SKP2A, an F-box protein that regulates cell division, is degraded via the ubiquitin pathway

Authors

  • Silvia Jurado,

    1. Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria Dpto. Biotecnología (INIA), Carretera de la Coruña Km 7 28 040 Madrid, Spain
    Search for more papers by this author
  • Sara Díaz-Triviño,

    1. Centro de Biología Molecular ‘Severo Ochoa’, Consejo Superior de Investigaciones Científicas, Universidad Autónoma de Madrid, Cantoblanco 28 049, Madrid, Spain
    Search for more papers by this author
  • Zamira Abraham,

    1. Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria Dpto. Biotecnología (INIA), Carretera de la Coruña Km 7 28 040 Madrid, Spain
    Search for more papers by this author
  • Concepción Manzano,

    1. Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria Dpto. Biotecnología (INIA), Carretera de la Coruña Km 7 28 040 Madrid, Spain
    Search for more papers by this author
  • Crisanto Gutierrez,

    1. Centro de Biología Molecular ‘Severo Ochoa’, Consejo Superior de Investigaciones Científicas, Universidad Autónoma de Madrid, Cantoblanco 28 049, Madrid, Spain
    Search for more papers by this author
  • Carlos del Pozo

    Corresponding author
    1. Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria Dpto. Biotecnología (INIA), Carretera de la Coruña Km 7 28 040 Madrid, Spain
      *(fax +34 913573107; e-mail pozo@inia.es).
    Search for more papers by this author

*(fax +34 913573107; e-mail pozo@inia.es).

Summary

Coordination between cell division and cell differentiation is crucial for growth and development of eukaryotic organisms. Progression through the different phases of cell division requires the specific degradation of proteins through the ubiquitin/proteasome 26S (Ub/26S) pathway. In plants, this pathway plays a key role in controlling several developmental processes and responses, including cell proliferation. SKP2A, an F-box protein, regulates the stability of the cell division E2FC-DPB transcription factor. Here, we show that the SKP2A forms a Skp, Cullin containing (SCF) complexin vivo that has E3 ubiquitin ligase activity. Interestingly, SKP2A is degraded through the Ub/26S pathway, and auxin regulates such degradation. SKP2A positively regulates cell division, at least in part by degrading the E2FC/DPB transcription repressor. Plants that overexpress SKP2A increase the number of cells in G2/M, reduce the level of ploidy and develop a higher number of lateral root primordia. Taken together, our results indicate that SKP2A is a positive regulator of cell division, and its stability is controlled by auxin-dependent degradation.

Introduction

The cell cycle consists of concatenated events through the four phases (G1, S, G2 and M) that lead to the generation of two identical daughter cells. Progression through these phases in different organisms is tightly regulated by the activity of cyclin-dependent kinases (CDKs), cyclins, CDK-inhibitors, transcription factors and by the specific degradation of proteins through the ubiquitin/proteasome 26S (Ub/26S) pathway (Bartek and Lukas, 2001; Hershko, 2005; del Pozo et al., 2004, 2006; Vodermaier, 2001). In human cells, SCFSKP2 controls the stability, and therefore the activity, of several cell-cycle regulators, such as p27 (Tsvetkov et al., 1999), cyclin E (Nakayama et al., 2000), E2F1 (Marti et al., 1999), CDK9 (Kiernan et al., 2001), pRB-related p130 (Tedesco et al., 2002), p57kip2 (Kamura et al., 2003) and CDT1 (Li et al., 2003). Because of the large number of cell-cycle proteins regulated by human SKP2, and the fact that this gene is frequently misregulated in several tumors, it is assumed that SKP2 is a critical regulator of cell proliferation (Dehan and Pagano, 2005; Krek, 1998).

Protein conjugation to Ub requires the sequential activity of the ubiquitin-activating (E1), ubiquitin-conjugating (E2) and ubiquitin-ligase (E3) enzymes (Deshaies, 1999). The E1 carries out the activation of Ub in an ATP-dependent manner. Once activated, Ub is transferred to an E2 enzyme. Specificity is largely determined by E3 ligases, which identify and bind substrates for ubiquitination (Deshaies, 1999). Once the targets are identified, they are labeled with a polyubiquitin chain, a signal that is recognized by the proteasome for their degradation (Deshaies, 1999; Smalle and Vierstra, 2004). In plants, the Ub/26S pathway regulates several processes during development in response to internal and/or external stimuli (Dreher and Callis, 2007; Moon et al., 2004; Smalle and Vierstra, 2004). Genomic analyses of the Arabidopsis genome has revealed that approximately 5% of its proteome functions in the Ub/26S pathway, suggesting that this pathway is involved in the control of the majority of plant biological processes. In fact, during the last few years multiple reports have implicated this pathway in the regulation of plant hormones signaling, light responses, circadian clock regulation, pathogen responses or cell division, among others (for a review, see Dreher and Callis, 2007; Moon et al., 2004; Smalle and Vierstra, 2004). Among the E3 ligases, Skp, Cullin, F-box containing (SCF) complexes are one of the best-characterized classes. The SCF is a multi-subunit complex composed of four proteins: CDC53 (Cullin1), RBX, SKP1 (ASK1 in plants) and an F-box protein (Deshaies, 1999; Lechner et al., 2006; del Pozo and Estelle, 2000). CUL1 functions as a structural scaffold for RBX and the ASK1/F-box proteins to form an active E3 ligase. The F-box protein is responsible for specific target recognition. Genomic analyses have identified more than seven hundred proteins containing an F-box motif in Arabidopsis and rice, although it is still unclear whether all of them are part of functional SCF complexes (Gagne et al., 2002; Jain et al., 2007; Risseeuw et al., 2003). In fact, recent works have shown that some F-box proteins might have functions that are not related to the SCF complex and ubiquitination of targets, such as transcriptional co-activator or enzymatic activities (for a review see Hermand, 2006).

In plants, two F-box proteins, SKP2A and SKP2B, have been identified by homology to human SKP2 (del Pozo et al., 2002b, 2006). SKP2A co-immunoprecipitates with CUL1, and targets the cell-cycle repressors E2FC and DPB for degradation (del Pozo et al., 2002b, 2006). The E2FC protein is highly unstable in light-stimulated hypocotyls, suggesting a possible role of this protein in the transition from the skoto- to photomorphogenetic development. In addition, E2FC seems to be important for the DNA endoreplication program, as plants with lower levels of E2FC showed reduced ploidy (del Pozo et al., 2006). Biochemical analyses have shown that E2FC and DPB accumulate in the skp2a mutant, but not in the skp2b mutant (del Pozo et al., 2006). On the other hand, SKP2B targets the CDK inhibitor KRP1 for degradation (Ren et al., in press). Despite the high degree of protein sequence homology (80%) between SKP2A and SKP2B, these data suggest that the two proteins might have differential target affinity.

Although we are starting to unravel the role of SKP2A and SKP2B in plant development, very little is known about their regulation. In this work we show that SKP2A forms an active SCF in vivo that has E3 ubiquitin ligase activity. The SKP2A gene is regulated at the transcriptional level, and the SKP2A protein itself is degraded through the ubiquitin pathway and is modified with N-myristic acid. Interestingly, auxin signaling regulates the degradation of SKP2A. Finally, we show that SKP2A stimulates cell division during plant development, probably through E2FC/DPB degradation in vivo.

Results

Expression of SKP2A is regulated through plant development

To better understand the role of SKP2A, we carried out expression analyses using transgenic plants that harbor the SKP2A::GUS construct. Histochemical analyses of several independent Arabidopsis transgenic lines exhibited a similar expression pattern, which is shown in Figure 1. We found that SKP2A is widely expressed during the first stages of germination (Figure 1a,b). Later, GUS staining was observed almost everywhere in the seedlings, but the staining was more intense in dividing areas, such as meristems and young primary leaves, and in the junction between the root and shoot, where adventitious roots are formed (Figure 1c,d). GUS expression was also detected in the hypocotyl (Figure 1c,d). In mature leaves we found GUS staining in vascular tissue and in stomatal cells (Figure 1e). GUS staining was also detected in young flower buds and mature flowers (Figure 1g). SKP2A was expressed in the root tip (Figure 1h,i), in the vascular cylinder of the distal part of the root (Figure 1h,j) and during the early events of lateral root primordial development (Figure 1j). Interestingly, SKP2A expression was very low and limited to a few cells in the meristem (Figure 1i, k and l).

Figure 1.

 Spatiotemporal expression of the SKP2A gene. The SKP2A promoter was fused to the GUS reporter gene (SKP2A::GUS). Arabidopsis transgenic plants harboring this reporter gene were analyzed by GUS staining at different stages of development.
(a) Imbibed seeds (1 day).
(b) Imbibed seedlings (2 days).
(c) Seedling (3.5-days old).
(d) Seedlings (7-days old).
(e) Mature rosette leaf (21-days old).
(f) Young flower buds.
(g) Mature flower.
(h) Main root of 3.5-days-old seedling.
(i) Higher magnification of the main root tip.
(j–l) Different developmental stages of a lateral root formation. Scale bars: 0.05 mm, except for 3 mm in (h).

Analysis by semi-quantitative RT-PCR showed that SKP2A was expressed in all the analyzed organs (Figure S1a). To analyze whether SKP2A is regulated through the cell cycle, its expression was analyzed by quantitative RT-PCR using RNA from synchronized Arabidopsis cultured cells. SKP2A shows two peaks of expression: one during the early S-phase and a second at G2/M (Figure S1b), indicating that SKP2A is also cell-cycle regulated.

SKP2A forms an SCF complex with ubiquitin ligase activity

F-box proteins interact with the ASK proteins through the F-box motif. We analyzed the interaction between SKP2A and different members of the ASK subfamilies using the two-hybrid assay. SKP2A strongly interacts with ASK1 and ASK2 and, to a lesser extent, with ASK18 (Figure 2a). However, interaction of SKP2A with ASK6, ASK8, ASK10 or ASK14 proteins was not detected (Figure 2a). In plants, several F-box proteins have been shown to form SCF complexes, but so far, their E3 ligase activity has not been reported. Therefore, we decided to analyze whether SKP2A has E3 ligase activity. To obtain a native SCFSKP2A complex, the MYC-SKP2A protein was immunopurified from protein extracts of MYC-SKP2AOE seedlings using anti-MYC agarose beads. SCFSKP2A was released from the beads by competition with the MYC peptide, and was then analyzed by immunoblotting. We found that ASK1 and, with less efficiency, CUL1 co-immunoprecipitated with MYC-SKP2A in vivo (Figure 2b). Despite the low quantity of CUL1 immunoprecipitated, when the immunopurified-SCFMYC-SKP2A was incubated with an E1 and Arabidopsis UBC8 in the presence of ATP, clear ubiquitin ligase activity was detected (Figure 2c). Interestingly, if the E2 enzyme was omitted from the reaction, ubiquitin ligase was still detected, although to a lesser extent (compare lanes 2 and 3), indicating that E2 activity might have been immunoprecipitated. No E3 ligase activity was observed when the proteins were incubated without ATP (Figure 2c). These results show that SKP2A forms an SCF complex that has E3 ligase activity.

Figure 2.

 SKP2A forms a Skp, Cullin containing (SCF) complex with E3 ligase activity.
(a) Two-hybrid analysis showing the interaction of SKP2A with ASK1, ASK2 or ASK18, but not with ASK6, ASK8, ASK10 and ASK14. The strength of the interaction was represented by the + symbol.
(b) Total protein extracts from wild-type (wt) or MYC-SKP2AOE plants were subjected to immunoprecipitation with anti-MYC beads, and the immunoprecipitated proteins were released by MYC-peptide competition. Then, these proteins were analyzed by immunoblotting with the antibodies against MYC, CUL1 or ASK1.
(c) The SCFMYC-SKP2A (E3) was immunopurified as in (b). These proteins were incubated with yeast E1, Arabidopsis GST-UBC8 and biotynilated-ubiquitin, as indicated. The reaction products were resolved by SDS-PAGE under reducing conditions to only detect covalent conjugates of ubiquitin. The proteins covalently modified with ubiquitin were detected by streptavidin-horseradish peroxidase (HRP) using the ECL system. As a control, similar reactions were carried out without ATP. The arrow indicates free biotinylated-ubiquitin.

SKP2A is a nuclear protein that is regulated by degradation

To determine the subcellular localization of SKP2A, we fused SKP2A to the GFP protein and expressed it in Arabidopsis plants under the control of the 35S promoter (GFPOE or SKP2A-GFPOE). The GFP protein alone localized both in the nucleus and in the cytoplasm, whereas SKP2A-GFP preferentially localized to the nucleus, although some SKP2A-GFP was also localized in the cytoplasm (Figure 3a). Although SKP2A-GFPOE plants expressed a high level of SKP2A-GFP mRNA (data not shown), the GFP signal was low, suggesting that SKP2A-GFP protein might be degraded. As it has been reported that some F-box proteins are degraded (Galan and Peter, 1999; Kipreos and Pagano, 2000), we wondered whether SKP2A is regulated by Ub/26S-dependent proteolysis. To test this, we treated SKP2A-GFPOE plants with the proteasome inhibitor MG132 or the DMSO solvent. When the proteasome activity is blocked, SKP2A-GFPOE plants accumulated a higher fluorescence signal (Figure S2). In addition, when MYC-SKP2AOE plants were treated with MG132 for 3 h, the MYC-SKP2A protein significantly accumulated (Figure 3b), whereas the mRNA expression remained similar (data not shown). To analyze the stability of SKP2A, we carried out cell-free degradation assays using MYC-SKP2AOE protein extract in the presence of DMSO or MG132. Figure 3(c) shows that the MYC-SKP2A protein almost disappeared after 80 min of incubation with DMSO, but remained stable when incubated with MG132. To analyze whether SKP2A is modified with Ub, we immunoprecipitated the MYC-SKP2A protein and analyzed this fraction by immunoblotting with an antibody against MYC or Ub. As shown in Figure 3(d), a smear of polyubiquitinated proteins was detected in the MYC-SKP2A fraction, but not in the control. It is remarkable that a defined anti-Ub reacting band ran to a similar position as the anti-MYC reacting band (Figure 3d, arrow). In addition, we carried out pull-down assays of MYC-GFPOE or MYC-SKP2AOE plant extracts using agarose-p62 beads or Q-matrix, both of which bind to ubiquitinated proteins. Both the agarose-p62 beads and the Q-matrix bound to an isoform, which probably corresponds to a monoubiquitinated MYC-SKP2A protein and to high-molecular-weight proteins that were detected with antibodies against MYC (Figure S3). Together, these results indicate that the MYC-SKP2A protein was ubiquitinated in vivo and was degraded through the UB/26S pathway.

Figure 3.

 SKP2A is a nuclear protein that is degraded by the ubiquitin pathway.
(a) Transgenic plants that overexpressed GFP or SKP2A-GFP fusion protein under the control of the 35S promoter were analyzed by confocal microscopy. The GFP protein was localized in both the nucleus and the cytosol, whereas the SKP2A-GFP protein was localized in the nucleus (arrowhead) and in the cytoplasm (arrow) of the root cells.
(b) SKP2A is degraded by the ubiquitin-proteasome pathway. Wild-type (wt) and MYC-SKP2AOE seedlings were treated with MG132 (+) or DMSO (−) for 3 h. Total protein extracts were analyzed by immunoblotting with antibodies against MYC. LC, loading control: blot stained with Coomassie blue.
(c) MYC-SKP2AOE protein extracts were incubated in an in vitro degradation assay with a proteasome inhibitor (MG132) or the DMSO solvent for different times. The MYC-SKP2A protein was detected by immunoblotting with antibodies against MYC. LC, loading control: blot stained with Coomassie blue.
(d) Protein extracts from wt and SKP2AOE plants were immunoprecipitated using anti-MYC beads. Half of the immunoprecipitated proteins were subjected to immunobloting using antibodies against MYC, and the other half with antibodies against ubiquitin (Ub). The open arrowhead points to the MYC-SKP2A protein, and the bracket indicates the MYC cross-reacting smear. The arrowhead indicates a smaller isoform of MYC-SKP2A protein that was specifically immunoprecipitated. The asterisks indicate unspecific cross-reacting bands. The arrow indicates a band that cross-reacts with anti-MYC and anti-Ubiquitin antibodies.
(e) The GFP and SKP2A proteins were translated in vitro in the presence of [35S]methionine or [9,10(n)-3H]-myristic acid. The arrow indicates the free methionine or myristic acid.

We carried out an in silico sanalysis on the SKP2A protein sequence and found several motifs for modification with N-myristic acid. To test if SKP2A is N-myristoylated, we translated the SKP2A and GFP proteins in vitro in the presence of radiolabeled N-myristic acid or methionine. As shown in Figure 3(e), SKP2A, but not GFP, is modified with the N-myristic acid. This result suggests that SKP2A can be myristoylated in vivo.

Auxin regulates the stability of SKP2A

To determine whether the proteolysis of SKP2A is regulated by hormones, we treated MYC-SKP2AOE seedlings with different plant hormones. We found only a slight reduction of MYC-SKP2A levels when auxin was added (Figure S4). However, this reduction was severe when we used the protein synthesis inhibitor cyclohexamide (Figure S4). Other hormones tested have little or no effect on SKP2A degradation (data not shown). To better analyze the influence of auxin, we carried out a time-course experiment. MYC-SKP2AOE seedlings were treated with cyclohexamide and with or without auxin. As shown in Figure 4(a), SKP2A levels were similar between seedlings treated with and without auxin after 30 min. However, auxin strikingly reduced the SKP2A levels after 60–120 min. As a control for protein degradation, we used MYC-GFPOE plants and, as shown in Figure 4(b), auxin did not change the GFP levels. These results indicate that SKP2A degradation does not require new protein and components of SKP2A degradation were already present in plants. As auxin seems to stimulate SKP2A degradation, we wanted to analyze the effect of impaired auxin response on SKP2A levels. To do this we used the chemical compound terfestatin-A, which blocks the auxin response (Yamazoe et al., 2005). We treated MYC-GFPOE or MYC-SKP2AOE seedlings with terfestatin-A over a period of 32 h, and then the level of both proteins were analyzed by immunoblotting. As shown in Figure 4(b), plants treated with terfestatin-A highly accumulated the MYC-SKP2A protein, but the level of MYC-GFP did not change. We also analyzed SKP2A stability in various auxin-response mutants. We found that axr2 and axr3 mutations severely reduced the levels of SKP2A (Figure 4c), suggesting that perturbations in the auxin response homeostasis destabilizes SKP2A protein. On the other hand, the axr1-12 mutation seems to slightly accumulate SKP2A (Figure 4d).

Figure 4.

 Auxin regulates the levels of SKP2A.
(a) MYC-SKP2AOE and MYC-GFPOE plants were treated with cyclohexamide and auxin for the indicated times. Total protein was analyzed by immunoblotting using an antibody against MYC. LC, loading control: blot stained with Coomassie blue.
(b) Total protein extracts from MYC-GFPOE and MYC-SKP2AOE seedlings treated with DMSO (−) or terfestatin-A (Terf. A) (+) were subject to immunoblotting using an antibody against MYC. LC, loading control: blot stained with Coomassie blue.
(c) Total protein extracts from 5-day-old MYC-SKP2AOE: two lines of axr2/MYC-SKP2AOE or axr3/MYC-SKP2AOE seedlings were subjected to immunoblotting using an antibody against MYC. LC, loading control: unspecific cross-reacting band. Right panel shows the expression level of the MYC-SKP2A gene analyzed by RT-PCR. As a control, ACTIN gene expression was analyzed.
(d) Total protein extracts from MYC-SKP2AOE or axr1-12/MYC-SKP2AOE 5-day-old seedlings were subjected to immunoblotting using an antibody against MYC. LC, loading control: unspecific cross-reactingband. Right panel shows the expression level of the MYC-SKP2A gene analyzed by RT-PCR. As a control, ACTIN gene expression was analyzed.
(e) Total protein extracts from 5-day-old wild-type (wt) seedlings treated with 2,4-d or ethanol were subjected to immunoblotting with affinity-purified antibodies against E2FC. LC, loading control: unspecific cross-reacting band.

As SKP2A targets E2FC/DPB for degradation, we analyzed the level of these factors after treatment with auxin. Plants treated with auxin for 6 h slightly accumulated E2FC (Figure 4e) and DPB (ZA and CdP, unpublished data).

SKP2A stimulates cell division in the meristems

We have previously shown that SKP2A targets E2FC and DPB, two cell-cycle proteins that repress cell division, for degradation through the Ub/26S pathway (del Pozo et al., 2002b, 2006). To better understand the role of SKP2A in cell division in vivo, we crossed the CYCB1-GUS line that harbors a labile cyclin B1;1 fused to a GUS protein, which only accumulates in G2/M cells (Colón-Carmona et al., 1999), with two independent lines of MYC-SKP2AOE plants. The results of both crosses were similar, and in this work we present the data for line MYC-SKP2A-22. Histochemical analysis of these seedlings revealed that overexpression of MYC-SKP2A strikingly increased the number of cells that accumulate CYCB1-GUS either in the shoot or in the root meristems, and in young expanding leaves (Figure 5a). As cell division in meristems is stochastic, we quantified the number of G2/M cells in the root meristems by counting CYCB1-GUS stained cells. As shown in Figure 5(b), MYC-SKP2AOE root meristems contained more dividing cells on average than the control plants. This higher number of GUS-positive cells in the root meristem might be a consequence of more dividing cells or more cells arrested at G2/M. To discern between these possibilities we analyzed the levels of the KNOLLE protein (Lauber et al., 1997), which accumulates in the division plate of mitotic cells. As shown in Figure 5(c), more KNOLLE-positive cells were detected in the root meristem of MYC-SKP2AOE plants, suggesting a higher rate of cell division. In addition, CYCB1-GUS/MYC-SKP2AOE seedlings developed longer roots than CYCB1-GUS seedlings (Figure 5d). As root cell size is similar between wild-type (wt) and SKP2AOE plants (Table 1), the difference in root growth is probably caused by the higher number of cells produced in the meristem. Taken together, these results indicate that overexpression of the MYC-SKP2A gene promotes cell division in proliferative areas.

Figure 5.

 Overexpression of SKP2A increases cell division in proliferative areas.
(a) GUS staining of 7-day-old CYCB1-GUS or CYCB1-GUS/MYC-SKP2AOE seedlings. The pictures show representative apical shoots or root tips of these seedlings. Scale bars: 0.5 mm.
(b) Distribution of GUS-positive cells per meristem of 3- or 7-day-old CYCB1-GUS or CYCB1-GUS/MYC-SKP2AOE seedlings.
(c) Confocal fluorescence images showing the KNOLLE protein in the root tips of wild-type (wt) or MYC-SKP2AOE seedlings.
(d) Root length of 3- or 7-day-old CYCB1-GUS or CYCB1-GUS/MYC-SKP2AOE seedlings. *Statistically significant difference, as determined by the Student’s t-test (P < 0.001).

Table 1.   Cell length of CYCB1-GUS and CYCB1-GUS/MYC-SKP2AOE roots.
 CYCB1-GUSCYCB1GUS/MYCSKP2AOE
  1. The cells were selected from the upper region of the elongation zone. At least 50 cells of 10 different roots were measured. There were no statistically significant differences between the values from CYCB1-GUS and CYCB1-GUS /MYC-SKP2A, as determined by the Student’s t-test (P < 0.001)

Average (mm)0.11340.1226
SD0.01090.0194

We have shown that E2FC-overexpressing plants showed a reduction in the number of cells that accumulated the cell-cycle marker CYCB1-GUS in the root meristem (del Pozo et al., 2006). As SKP2A seems to target E2FC for degradation, we decided to analyze whether the level of CYCB1-GUS could be restored by overexpressing SKP2A. Thus, we crossed CYCB1-GUS/E2FCOE with MYC-SKP2AOE or with wt plants. GUS staining of the F1 seedlings of these crosses showed that the level of GUS-positive cells was restored by the overexpression of SKP2A (Figure 6a,b). These data suggest that SKP2A is likely to degrade the excess of the E2FC protein. This idea was corroborated by immunoblotting analyses, because MYC-SKP2AOE plants accumulated less E2FC protein in the roots than wt plants (Figure 6c). Plants with reduced levels of E2FC (e2fc-R) reach lower ploidy levels than wt cells (del Pozo et al., 2006). To test the effect of overexpressing MYC-SKP2A on ploidy level, we quantified the DNA content in wt and MYC-SKP2AOE leaves. Consistent with the ploidy level of e2fc-R plants, overexpressing SKP2A leads to a reduction in DNA content (Figure 6d).

Figure 6.

 SKP2A targets E2FC to regulate cell division.
(a) Representative pictures of root meristems of 5-day-old CYCB1-GUS, CYCB1-GUS/E2FCOE × WT and CYCB1-GUS/E2FCOE × MYC-SKP2AOE seedlings stained for GUS activity.
(b) Quantification of the GUS-positive cells in the root meristem of 5-day-old CYCB1-GUS, CYCB1-GUS/E2FCOE × WT and CYCB1-GUS/E2FCOE × MYC-SKP2AOE seedlings. The values represent the average and SD of at least 25 meristems.
(c) Total root protein extracted from wild-type (wt) and MYC-SKP2AOE plants subjected to direct immunoblotting with affinity-purified antibodies against E2FC. LC, loading control: unspecific cross-reacting band.
(d) Ploidy distribution of 14-day-old rosette leaf nuclei of wt and MYC-SKP2AOE plants.

Overexpression of SKP2A induces lateral root primordia formation

Expression analyses have shown that SKP2A was expressed in the vascular cylinder of the root, and that during the early event of lateral root primordia (LRP) formation. To analyze whether SKP2A functions during LRP development, we took advantage of the CYCB1-GUS/MYC-SKP2AOE lines because GUS activity labels spots that correspond to LRP. GUS-staining analyses of CYCB1-GUS and CYCB1-GUS/MYC-SKP2AOE seedlings showed that over-expression of MYC-SKP2A increased the number of LRP formed per centimeter of main root (Figure 7a). It is well known that auxin signaling is needed to trigger LRP formation. To analyze whether the increase in LRP is dependent on auxin response, we crossed the slr1 mutant, which is impaired in the auxin response and does not develop lateral roots (Fukaki et al., 2002), with the MYC-SKP2AOE plants. We did not find any lateral root in either slr1 or slr1/MYC-MYC2AOE plants grown in MS medium over a period of 10 days (Figure 7b), suggesting that the effect of SKP2A on LRP production is auxin dependent. We found that overexpression of MYC-SKP2A increases cell division in the LRP, but also stimulates the auxin response. To test this we crossed a MYC-SKP2AOE line with DR5::GUS, an auxin-response marker line. When we analyzed the F1 of these crosses, we found that overexpression of MYC-SKP2A locally increased the GUS staining of the roots in the differentiation/elongation zone (Figure 7c). In addition, when these lines were treated with auxin for 3 or 6 h, overexpression of MYC-SKP2A significantly increased the signal of the auxin-response marker in the meristem and in the differentiation zone. After 24 h of auxin treatment, the DR5::GUS/MYC-SKP2AOE seedling showed stronger GUS staining in the root and in the shoot meristem than the DR5::GUS seedling (Figure S5). This data suggests that SKP2A might couple the auxin response and cell division to form LRP.

Figure 7.

 Overexpression of SKP2A increases the number of lateral root primordia.
(a) Number of lateral root primordia (LPR), root length (RL) and LRP per centimetre of primary root length (LPR/RL). CYCB1-GUS or the CYCB1-GUS/MYC-SKP2AOE seedlings were grown in MS medium for 5 days and were then stained for GUS activity. The LPRs were measured as GUS-stained spots. *Statistically significant difference, as determined by Student’s t-test (P < 0.001).
(b) Picture of wild-type (wt), MYC-SKP2AOE, slr1 and slr1/MYC-SKP2AOE seedlings grown for 10 days in MS medium. slr1 and slr1/ MYC-SKP2AOE seedlings did not show any LRP.
(c) GUS staining of DR5::GUS or DR5::GUS/MYC-SKP2AOE seedlings grown in MS for 5 days and for 3 or 6 h with or without 0.1 μm of 2,4-d. Arrows indicate points of GUS staining in the DR5::GUS/MYC-SKP2AOE roots.

Discussion

SKP2A forms an SCF complex with E3 ligase activity

The SCF class of ubiquitin ligases are composed of four subunits; Cullin1, SKP1 (ASK in plants) and RBX form a common scaffold in which different F-box proteins are assembled. F-box proteins interact with ASKs through the F-box motif and are responsible for specific recognition of the target (Jackson and Eldridge, 2002; Patton et al., 1998; Tan et al., 2007). In the case of plants, different F-box proteins interact with members of the ASK family with different affinity (Risseeuw et al., 2003; Takahashi et al., 2004). Here, using the two-hybrid assay, we found that SKP2A strongly interacts with ASK1 and ASK2, and weakly interacts with ASK18. In addition, we showed that SKP2A interacts in vivo with ASK1, and with CUL1 and CUL1-RUB modified isoforms. This data suggests that, similarly to human SKP2, SKP2A might form distinct types of SCF complexes and, consequently, might target several proteins for degradation.

Several reports have described E3s involved in diverse aspects of plant development and responses to environmental signals (Dreher and Callis, 2007; Moon et al., 2004; Smalle and Vierstra, 2004). Ubiquitin ligase activity has been shown in vitro for some types of E3 ligases (Dong et al., 2006; Kraft et al., 2005; Yang et al., 2006; Zhang et al., 2005), but, so far, ubiquitin ligase activity for a plant SCF complex has not been demonstrated biochemically. This point is important as recent works have described some F-box proteins that did not form SCF complexes, and are not involved in target ubiquitination (for a review see Hermand, 2006). Here, we developed an in vivo–in vitro assay that demonstrated SKP2A forms an SCF complex in vivo and catalyzes the ubiquitination, confirming that SKP2A forms part of a functional SCF. The identities of these ubiquitinated proteins are unknown, although they might be the E2, MYC-SKP2A or other proteins that co-immunoprecipitate with MYC-SKP2A. As mentioned before, some E2 enzymes might have target and/or E3 specificity (Ardley and Robinson, 2005). SCFSKP2A showed higher activity when incubated with UBC8 rather than with other Arabidopsis E2s (data not shown), suggesting that UBC8 and SKP2A might function together to promote ubiquitination. Interestingly, when we carried out this assay without UBC8 we also observed E3 activity, indicating that it is likely that an endogenous E2 is immunoprecipitated with MYC-SKP2A. Taken together, we have been able to show that the SCFSKP2A complex has E3 Ub-ligase activity. The development of this kind of assay will be important to further address whether an F-box containing protein functions as an E3 ligase or has other functions.

SKP2A is regulated by Ub/26S-dependent degradation

Human SKP2, which plays a key role in the control of the cell cycle, is regulated by different mechanisms, such as gene transcription (Lisztwan et al., 1998;Michel and Xiong, 1998), acetylation and ubiquitin-dependent proteolysis (Wirbelauer et al., 2000). Arabidopsis SKP2A seems to be regulated, at least, at two levels: (i) at the transcriptional level, because SKP2A is cell-cycle regulated and highly expressed in dividing areas, and (ii) at the post-translational level, because SKP2A protein is degraded through the Ub/26S pathway and modified with N-myristic acid. The degradation of F-box proteins by the Ub/26S pathway has been previously described (Galan and Peter, 1999; Kipreos and Pagano, 2000). It is thought that during evolution the F-box motif was first used as a destruction mark, which was later incorporated in the F-box proteins to selectively recruit target proteins. This strategy leads to the possibility in which the target or both the target and the F-box protein are degraded. In mammals, SKP2 is degraded by the APC(Cdh1) complex (Bashir et al., 2004). In the case of SKP2A, its degradation might be achieved by a specific E3 or might be a consequence of autoubiquitination. Interestingly, we have found that SKP2A stability is regulated by auxin, which stimulates its degardation. In addition, SKP2A is highly unstable in the axr2 or axr3 mutants, but not into the axr1-12 or slr1 mutants. axr3 mutants showed an increase in the magnitude of the auxin response, but the axr2 phenotype is consistent with lower auxin-response activity (Leyser et al., 1996; Timpte et al., 1994). These results showed that auxin response implies complex network pathways in which AUX/IAA proteins have different functions. Our data clearly indicate that a correct auxin-response homeostasis is important to control SKP2A stability. Terfestatin-A is a compound that led to the accumulation of IAA proteins, thereby blocking the auxin response. Interestingly, plants treated with terfestatin-A also accumulated SKP2A, supporting the idea that somehow the auxin response controls SKP2A stability. Similarly, SKP2A slightly accumulated in the axr1-12 mutant, which impaired the modification of CUL1 with RUB. It is possible that this lack of RUB modification reduces the SCF-dependent degradation of SKP2A. However, another possibility is that SKP2A does not assemble into an SCF in the axr1-12 mutant, thereby blocking its autoubiquitination and degradation. Although we do not have scientific evidence we favor the second possibility, as E2FC accumulated in the axr1-12 mutant, suggesting that SKP2A, which also accumulated in the mutant, is not functional. Taken together, our data suggests that a balance exists between SKP2A expression and protein level, generating a functional turnover that might avoid an overfunctioning of the SCFSKP2A complex.

Likewise, auxin treatment led to the accumulation of E2FC and DPB, probably via the destabilization of SKP2A. Similarly, auxin mediates the stabilization of E2FB in BY2 cells (Magyar et al., 2005). Thus, it is tempting to speculate that one of the connections between auxin response and cell division is controlled by SKP2A activity and the stabilization of E2FC/DPB and/or E2FB. Furthermore, it will be very interesting to analyze whether SKP2A targets E2FB for degradation.

SKP2A is modified with N-myristic acid. It has been described that this modification is related to the membrane attachment of proteins, and is likely to be involved with the regulation of protein–protein interactions (Johnson et al., 1994; Matsubara et al., 2003). Owing to the nature of the F-box proteins that interact with the targets for their ubiquitination, it is possible that myristoylation of SKP2A plays a role in regulating the recruitment of its targets.

SKP2A, a cell-cycle gene

The results presented in this manuscript illustrate that SKP2A functions to control cell division in plants. Based on the SKP2A expression profile, it is likely that it functions during late S-phase and mitosis. In keeping with this, overexpression of SKP2A increases the number of dividing cells in the meristems, indicating that SKP2A might function in promoting cell-cycle progression, but it is not sufficient to induce ectopic proliferation by itself in non-dividing zones. How does SKP2A control cell division? Certainly, SKP2A targets different cell-cycle regulators for proteolysis through the Ub/26S pathway. Recently, we have shown that the cell-cycle repressors E2FC and DPB are targeted by the SCFSKP2A for degradation (del Pozo et al., 2002b, 2006). Interestingly, MYC-SKP2AOE plants show a phenotype similar, in some aspects, to that found in plants with lower levels of E2FC (e2fc-R), such as the increased expression of the CYCB1-GUS marker, the higher number of LRP, and the lower DNA ploidy levels (del Pozo et al., 2006). In addition, we found that the number of cells showing CYCB1-GUS, which were reduced in the E2FCOE plants, were re-established by overexpressing MYC-SKP2A. Taken together, our data indicate that SCFSKP2A promotes cell division by degrading E2FC and DPB. However, the function of SKP2A might not be limited to targeting these two proteins. In fact, human SKP2 targets several regulators of cell division for degradation (see Introduction), but also targets other non-cell-cycle related proteins, such as the ISG15 isopeptidase (Tokarz et al., 2004) or E2A (Nie et al., 2003). Thus, it would not be surprising that Arabidopsis SKP2A might also target diverse proteins for degradation.

Lateral root primordia are derived from G1-arrested pericycle cells (for review see Casimiro et al., 2003; De Smet et al., 2006). It has been shown that auxin triggers the re-entry of the pericycle cells in proliferation to form LRP (Benkova et al., 2003; Himanen et al., 2002; Vanneste et al., 2005). Lateral root formation requires the auxin response pathway and specific proteolysis of key regulatory proteins, such as the IAA proteins by the TIR1 family of F-box proteins (Dharmasiri and Estelle, 2002; Gray et al., 2001) and NAC1 by SINAT5 (Xie et al., 2002). Reduced activity of the SCFTIR1 complex or the AXR1-RUB pathway impairs the auxin response and reduces the number of lateral roots developed (Lincoln et al., 1990; del Pozo et al., 2002a; Ruegger et al., 1998). In addition, mutations in another E3, XBAT32, also reduces lateral root formation, but in this case, the target proteins of this E3 are unknown (Nodzon et al., 2004). These observations clearly indicate that the Ub/26S pathway plays a key role in controlling LRP formation and development. We have found that overexpression of MYC-SKP2A increased the number of LPRs, and that this increase is dependent on the auxin response. Recently, it has been proposed that the auxin response in this zone is critical for the priming of pericycle cells for lateral root initiation (De Smet et al., 2007). It is remarkable that overexpression of SKP2A increases the DR5::GUS expression in the meristem and differentiation zone, suggesting that SKP2A might be involved in such priming. This higher level of DR5::GUS expression can be explained for two reasons: overexpression of SKP2A might increase the auxin response, or might accumulate more auxin in the root. We favor the first possibility, as when the seedlings are treated with high levels of auxin the MYC-SKP2A-overexpression plants showed higher expression of the DR5::GUS marker. The slr1 mutant does not develop any visible lateral roots (Fukaki et al., 2002). SLR encodes the IAA14 transcription factor that regulates the auxin response, and its degradation is required for lateral root formation. We found that slr1/MYC-SKP2AOE seedlings did not develop any visible lateral roots, indicating that SKP2A overexpression is not sufficient to induce lateral root formation and requires the correct auxin response. Overexpression of activators of cell division in the pericycle cells can induce cell division, but this is not sufficient to develop an LRP (Vanneste et al., 2005), probably resulting from the lack of auxin-response activation. In our case, SKP2A stimulates cell division, probably by degradation of E2FC and DPB, but also increased an auxin response locally. This data suggests that SKP2A might be coupling cell division and auxin response to promote LRP formation.

Experimental procedures

Plant material and growth conditions

All studies were carried out using the Arabidopsis thaliana ecotype Columbia. The MYC-SKP2A overexpressing plants (MYC-SKP2AOE) were described in del Pozo et al. (2004, 2006), expressing the cDNA of the SKP2A gene (At1g21410) under the control of the constitutive promoter 35S. The GFP coding region was cloned using Gateway technology into the C-TAP vector (Rubio et al., 2005). The SKP2A cDNA was cloned into the gateway vector pDONR221 (Invitrogen, http://www.invitrogen.com), and was then transferred into the Gateway binary vector pGWB5 (http://bio2.ipc.shimane-u.ac.jp/pgwbs/index.htm) by LR recombination (Invitrogen) to generate the 35S::SKP2A-GFP construct. To generate the SKP2A::GUS construct, we cloned a 1.8-kb DNA fragment upstream from the ATG of the SKP2A gene into the pBI101.1 vector (Clontech, http://www.clontech.com). These SKP2A-GFP and SKP2A::GUS constructs were introduced into Agrobacterium tumefaciens (C58C1 strain), and were used for transformation of Arabidopsis plants by the infiltration method (Clough and Bent, 1998). Independent stable transgenic lines were selected for further studies. To generate the axr1-12/MYC-SKP2AOE, slr1//MYC-SKP2AOE, axr2//MYC-SKP2AOE or axr3/MYC-SKP2AOE plants we crossed the axr1-12, slr1-1, axr2-1 or axr3-1 mutants with MYC-SKP2AOE plants. Afterwards we performed back-crosses until we obtained plants expressing similar levels of MYC-SKP2A mRNA as the parental lines.

RT-PCR

Total RNA was extracted using the Trizol procedure, as described by the manufacter (Invitrogen). For the semiquantitative RT-PCR, 200 ng of total RNA was amplified using the one-step RT-PCR kit (Invitrogen), and the amplified bands were analyzed by agarose electrophoresis. The β-ACTINE gene was amplified as a loading control.

Immunoprecipitation and western-blot analyses

Wild-type, MYC-GFPOE or MYC-SKP2AOE seedlings were grown for 4 days in MS medium (16-h light/8-h dark). Seedlings were ground in liquid nitrogen and total proteins were extracted in TBS-0.5 buffer [100 mm Tris–HCl, pH 7.5, 150 mm NaCl and 0.5% NP-40 supplemented with 1 mm PMSF (phenyl-methyl-sulfonyl-fluoride) and plant protease cocktail inhibitor (Sigma, http://www.sigmaaldrich.com), and, when indicated, plus proteasome inhibitors (4 nm of epoxomicin or 50 μm of MG132; Biomol, http://www.biomol.com)]. The extracts were sonicated for 10 s three times and then incubated for 15 min on ice. Afterwards, the extracts were cleared by centrifugation and the supernatants filtrated through 0.80-μm filters. To immunoprecipitate MYC-SKP2A protein, total protein extract was incubated with 0.4 μg of 9E10-biotynilated antibody (Sigma) for 1 h at 4°C. Streptavidin-agarose beads (Sigma) were added and then incubated for one additional hour. Beads were washed four times for 15 min each with the extraction buffer. To release the immunopurified proteins, the beads were incubated with the MYC peptide at 0.5 μg μl−1 (Sigma) for 90 min in elution buffer (50 mm Tris–HCl, pH 7, 100 mm NaCl) at 20°C, and then analyzed by immunoblotting using the monoclonal antibody (9E10) against MYC (Santa Cruz Biotechnology, Inc., http://www.scbt.com). The CUL1 antibody was used as described by Gray et al. (1999). The ASK1 antibody was obtained from W. Crosby’s lab.

Yeast transformation and LacZ and HIS3 assays

For the two hybrid experiments we used the pGAD and pGBT8 plasmids (Clontech) adapted to GatewayTM technology. SKP2A was cloned into pGBT8, and the cDNAs of ASK1, ASK2, ASK6, ASK8, ASK10, ASK14, ASK18 and SKP2A were introduced into the pGAD by LR reaction. In these experiments we used the haploid strain of Saccharomyces cerevisiae HF7C (Clontech) carrying the LacZ and HIS3 reporter genes under the control of a truncated Gal1 promoter, and yeast transformation was performed by the polyethylene glycol method, and transformants were screened for β-galactosidase production using X-gal (DUCHEFA, http://www.duchefa.com) as substrate and growth in absence of histidine supplemented with 10 or 30 mm of 3AT (3-amino-1,2,4-triazole; Sigma).

E3 ubiquitin ligase activity assays

The UBC8 (At5g41700) cDNA was cloned into the pGEX vector (Amersham, http://www.amersham.com). The recombinant GST-UBC8 was expressed in the BL21 bacterial strain in standard conditions, and was purified following the manufacture’s instructions. Recombinant protein was released from beads with 50 mm of reduced glutathione in 20 mm Tris–HCl, pH 7, 20 mm NaCl, 5% of glycerol, and kept at −80°C.

MYC-SKP2A was immunoprecipitated, as described above, using 1 mg of total protein extract of 4-day-old MYC-SKP2AOE plants. The SCFMYC-SKP2A complex was released from the streptavidin-agarose beads by competition with the MYC peptide in a final volume of 20 μl of 50 mm Tris–HCl, pH 7, 100 mm NaCl. For each E3 ligase assay we mixed 2 μl of the SCFMYC-SKP2A complex with 100 nm yeast E1, 0.1 μg of GST-UBC8, 0.1 mm DTT, 5 mm MgCl2 and 0.5 μg of biotynilated-ubiquitin, and 5 mm ATP. In the control reaction ATP was avoided. The reactions were incubated at 30°C for 45 min and then stopped by adding loading buffer with 100 mm DTT, and then incubated at 100°C for 10 min. In both cases, the proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes (GE Healthcare, http://www.gehealthcare.com). These blots were blocked with 3% BSA in PBS-T (PBS1x with 0.1% of tween-20) for 1 h. Afterwards, the blot was incubated with Streptavidin-horseradish peroxidase (Sigma) at 1:5000 for 45 min to detect the Ub-biotin conjugated proteins. The blot was washed three times for 15 min each and then revealed with the ECL plus kit (Amersham).

Chemical treatments and degradation assays

MYC-SKP2AOE or MYC-GFPOE seedlings were grown on solid MS medium for 6 days and then treated with or without 100 μm cyclohexamide (chx) in MS liquid medium. When indicated, chx treatment was combined with the addition of proteasome inhibitors (100 μm of MG132) or with 1 μm of 2,4-D (2,4-Dichlorophenoxyacetic acid). Afterwards, total protein was extracted and analyzed by immunoblotting using an antibody against MYC.

To analyze the stability of the MYC-SKP2A protein, total protein from 6-day-old MYC-SKP2AOE plants was extracted, and the degradation was analyzed as described by del Pozo et al. (2004, 2006). When indicated, proteasome inhibitor (100 μm of MG132) or the solvent DMSO as control was added to the reactions. The reactions were stopped by adding SDS-Laemmli sample buffer and boiling the sample for 5 min. The MYC-SKP2A protein was detected by immunobloting using the anti-MYC antibody.

MYC-GFPOE and MYC-SKP2AOE plants were grown in MS liquid for 6 days and then treated with terfestatin-A (Yamazoe et al., 2005) or the DMSO solvent for 32 h. The levels of MYC-tagged proteins were analyzed by immunoblotting with antibodies against MYC.

In vitro translation

The cDNA encoding the SKP2A gene was cloned into the pBS vector (Stratagene, http://www.stratagene.com) under the control of the T7 promoter. The GFP protein was cloned into pDONOR221 using gateway technology (Invitrogen). Later, the GFP was cloned into the Gateway binary vector Destiny17 by LR recombination (Invitrogen), to create a plasmid that incorporated a histidine tag to the GFP, which increases its molecular weight. A 1-μg sample of each plasmid was used for in vitro translation using the Wheat germ system (Promega, http://www.promega.com), in the presence of [35S]methionine or [9,10(n)-3H]myristic acid (GE Healthcare). The reactions were incubated at 30°C for 2 h. The translated proteins were separated by SDS-PAGE, the gel was dried and the radioactive signal was detected by autoradiography.

Lateral root primordia and root-length analyses

To quantify the number of LRP, we used the CYCB1-GUS (Colón-Carmona et al., 1999) marker and the CYCB1-GUS/MYC-SKP2AOE obtained by genetic cross. Both lines were grown on MS for 5 days and then stained for GUS activity in duplicated independent experiments. Lateral root primordia were counted as GUS-stained spots (those that did not break the epidermal cell layer) in about 40 seedlings using a Leica stereomicroscope MZ9,5 (Leica, http://www.leica-microsystems.com). The primary root length of the seedlings was measured using NIH image 1.61. The number of LRP was divided by the length of the primary root. The result shown in Figure 7(a) corresponds to the average and the standard deviation of all the measurements. To analyze the number of LRP in the slr1 or the slr1//MYC-SKP2AOE mutants, we grew the seedlings for 8 days in MS medium. Afterwards, the seedlings were treated as described by Malamy and Benfey (1997), and then stained with propidium iodide (0.1 μg ml−1) in PBS, and the LRP were counted in a fluorescence microscope.

Flow cytometry analyses

Plants were grown in MS medium for 14 days, and then the first two true leaves were used for ploidy measurements. Ploidy measurements were performed as described by del Pozo et al. (2004, 2006).

GUS assays

GUS staining was carried out as described by del Pozo et al. (2004, 2006).

Immunolocalization assays

Wild-type and MYC-SKP2AOE seedlings were grown vertically for 7 days. Immunolocalizations were performed using the KNOLLE antibody, as described in Lauber et al. (1997). The pictures shown are representatives of four independent experiments, and correspond to the merged stack of 10 slices of 2-μm thick.

Acknowledgements

The authors are indebted to N. Cisneros and V. Fernandez for technical assistance. We also want to thank P. Doerner for the CYCB1-GUS line, W. Crosby and E. Risseeuw for the ASK1 antibody, M. Estelle for the CUL1 antobody, G. Jürgens for the KNOLLE antibody and D. Nozaki for the terfestatin-A compound. This work has been supported by grants BMC2001-2292 and BIO2004-01749 (Spanish Ministery of Science and Technology) to JCP, and by grant BFU2006-05662 to CG. The institutional support of Fundación Ramón Areces to the CBM is also acknowledged.

Ancillary

Advertisement