emission filter. Images of leaf trichomes were acquired with a TM-1000 Tabletop electron microscope (Hitachi).

Quantification of Trichome Nuclear DNA Content

For DAPI staining, 3-week-old mature leaves were fixed using acetic acid (75% acetic acid, 25% ethanol) for 2 h and washed for 1 h with 70% ethanol. Leaves were briefly submerged in EDTA (0.5 M), and trichomes were removed using forceps. DNA was stained using 20 μg/mL DAPI in McIlvaine’s buffer (60 mM citric acid, 80 mM Na₂HPO₄, pH 4.1). Trichomes were mounted in Vectashield mounting medium for fluorescence H1000 (Vector laboratories; no. CA94010) and observed via epifluorescence on an Axioscope Imager microscope from Zeiss. Nuclear size and epifluorescence signal were analyzed using the ImageJ 1.41 software. The integrated density was calculated by multiplication of nuclear size and fluorescence intensity, normalized against the integrated density of wild-type Col-0 trichome nuclei. Wild-type trichome nuclear size was arbitrarily set to 32C.

RT-qPCR Analysis

RNA was extracted from the respective tissues with the RNeasy Kit (Qiagen). After DNase treatment with the RQ1 RNase-Free DNase (Promega), cDNA was synthesized with the iScript cDNA Synthesis Kit (Bio-Rad). Relative expression levels were determined with the LightCycler 480 Real-Time SYBR green PCR System (Roche). Two different reference genes were used for normalization, depending on the source of the tissue. ACT (At3G18780) and UBQ10 (At4G05320) were used for cell suspension cultures, ACT and CKII (At3G50000) for plant material, and PAC1 (At3G22110) and EMB2386 (At1G02780) for root tip synchronization. Primer sequences used for RT-qPCR analyses are given in Supplemental Table 3 online.

Flow Cytometry

Leaves were chopped with a razor blade in 200 μL CyStain UV Precise nuclei extraction buffer (Partec), and DNA was stained by adding 800 μL staining buffer (Partec). Nuclei were measured with a CyFlow Flow Cytometer (Partec) and analyzed with the FloMax software (Partec). Unless stated otherwise, three or more leaves were analyzed for each technical repeat.

Accession Numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: UVI4 (At2G42260); OSD1 (At3G57860); CCS52A1 (At4G29910); CYCA2;3 (At1g15570); CDKB1;1 (At3G54180); E2Fa (At2G36010); E2Fb (At5G22220); DPa (At5G02470).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure 1. Quantification of the Wild-Type (Col-0) and uvi4, ccs52a1-1, and uvi4 ccs52a1-1 Mutant Trichome Branch Number.

• Supplemental Figure 2. Quantification of Nuclear DNA Content of the Wild-Type (Col-0) and uvi4, ccs52a1-1, and uvi4 ccs52a1-1 Mutant Trichomes.

• Supplemental Figure 3. Cell Cycle Phase-Dependent Accumulation of the UVI4-GFP Protein.

• Supplemental Figure 4. Copurification of UVI4 with CCS52A1 and CCS52A2.

• Supplemental Figure 5. Binding of E2Fa to the UVI4 Promoter.

• Supplemental Figure 6. Graphical Representation of the T-DNA Insertion in the osd1-3 Line.

• Supplemental Table 1. Ploidy Distribution of 3-Week-Old First Leaves of the Wild Type (Col-0) and the uvi4, ccs52a1-1, and uvi4 ccs52a1-1 Mutants.

• Supplemental Table 2. Ploidy Distribution of 3-Week-Old First Leaves of the Wild Type (Col-0) and the uvi4, CDKB1;1.N161, and uvi4 CDKB1;1.N161 Mutants.

• Supplemental Table 3. Primer Sequences Used for Cloning, Genotyping, RT-qPCR, and ChIP Analysis.

• Supplemental Data Set 1. Pairwise Y2H Interactions Tested between Arabidopsis APC/C Proteins.