Specialization of CDC27 function in the Arabidopsis thaliana anaphase-promoting complex (APC/C)

Authors

  • José M. Pérez-Pérez,

    1. Department of Molecular Genetics, Utrecht University, Padualaan 8, 3584 CH, Utrecht, The Netherlands
    Search for more papers by this author
    • Current address: División de Genética and Instituto de Bioingeniería, Universidad Miguel Hernández, Edificio Vinalopó, Avda. de la Universidad s/n, 03202 Elche (Alicante), Spain.

  • Olivier Serralbo,

    1. Department of Molecular Genetics, Utrecht University, Padualaan 8, 3584 CH, Utrecht, The Netherlands
    Search for more papers by this author
    • Current address: Laboratoire de Génétique et de Physiologie du Développement (LGPD), Developmental Biology Institute of Marseille (IBDM), CNRS UMR6545, University Aix-Marseille II, 13288 Marseille cedex 09, France.

  • Marleen Vanstraelen,

    1. Institut des Sciences du Végétal, Centre National de la Recherche Scientifique, Unité Propre de Recherche 2355, Avenue de la Terrasse, 91198 Gif-sur-Yvette, France
    Search for more papers by this author
  • Cristina González,

    1. Department of Molecular Genetics, Utrecht University, Padualaan 8, 3584 CH, Utrecht, The Netherlands
    Search for more papers by this author
  • Marie-Claire Criqui,

    1. Institut de Biologie Moleculaire des Plantes du CNRS, 12, rue du General Zimmer, 67084 Strasbourg cedex, France
    Search for more papers by this author
  • Pascal Genschik,

    1. Institut de Biologie Moleculaire des Plantes du CNRS, 12, rue du General Zimmer, 67084 Strasbourg cedex, France
    Search for more papers by this author
  • Eva Kondorosi,

    1. Institut des Sciences du Végétal, Centre National de la Recherche Scientifique, Unité Propre de Recherche 2355, Avenue de la Terrasse, 91198 Gif-sur-Yvette, France
    Search for more papers by this author
  • Ben Scheres

    Corresponding author
    1. Department of Molecular Genetics, Utrecht University, Padualaan 8, 3584 CH, Utrecht, The Netherlands
    Search for more papers by this author

(fax +31 0 30 253 2837; e-mail b.scheres@uu.nl).

Summary

To investigate the specialization of the two Arabidopsis CDC27 subunits in the anaphase-promoting complex (APC/C), we analyzed novel alleles of HBT/CDC27B and CDC27A, and characterized the expression of complementing HOBBIT (HBT) protein fusions in plant meristems and during the cell cycle. In contrast to other APC/C mutants, which are gametophytic lethal, phenotypes of weak and null hbt alleles indicate a primary role in the control of post-embryonic cell division and cell elongation, whereas cdc27a nulls are phenotypically indistinguishable from the wild type. However, cdc27a hbt double-mutant gametes are non-viable, indicating a redundant requirement for both CDC27 subunits during gametogenesis. Yeast-two-hybrid and pulldown studies with APC/C components suggest that the two Arabidopsis CDC27 subunits participate in several complexes that are differentially required during plant development. Loss-of-function analysis, as well as cyclin B reporter protein accumulation, indicates a conserved role for the plant APC/C in controlling mitotic progression and cell differentiation during the entire life cycle.

Introduction

New cells for plant growth are continuously provided by meristems, which contain mitotically active cells derived from small stem cell populations (Weigel and Jürgens, 2002). A pivotal question is how cell division and patterning are coordinated during organ growth (Jakoby and Schnittger, 2004). In the Arabidopsis root meristem (RM), the quiescent center (QC) maintains surrounding stem cells (van den Berg et al., 1997). QC and stem cell specification requires a combination of AP2-domain and GRAS transcription factors (Aida et al., 2004; Di Laurenzio et al., 1996; Heidstra et al., 2004; Helariutta et al., 2000; Sabatini et al., 2003). Crosstalk between stem cell specification factors with cell cycle and cell differentiation has been recently reported: reduction of RETINOBLASTOMA-RELATED1 (RBR1) activity in the RM increases the size of the stem cell pool downstream of the SCARECROW patterning gene (Wildwater et al., 2005). Although RBR1 and other members of the G1/S checkpoint are beginning to be connected with meristem maintenance factors (Dewitte et al., 2003; Wyrzykowska et al., 2006), crosstalk is less obvious for other cell cycle phases.

The anaphase-promoting complex (APC/C) is a multisubunit ubiquitin ligase that has a crucial role in cell cycle progression and mitotic exit by inducing the proteolysis of several cell cycle regulators, including mitotic cyclins. Activation and substrate specificity of the APC/C during the cell cycle are determined by two adaptor WD40-containing activators: Cdc20/Fizzy and Cdh1/Fizzy-related (Castro et al., 2005; Clute and Pines, 1999; Harper et al., 2002; Peters, 2002). All vertebrate APC/C subunits, Fizzy and Fizzy-related proteins have counterparts in plants (Capron et al., 2003a; Cebolla et al., 1999; Fulop et al., 2005; Tarayre et al., 2004). Arabidopsis apc2 and nomega/cdc16 null mutants are impaired in female gametogenesis and accumulate A- and B-type cyclin reporter proteins, supporting an essential role of the APC/C complex for cell division control during early plant development (Capron et al., 2003b; Kwee and Sundaresan, 2003).

The HOBBIT (HBT) gene is required for post-embryonic cell division and for the differentiation of distal tissues in the root, and encodes an Arabidopsis homologue of the CDC27 subunit of the APC/C (Blilou et al., 2002; Willemsen et al., 1998). HBT is unique among Arabidopsis core APC/C encoding genes: (i) it is the only gene in which transcription is cell cycle regulated (Blilou et al., 2002; Capron et al., 2003b; Fulop et al., 2005); (ii) its function is required predominantly post-embryonically for cell division and cell differentiation (Blilou et al., 2002; Willemsen et al., 1998); and (iii) a homologous gene, CDC27A, resides in the Arabidopsis genome (Capron et al., 2003a). These observations lead us to pose the question of whether CDC27 proteins have specialized roles during plant development.

To address specialization of the two CDC27 homologues we studied new hobbit (hbt) and cdc27a alleles, and the expression of complementing HBT protein fusions. Furthermore, we demonstrate that HBT can bind to different APC/C activators and show in vivo interactions of HBT and CDC27A with APC2, a core component of the plant APC/C complex. Our results reveal both redundancy and functional divergence between CDC27A and HBT in Arabidopsis APC/C function during different stages of development, and indicate that different APC/C complexes have specialized functions.

Results

Strong HOBBIT (CDC27B) and CDC27A alleles reveal different functions during development

We screened available mutant collections to identify novel loss-of-function alleles of the two Arabidopsis CDC27 genes: CDC27A and HBT/CDC27B.

The hbt-11 allele obtained from the SIGNAL collection (Alonso et al., 2003) carries a single T-DNA insertion in the first intron of the HBT gene, 399-bp downstream of the HBT (At2g20000) initiation codon (Figure 1a; Table 1). hbt-11 heterozygous plants appear as wild type (WT), and a quarter of their progeny showed the strong phenotype described previously (Willemsen et al., 1998; Figure 1b; Table 1). RMs in mature hbt-11 embryos have aberrant cell numbers, cell morphology and irregular cell arrangements (Figure 1c). Genotyping for the T-DNA insertion in the progeny of hbts021513 heterozygotes with a WT phenotype suggests a slight decrease in the transmission of the mutant allele through the gametes (Table 1a), which was previously observed in other strong hbt alleles (Willemsen et al., 1998; V. Willemsen and B.S., unpubl. data). RT-PCR using total RNA reveals a strong reduction of HBT expression in hbt-11 homozygotes, but the upregulation in heterozygotes might suggest compensatory regulation (Figure 1d). The CDC27A transcript is upregulated in hbt-11 homozygotes, and whether this is also a compensation effect is not clear. Previously described hbt2311 mutants, encoding an HBT protein lacking seven of the 10 TPR domains (Blilou et al., 2002), have been renamed as hbt-1. hbt-1 and hbt-11 homozygotes are phenotypically indistinguishable, and HBT expression is highly downregulated in hbt-11/hbt-11. Thus, both alleles are likely to be null, indicating that HBT is not essential for gametophytic development, in contrast to core APC/C subunits (Capron et al., 2003b; Kwee and Sundaresan, 2003).

Figure 1.

 Isolation of novel HBT and CDC27 alleles. (a) Diagram of HBT and CDC27A genomic regions and molecular defects of the alleles used (see Experimental procedures). Full boxes represent exons, open reading frames are depicted in red with TRP domains in blue. (b) Progeny of HBT/hbt-11 plants. (c) Aniline blue staining of wild-type (WT) and hbt-11 roots in mature embryos. (d) RT-PCR using HBT and CDC27A specific oligos. A and B represent CDC27A and HBT/CDC27B WT alleles, whereas a and b represent their mutant alleles. Amplification of the EF1α gene was used as a reference. (e, f) Genetic interaction between cdc27a and hbt-1 during gamete development. (e) Opened siliques with the arrowheads pointing to the aborted ovules. (f) Microscopic magnification of the aborted ovule from the inset with immature nuclei indicated with asterisks. (g) Genetic interaction hypothesis between cdc27a and hbt-1. Same notation as in (d). Presumptive gametic lethality is indicated by the blue background. hbt homozygotes are depicted in red. Scale bars: 2 mm (b, e) and 25 μm (c, f).

Table 1.   Analysis of CDC27 alleles and their interactions. (a) Genetic analysis and transmission of the novel cdc27a and hbt alleles. (b) Interaction between cdc27a and hbt-1 alleles during gametogenesis. Between 10 and 15 siliques from the main inflorescence stem of 5-week-old plants were manually dissected and scored for aborted ovules. Data represent averages ± SD. (c) Genetic interaction between cdc27a and hbt-1 alleles after embryogenesis.
Female × MaleGenotypeTEaP value (hypothesis)b
AAAaaa
(a)
 HBT/hbt-11 selfed133240106c47.20.2176 (1:2:1)
 CDC27A/cdc27a-1 selfed48103 3245.60.0581 (1:2:1)
 CDC27A/cdc27a-1 ×  WT3121040.40.1659 (1:1)
 WT × CDC27A/cdc27a-12114040.00.2367 (1:1)
Genotype of the F1 parentsSeeds per siliqueAborted ovules per siliqueP-valuea
(b)
 CDC27A/CDC27A; HBT/HBT46.3 ± 5.01.8 ± 0.80.0007
 CDC27A/CDC27A; HBT/hbt-145.2 ± 6.22.1 ± 1.90.0011
 CDC27A/cdc27a141413;HBT/HBT42.0 ± 6.31.3 ± 1.30.0008
 CDC27A/cdc27a059166;HBT/HBT44.7 ± 6.71.2 ± 1.90.0005
 CDC27A/cdc27a141413;HBT/hbt-130.8 ± 7.512.8 ± 3.40.5071
 CDC27A/cdc27a059166;HBT/hbt-133.2 ± 2.910.9 ± 3.11.0000
Genotype of the F1 parentsWTHBTP valuea
5:18:1
  1. aTE, transmission efficiency of the mutant allele = number of mutant alleles/number of total alleles, as determined by PCR.

  2. bTwo-tailed P-values represent the fit of the data to the expected 1:2:1 (AA:Aa:aa) or 1:1 (AA:Aa) segregations.

  3. cThe Hbt phenotype was genotyped for the hbt allele by pooling.

  4. aTwo-tailed P values represent the fit of the data to an expected segregation of 3:1 wild type (WT) : aborted.

  5. aTwo-tailed P values represent the fit of the data to expected segregations of 5:1 wild type (WT):HBT (if only the double mutant female gametes are lethal) or 8:1 WT:HBT (when both double mutant gametes are lethal). The statistically significant P values are given in italics.

(c)
 CDC27A/cdc27a141413;HBT/hbt-1162180.01640.6315
 CDC27A/cdc27a059166;HBT/hbt-1140160.03160.7290

Three T-DNA insertional mutants in CDC27A (At3g16320; see Experimental procedures), cdc27a-1, cdc27a-2 and cdc27a-3, disrupt the predicted coding region at 1028, 3024 and 3721 bp in the genomic DNA and downstream of the initiation codon, respectively (Figure 1a). Homozygotes for all of these alleles were indistinguishable from the WT in our experimental conditions (Table 1a and data not shown). CDC27A expression is strongly reduced in the cdc27a homozygotes, suggesting that they are nulls (Figure 1d and data not shown). To rule out subtle gametophytic defects of the cdc27a-1 heterozygotes, the progenies of their crosses with WT were genotyped for the T-DNA insertion (Table 1a). Similar results were obtained from crosses using cdc27a-2 and cdc27a-3 heterozygotes (data not shown). Our results indicate that gametic transmission of the cdc27a alleles is 20% reduced (Table 1a).

As homozygotes for null mutations in single-copy subunits of the Arabidopsis APC/C display lethality, resulting from an early arrest in cell division during gametogenesis (Capron et al., 2003b; Kwee and Sundaresan, 2003), we wondered whether HBT and CDC27A play redundant roles during gametogenesis. Although the number of aborted ovules in either CDC27A/cdc27a or HBT/hbt-1 siliques was not significantly different from those found in the WT (Figure 1e; Table 1b), 25% of the ovules from CDC27A/cdc27a;HBT/hbt-1 plants are arrested in their growth (Figure 1e,f; Table 1b), which is consistent with full lethality of the cdc27a hbt-1 female gametes. Arrested female gametophytes always (n = 14) display fewer than eight nuclei (Figure 1f, asterisks), indicating a reduction of the three mitotic division rounds that occur in the WT. Both the segregation for hbt-1 seedlings (Table 1c), and the genotyping for the mutant alleles (see Experimental procedures) in the viable progeny of several CDC27A/cdc27a;HBT/hbt-1 double heterozygotes, confirms the synergistic and lethal phenotype of the cdc27a hbt-1 double mutants during female gametophyte development (Figure 1g). Two additional evidences suggest that male gametes are also affected by the reduction of CDC27 function. First, the proportion of cdc27a/cdc27a;HBT/hbt F2 plants was much lower than that expected considering cdc27a;hbt gametes are fully viable (1/43 vs. 1/10, P = 0.0937), and second, frequencies for CDC27A alleles were biased towards the WT one (0.75 and 0.25 for CDC27A and cdc27a alleles, respectively).

Taken together, our results reveal redundant roles for HOBBIT and CDC27A genes during gametophyte development, and a specialized role for HBT in post-embryonic growth and cell division.

Hypomorphic hobbit alleles specifically affect cell division and cell expansion

Previously described hbt mutant alleles affect maintenance of cell specification, responses to the phytohormone auxin, cell division and cell size (Blilou et al., 2002; Willemsen et al., 1998). Analysis of hbt loss-of-function clones indicated that changes in auxin response were not primary defects (Serralbo et al., 2006). To further distinguish between primary and secondary defects, we analyzed hypomorphs to identify the processes most sensitive to HBT reduction. We identifed weak recessive alleles from a collection of Arabidopsis TILLING lines (Till et al., 2003; Figure 1a). hbt-12 and hbt-13 carry missense mutations in the central domain of the HBT protein, between the first and the second TPR domains (Figure 1a; see Experimental procedures). hbt-12 and hbt-13 homozygotes are phenotypically similar dwarfs with small leaves and stunted growth (Figure 2a), and hbt-12 was further characterized. hbt-12 roots are significantly shorter compared with the WT but continue to grow (data not shown). The hbt-12 RM is smaller than in WT (Figure 2b,c) and contains fewer cells (Figure 2d, top graph) of normal size (Table 2). Columella stem cells are present (Figure 2c, asterisk), but the one or two remaining tiers of differentiated columella cells in hbt-12 are significantly smaller than in WT (Figure 2d, middle graph; Table 2). hbt-12 homozygotes have a smaller elongation zone (EZ) compared with WT, as indicated by the location of maximally expanded cells, and the onset of xylem and epidermal cell differentiation (arrowhead in Figure 2c). Length of differentiated epidermal cells along the main plant axis decreases 2.1-fold in hbt-12 mutants compared with WT (Figure 2d, bottom graph; Table 2), whereas mature root hair length is not significantly altered (Figure 2c and data not shown).

Figure 2.

 Analysis of hypomorphic hbt alleles.
(a) Wild-type (WT) (top) and hbt-12 (bottom) mature plants. Details of (b) WT and (c) hbt-12 roots: an arrowhead points to the first epidermal differentiated cell, and an asterisk marks the quiescent center. (d) Meristem size and lengths (in μm) of the differentiated epidermal and columella cells. NoC: number of cortex cells in the meristem. All measurements were taken 7 days after sowing (7 das). (e) Ploidy analysis of sorted nuclei from root tips (WT, hbt-12). Pictures were taken 7 das for seedlings and 28 das for mature plants and siliques. See Experimental procedures for ploidy and imaging analyses. Scale bars: 10 mm (a) and 25 μm (b).

Table 2.   Root morphometry of weak hbt alleles
 Wild typehbt-12
  1. Measurements were taken from stored Nomarsky images of 7-day-old roots. The number of samples analyzed in each case is indicated between brackets.

Number of meristematic cells25.9 ± 3.7 (12)13.5 ± 1.7 (15)
Cell length in the root meristem (μm)7.5 ± 1.2 (170)7.8 ± 1.5 (108)
Number of columella layers5.1 ± 0.2 (11)3.4 ± 0.5 (12)
Cell length in distal columella (μm)24.6 ± 3.9 (18)18.7 ± 4.5 (20)
Number of epidermal cells in the elongation zone8.2 ± 2.5 (14)4.0 ± 0.8 (10)
Length of differentiated epidermal cells (μm)130.4 ± 22.5 (43)55.2 ±  (65)

Ploidy analyses on sorted nuclei from hbt-12 roots (Figure 2e) show a 3.6 ± 2.1% reduction of 2C and a 10.0 ± 2.1% increase of 4C cells, suggesting either over-representation of G2 cells during division or defective endoreduplication. Interestingly, the proportion of nuclei with high ploidy levels (8C and 16C) is reduced by 13.6 ± 4.1%, which also indicates defects during endoreduplication.

Taken together, our data indicate that the HBT reduction primarily affects cell division and cell expansion in all plant organs.

HOBBIT protein fusions localize to dividing and elongating cells

HBT gene transcription is cell cycle regulated, unlike other genes encoding APC/C subunits (Blilou et al., 2002; Fulop et al., 2005), which prompted us to study its protein accumulation. An HBTg-GUS protein fusion was used to survey HBT expression in different tissues. Low HBT expression was found in unfertilized ovules, although their low levels preclude accurate localization (Figure 3a). In young embryos, up to the globular stage, HBTg-GUS is restricted to proliferative tissues of the chalazal bulb (Figure 3b, inset). From the heart stage (Figure 3b) to the early torpedo stage, HBTg-GUS is expressed ubiquitously, but at later stages of embryogenesis it becomes restricted to the shoot and root primordia, and variable cell patches in cotyledons and hypocotyl (Figure 3c), consistent with cell cycle control of HBT transcription. After germination, HBTg-GUS localizes to dividing cells within the RM (Figure 3d), lateral root primordia (Figure 3e,e′), leaf primordia at the shoot apical meristem (Figure 3f) and flower primordia (Figure 3g and inset). In roots, HBTg-GUS is excluded from the differentiation zone (DZ) and differentiated columella cells, and it accumulates in the QC and columella stem cells at lower levels than that of the meristem cells (Figure 3d and inset). In lateral root primordia, HBTg-GUS is expressed from stage III onwards (Figure 3e,e′) when periclinal divisions occur (Casimiro et al., 2003). HBTg-GUS is absent in adult leaves (data not shown) and mature flowers (Figure 3g), consistent with its exclusion from differentiated tissues. Our data reveal a strong correlation between mitotically active cells and HBTg-GUS expression.

Figure 3.

 HOBBIT (HBT) protein expression.
(a) HBTg-GUS protein fusion in the ovules. (b, c) HBT is ubiquitously expressed during embryogenesis: (b) heart stage and (c) late torpedo stage (arrowheads point to meristem anlagens). Post-embryonically, HBTg-GUS is expressed in the root meristem (d, inset shows stem cell area), lateral root primordia (e, stage III; e′, stage VII), shoot apical meristem (f, asterisk), young leaves (f) and flower primordia (g, bottom-right inset shows flower buds). HBTg-GFP (h–r) is expressed in meristematic cells (h) and is mostly absent from columella (col, i) and lateral root (lrc, i′) cap cells within the root tip. HBT-GFP is weakly expressed in epidermal differentiated cells (j) and becomes stabilized after MG132 proteosome inhibitor treatment (k). (l) Detailed subcellular structure of an epidermal cell in which the nucleus (nu) and nucleolus (nl) are clearly distinguishable. HBTg-GFP subcellular localization is dynamic during the cell cycle: (m) interphase, (n) prophase, (o) metaphase, (p) anaphase, (q) early telophase and (r) cytokinesis. (s) αCDC27A (green), contrasted in the upper panel with α-tubulin (red), reveals the absence of CDC27A in the anaphase (third cell from left). Scale bars: 10 μm (l–s), 25 μm (b–e′ and h–k) and 100 μm (f, g).

The HBTg-GFP protein completely rescues hbt-1 mutants, allowing detailed localization studies. HBTg-GFP is, in contrast to HBT transcript, homogeneously distributed in the RM (Figure 3h). Similar to HBTg-GUS, HBTg-GFP levels are significantly reduced in the differentiated columella (Figure 3i, col), lateral root cap (Figure 3i′, lrc) and epidermal cells (Figure 3j). High HBT protein levels in mitotically active cells gradually decrease when cells cease division and pass through the EZ (Figure 3h). Treatment of HBTg-GFP plants with the proteasome inhibitor MG132, which blocks the 26S proteasome-mediated degradation of ubiquitin-targeted proteins (Callis and Vierstra, 2000), elevates HBTg-GFP levels in epidermal cells at EZ and DZ in the presence of the protein biosynthesis inhibitor cycloheximide (Figure 3k). As RT-PCR did not reveal differences in HBT mRNA levels by this treatment (data not shown), the HBTg-GFP increase appears to be the result of increased protein stability. Our results indicate that athough HBT transcript is only produced in dividing cells (Blilou et al., 2002), proteolytic degradation regulates HBT and its persistence in expanding cells.

During the interphase (Figure 3m) HBTg-GFP localizes mainly to the nucleus (nu), and is excluded from the nucleolus (nl; see Figure 3l for nuclear structure). At the prophase (Figure 3n) HBTg-GFP associates with the prophase spindle, and weakly stains the pre-prophase band region (arrowheads in Figure 3n). After nuclear membrane disintegration at the metaphase HBTg-GFP localization is diffuse within the cell (Figure 30), and during the anaphase the protein is particularly concentrated at the mitotic spindle (Figure 3p). In the early telophase, HBTg-GFP becomes restricted to the newly formed nuclei (asterisks in Figure 3q) and to the cell plate (Figure 3q), and during the late telophase its signal is again restricted to the nucleus (asterisks in Figure 3r). The prominent co-localization of HBTg-GFP with the mitotic spindle contrasts with the lack of detectable spindle signal with CDC27A antibodies in anaphase (Figure 3s; Capron et al., 2003b).

Biochemical characterization of CDC27 subunits of the APC/C complex

It has been established that APC/C activators form a complex with CDC27A in Arabidopsis (Fulop et al., 2005). We assayed whether HBT/CDC27B interacts with the Arabidopsis APC/C activators by performing yeast two-hybrid assays with Ccs52A1, Ccs52A2 and Ccs52B, and the five isoforms of Cdc20 (Figure 4a).

Figure 4.

 HOBBIT (HBT) protein interaction studies.
(a) Interaction of the anaphase-promoting complex (APC/C) activators with HBT in a yeast two-hybrid assay. Autoactivation and repression of autoactivation of yeast strains containing the APC/C activators and the empty vector are presented on the left-hand control panel. Growth of yeast strains was followed in a 10-fold dilution series on selective medium in the absence of 3-AT as well as at 10 mm 3-amino-1,2,4-triazole (3-AT) in the case of Ccs52A1 and Ccs52B. Activator–HBT interactions are shown on the right-hand panel, where yeast strains were grown under identical conditions as the controls. White numbers indicate the highest 3-AT concentrations in millimolar that still allow yeast growth. (b) Yeast two-hybrid assay for CDC27A and HBT interaction with APC10, and for homo- and heterodimerization of CDC27A and HBT. Experimental conditions are the same as in Figure 4(a) except that interactions are shown at 5 mm 3-AT. (c) Detection of hemaglutinin (HA)-tagged HBT (HA-HBT) in protein extracts using monoclonal mouse anti-HA antibodies. APC2 detection in HA immunoprecipitates from pHBT::HA-HBT-expressing plants. Immunoprecipitation of the APC/C complex using anti-APC2 antibodies and detection of its association with functional HA-HBT and CDC27A. I, input; –, specific antibody not added.

The strongest binding of HBT occurs with Cdc20.1 and Cdc20.2, followed by Ccs52A1, Ccs52B and Cdc20.5. Interaction of HBT with Cdc20.4 and Ccs52A2 is clear but detectable only in the absence of the quantitative inhibitor (see Experimental procedures). The interaction of HBT with Cdc20.3 is at the background level. These results demonstrate that HBT, similarly to CDC27A, is able to interact in yeast with the APC/C activators with a preference for Cdc20.1, Cdc20.2 and Ccs52A1.

Pairwise interactions with the APC/C activators do not reveal whether HBT is included in the core APC/C complex. In yeast two-hybrid assays both CDC27A and HBT interact with APC10, suggesting their incorporation into the APC/C complex (Figure 4b). CDC27 subunits can dimerize in other eukaryotes (Passmore et al., 2005), which raised the possibility for homo- and/or heterodimerization of CDC27A and HBT. However, although HBT interacts with itself, neither CDC27A homodimerization nor heterodimerization of CDC27A with HBT was detected in yeast two-hybrid assays (Figure 4b), indicating that the core Arabidopsis APC/C could exist in two forms, one containing CDC27A and the other containing HBT.

To substantiate the presence of distinct isoforms in vivo, we used immunoaffinity purified antibodies against the Arabidopsis APC2 subunit (Capron et al., 2003b) in order to pull down APC/C complexes from root protein extracts (see Experimental procedures). APC2 and APC11 are required to form the minimal ubiquitin-ligase module of the human APC/C (Tang et al., 2001). Accordingly, tagged versions of APC2 and APC11 co-immunoprecipitate after their transient expression in Arabidopsis protoplasts (Capron et al., 2003b).

We observed co-immunoprecipitation of functional hemaglutinin (HA)-tagged HBT, as well as of CDC27A (Figure 4c) with APC2, with the latter being detected with a specific polyclonal antibody (Capron et al., 2003b). Conversely, we detected APC2 protein in HA immunoprecipitates from protein extracts of pHBT::HA-HBT-expressing plants (Figure 4c), confirming the association between APC2 and HA-HBT. However, we were unable to detect CDC27A protein in HA-HBT immunoprecipitates (data not shown).

Our results suggest the existence of separate HBT and CDC27A APC/C isoforms in vivo, but at this point we cannot exclude the possibility that HBT and CDC27A reside in the same APC/C multimer in vivo if some epitopes are masked in such a complex.

hbt mutants accumulate cyclin B-GUS reporter proteins

Because targeted degradation of cyclin B and securin are major functions of the APC/C complex during the cell cycle (Thornton and Toczyski, 2003), and because plants lack the securin homologue (Capron et al., 2003a), we analyzed stabilization of cyclin B reporters in hbt mutants to address whether an HBT-containing APC/C complex served canonical roles. A constitutively expressed form of CycB1;1 from tobacco fused to GUS (35S::NtCycB1;1-GUS) highly accumulates in the hbt-1 (Figure 5b,c), but not in WT seedlings (Figure 5a), mirroring the expression dynamics of the D-box mutated NtCycB;1-GUS that is resistant to APC/C-mediated degradation (Weingartner et al., 2004; data not shown). In addition, pCycB1;1::D-boxCycB1;1-GUS, which marks cells in the G2/M phase (Colon-Carmona et al., 1999; Figure 5d), is ectopically expressed in hbt-1 hypocotyls and cotyledons (Figure 5e), as well as highly expressed in embryos (Figure 5f,g). To exclude the possibility that the enhanced CycB1;1-GUS levels reflected upregulation of B-type cyclin gene expression in the hbt mutants, rather than protein stabilization, transcript levels of CycB1;1 and other mitotic cyclin genes were measured by real-time RT-PCR. Several mitotic cyclin genes were downregulated rather than upregulated in the hbt-1 mutants (Figure 5h). Therefore, our data indicate that hbt-1 seedlings are defective in targeted proteolysis of mitotic cyclins, in line with HBT activity in a canonical APC/C complex.

Figure 5.

 Cyclin B expression in hbt-1 mutants.
35S::NtCycB1;1-GUS expression in wild-type (WT) roots (a) and hbt-1 (b, c) seedlings. pCycB1;1::DBCycB1;1-GUS expression pattern in young seedlings (d, WT; e, hbt-1) and embryos (f, WT; g, hbt-1). (h) Real-time PCR quantization of mitotic cyclin gene expression. Scale bars: 25 μm.

Discussion

The initial discovery that HBT/CDC27B encoded a potential APC/C subunit left the question of how interference with the function of the APC/C complex leads to specific embryonic and post-embryonic developmental phenotypes open for investigation (Blilou et al., 2002; Willemsen et al., 1998). Recently, clonal analysis studies suggested that the primary role of HBT was in cell division and cell expansion (Serralbo et al., 2006). Here, we addressed the apparent discrepancy between a canonical role for HBT as cell cycle regulator, and the specific phenotypes, by examining the redundancy between HBT/CDC27B and its homolog CDC27A. Whereas CDC27A and core Arabidopsis APC/C subunits are constitutively expressed (Blilou et al., 2002; Capron et al., 2003b; Kwee and Sundaresan, 2003), we show that HBT/CDC27B protein is restricted to mitotically active and elongating cells, and is mostly excluded from differentiated tissues. Strong hbt mutants impair meristematic cell division and cell elongation post-embryonically, and cdc27a null mutants display at most a very mild gametophytic transmission defect in our experimental conditions. However, the double mutant is synergistic, displaying a gametophytic lethal phenotype. Our data suggest that both proteins are redundantly required for the full APC/C function during cell cycle progression, but that HBT/CDC27 is essential to drive cell division-associated post-embryonic growth. In line with specialized roles of different APC/C subunits, HBT protein fusions, but not CDC27A antibodies, stain the mitotic spindle at both the prophase and the anaphase, and we were unable to detect complexes containing both CDC27 homologues after immunoprecipitation with the core APC2 protein.

Complex roles for the plant APC/C

Cryo-electron microscopy and quantitative determination of subunit stoichiometry of yeast APC/C complexes indicates multiple copies of each TPR-containing protein (Cdc27, Cdc23, Cdc16 and Apc5) (Passmore et al., 2005). However, similar studies on vertebrate APC/C suggest that Cdc27 is present in a single copy per complex (Dube et al., 2005). The latter stoichiometry would be consistent with the formation of CDC27-variant specific APC/C complexes in Arabidopsis, but does not readily explain the ability of HBT to dimerize. Therefore, isolation and subunit analysis will be required to compare plant and non-plant APC/C complexes.

Recent results suggest that the three AtCcs52 activator proteins specifically interact in vitro with different subset of A- and B-type cyclins (Fulop et al., 2005), highlighting the complexity of the G2/M transition in plants. Our findings that CDC27A and HBT differentially interact with a subset of activators indicate that the versatile control of degradation through APC/C may not only be regulated by its activators, but also by its subunit composition. Although apc2 nulls are only impaired in female gametogenesis (Capron et al., 2003b), nomega mutants defective in the CDC16 subunit of the APC/C complex show both female and male gametophyte transmission defects (Kwee and Sundaresan, 2003), similar to cdc27a hbt (this study). Potential targets for such a differential regulation are suggested by Arabidopsis tardy asynchronous meiosis (tam) mutants, which carry a missense mutation in the CycA1;2 gene that reduces cell cycle progression specifically during male meiosis (Wang et al., 2004).

A primary role for HOBBIT in cell cycle control

Our analysis of weak alleles indicates that the primary HBT function is to mediate cell division and endoreduplication, which contributes to meristem activity and cell expansion. Accordingly, post-embryonic removal of a single complementing HBT gene between recombination sites affects cell division and cell elongation, prior to other defects in auxin response and cell differentiation (Serralbo et al., 2006). These data are in line with canonical roles for APC/C components, and suggest that effects on cell division planes and maintenance of cell identity in strong alleles are secondary consequences of cell division and cell expansion defects.

hbt mutants were selected because of a defined embryonic defect in RM founder cells, which at the time was taken to indicate a primary role in cell specification (Willemsen et al., 1998). Our findings indicate that cell cycle control is a primary HBT function, which suggests that alterations in cell cycle can interfere with cell fate determination in plants, for example by disrupting specific signalling events as a consequence of altered cell division patterns. Mutations in DNA polymerase ε, which lengthen the cell cycle, cause similar specific defects in the founder cells for the RM (Jenik et al., 2005). This strengthens the notion that subtle cell cycle alterations in the embryo can interfere with developmental programming in RM founder cells. Interestingly, callus can be induced from hbt mutants, which was, at the time, taken to suggest that the cell cycle was not primarily controlled by HBT (Willemsen et al., 1998). However, cell cycle rates in such callus were not measured, and tissue culture might relax the demand for high HBT activity.

Regulation of APC/C activity through the HBT subunit may help defining cell division and cell EZs

Cells in roots of weak hbt TILLING alleles reached smaller sizes and their endoreduplication levels were reduced, and effects on cell expansion and ploidy levels were also observed upon clonal deletion of HBT in roots and leaves (Serralbo et al., 2006). The HBT gene is transcriptionally active in mitotic cells, but the protein persists in the EZ. Interestingly, plant Fizzy-related and CycA2;3 proteins have been implicated in ploidy control (Cebolla et al., 1999; Imai et al., 2006). In addition, the Drosophila APC2 subunit of the APC/C complex controls both mitotic and endoreduplication cycles (Kashevsky et al., 2002). It is therefore an attractive hypothesis that HBT protein limits APC/C controlled endocycles in the early EZ, and thereby determines final cell size, and future experiments should probe this scenario.

Experimental procedures

Plant material and growth conditions

hbt-11 was found in the progeny of the Salk_021513 line. The hbt-12 and hbt-13 mutants were identified in a collection of Arabidopsis TILLING EMS-induced mutants (Till et al., 2003). hbt-12 carries a V328–M point mutation and the hbt-13 mutation changes G335–R. cdc27a-1, cdc27a-2 and cdc27a-3 mutants in the CDC27A gene were obtained from the Salk_0141413, Salk_059166 and Salk_0872250 insertion lines, respectively, from the SIGNAL collection (Alonso et al., 2003). Primer pairs used for genotyping are displayed in Table 3. After sterilization seeds were stratified for 2 days at 4°C in the dark, and were then transferred to near vertically oriented agar plates containing 0.5 × MS salt mixture (Duchefa, http://www.duchefa.com) and 1% sucrose (pH 5.8), maintained at 22°C with a 16-h light and 8-h dark cycle.

Table 3.   Primer pairs used in this work
Gene (ATG number)Oligo sequences (5′→3′)PCR product size (bp)
ForwardReverse
HBT (At2g20000)
 RT-qPCRATGGATATATACTCTACGGTCCTCAGTGTCTTGTATCTACACGAAGTG 298
 Salk_021513ATTAAATACCTCCGCACCAGGGGAAGCTATGCTTGTGGACTG1063
 TillingAGTTCTCCAAAGTCCACTGTTAACGAAGTTTCATATACGTATCCAGTGC 896
CDC27A (At3g16320)
 RT-qPCRTGCTTTTACAGGTAATGGATGCTCGCACTCCATGTACAGACATC  82
 Salk_141413TTGTCGGAATCTAGAGGATGCTGCAGAGAATTGAATTTGTTGC1054
 Salk_059166GAAATCCTTGCAGGTCAATCAGACGTCAGGCCAGTCTGTAAG 738
 Salk_087225AACGCCTCATCGTTTCTCTGTGATTGACCTGCAAGGATTTC1032
T-DNA
 LBa1TGGTTCACGTAGTGGGCCATCG

Flow cytometry

Between 30 and 40 root tips (500 μm) of 7-day-old seedlings were chopped in 500 μl of cold nuclear isolation buffer [45 mm MgCl2, 30 mm sodium citrate, 20 mm (4-morpholino)propanesulfonate, pH 7.0, 0.1% (w/v) Triton X-100; Galbraith et al., 1983; ] containing 2.5 μg ml−1 4′,6-diamidino-2-phenylindole (DAPI; Roche, http://www.roche.com). The crude preparation of isolated nuclei was filtered (48 μm) and immediately analysed on an ELITE ESP cytometer (Beckman-Coulter, http://www.beckman.com) using UV excitation and gates to eliminate debris or doublets as described in Coba de la Peña and Brown (2001). DNA histograms corresponding to 5000 isolated nuclei were drawn, and the frequency of ploidy levels was calculated using winmdi 2.8 software (Joe Troter, The Scripps Research Institute, http://www.scripps.edu).

HBTg-GUS and HBTg-GFP protein fusions

A rescuing 9-kb genomic fragment of the HBT gene (Blilou et al., 2002) digested with ClaI was fused in frame to the GUS or to the GFP sequences. The resulting pHBT::HBTg-GUS or pHBT::HBTg-GFP constructs were introduced into Agrobacterium tumefaciens strain C58C1 (GV3101) by electroporation, and were used to transform HBT/hbt-1 heterozygotes with the floral-dip method (Clough and Bent, 1998). Whereas HBTg-GUS protein fusion partially rescued the strong hbt-1 allele, rescue of the hbt phenotype in pHBT::HBTg-GFP-expressing plants confirms that HBTg-GFP is fully functional. Histochemical analysis of GUS activity was performed as described in Willemsen et al. (1998). HBTg-GFP seedlings stained with propidium iodide (10 μg ml−1 in distilled water) were visualized by laser scanning confocal microscopy (LSCM) using a Leica (http://www.leica.com) SP2 inverted confocal microscope equipped with GFP filters.

Proteosome inhibitor treatment

pHBT::HBTg-GFP and pHBT::HBTg-GUS seedlings were grown on near vertically oriented plates for 4 days and were then transferred to 24-well Microtiter plates containing 1 ml of 0.5 × MS salt mixture and 1% sucrose (pH 5.8), with 10 μm MG132 proteosome inhibitor and 5 μm cicloheximide. Seedlings were kept for 24 h in the growth chamber and roots were examined by LSCM as described above.

Quantitative RT-PCR

RNA was extracted from frozen samples of 50–100 mg using the Qiagen RNeasy Mini Kit (Qiagen, http://www.qiagen.com), following the instructions of the manufacturer, and chromosomal DNA contamination was removed using the DNA-free kit (Ambion, http://www.ambion.com). Each RNA sample (4 μg) was reverse transcribed using an oligo-dT12-18 primer (Amersham, http://www.amersham.com) and SuperScriptIII (Invitrogen, http://www.invitrogen.com), following the instructions of the manufacturer. The cDNA obtained in this way was diluted by adding 60 μl of distilled water. For the real time PCR, 25-μl reaction mixes were prepared including 12.5 μl of the SYBR Green PCR Master Kit (Applied Biosystems, http://www.appliedbiosystems.com), 0.4 pmol of a primer pair (Table 3) and 1 μl of cDNA. PCR amplifications were carried out in 96-well optical reaction plates on the ABI PRISM 7700 Sequence Detection System (Applied Biosystems). At least three independent amplifications were performed from each cDNA sample. The thermal cycling program started with a step of 2 min at 50°C and 10 min at 95°C, followed by 40 cycles (15 sec at 95°C and 1 min at 60°C). Dissociation kinetics analyses of the amplification products confirmed that only the expected products were amplified.

Relative quantization of gene expression was carried out using the 2–ΔΔCT method (Livak and Schmittgen, 2001). Expression levels of the target genes were normalized using the EF1α (At5g60390) gene as a reference. The relative levels of target gene transcripts and confidence intervals were calculated following the method described by Pérez-Pérez et al. (2004). Data were represented as the relative gene expression normalized to the internal reference (EF1α) and relative to gene expression in WT.

Yeast two-hybrid analysis

GATEWAYTM (Invitrogen) compatible yeast two-hybrid vectors were designed by inserting the GATEWAYTM cassette into the pGADT7 and pGBKT7 backbone. The open reading frames (ORFs) of APC10, CDC27A and CDC27B/HBT were amplified using attB-flanked gene-specific primers, and were transferred via BP and LR reactions to the pGBKgtw and pGADgtw vectors mentioned above. The yeast strain AH109 (Clontech, http://www.clontech.com) was co-transformed with pGADgtw and pGBKgtw constructs containing the different inserts. Plates were incubated for 3 days at 30°C on medium without leucine and tryptophan. For each interaction tested, three individual colonies were mixed in 100 μl H2O and diluted 10-, 100- and 1000-fold. Each dilution series (4 μl) was spotted on medium lacking leucine, tryptophan and histidine. Growth was scored 3–4 days after incubation at 30°C. Cdc20.1, Cdc20.2, Ccs52A1 and Ccs52B exhibited autoactivation. The 10-fold dilution series of the yeast cultures containing Cdc20.1 and Cdc20.2 with the empty vector reduced the growth on the selective medium. This was insufficient in the case of Ccs52A1 and Ccs52B, which grew even in 1000-fold dilution. Autoactivation was controlled by the quantitative inhibitor 3-amino-1,2,4-triazole (3-AT). The strength of the interactions was measured by the ability of yeast strains to grow on histidine-free medium supplemented with 0, 5 and 10, or 25, 50 and 100 mm 3-AT. The lowest concentration that inhibited autoactivation of Ccs52A1 and Ccs52B was 10 mm.

Immunoprecipitation and Western blots

Proteins were extracted from young seedlings using the buffer described in Vodermaier et al. (2003) containing a complete protease-inhibitor cocktail (Roche). For the immunoprecipitation of the APC/C complex, 1 mg of protein extract was pre-cleared with Protein A Sepharose™ 4 Fast Flow (Amersham) for 1 h at 4°C. The supernatant was incubated with 1:300 affinity-purified mouse anti-APC2 (Capron et al., 2003b) for 2 h at 4°C, followed by incubation with 30 μl of Protein A Sepharose™ 4 Fast Flow (Amersham) for another 2 h at 4°C. The washing and elution of immune complexes were performed according to the manufacturer’s recommendations. The eluates were loaded on 10% SDS-PAGE gels, and proteins were transferred to Hybond-ECL membranes (Amersham). The efficiency of the immunoprecipitation was verified using 1:3000 mouse monoclonal 12CA5 anti-HA antibody (Roche) or 1:2000 affinity-purified rabbit anti-CDC27A (Capron et al., 2003b). Secondary goat anti-mouse or goat anti-rabbit horseradish peroxidase (HRP)-conjugated antibodies (Amersham) were used at 1:5000 dilution. Detection was performed using the ECL™ Western Blotting Detection kit following the instructions of the manufacturer (Amersham).

Microscopy and histology

For whole-mount, starch granules visualization and GUS staining, seedlings were cleared and mounted according to the method described by Willemsen et al. (1998). Root length was measured as described (Willemsen et al., 1998). The number of root meristematic cells was obtained by counting cortex cells showing no signs of rapid elongation.

Images were taken using a Zeiss Axioskop microscope equipped with a Nikon DXM1200 digital camera (Carl Zeiss, Inc., http://www.zeiss.com) and were digitally processed with the Adobe photoshop 7.0 program (Adobe Systems Incorporated, http://www.adobe.com).

Aniline blue staining on WT and hbt-11 imbibed seeds was performed as described in Bougourd et al. (2000).

Acknowledgements

We thank Maarten Terlou for help with morphometric analysis and Fritz Kindt, Ronald Leito and Piet Brouwer for photography. This work was supported by an EC-RTN contract to JMPP (HPRN-CT-2002-00333) and an EC-MCF fellowship to OS (HPMF-CT-1999-00013).

Ancillary