Title: Identification of XDRP1; a Xenopus protein related to yeast Dsk2p binds to the N-terminus of cyclin A and inhibits its degradation
Abstract: Article15 September 1999free access Identification of XDRP1; a Xenopus protein related to yeast Dsk2p binds to the N-terminus of cyclin A and inhibits its degradation Minoru Funakoshi Minoru Funakoshi Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Stephan Geley Stephan Geley ICRF Clare Hall, Herts, EN6 3LD UK Search for more papers by this author Tim Hunt Tim Hunt ICRF Clare Hall, Herts, EN6 3LD UK Search for more papers by this author Takeharu Nishimoto Takeharu Nishimoto Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Hideki Kobayashi Corresponding Author Hideki Kobayashi Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Minoru Funakoshi Minoru Funakoshi Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Stephan Geley Stephan Geley ICRF Clare Hall, Herts, EN6 3LD UK Search for more papers by this author Tim Hunt Tim Hunt ICRF Clare Hall, Herts, EN6 3LD UK Search for more papers by this author Takeharu Nishimoto Takeharu Nishimoto Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Hideki Kobayashi Corresponding Author Hideki Kobayashi Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan Search for more papers by this author Author Information Minoru Funakoshi1, Stephan Geley2, Tim Hunt2, Takeharu Nishimoto1 and Hideki Kobayashi 1 1Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka, 812-8582 Japan 2ICRF Clare Hall, Herts, EN6 3LD UK *Corresponding author. E-mail: [email protected] The EMBO Journal (1999)18:5009-5018https://doi.org/10.1093/emboj/18.18.5009 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Using the N-terminus of cyclin A1 in a two-hybrid screen as a bait, we identified a Xenopus protein, XDRP1, that contains a ubiquitin-like domain in its N-terminus and shows significant homology in its C-terminal 50 residues to Saccharomyces cerevisiae Dsk2 and Schizosaccharomyces pombe dph1. XDRP1 is a nuclear phosphoprotein in Xenopus cells, and its phosphorylation is mediated by cyclin A-dependent kinase. XDRP1 binds to both embryonic and somatic forms of cyclin A (A1 and A2) in Xenopus cells, but not to B-type cyclins. The N-terminal ubiquitin-like domain of XDRP1, but not the C-terminal Dsk2-like domain, is required for interaction with cyclin A. XDRP1 requires residues 130–160 of cyclin A1 for efficient binding, which do not include the destruction box of cyclin A. The addition of bacterially expressed XDRP1 protein to frog egg extract inhibited the Ca2+-induced degradation of cyclin A, but not that of cyclin B. The injection of XDRP1 protein into fertilized Xenopus eggs blocked embryonic cell division. Introduction Proteolysis plays an essential role in the periodic events of cell-cycle progression through the irreversibility caused by the degradation of cyclin and other proteins, such as Pds1/Cut2 (see review by King et al., 1996a). Mitotic cyclins are rapidly degraded at the exit of mitosis and their degradation is necessary to enter the next cell cycle. The degradation of mitotic cyclins is mediated by the ubiquitin–proteasome machinery (Sudakin et al., 1995; Hochstrasser, 1996; King et al., 1996b), and this process requires an intact destruction box; a conserved nine-residue motif in their N-terminus (Glotzer et al., 1991). Several other proteins involved in the regulation of mitotic events, such as chromosome segregation and spindle formation, possess a similar motif, and their degradation is also controlled by the ubiquitin-proteolytic machinery, activation of which is determined differently at least in part by members of the fizzy family of proteins including Cdc20/slp1 and Cdh1/Hct1 (Cohen-Fix et al., 1996; Funabiki et al., 1996; Juang et al., 1997; Sigrist and Lehner, 1997). These proteins activate the selective ubiquitin ligase E3, known as the anaphase-promoting complex (APC) or cyclosome (King et al., 1995; Sudakin et al., 1995; Schwab et al., 1997; Visintin et al., 1997). The initiation of cyclin B degradation is also triggered by a direct cleavage of the N-terminal domain of cyclin by the 26S proteasome (Tokumoto et al., 1997). During the rapid cleavage cell cycles of early embryos, mitotic cyclins are destroyed only during a narrow time window (Hunt et al., 1992; Sudakin et al., 1995). Moreover, cyclins A and B are destroyed at slightly different times during mitosis (Lehner and O'Farrell, 1989; Minshull et al., 1990; Pines and Hunter, 1990; Hunt et al., 1992), with cyclin A being degraded prior to cyclin B, and the cyclin A-associated kinase activity increases and decreases earlier in the cell cycle than the cyclin B-associated kinase activity (Minshull et al., 1990), suggesting that functional differences exist between cyclin A and cyclin B at distinct times of mitosis (Luca et al., 1991). Mitotic cyclins are characterized by two conserved motifs in the amino acid sequence; the destruction box and the cyclin box. The degradation of mitotic cyclins requires the destruction box of the N-terminus for its recognition by the ubiquitin-mediated proteolytic machinery (Glotzer et al., 1991; Kobayashi et al., 1992; Lorca et al., 1992). Apart from the destruction box, the N-termini of mitotic cyclins contain other conserved motifs upstream of the start of the conserved cyclin box, which is required for association with Cdks. For example, the N-terminus of cyclin B1 contains a nuclear export signal LCQAFS, and both cyclins A and B contain autophosphorylation sites (Pines and Hunter, 1994; Klotzbucher et al., 1996; Hagting et al., 1998; Toyoshima et al., 1998). Cyclin A contains a conserved motif, FxxVDE, between the destruction box and the start of cyclin box, which has recently been shown to be a cleavage site of ICE-like caspase for apoptotic proteolysis of cyclin (Stewart et al., 1994; Stack and Newport, 1997). These imply that the conserved motifs in the N-terminus of mitotic cyclins between the stretch of the destruction box and the cyclin box could specify the distinct unknown functions of mitotic cyclins A and B during mitosis. To investigate the role(s) of these motifs with the aim of understanding the selective timing of degradation of cyclins A and B, we performed a yeast two-hybrid screen to isolate the cyclin A-specific interacting factors using the N-terminal domain of cyclin A as a bait. The major interacting species, XDRP1, was a Xenopus Dsk2-related protein in yeast that appears to be involved in the spindle checkpoint pathway. XDRP1 interacts with A-type cyclins, but not with cyclin B. Surprisingly, high levels of XDRP1 inhibited the degradation of cyclin A, but not that of cyclin B, and blocked cleavage cell division of embryos. Results Xenopus XDRP1 is a novel N-terminal ubiquitin-like protein To search for factors that interacted with conserved domains in the N-terminus of cyclin A, we used the N-terminal 160 residues of Xenopus cyclin A1 fused to the Gal4 DNA-binding domain as a bait in a two-hybrid screen. Twenty out of 41 strongly interacting clones were found to be identical in sequence, containing a 2.3 kb insert of a full-length cDNA clone (Figure 1). The putative ORF of this clone, Xenopus Dsk2-related protein 1 (XDRP1), was composed of 585 residues with a calculated molecular mass of 62 kDa. A search of the databases revealed that the N-terminus of XDRP1 had strong homology to residues 39–70 of ubiquitin (53% identity), and that the C-terminus showed a high similarity to the C-terminus of budding yeast Dsk2 (Biggins et al., 1996) and fission yeast dph1 (He et al., 1998) (42 and 51% identity, respectively, in the 47 C-terminal residues), although the yeast proteins are smaller (39 and 36 kDa) (Figure 1). Nematode, mouse and human homologues were identified in the databases, although no full-length clones have yet been reported. The middle of the protein contains a number of short conserved repeated sequences, in which the sequence MXNPD/E/Q appeared, that were also observed in the yeast proteins (see Figure 1), and the sequences around residues 360 and 530 were unique to this protein. An alternative spliced form of XDRP1 consisting of 600 amino acid residues are isolated (data not shown), where 15 amino acid residues, EEIVIGETACVLEFK, are inserted into the ubiquitin-like domain, between amino acids 33 and 34 (Figure 1, indicated by a dot in the sequence). Figure 1.Deduced amino acid sequence of Xenopus XDRP1. The deduced amino acid sequence of ORF of XDRP1 was shown, and its numbers are indicated at the right of the sequence. The N-terminal ubiquitin-like sequences (7–78) are shaded and seven identical residues with ubiquitin and four XDRP1-related proteins are shown by letters in a black rectangle. Similarly, the C-terminal yeast Dsk2p-like sequence (539–585) and identical residues are shown. Seven conserved repeats in the middle (171–240) and three conserved repeats (around 350–450) are underlined. The M-NPX sequences scattered are also shaded. The structures of frog XDRP1, budding yeast Dsk2, fission yeast dph1 and Nematoda F15C11.2 are schematically shown on the right of the sequence. Download figure Download PowerPoint XDRP1 persists in Xenopus cells as a nuclear phosphoprotein We raised polyclonal antibodies against bacterially expressed XDRP1. Figure 2A shows that the antiserum against XDRP1 easily detected a single polypeptide with an apparent molecular weight of ∼68 kDa by immunoblotting Xenopus egg extracts. The concentration of XDRP1 in eggs was estimated as 0.5–1 ng/egg, ∼8–16 nM (data not shown), using immunoblotting with bacterially expressed GST–XDRP1 protein as a standard. XDRP1 protein was detected as a faster-migrating band through the early stages of oogenesis (Figure 2A). Two bands were detected in Xenopus embryos and cultured cells (lanes 3–5). The slower-migrating band detected in eggs (Figure 2B, lane 1) was converted to the faster-migrating one by treatment with lambda phosphatase (lane 3) and phosphatase inhibitors blocked this band shift (lane 4), from which we conclude that the upper XDRP1 band was a phosphorylated form of the lower one. A non- or hypo-phosphorylated form of XDRP1 was present in immature oocytes (Figure 2A, lane 1) and phosphorylated forms appeared during oocyte maturation (lane 2). At later stages of development and in cultured cells, both forms were detected (lanes 3–5). Dephosphorylated form appeared at 40–50 min after activating egg to drive the cell cycle (Figure 2C), although no typical cell-cycle pattern of phosphorylation was observed at its protein level. When 35S-labelled XDRP1 in egg extracts was immunoprecipitated with anti-XDRP1 antibody, however, phosphorylated form was detected mainly at CSF-arrested extracts and non-phosphorylated form was detected in interphase (Figure 2D). Therefore, XDRP1 is phosphorylated at M phase, but not in interphase. Figure 2.XDRP1 is a phosphoprotein. (A) Immunoblotting of XDRP1 during Xenopus development. Xenopus oocytes, eggs, embryos (cleavage and blastula) and kidney culture cells (WAK) were taken and the samples were immunoblotted with anti-XDRP1 antibody. (B) XDRP1 is phosphorylated. XDRP1 was detected by immunoblotting with anti-XDRP1 antibody. Lane 1, CSF-arrested extracts; lane 2, WAK cells; lane 3, CSF-arrested extracts treated with λ phosphatase; lane 4, CSF-arrested extracts with λ phosphatase and phosphatase inhibitors; lane 5, CSF-arrested extracts diluted 7-fold with the buffer. (C) XDRP1 abundance during the cell cycle. Xenopus eggs activated by calcium ionophore were sampled at the times indicated. XDRP1 was detected by immunoblotting with anti-XDRP1 antibody. The M phase is indicated as an asterisk (*) at the bottom, where the H1 kinase activity is high (data not shown). (D) Phosphoryla- tion of XDRP1 at M phase. Xenopus CSF-arrested extracts or interphase extracts were labelled for 60 min with 1 μCi/μl of [35S]methionine in the presence of cyclin A1 mRNA. 35S-labelled XDRP1 was immunoprecipitated from Xenopus egg extracts with anti-XDRP1 antibody, and analyzed by SDS–PAGE and autoradiography. M, Xenopus CSF-arrested extracts (lanes 1–2); and Int, interphase extracts (lanes 3–4). PI, preimmune serum (lanes 1 and 3); and I, immune serum (lanes 2 and 4). Control, the precipitates with anti-cyclin A1 antibody K3 (lane 5). Note a weak band in lane 2 (arrowhead). Download figure Download PowerPoint To investigate the subcellular localization of XDRP1, cultured Xenopus cells were stained using anti-XDRP1 antiserum. XDRP1 protein was predominantly localized to the nucleus in interphase cells (Figure 3A), while in mitotic cells, no particular subcellular location was observed. XDRP1 was detected in both nuclear and cytoplasmic fractions of cell lysates (Figure 3B). The nuclear form had a lower mobility than the cytoplasmic one (lane 1), thus suggesting that the phosphorylated form of XDRP1 is located in the nucleus. Figure 3.XDRP1 locates at the nucleus. (A) Immunofluorescence of XDRP1. Asynchronous Xenopus cultured WAK cells were immuno-stained with anti-XDRP1 antibody (left columns). The same field was observed by counterstaining with DAPI for nuclear DNA (middle columns) and by a microscope to demonstrate the cell morphology (right columns). The asterisk indicates a mitotic cell. Bar, 20 μm. (B) Subcellular fractionation. WAK cells were fractionated into the cytoplasmic and the nuclear fractions as described in the Materials and methods. Three micrograms of total proteins in each fraction were applied on the gel and were co-immunoblotted with both anti-XDRP1 antibody and anti-RCC1 antibody. Nuclear protein RCC1 was used as a control to check the fractionation. Lane 1, nuclear fraction; lane 2, cytosol; lane 3, total cell fraction. Download figure Download PowerPoint XDRP1 binds to full-length cyclin A1 and A2, but not to B-type cyclins We tested binding of XDRP1 with cyclin A by growth and β-galactosidase in the yeast two-hybrid system, since XDRP1 was originally isolated by yeast two-hybrid screen using the N-terminal domain of cyclin A1 as bait (positions 1–160). Both assays showed XDRP1 to bind to the N-terminal domain of cyclin A1 (Figure 4A). Figure 4.XDRP1 binds to cyclin A, but not to cyclin B. (A) Binding of XDRP1 to cyclin A1 in yeast two-hybrid system. The N-terminal domain of Xenopus cyclin A1 (1–160) cloned into PAS404 was expressed in yeast and their histidine-prototrophic growth (left) and increased β-galactosidase activity (right) was tested. (B) Binding of XDRP1 to cyclin A1 in Xenopus extracts in vitro. 35S-labelled Xenopus cyclin A1 was precipitated with GST (lanes 1 and 2), Xenopus GST–CDK2 (lanes 3 and 4) and GST–XDRP1 (lanes 5 and 6) in the CSF/Retic-mixed extracts, followed by SDS–PAGE and autoradiography. Top panel, translation; bottom panel, GST-bound cyclin A1. (C) Binding of XDRP1 to other A-type cyclins. Xenopus cyclins A1 and A2 and human cyclin A2 were 35S-labelled in the extracts and their binding abilities to XDRP1 were tested by GST–XDRP1 pull-down. Exposure time, overnight. Translation (top panel) and GST–XDRP1 precipitates (bottom panel). (D) Binding of XDRP1 to cyclin A2 in transfected cells. T7-His6-tagged XDRP1 were precipitated from the WAK cell extracts as described in the Materials and methods. The precipitates were immunoblotted with either anti-T7 antibody (lanes 5–8, top panel) or anti-HA antibody (lanes 5–8, bottom panel). Expression was checked by immunoblotting of the extracts (15–30 μl) with anti-T7 antibody for XDRP1 (lanes 1–4, top) or anti-HA antibody for cyclin A2 (lanes 1–4, bottom). (E) Binding of XDRP1 to cyclin A, but not cyclin B. Xenopus cyclin A1, cyclin B1 and cyclin B2 mRNAs were translated in the CSF/Retic-mixed extracts and assayed for XDRP1 binding. Left panel, translation; right panel, GSH–Sepharose beads bound. Exposure time, 2 days. Download figure Download PowerPoint Next we tested the binding of cyclin A1 to XDRP1 in Xenopus egg extracts. The full-length cyclin A1 mRNA was translated in the presence of [35S]methionine in frog egg extracts, as described in Materials and methods. Recombinant GST–XDRP1 was then added, followed by affinity chromatography on GSH–Sepharose to isolate the proteins that bound to the bacterially expressed GST–XDRP1. Figure 4B shows cyclin A1 was retained when GST–XDRP1 was added (Figure 4B, lane 6), and not with GST alone (lane 2). The binding of cyclin A to GST–XDRP1 was significantly lower than the binding of cyclin A to GST–CDK2, which acted as a positive control in this experiment (Figure 4B, lane 4). GST–XDRP1 also bound to Xenopus and human cyclin A2, the form of cyclin A found in somatic cells (Figure 4C, lanes 5 and 6). We also checked the binding of XDRP1 to cyclin A1 in yeast cells, where a full-length Xenopus cyclin A1 overproduced from a yeast-inducible vector could be isolated by addition of GST–XDRP1 to the lysate followed by affinity chromatography on GSH–Sepharose and immunoblotting with anti-cyclin A1 antibody (data not shown). Next we tested interactions between XDRP1 and cyclin A in Xenopus cells. Constructs of T7-His-tagged XDRP1 and HA-tagged cyclin A2 were transfected into WAK cells and both proteins were expressed in vivo (Figure 4D, lane 4). Then, T7-His-XDRP1 proteins were precipitated with TALON metal affinity resin, and XDRP1-bound materials were immunoblotted with anti-HA antibody against HA–cyclin A2. Figure 4D shows that XDRP1 interacted with cyclin A2 in cells (lane 8). When immunoprecipitation using anti-XDRP1 antibody was carried out after translating cyclin A1 mRNA in the Xenopus egg extracts, labelled cyclin A1 was weakly co-precipitated with anti-XDRP1 antibody (see arrowhead in Figure 2D). It proved difficult to demonstrate interactions between them in Xenopus cell extracts using immunoprecipitation with either anti-XDRP1 or anti-cyclin A1 antibodies, followed by immunoblotting. It was clear that the interaction between XDRP1 and cyclin A was weaker than that between CDK2 and cyclin A (see Figure 4B). Next, to test whether XDRP1 bound to B-type cyclins, cyclins A and B were labelled with [35S]methionine in the egg extracts and tested for their ability to bind to GST–XDRP1. XDRP1 bound to Xenopus cyclin A1 (Figure 4E, lane 5) in the frog egg extract, but not to cyclin B1 (lane 6) or cyclin B2 (lane 7). Even when both cyclins A and B were co-translated, neither cyclin B1 nor cyclin B2 bound to GST–XDRP1 (Figure 4E, lanes 8 and 9). Therefore, XDRP1 appears to bind to cyclin A specifically and not cyclin B. The same results were obtained by two-hybrid assay (both histidine-prototrophic growth and increased β-galactosidase activity) using the equivalent peptides of the N-terminal cyclin B (positions 1–132 in cyclin B1 and 1–127 in cyclin B2) (data not shown). XDRP1 binds to the cyclin A–CDK complex and XDRP1 phosphorylation depends upon cyclin A-dependent kinase To test whether XDRP1 could bind to cyclin A–CDK complexes, Cdc2 was labelled with [35S]methionine in egg extracts, and affinity chromatography with GST–XDRP1 was performed as before. Both cyclin A1 and Cdc2 were retained on GSH–Sepharose in the presence of GST–XDRP1 (Figure 5A, lane 6). A similar result was obtained for CDK2 (data not shown). To confirm that XDRP1 actually binds to the cyclin–CDK complex, we examined whether the GST–XDRP1 displayed histone H1 kinase activity. Figure 5B shows that the GSH–Sepharose beads carrying GST–XDRP1 possessed H1 kinase activity (Figure 5B, lane 4) only when cyclin A1 mRNA was translated in the extract (lanes 3 and 4). GST alone did not retain cyclin A or significant kinase activity (lanes 1 and 2). Figure 5.XDRP1 is phosphorylated by cyclin A-dependent kinase. (A) Co-precipitation of cyclin A1 and Cdc2 with XDRP1. Cyclin A1 and Cdc2 mRNAs were co-translated in CSF/Retic-mixed extracts with [35S]methionine, and the precipitates with GST–XDRP1 were subjected to autoradiography. (B) H1 kinase activity of the GST–XDRP1 precipitates. Cyclin A1 mRNA was translated in CSF/Retic extracts and the precipitates with GST–XDRP1 was assayed for histone H1 kinase activity. Expression of cyclin A1 (top panel), cyclin A1 bound to GST–XDRP1 (middle panel) and histone H1 kinase activity (bottom panel). (C) Phosphorylation of XDRP1 depends on cyclin A. GST–XDRP1 was incubated in cyclin A1-expressed CSF/Retic-mixed extracts and detected by immunoblotting with anti-XDRP1 antibody as described in the Materials and methods (top panel). The expression of cyclin A1 were checked by immunoblotting with anti-cyclin A1 antibody (bottom panel). (D) Phosphorylation of XDRP1 is blocked by butyrolactone I. GST–XDRP1 was incubated with different concentrations of butyrolactone I; 0 nM (lane 2), 50 nM (lane 3) and 200 nM (lane 4). GST–XDRP1 in the absence of cyclin A1 (lane 1). For details see Materials and methods. Download figure Download PowerPoint Since XDRP1 binds to the cyclin A–CDK complex, whether XDRP1 phosphorylation is mediated by cyclin A-dependent kinase or not was tested (Figure 5C). GST–XDRP1 was phosphorylated when cyclin A1 was translated in egg extracts (Figure 5C, lane 2), where CDK activity was elevated by cyclin A1 expression (data not shown). Increasing amounts of butyrolactone I added to the extracts inhibited phosphorylation of GST–XDRP1 (Figure 5D, lanes 3 and 4). GST–XDRP1 was directly phosphorylated by the immuno-purified cyclin A–cdc2 complex in vitro (data not shown). Therefore, cyclin A-dependent kinase contributes to phosphorylation of XDRP1. The ubiquitin-like domain of XDRP1 is necessary but not sufficient for interaction with cyclin A We tested which domains of XDRP1 were required for the interaction with cyclin A. N- and C-terminal deletions were made and their binding abilities for cyclin A1 were tested by two-hybrid assay (Figure 6A). Deletions of the ubiquitin-like domain in the N-terminus (76–585) abolished its binding with cyclin A1, but that of the C-terminal conserved Dsk2-like domain did not (1–538), showing a requirement for the N-terminal ubiquitin-like domain for XDRP1 to interact with cyclin A. The non-conserved middle region of XDRP1 was also required for its binding to cyclin A1, however, because the N-terminus alone showed no interaction. Figure 6.Analysis of motifs necessary for interaction between XDRP1 and cyclin A. (A) Abilities of cyclin A1 binding to XDRP1 mutants. Deletion mutants of XDRP1 (Δ1–76, Δ1–538, Δ76–585 and Δ538–585) were cloned into pGAD10, were expressed in yeast, and their histidine-prototrophic growth (left) and increased β-galactosidase activity (right) was tested in yeast two-hybrid system. (B) Abilities of XDRP1 binding to cyclin A1 mutants. The mutants of Xenopus cyclin A1 (1–160, destruction box mutants ATVA, Δ88–144, FxxVDE, 134–418 and 161–418) were translated and the binding to GST–XDRP1 was tested in the CSF/Retic-mixed extracts. A diagram of the constructs and their binding abilities were shown on the bottom. 1–160 with GST was used as a negative control (lane 1). Exposure time, 3 days. Note the slow migration of 1–160 (lane 2), probably due to the misfolding by the conformational change of the N-terminal fragment. (C) Putative binding motif of XDRP1 in cyclin A1. The sequences with high homology in the alignment of A-type cyclins among frog, human, mouse and chicken are shown in shaded boxes. Identical residues are indicated by an asterisk and conserved ones are shown by a double dot at the bottom. A schematic drawing of the conserved motifs in cyclin A protein is shown at the bottom. A bar at the top of alignment indicates the predicted binding region obtained by the result of Figure 6B. Download figure Download PowerPoint Residues surrounding the major phosphorylation site in cyclin A are necessary for interaction with XDRP1 To define the region of cyclin A1 that was required for XDRP1 binding, we prepared a number of mutants and synthesized the corresponding 35S-labelled proteins in egg extracts (Figure 6B, upper panel) followed by affinity chromatography using GST–XDRP1 (Figure 6B, lower panel). The N-terminal fragment (1–160), which was originally used as a bait for two-hybrid screening, bound to XDRP1 as expected, but its binding was rather poor in comparison with full-length cyclin A1 (Figure 6B, lanes 2 and 3). No signal was detected with GST alone (lane 1). The construct comprising residues 133–418 of cyclin A1 bound to XDRP1 as well as the full-length protein (lanes 3 and 7), whereas residues 161–418 did not bind at all (lane 8). The construct with an internal deletion, Δ88–144, bound to XDRP1 with slightly reduced efficiency (lane 5) compared with full-length cyclin A1 (lane 3). These results suggest that the residues between 133 and 160 of cyclin A1 are required for XDRP1 binding. A destruction-box mutant ATVA bound to XDRP1 normally in egg extracts (lane 4). Therefore, the destruction box, which is necessary for the ubiquitin-mediated proteolysis of mitotic cyclins, was not required for XDRP1 binding. The sequence around 133–160 contains a motif conserved among A-type cyclins, in which phosphorylation site SP at 136/137 surrounded by other partially conserved residues around 130–145 is reasonably well conserved (Figure 6C). It lies just upstream of the N-terminal helix that interacts with CDK2. High levels of XDRP1 selectively inhibit the degradation of cyclin A and block the cell cycle in Xenopus embryo To test if any function in relation to cyclin destruction could be detected, we investigated the effects of XDRP1 on cyclin degradation in Xenopus egg extracts. Cyclin A1 and/or B2 was first labelled with [35S]methionine in a 1:1 mixture of reticulocyte lysate and frog egg extract, and this labelled substrate was then added to fresh egg extract, to which CaCl2 was added to trigger the exit from M phase. The degradation of cyclin A1 was strongly inhibited by the addition of GST–XDRP1, but that of cyclin B was not (Figure 7A, panel b). However, cyclin A1 or cyclin B2 proteolysis was not inhibited in extracts to which GST or XB buffer alone was added (panel a). Because cyclin B was degraded normally in the presence of XDRP1, XDRP1 addition to the extracts did not affect the ubiquitin-dependent proteolytic machinery itself. Since the egg extracts contain endogenous XDRP1, we next analysed the effect of XDRP1 immunodepletion in the extracts. As shown in Figure 7A, panel c, cyclins A and B were both degraded normally in the XDRP1-immunodepleted extracts. In this experiment, >95% of endogenous XDRP1 was depleted, when checked by immunoblotting (data not shown). Figure 7.GST–XDRP1 selectively inhibits degradation of cyclin A, and blocked the cleavage cell division in Xenopus embryo. (A) Selective inhibition of degradation of cyclin A, but not that of cyclin B. Degradation assays were carried out in Xenopus CSF-arrested extracts as described in Materials and methods. Cyclin degradation was induced by adding 0.4 mM calcium to the extracts. In this experiment, 100 ng/μl of purified GST–XDRP1 was added to the extracts. The addition of GST alone (100 ng/μl) had no inhibitory effect on cyclin destruction. Panel a, XB buffer-added extracts; panel b, GST–XDRP1-added extracts; panel c, XDRP1-depleted extracts. (B) Microinjected GST–XDRP1 blocked embryonic cell division in Xenopus. GST–XDRP1 (60 ng/embryo) was injected into the fertilized eggs as described in Materials and methods. GST was injected as a control (60 ng/embryo). The eggs injected the lower concentration of GST-XDRP1 (30 ng/embryo) cleaved slowly, and as a result, an injected embryo induced incomplete gastrulation (data not shown). Download figure Download PowerPoint To examine the effect of elevated levels of GST–XDRP1 on the cell cycle in Xenopus embryos, we injected 60 ng of GST–XDRP1 into fertilized eggs and observed its effects on cleavage (Figure 7B). The injection of GST–XDRP1 into one-cell stage embryo blocked cell division. The injection of similar amounts of GST alone had no obvious effects on the cleavage of embryos. The percentage showing blocking of the cell cycle was 93% (14/15 embryos) in GST–XDRP1 and 11% (2/18 embryos) in GST alone. Therefore, the increasing GST–XDRP1 protein appears to induce an inhibitory effect on the cell division in cleaving embryos, probably because of failure to degrade cyclin A. Discussion In this paper we identified a Xenopus protein XDRP1 that shows homology to a region of ubiquitin in its N-terminus and to the yeast proteins Dsk2 and dph1 in its C-terminus. XDRP1 was identified by virtue of its binding to a site found in the N-terminus of several kinds of cyclin A, and does not bind to B-type cyclins. XDRP1 is a nuclear phosphoprotein in Xenopus somatic cell cultures. Elevated levels of GST–XDRP1 inhibited degradation of cyclin A, but not that of cyclin B, and also blocked cell division in cleaving Xenopus embryos. Xenopus XDRP1 contains a ubiquitin-like domain in the N-terminus and shares the C-terminal conserved domains with yeast Dsk2 XDRP1