- Email: [email protected]
β3-Endonexin as a Novel Inhibitor of Cyclin A-Associated Kinase
Biochemical and Biophysical Research Communications 267, 947–952 (2000) doi:10.1006/bbrc.1999.2007, available online at http://www.idealibrary.com on
3-Endonexin as a Novel Inhibitor of Cyclin A-Associated Kinase Akihira Ohtoshi,* Tatsuya Maeda,* Hideaki Higashi,* Satoshi Ashizawa,* Masafumi Yamada,* and Masanori Hatakeyama* ,† ,1 *Department of Viral Oncology, Cancer Institute, Japanese Foundation for Cancer Research, Tokyo 170-8455, Japan; and †Research Section of Chemistry, Institute of Immunological Science, Hokkaido University, Sapporo 060-0815, Japan
Received December 2, 1999
Cyclin A is indispensable for S phase cell cycle progression and is suggested to be a crucial target of cell adhesion signals. In this study, we demonstrate that 3-endonexin, a molecule known to associate with the integrin 3 cytoplasmic domain, specifically binds cyclin A. Deletion of the amino-terminal 52-amino-acid residues including the cyclin-binding RxL motif abolishes the ability of 3-endonexin to interact with cyclin A. In an in vitro kinase assay, 3-endonexin inhibits pRB kinase activity associated with cyclin A-Cdk2 while leaving its histone H1 kinase activity unaffected. Coexpression of 3-endonexin in yeast cells overcomes growth suppression caused by an activation of cyclin A-associated kinase. Our results indicate that 3-endonexin is a novel cyclin A-binding molecule that regulates cyclin A-associated pRB kinase activity. © 2000 Academic Press Key Words: 3-endonexin; cyclin A; Cdk inhibitor; pRB; cell adhesion.
Sequential activation of various combinations of cyclins and cyclin-dependent kinases (Cdks) provides a molecular framework for coordinated cell cycle progression (1). In mammalian cells, growth stimulation leads to elevated expression of D-type cyclins and activates cyclin D-Cdk4/6 complex to phosphorylate the pRB (2– 4). The subsequent activation of cyclin E-Cdk2 in late G1 further hyperphosphorylates and extensively inactivates the pRB, thereby relieving pRBcontrolled E2F transcriptional activities that are essentially required for G1 to S phase transition (1). In the M phase of the mitotic cell cycle, a distinct cyclinCdk complex, cyclin B-Cdc2, works as the driving force of cell cycle progression leading to mitosis. Cyclin A, in
contrast with other cyclins, functions twice in each cell cycle by forming a complex with Cdk2 in the late G1 and S phases and with Cdc2 in G2 and M phases (5–7). Cyclin A is transcriptionally up-regulated in late G1 following cyclin E-Cdk activation and accumulates in S phase. A number of studies suggest that cyclin A-Cdk2 is involved in controlling DNA replication. Indeed, whereas cyclin A is present in both nucleus and cytoplasm, it accumulates at sites of DNA replication during S phase, and in a cell-free system cyclin A is capable of inducing DNA replication (8 –10). In addition, recent studies have implicated a unique connection between cyclin A and adhesion-dependent cell proliferation (11–15). Cells arrested by suspension fail to express cyclin A (14). Conversely, cells that ectopically express cyclin A become anchorage-independent for growth (12). These observations indicate that cell adhesion signals eventually regulate the activity of cyclin A. Cell adhesions such as anchorage of cells to the extracellular matrix are mediated by a large family of heterodimeric cell surface adhesion receptors termed integrins. They are composed of ␣ and  subunits, and evidence indicates that cytoplasmic domains of the integrin  subunits are responsible for the generation of intracellular signals following interaction with the extracellular matrix (11). However, molecular mechanisms that couple the adhesion signals and cyclin A remain to be elucidated. In this work, we demonstrate that 3-endonexin (16), which was originally isolated as a molecule that specifically interacts with the cytoplasmic domain of integrin 3, binds to cyclin A and, through the interaction, inhibits cyclin A-associated kinase activity. Our results provide a molecular link between the cell adhesion system and the cell cycle machinery. MATERIALS AND METHODS
1
To whom correspondence should be addressed at the Department of Viral Oncology, Cancer Institute, Japanese Foundation for Cancer Research 1-37-1 Kami-Ikebukuro, Toshima-ku, Tokyo 170-8455, Japan. Fax: ⫹81-3-5394-3816. E-mail: [email protected].
Yeast strains, media, and plasmid DNAs. The following yeast strains were used: KSC9 (MATa cdc28-4 ura3⌬ns ade1 his2 leu23,112 trp1-1a), Y187 (MAT␣ ura3-52 his3-200 ade2-101 trp1-901
947
0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
leu2-3,112 gal4⌬ met ⫺ gal80⌬ GAL1 UAS-GAL1TATA-LacZ) (CLONTECH), and Y190 (MATa ura3-52 his3-200 ade2-101 lys2-801 trp1-901 leu23,112 gal4⌬ gal80⌬ cyh r2 LYS2::GAL1 UAS-HIS3 TATA-HIS3 URA3:: GAL1 UAS-GAL1 TATA-LacZ) (17). pAS2.1C-cycA is a single-copy vector derived from pAS2.1 and carries a fusion gene encoding GAL4 DNA binding domain-human cyclin A chimeric protein under ADH1 promoter. pAS2.1C-cycA⌬C20 was constructed by a sequential Exonuclease III and SI nuclease treatment. The plasmid expresses a truncated cyclin A, in which the carboxy-terminal 20 amino acid residues were replaced by a serine residue derived from the vector. To conditionally express human cyclin A in yeast, p416GAL1-cycA was made by inserting a human cyclin A cDNA downstream of the Gal1 promoter in a single-copy yeast expression vector, p416GAL1 (18). A multicopy vector, p424GAL1 (18), was used to express 3-endonexin in yeast. For two-hybrid screening, a cDNA library derived from a human WI-38 lung fibroblast cell (CLONTECH) was used. Purification of GST-fusion proteins in E. coli. A GST gene fusion vector, pGEX-5X-1, was used to express GST fusion proteins in an E. coli strain, BL21. Growing BL21 cells carrying the expression plasmids were exposed to 1 mM IPTG to induce GST-fusion protein production. The cells were then harvested, washed with phosphatebuffered saline (PBS), and resuspended in NETN Buffer (20 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% NP-40) supplemented with protease inhibitors (5 g/ml PMSF, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml trypsin inhibitor). The cell suspension was then subjected to sonication and the clear lysate was collected for further purification according to manufacturer’s instructions (Pharmacia Biotech). In vitro GST pull-down assay. 35S-methionine-labeled cyclin A, cyclin E or cyclin B1 was prepared by in vitro transcription and translation using TNT Coupled Reticulocyte Lysate Systems (Promega). 35S-methionine-labelled cyclin protein and GST-fusion protein attached Glutathione Sepharose 4B were then mixed in a 200 l of NETN Buffer containing protease inhibitors and 0.5% skim milk. After washing the beads with NETN buffer, one-fifth of the aqueous fraction and a half of beads fraction were resuspended in SDS sample buffer and electrophoresed to separate proteins. Preparation of COS cell lysates and immunoprecipitation. Mammalian expression vectors for carboxy-terminal triple hemagglutinin (HA)-tagged cyclin A (pSR␣-cycA-3xHA) and the amino-terminal 6xMyc-tagged 3-endonexin (pSR␣-6xMyc-3-endonexin) were transfected into COS cells using a DEAE-dextran method. Cell lysis and immunoprecipitation were performed as described previously (19). Immunoblot analysis. Yeast cells were harvested, washed with sterilized water, and resuspended in SDS sample buffer. Extracts or immunoprecipitates from COS cells were prepared as described above. Separated proteins by SDS-PAGE were electrotransferred to a PVDF transfer membrane (Millipore), incubated with the detection antibody, 9E10 for Myc-tagged proteins, 12CA5 for HA-tagged proteins, or anti-human cyclin A (sc-751, Santa Cruz). Proteins were visualized by using enhanced chemiluminescence reagents. In vitro kinase assay. Purified cyclin A-Cdk2 complex produced in insect cells using a baculovirus system was prepared as described previously (20). GST-fusion proteins were prepared as described above. Truncated pRB proteins for pRB kinase assay, Rb-C and Rb-N, were purchased from New England Biolabs. Kinase reaction was performed at 30°C for 10 min in kinase buffer (20 mM Tris-HCl, pH 7.4, 10 mM MgCl 2, 4.5 mM -mercaptoethanol, 1 mM EGTA). The reaction solution contained 50 M of cold ATP and 0.4 l of [␥- 32P]ATP (6000 Ci/mmol, Dupont).
-Galactosidase assay. Proliferating yeast cells in SD media were harvested, transferred to YPAD media and cultured at 30°C. The grown cells were resuspended in Z buffer, frozen in liquid nitrogen and thawed at room temperature. -mercaptoethanol and
O-nitrophenyl -D-galactopyranoside (ONPG) were added to a final concentration, 38.9 mM and 0.16 mg/ml, respectively, to measure -galactosidase activity.
RESULTS To identify molecules that specifically interact with cyclin A, a yeast two-hybrid screening was performed. Because high level expression of wild-type human cyclin A is toxic to yeast, we generated a series of cyclin A-deletion derivatives and one in which the carboxyterminal 20 amino acid residues of cyclin A were replaced by a serine residue was found to be non-toxic in Y190 yeast strain and yet complemented the growth defect of a triple cln ⫺ (cln1 ⫺ cln2 ⫺ cln3 ⫺) yeast strain (data not shown) (21). Using the truncated cyclin A as a bait, a human lung fibroblast library (1 ⫻ 10 6 independent yeast transformants) was screened and three positive clones were isolated. They respectively contained cDNAs fragment encoding p21 Cip1, p55CDC, and 3-endonexin. p21 Cip1 is a Cdk inhibitor (22–24) and p55CDC is a WD repeat protein which has a homology to the budding yeast Cdc20 protein and the Drosophila fizzy protein (25, 26). In this report, we focus on 3endonexin. 3-Endonexin was originally isolated as a molecule that specifically interacts with the cytoplasmic domain of the 3 subunit of integrin (16). There are long and short forms of 3-endonexin, which are translated from alternatively spliced mRNAs, and the one isolated in our screening was the long form. Physical interaction between 3-endonexin and cyclin A was first examined in yeast two hybrid -galactosidase assay (Fig. 1). p21 Cip1, which is known to bind cyclin A, was employed as a positive control. Fusion proteins consisting of GAL4 activation domain (AD) and p21 Cip1 or wild-type 3-endonexin induced transcription of -galactosidase reporter gene and scored positive -galactosidase activity in a prey dependent manner. The production of each GAL4-AD fusion protein was confirmed by anti-HA immunoblotting (Fig 1; the bottom panel). Similarly, comparable amounts of GAL4 DNA binding domain-cyclin A fusion proteins were detected in each yeast lysate (Fig 1; the middle panel). Hence, the result indicates that, like p21 Cip1, 3-endonexin specifically binds cyclin A. To examine a potential interaction between 3endonexin and cyclins other than cyclin A, in vitro pull-down assay was performed using GST-3endonexin and in vitro translated, radio-labelled cyclin A, cyclin E or cyclin B1. As shown in Fig. 2, a significant amount of the in vitro translated cyclin A was specifically bound to GST-3-endonexin. In contrast, no specific binding was observed between GST-3endonexin and cyclin E or cyclin B1. The observation indicates that 3-endonexin specifically interacts with cyclin A.
948
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Two-hybrid interaction between 3-endonexin and cyclin A. Y187 yeast cells harboring plasmids indicated were subjected to -galactosidase liquid assay. The values are shown in Miller’s unit. Middle and bottom panels show the expression levels of GAL4 DNA binding domain-cyclin A (cyclin A), GAL4 activation domain-p21 Cip1 (p21) and GAL4 activation domain-3-endonexin (3-endonexin), respectively, in the same samples used in the top panel.
To address physical interaction between 3endonexin and cyclin A in vivo, we constructed mammalian expression vectors for the long and short forms of Myc epitope-tagged 3-endonexin (Myc-3-endonexin) and hemagglutinin (HA) epitope-tagged cyclin A
FIG. 2. Specificity of interaction between 3-endonexin and cyclins. GST pull-down assay was performed using GST-p21 Cip1 (p21) or GST-3-endonexin (3-endonexin). In vitro translated, 35Smethionine-labeled cyclin A, E or B1 was incubated with the GSTfusion protein. One-fifth of supernatant fraction and a half volume of beads fraction were loaded on odd-numbered lanes marked “S” and even numbered lanes marked “B,” respectively.
FIG. 3. Complex formation between 3-endonexin and cyclin A in mammalian cells. (A) Total cell lysates (lanes 6 –10) and anti-Myc immunoprecipitates (lanes 1–5) were prepared from COS cells expressing Myc-tagged 3-endonexin (long form) (lanes 1, 3, 6 and 8), Myc-tagged 3-endonexin (short form) (lanes 4 and 9), Myc-tagged ⌬N52 (lanes 5, 10) together with HA-tagged cyclin A (HA-cyc A) (lanes 2–5, 7–12), separated by SDS-PAGE (lanes 1– 6) and were immunoblotted with anti-HA (upper panel) or anti-Myc (lower panel). (B) Total cell lysates (lanes 6 –10) and anti-HA immunoprecipitates (lanes 1–5) were prepared from COS cells expressing the long form of Myc-tagged 3-endonexin (lanes 1, 3, 6 and 8), the short form of Myc-tagged 3-endonexin (lanes 4 and 9), Myc-tagged ⌬N52 (lanes 5, 10) together with HA-tagged cyclin A (HA-cyc A) (lanes 2–5, 7–12), separated by SDS-PAGE (lanes 1– 6) and were immunoblotted with anti-Myc (upper panel) or anti-HA (lower panel).
(HA-cyclin A). They were transiently cotransfected into monkey COS cells. Lysates were then prepared from the transfected cells and subjected to sequential immunoprecipitation-immunoblotting analyses (Figs. 3A and 3B). When both Myc-3-endonexin and HAcyclin A were co-expressed in COS cells, anti-Myc or anti-HA immunoprecipitates contained both Myc-3endonexin and HA-cyclin A. In contrast, when either Myc-3-endonexin or HA-cyclin A was singly expressed in COS cells, anti-HA and anti-Myc failed to coprecipitate Myc-3-endonexin and HA-tagged cyclin A, respectively. These results indicate that 3-endonexin and cyclin A exist in a physical complex in mammalian cells. Both long and short forms of 3-endonexin possess an RxL motif (in the single letter code for amino acid, where x is any amino acid) at its amino-terminal region
949
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(residues 5 and 7). This motif is known to constitute cyclin docking sites in other cyclin binding molecules (27–31). To determine whether the sequences involving the RxL motif are essential for the interaction of 3endonexin with cyclin A, we generated a cDNA encoding a mutant molecule, ⌬N52, which lacks the aminoterminal 52 amino acid residues including the RxL motif from the long form of 3-endonexin. When expressed in COS cells, the ⌬N52 mutant did not bind cyclin A (Fig. 3). The result thus implicates that the amino-terminal region of 3-endonexin, which contains the RxL motif, is indispensable for the interaction of 3-endonexin with cyclin A. To pursue the effect of 3-endonexin on cyclin A-associated kinase activities, we next performed in vitro kinase assay using either histone H1 or the retinoblastoma protein (pRB) as a substrate in the presence of increased amounts of 3-endonexin. As shown in Fig. 4A, purified cyclin A-Cdk2 complex efficiently phosphorylated histone H1 (compare lane 1 to 2). This histone H1 kinase activity was inhibited by adding GST- p21 Cip1 fusion protein (lanes 3–5). However, when comparable amount of GST-3-endonexin was added instead of GST-p21 Cip1, little inhibition of the histone H1 kinase activity was observed (lanes 6 – 8). This indicates that 3-endonexin may not be a bona fide inhibitor of cyclin A-Cdk2 as examined by its histone H1 kinase activity. Next, cyclin A-associated pRB kinase activity was examined with the use of either the carboxy-terminal half (Rb-C; Fig. 4B) or the aminoterminal half of pRB (Rb-N; Fig. 4C). As reported, cyclin A-Cdk2 was capable of phosphorylating pRB. However, addition of GST-3-endonexin to the reaction gave rise to a marked inhibition of pRB phosphorylation in a dose-dependent manner (Fig. 4B and C, lanes 3–5). The observation indicates that 3-endonexin selectively inhibits cyclin A-associated pRB kinase activity. Cyclin A is growth-suppressive when expressed in a wild type yeast strain but not in temperature-sensitive cdc28 strains (32). This toxic effect is caused by the formation of an active cyclin A-Cdc28 kinase in yeast, which is relieved by a mutation in Cdc28 (32). We wondered whether 3-endonexin is able to counteract the toxic effects of human cyclin A on yeast by acting as an inhibitor of cyclin A-associated kinase. To this end, we established a yeast strain, KSC9, whose growth is inhibited by a conditional expression of cyclin A under the control of galactose-inducible GAL1 promoter. Colony formation of KSC9 cells was severely impaired on 2% galactose plates where human cyclin A was induced (Fig. 5A, lane 1; Fig. 5B, lanes 1 and 2). However, by coexpressing either 3-endonexin or p27 Kip1 Cdk inhibitor, the growth of yeast cells on the galactose plate was restored (Fig. 5A, lanes 2– 4). Anti-cyclin A immunoblotting confirmed that induction of cyclin A was not disturbed by the presence of 3-endonexin (Fig. 5B,
FIG. 4. Effect of 3-endonexin on kinase activities of cyclin A-Cdk2. In vitro kinase reaction by cyclin A-Cdk2 was performed using histone H1 (A) or the retinoblastoma protein (B, C) as a substrate in the presence of various amounts of GST-3-endonexin (3-endonexin) or GST-p21 Cip1 (p21). (A, B, C) Upper panels show 32P incorporated radioactive bands. The position of histone H1, Rb-C, Rb-N or GST-3-endonexin is indicated by arrow. Lower panels show the amount of GST-fusion protein added in the kinase reaction by Coomassie staining.
lanes 3– 8). Accordingly, the observation suggests that the toxic effects of cyclin A is neutralized due to the reduced cyclin A-Cdc28 kinase activity by p27 Kip1 or 3-endonexin. DISCUSSION In this work, we demonstrate that 3-endonexin interacts with cyclin A. The amino-terminal 52 amino acids of 3-endonexin are indispensable for its binding with cyclin A. This region contains the RxL motif that is conserved among substrates and inhibitors of cyclin-Cdk such as p107, E2F1, p21Cip1 and p27 Kip1 (27–31). Accordingly, the amino-terminal region of 3-endonexin includ-
950
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. 3-Endonexin neutralizes cyclin A toxicity in yeast. (A) Yeast cells (KSC9) that conditionally express human cyclin A in the presence of galactose were transformed by an empty plasmid or a plasmid carrying a 3-endonexin or a murine p27 Kip1 (p27) cDNA and were spotted on media containing glucose (Glc) or galactose (Gal) and sucrose (Suc) as carbon sources. After 4 days incubation at 30°C, colony formation was scored. (B) Yeast cells harboring a various combination of plasmids used in A were transferred from noninduction media (2% Glc) to induction media (2% Gal) so as to induce cyclin A expression. Before and after 7 h protein induction at 30°C, cells were harvested and cyclin A expression was examined by immunoblotting using anti-cyclin A.
ing the RxL motif is likely to constitute a docking site of cyclin A. Two lines of evidence suggest that 3-endonexin acts as a specific inhibitor of cyclin A-dependent kinase. First, 3-endonexin specifically binds cyclin A and inhibits cyclin A-Cdk2 kinase activity. Second, growth inhibition of a budding yeast caused by ectopic expression of cyclin A is relieved by co-expressing 3endonexin. To our knowledge, 3-endonexin is the first Cdk inhibitor that specifically regulates cyclin A-associated kinase. This raises an intriguing possibility that 3-endonexin may have a unique role in S phase when cyclin A is present. Furthermore, it selectively inhibits the pRB kinase activity but not the histone H1 kinase activity of cyclin A-Cdk2. Recently, evidence has been provided that the functional, hypophosphorylated form of pRB persist to exist and plays a role in the regulation of DNA synthesis during S phase (33, 34). In addition, the pRB related p107 is upregulated from late G1 phase and is abundantly expressed in S phase cells as a form of complex with
E2Fs (35, 36). Given these observations, significant amounts of pRB and p107 molecules should be kept active not only in G1 phase but also in S phase as the hypophosphorylated forms. 3-Endonexin may be involved in the functional regulation of the pRB family proteins in S phase by controlling cyclin A-associated pRB kinase activity. Recent studies have implicated that cell adhesion signals regulate the function of cyclin A (11–15). Our study may provide a molecular link between cell adhesion systems and cyclin A because 3-endonexin was originally identified as a molecule which interacts with the cytoplasmic domain of integrin 3 (16). Binding of 3-endonexin to the cytoplasmic domain of integrin 3 is reported to modulate the binding affinity of the ␣IIb3 integrin to its ligand (37), suggesting that 3-endonexin serves as an “affinity converter” of integrins. However, while 3-endonexin is expressed in various tissues and is localized in both cytoplasm and nucleus, expression of integrin 3 is restricted to certain cell types (37, 38). Hence, it is highly likely that 3-endonexin has additional biological functions other than simply acting as an affinity modulator of integrins. Indeed, our findings indicate that it regulates cyclin A-associated kinase activities. Hence, an intriguing possibility is that the complex formation of 3-endonexin with cyclin A is regulated through sequestration of 3-endonexin by integrins. Obviously, the roles of 3-endonexin in S phase progression as well as in cell adhesion warrant further investigation. ACKNOWLEDGMENTS We thank Dr. D. Cobrinik for reading the manuscript. We are grateful to Drs. M. Kawabata and J.-I. Hanai for plasmids. We also thank Dr. S. Ogawa for helpful discussions. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture, Japan and by a Research Grant from Nippon Boehringer Ingelheim Co., Ltd.
REFERENCES
951
1. Weinberg, R. A. (1995) Cell 81, 323–330. 2. Matsushime, H., Roussel, M. F., Ashmun, R. A., and Sherr, C. J. (1991) Cell 65, 701–713. 3. Dawdy, S. F., Hinds, P. W., Louie, K., Reed, S. I., Arnold, A., and Weinberg, R. A. (1993) Cell 73, 499 –511. 4. Ewen, M. E., Sluss, H. K., Sherr, C. J., Matsushime, H., Kato, J., and Livingston, D. M. (1993) Cell 73, 487– 497. 5. Elledge, S. J., Richman, R., Hall, F. L., Williams, R. T., Lodgson, N., and Haper, J. W. (1992) Proc. Natl. Acad. Sci. USA 89, 2907–2911. 6. Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992) EMBO J. 11, 961–971. 7. Rosenblatt, J., Gu, Y., and Morgan, D. O. (1992) Proc. Natl. Acad. Sci. USA 89, 2824 –2828.
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
8. Cardoso, M. C., Leonhardt, H., and Nadal-Ginard, B. (1993) Cell 74, 979 –992. 9. D’Urso, G., Marraccino, R. L., Marshak, D. R., and Roberts, J. M. (1990) Science 250, 786 –791. 10. Krude, T., Jackman, M., Pines, J., and Laskey, R. A. (1997) Cell 88, 108 –119. 11. Clark, E. A., and Brugge, J. S. (1995) Science 268, 233–239. 12. Guadagno, T. M., Ohtsubo, M., Roberts, J. M., and Assoan, K. A. (1993) Science 262, 1572–1575. 13. Radeva, G., Petrocelli, T., Behrend, E., Leung-Hagesteijn, C., Films, J., Slingerland, J., and Dedhar, S. (1997) J. Biol. Chem. 272, 13937–13944. 14. Shulze, A., Zerfass-Thome, K., Berges, J., Middendorp, S., Jansen-Dr, P., and Henglein, B. (1996) Mol. Cell. Biol. 16, 4632– 4638. 15. Symington, B. E. (1992) J. Biol. Chem. 267, 25744 –25747. 16. Shattil, S. J., O’Toole, T., Eigenthaler, M., Thon, V., Williams, M., Babior, B. M., and Ginsberg, M. H. (1995) J. Cell Biol. 131, 807– 816. 17. Hatakeyama, M., Brill, J. A., Fink, G. R., and Weinberg, R. A. (1994) Genes Dev. 8, 1759 –1771. 18. Mumberg, D., Muller, R., and Funk, M. (1994) Nucleic Acids Res. 22, 5767–5768. 19. Shinobu, N., Maeda, T., Aso, T., Ito, T., Kondo, T., Koike, K., and Hatakeyama, M. (1999) J. Biol. Chem. 274, 17003–17010. 20. Kitagawa, M., Higashi, H., Suzuki-Takahashi, I., Segawa, K., Hanks, S. K., Taya, Y., Nishimura, S., and Okuyama, A. (1995) Oncogene 10, 229 –236. 21. Lew, D. J., Dulic, V., and Reed, S. I. (1991) Cell 66, 1197–1206. 22. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., and Vogelstein, B. (1993) Cell 75, 817– 825. 23. Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K., and Elledge, S. J. (1993) Cell 75, 805– 816.
24. Xiong, Y., Hannon, G. J., Zhang, H., Casso, D., Kobayashi, R., and Beach, D. (1993) Nature 366, 701–704. 25. Dawson, I. A., Roth, S., and Artavanis-Tsakonas, S. (1995) J. Cell Biol. 129, 725–737. 26. Weinstein, J., Jacobsen, F. W., Hsu-chen, J., Wu, T., and Baum, L. G. (1994) Mol. Cell. Biol. 14, 3350 –3363. 27. Adams, P. D., Sellers, W. R., Sharma, S. K., Wu, A. D., Nalin, C. M. and Kaelin, W. G., Jr. (1996) Mol. Cell. Biol. 16, 6623– 6633. 28. Castano, E., Klyner, Y., and Dynlacht, B. D. (1998) Mol. Cell. Biol. 18, 5380 –5391. 29. Shulman, B., A., Lindstrom, D. L., and Harlow, E. (1998) Proc. Natl. Acad. Sci. USA 95, 10453–10458. 30. Vlach, J., Hennecke, S., and Amati, B. (1997) EMBO J. 16, 5334 –5344. 31. Woo M. S.-A., Sanchez, I., and Dynlacht, B. D. (1997) Mol. Cell. Biol. 16, 4632– 4638. 32. Funakoshi, M., Sikder, H., Ebihara, H., Irie, K., Sugimoto, K., Matsumoto, K., Hunt, T., Nishimoto, T., and Kobayashi, H. (1997) Genes Cells 2, 329 –343. 33. Chew, Y. P., Ellis, M., Wilkie, S., and Mittenacht, S. (1998) Oncogene 17, 2177–2186. 34. Knudsen, E. S., Buckmaster, C., Chen, T.-T., Feramisco, J. R., and Reed, S. I. (1998) Genes Dev. 12, 2278 –2292. 35. Ikeda, M.-A., Jakoi, L., and Nevins, J. R. (1996) Proc. Natl. Acad. Sci. USA 93, 3215–3220. 36. Moberg, K., Starz, M. A., and Lees, J. A. (1996) Mol. Cell. Biol. 16, 1436 –1449. 37. Kashiwagi, H., Schwartz, M. A., Eigenthaler, M., Davis, K. A., Ginsberg, M. H., and Shattil, S. J. (1997) J. Cell Biol. 137, 1433–1443. 38. Eigenthaler, M., Hoferer, L., Shattil, S. J., and Ginsberg, M. H. A (1997) J. Biol. Chem. 272, 7693–7698.
952
3-Endonexin as a Novel Inhibitor of Cyclin A-Associated Kinase Akihira Ohtoshi,* Tatsuya Maeda,* Hideaki Higashi,* Satoshi Ashizawa,* Masafumi Yamada,* and Masanori Hatakeyama* ,† ,1 *Department of Viral Oncology, Cancer Institute, Japanese Foundation for Cancer Research, Tokyo 170-8455, Japan; and †Research Section of Chemistry, Institute of Immunological Science, Hokkaido University, Sapporo 060-0815, Japan
Received December 2, 1999
Cyclin A is indispensable for S phase cell cycle progression and is suggested to be a crucial target of cell adhesion signals. In this study, we demonstrate that 3-endonexin, a molecule known to associate with the integrin 3 cytoplasmic domain, specifically binds cyclin A. Deletion of the amino-terminal 52-amino-acid residues including the cyclin-binding RxL motif abolishes the ability of 3-endonexin to interact with cyclin A. In an in vitro kinase assay, 3-endonexin inhibits pRB kinase activity associated with cyclin A-Cdk2 while leaving its histone H1 kinase activity unaffected. Coexpression of 3-endonexin in yeast cells overcomes growth suppression caused by an activation of cyclin A-associated kinase. Our results indicate that 3-endonexin is a novel cyclin A-binding molecule that regulates cyclin A-associated pRB kinase activity. © 2000 Academic Press Key Words: 3-endonexin; cyclin A; Cdk inhibitor; pRB; cell adhesion.
Sequential activation of various combinations of cyclins and cyclin-dependent kinases (Cdks) provides a molecular framework for coordinated cell cycle progression (1). In mammalian cells, growth stimulation leads to elevated expression of D-type cyclins and activates cyclin D-Cdk4/6 complex to phosphorylate the pRB (2– 4). The subsequent activation of cyclin E-Cdk2 in late G1 further hyperphosphorylates and extensively inactivates the pRB, thereby relieving pRBcontrolled E2F transcriptional activities that are essentially required for G1 to S phase transition (1). In the M phase of the mitotic cell cycle, a distinct cyclinCdk complex, cyclin B-Cdc2, works as the driving force of cell cycle progression leading to mitosis. Cyclin A, in
contrast with other cyclins, functions twice in each cell cycle by forming a complex with Cdk2 in the late G1 and S phases and with Cdc2 in G2 and M phases (5–7). Cyclin A is transcriptionally up-regulated in late G1 following cyclin E-Cdk activation and accumulates in S phase. A number of studies suggest that cyclin A-Cdk2 is involved in controlling DNA replication. Indeed, whereas cyclin A is present in both nucleus and cytoplasm, it accumulates at sites of DNA replication during S phase, and in a cell-free system cyclin A is capable of inducing DNA replication (8 –10). In addition, recent studies have implicated a unique connection between cyclin A and adhesion-dependent cell proliferation (11–15). Cells arrested by suspension fail to express cyclin A (14). Conversely, cells that ectopically express cyclin A become anchorage-independent for growth (12). These observations indicate that cell adhesion signals eventually regulate the activity of cyclin A. Cell adhesions such as anchorage of cells to the extracellular matrix are mediated by a large family of heterodimeric cell surface adhesion receptors termed integrins. They are composed of ␣ and  subunits, and evidence indicates that cytoplasmic domains of the integrin  subunits are responsible for the generation of intracellular signals following interaction with the extracellular matrix (11). However, molecular mechanisms that couple the adhesion signals and cyclin A remain to be elucidated. In this work, we demonstrate that 3-endonexin (16), which was originally isolated as a molecule that specifically interacts with the cytoplasmic domain of integrin 3, binds to cyclin A and, through the interaction, inhibits cyclin A-associated kinase activity. Our results provide a molecular link between the cell adhesion system and the cell cycle machinery. MATERIALS AND METHODS
1
To whom correspondence should be addressed at the Department of Viral Oncology, Cancer Institute, Japanese Foundation for Cancer Research 1-37-1 Kami-Ikebukuro, Toshima-ku, Tokyo 170-8455, Japan. Fax: ⫹81-3-5394-3816. E-mail: [email protected].
Yeast strains, media, and plasmid DNAs. The following yeast strains were used: KSC9 (MATa cdc28-4 ura3⌬ns ade1 his2 leu23,112 trp1-1a), Y187 (MAT␣ ura3-52 his3-200 ade2-101 trp1-901
947
0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
leu2-3,112 gal4⌬ met ⫺ gal80⌬ GAL1 UAS-GAL1TATA-LacZ) (CLONTECH), and Y190 (MATa ura3-52 his3-200 ade2-101 lys2-801 trp1-901 leu23,112 gal4⌬ gal80⌬ cyh r2 LYS2::GAL1 UAS-HIS3 TATA-HIS3 URA3:: GAL1 UAS-GAL1 TATA-LacZ) (17). pAS2.1C-cycA is a single-copy vector derived from pAS2.1 and carries a fusion gene encoding GAL4 DNA binding domain-human cyclin A chimeric protein under ADH1 promoter. pAS2.1C-cycA⌬C20 was constructed by a sequential Exonuclease III and SI nuclease treatment. The plasmid expresses a truncated cyclin A, in which the carboxy-terminal 20 amino acid residues were replaced by a serine residue derived from the vector. To conditionally express human cyclin A in yeast, p416GAL1-cycA was made by inserting a human cyclin A cDNA downstream of the Gal1 promoter in a single-copy yeast expression vector, p416GAL1 (18). A multicopy vector, p424GAL1 (18), was used to express 3-endonexin in yeast. For two-hybrid screening, a cDNA library derived from a human WI-38 lung fibroblast cell (CLONTECH) was used. Purification of GST-fusion proteins in E. coli. A GST gene fusion vector, pGEX-5X-1, was used to express GST fusion proteins in an E. coli strain, BL21. Growing BL21 cells carrying the expression plasmids were exposed to 1 mM IPTG to induce GST-fusion protein production. The cells were then harvested, washed with phosphatebuffered saline (PBS), and resuspended in NETN Buffer (20 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% NP-40) supplemented with protease inhibitors (5 g/ml PMSF, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml trypsin inhibitor). The cell suspension was then subjected to sonication and the clear lysate was collected for further purification according to manufacturer’s instructions (Pharmacia Biotech). In vitro GST pull-down assay. 35S-methionine-labeled cyclin A, cyclin E or cyclin B1 was prepared by in vitro transcription and translation using TNT Coupled Reticulocyte Lysate Systems (Promega). 35S-methionine-labelled cyclin protein and GST-fusion protein attached Glutathione Sepharose 4B were then mixed in a 200 l of NETN Buffer containing protease inhibitors and 0.5% skim milk. After washing the beads with NETN buffer, one-fifth of the aqueous fraction and a half of beads fraction were resuspended in SDS sample buffer and electrophoresed to separate proteins. Preparation of COS cell lysates and immunoprecipitation. Mammalian expression vectors for carboxy-terminal triple hemagglutinin (HA)-tagged cyclin A (pSR␣-cycA-3xHA) and the amino-terminal 6xMyc-tagged 3-endonexin (pSR␣-6xMyc-3-endonexin) were transfected into COS cells using a DEAE-dextran method. Cell lysis and immunoprecipitation were performed as described previously (19). Immunoblot analysis. Yeast cells were harvested, washed with sterilized water, and resuspended in SDS sample buffer. Extracts or immunoprecipitates from COS cells were prepared as described above. Separated proteins by SDS-PAGE were electrotransferred to a PVDF transfer membrane (Millipore), incubated with the detection antibody, 9E10 for Myc-tagged proteins, 12CA5 for HA-tagged proteins, or anti-human cyclin A (sc-751, Santa Cruz). Proteins were visualized by using enhanced chemiluminescence reagents. In vitro kinase assay. Purified cyclin A-Cdk2 complex produced in insect cells using a baculovirus system was prepared as described previously (20). GST-fusion proteins were prepared as described above. Truncated pRB proteins for pRB kinase assay, Rb-C and Rb-N, were purchased from New England Biolabs. Kinase reaction was performed at 30°C for 10 min in kinase buffer (20 mM Tris-HCl, pH 7.4, 10 mM MgCl 2, 4.5 mM -mercaptoethanol, 1 mM EGTA). The reaction solution contained 50 M of cold ATP and 0.4 l of [␥- 32P]ATP (6000 Ci/mmol, Dupont).
-Galactosidase assay. Proliferating yeast cells in SD media were harvested, transferred to YPAD media and cultured at 30°C. The grown cells were resuspended in Z buffer, frozen in liquid nitrogen and thawed at room temperature. -mercaptoethanol and
O-nitrophenyl -D-galactopyranoside (ONPG) were added to a final concentration, 38.9 mM and 0.16 mg/ml, respectively, to measure -galactosidase activity.
RESULTS To identify molecules that specifically interact with cyclin A, a yeast two-hybrid screening was performed. Because high level expression of wild-type human cyclin A is toxic to yeast, we generated a series of cyclin A-deletion derivatives and one in which the carboxyterminal 20 amino acid residues of cyclin A were replaced by a serine residue was found to be non-toxic in Y190 yeast strain and yet complemented the growth defect of a triple cln ⫺ (cln1 ⫺ cln2 ⫺ cln3 ⫺) yeast strain (data not shown) (21). Using the truncated cyclin A as a bait, a human lung fibroblast library (1 ⫻ 10 6 independent yeast transformants) was screened and three positive clones were isolated. They respectively contained cDNAs fragment encoding p21 Cip1, p55CDC, and 3-endonexin. p21 Cip1 is a Cdk inhibitor (22–24) and p55CDC is a WD repeat protein which has a homology to the budding yeast Cdc20 protein and the Drosophila fizzy protein (25, 26). In this report, we focus on 3endonexin. 3-Endonexin was originally isolated as a molecule that specifically interacts with the cytoplasmic domain of the 3 subunit of integrin (16). There are long and short forms of 3-endonexin, which are translated from alternatively spliced mRNAs, and the one isolated in our screening was the long form. Physical interaction between 3-endonexin and cyclin A was first examined in yeast two hybrid -galactosidase assay (Fig. 1). p21 Cip1, which is known to bind cyclin A, was employed as a positive control. Fusion proteins consisting of GAL4 activation domain (AD) and p21 Cip1 or wild-type 3-endonexin induced transcription of -galactosidase reporter gene and scored positive -galactosidase activity in a prey dependent manner. The production of each GAL4-AD fusion protein was confirmed by anti-HA immunoblotting (Fig 1; the bottom panel). Similarly, comparable amounts of GAL4 DNA binding domain-cyclin A fusion proteins were detected in each yeast lysate (Fig 1; the middle panel). Hence, the result indicates that, like p21 Cip1, 3-endonexin specifically binds cyclin A. To examine a potential interaction between 3endonexin and cyclins other than cyclin A, in vitro pull-down assay was performed using GST-3endonexin and in vitro translated, radio-labelled cyclin A, cyclin E or cyclin B1. As shown in Fig. 2, a significant amount of the in vitro translated cyclin A was specifically bound to GST-3-endonexin. In contrast, no specific binding was observed between GST-3endonexin and cyclin E or cyclin B1. The observation indicates that 3-endonexin specifically interacts with cyclin A.
948
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Two-hybrid interaction between 3-endonexin and cyclin A. Y187 yeast cells harboring plasmids indicated were subjected to -galactosidase liquid assay. The values are shown in Miller’s unit. Middle and bottom panels show the expression levels of GAL4 DNA binding domain-cyclin A (cyclin A), GAL4 activation domain-p21 Cip1 (p21) and GAL4 activation domain-3-endonexin (3-endonexin), respectively, in the same samples used in the top panel.
To address physical interaction between 3endonexin and cyclin A in vivo, we constructed mammalian expression vectors for the long and short forms of Myc epitope-tagged 3-endonexin (Myc-3-endonexin) and hemagglutinin (HA) epitope-tagged cyclin A
FIG. 2. Specificity of interaction between 3-endonexin and cyclins. GST pull-down assay was performed using GST-p21 Cip1 (p21) or GST-3-endonexin (3-endonexin). In vitro translated, 35Smethionine-labeled cyclin A, E or B1 was incubated with the GSTfusion protein. One-fifth of supernatant fraction and a half volume of beads fraction were loaded on odd-numbered lanes marked “S” and even numbered lanes marked “B,” respectively.
FIG. 3. Complex formation between 3-endonexin and cyclin A in mammalian cells. (A) Total cell lysates (lanes 6 –10) and anti-Myc immunoprecipitates (lanes 1–5) were prepared from COS cells expressing Myc-tagged 3-endonexin (long form) (lanes 1, 3, 6 and 8), Myc-tagged 3-endonexin (short form) (lanes 4 and 9), Myc-tagged ⌬N52 (lanes 5, 10) together with HA-tagged cyclin A (HA-cyc A) (lanes 2–5, 7–12), separated by SDS-PAGE (lanes 1– 6) and were immunoblotted with anti-HA (upper panel) or anti-Myc (lower panel). (B) Total cell lysates (lanes 6 –10) and anti-HA immunoprecipitates (lanes 1–5) were prepared from COS cells expressing the long form of Myc-tagged 3-endonexin (lanes 1, 3, 6 and 8), the short form of Myc-tagged 3-endonexin (lanes 4 and 9), Myc-tagged ⌬N52 (lanes 5, 10) together with HA-tagged cyclin A (HA-cyc A) (lanes 2–5, 7–12), separated by SDS-PAGE (lanes 1– 6) and were immunoblotted with anti-Myc (upper panel) or anti-HA (lower panel).
(HA-cyclin A). They were transiently cotransfected into monkey COS cells. Lysates were then prepared from the transfected cells and subjected to sequential immunoprecipitation-immunoblotting analyses (Figs. 3A and 3B). When both Myc-3-endonexin and HAcyclin A were co-expressed in COS cells, anti-Myc or anti-HA immunoprecipitates contained both Myc-3endonexin and HA-cyclin A. In contrast, when either Myc-3-endonexin or HA-cyclin A was singly expressed in COS cells, anti-HA and anti-Myc failed to coprecipitate Myc-3-endonexin and HA-tagged cyclin A, respectively. These results indicate that 3-endonexin and cyclin A exist in a physical complex in mammalian cells. Both long and short forms of 3-endonexin possess an RxL motif (in the single letter code for amino acid, where x is any amino acid) at its amino-terminal region
949
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(residues 5 and 7). This motif is known to constitute cyclin docking sites in other cyclin binding molecules (27–31). To determine whether the sequences involving the RxL motif are essential for the interaction of 3endonexin with cyclin A, we generated a cDNA encoding a mutant molecule, ⌬N52, which lacks the aminoterminal 52 amino acid residues including the RxL motif from the long form of 3-endonexin. When expressed in COS cells, the ⌬N52 mutant did not bind cyclin A (Fig. 3). The result thus implicates that the amino-terminal region of 3-endonexin, which contains the RxL motif, is indispensable for the interaction of 3-endonexin with cyclin A. To pursue the effect of 3-endonexin on cyclin A-associated kinase activities, we next performed in vitro kinase assay using either histone H1 or the retinoblastoma protein (pRB) as a substrate in the presence of increased amounts of 3-endonexin. As shown in Fig. 4A, purified cyclin A-Cdk2 complex efficiently phosphorylated histone H1 (compare lane 1 to 2). This histone H1 kinase activity was inhibited by adding GST- p21 Cip1 fusion protein (lanes 3–5). However, when comparable amount of GST-3-endonexin was added instead of GST-p21 Cip1, little inhibition of the histone H1 kinase activity was observed (lanes 6 – 8). This indicates that 3-endonexin may not be a bona fide inhibitor of cyclin A-Cdk2 as examined by its histone H1 kinase activity. Next, cyclin A-associated pRB kinase activity was examined with the use of either the carboxy-terminal half (Rb-C; Fig. 4B) or the aminoterminal half of pRB (Rb-N; Fig. 4C). As reported, cyclin A-Cdk2 was capable of phosphorylating pRB. However, addition of GST-3-endonexin to the reaction gave rise to a marked inhibition of pRB phosphorylation in a dose-dependent manner (Fig. 4B and C, lanes 3–5). The observation indicates that 3-endonexin selectively inhibits cyclin A-associated pRB kinase activity. Cyclin A is growth-suppressive when expressed in a wild type yeast strain but not in temperature-sensitive cdc28 strains (32). This toxic effect is caused by the formation of an active cyclin A-Cdc28 kinase in yeast, which is relieved by a mutation in Cdc28 (32). We wondered whether 3-endonexin is able to counteract the toxic effects of human cyclin A on yeast by acting as an inhibitor of cyclin A-associated kinase. To this end, we established a yeast strain, KSC9, whose growth is inhibited by a conditional expression of cyclin A under the control of galactose-inducible GAL1 promoter. Colony formation of KSC9 cells was severely impaired on 2% galactose plates where human cyclin A was induced (Fig. 5A, lane 1; Fig. 5B, lanes 1 and 2). However, by coexpressing either 3-endonexin or p27 Kip1 Cdk inhibitor, the growth of yeast cells on the galactose plate was restored (Fig. 5A, lanes 2– 4). Anti-cyclin A immunoblotting confirmed that induction of cyclin A was not disturbed by the presence of 3-endonexin (Fig. 5B,
FIG. 4. Effect of 3-endonexin on kinase activities of cyclin A-Cdk2. In vitro kinase reaction by cyclin A-Cdk2 was performed using histone H1 (A) or the retinoblastoma protein (B, C) as a substrate in the presence of various amounts of GST-3-endonexin (3-endonexin) or GST-p21 Cip1 (p21). (A, B, C) Upper panels show 32P incorporated radioactive bands. The position of histone H1, Rb-C, Rb-N or GST-3-endonexin is indicated by arrow. Lower panels show the amount of GST-fusion protein added in the kinase reaction by Coomassie staining.
lanes 3– 8). Accordingly, the observation suggests that the toxic effects of cyclin A is neutralized due to the reduced cyclin A-Cdc28 kinase activity by p27 Kip1 or 3-endonexin. DISCUSSION In this work, we demonstrate that 3-endonexin interacts with cyclin A. The amino-terminal 52 amino acids of 3-endonexin are indispensable for its binding with cyclin A. This region contains the RxL motif that is conserved among substrates and inhibitors of cyclin-Cdk such as p107, E2F1, p21Cip1 and p27 Kip1 (27–31). Accordingly, the amino-terminal region of 3-endonexin includ-
950
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. 3-Endonexin neutralizes cyclin A toxicity in yeast. (A) Yeast cells (KSC9) that conditionally express human cyclin A in the presence of galactose were transformed by an empty plasmid or a plasmid carrying a 3-endonexin or a murine p27 Kip1 (p27) cDNA and were spotted on media containing glucose (Glc) or galactose (Gal) and sucrose (Suc) as carbon sources. After 4 days incubation at 30°C, colony formation was scored. (B) Yeast cells harboring a various combination of plasmids used in A were transferred from noninduction media (2% Glc) to induction media (2% Gal) so as to induce cyclin A expression. Before and after 7 h protein induction at 30°C, cells were harvested and cyclin A expression was examined by immunoblotting using anti-cyclin A.
ing the RxL motif is likely to constitute a docking site of cyclin A. Two lines of evidence suggest that 3-endonexin acts as a specific inhibitor of cyclin A-dependent kinase. First, 3-endonexin specifically binds cyclin A and inhibits cyclin A-Cdk2 kinase activity. Second, growth inhibition of a budding yeast caused by ectopic expression of cyclin A is relieved by co-expressing 3endonexin. To our knowledge, 3-endonexin is the first Cdk inhibitor that specifically regulates cyclin A-associated kinase. This raises an intriguing possibility that 3-endonexin may have a unique role in S phase when cyclin A is present. Furthermore, it selectively inhibits the pRB kinase activity but not the histone H1 kinase activity of cyclin A-Cdk2. Recently, evidence has been provided that the functional, hypophosphorylated form of pRB persist to exist and plays a role in the regulation of DNA synthesis during S phase (33, 34). In addition, the pRB related p107 is upregulated from late G1 phase and is abundantly expressed in S phase cells as a form of complex with
E2Fs (35, 36). Given these observations, significant amounts of pRB and p107 molecules should be kept active not only in G1 phase but also in S phase as the hypophosphorylated forms. 3-Endonexin may be involved in the functional regulation of the pRB family proteins in S phase by controlling cyclin A-associated pRB kinase activity. Recent studies have implicated that cell adhesion signals regulate the function of cyclin A (11–15). Our study may provide a molecular link between cell adhesion systems and cyclin A because 3-endonexin was originally identified as a molecule which interacts with the cytoplasmic domain of integrin 3 (16). Binding of 3-endonexin to the cytoplasmic domain of integrin 3 is reported to modulate the binding affinity of the ␣IIb3 integrin to its ligand (37), suggesting that 3-endonexin serves as an “affinity converter” of integrins. However, while 3-endonexin is expressed in various tissues and is localized in both cytoplasm and nucleus, expression of integrin 3 is restricted to certain cell types (37, 38). Hence, it is highly likely that 3-endonexin has additional biological functions other than simply acting as an affinity modulator of integrins. Indeed, our findings indicate that it regulates cyclin A-associated kinase activities. Hence, an intriguing possibility is that the complex formation of 3-endonexin with cyclin A is regulated through sequestration of 3-endonexin by integrins. Obviously, the roles of 3-endonexin in S phase progression as well as in cell adhesion warrant further investigation. ACKNOWLEDGMENTS We thank Dr. D. Cobrinik for reading the manuscript. We are grateful to Drs. M. Kawabata and J.-I. Hanai for plasmids. We also thank Dr. S. Ogawa for helpful discussions. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture, Japan and by a Research Grant from Nippon Boehringer Ingelheim Co., Ltd.
REFERENCES
951
1. Weinberg, R. A. (1995) Cell 81, 323–330. 2. Matsushime, H., Roussel, M. F., Ashmun, R. A., and Sherr, C. J. (1991) Cell 65, 701–713. 3. Dawdy, S. F., Hinds, P. W., Louie, K., Reed, S. I., Arnold, A., and Weinberg, R. A. (1993) Cell 73, 499 –511. 4. Ewen, M. E., Sluss, H. K., Sherr, C. J., Matsushime, H., Kato, J., and Livingston, D. M. (1993) Cell 73, 487– 497. 5. Elledge, S. J., Richman, R., Hall, F. L., Williams, R. T., Lodgson, N., and Haper, J. W. (1992) Proc. Natl. Acad. Sci. USA 89, 2907–2911. 6. Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992) EMBO J. 11, 961–971. 7. Rosenblatt, J., Gu, Y., and Morgan, D. O. (1992) Proc. Natl. Acad. Sci. USA 89, 2824 –2828.
Vol. 267, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
8. Cardoso, M. C., Leonhardt, H., and Nadal-Ginard, B. (1993) Cell 74, 979 –992. 9. D’Urso, G., Marraccino, R. L., Marshak, D. R., and Roberts, J. M. (1990) Science 250, 786 –791. 10. Krude, T., Jackman, M., Pines, J., and Laskey, R. A. (1997) Cell 88, 108 –119. 11. Clark, E. A., and Brugge, J. S. (1995) Science 268, 233–239. 12. Guadagno, T. M., Ohtsubo, M., Roberts, J. M., and Assoan, K. A. (1993) Science 262, 1572–1575. 13. Radeva, G., Petrocelli, T., Behrend, E., Leung-Hagesteijn, C., Films, J., Slingerland, J., and Dedhar, S. (1997) J. Biol. Chem. 272, 13937–13944. 14. Shulze, A., Zerfass-Thome, K., Berges, J., Middendorp, S., Jansen-Dr, P., and Henglein, B. (1996) Mol. Cell. Biol. 16, 4632– 4638. 15. Symington, B. E. (1992) J. Biol. Chem. 267, 25744 –25747. 16. Shattil, S. J., O’Toole, T., Eigenthaler, M., Thon, V., Williams, M., Babior, B. M., and Ginsberg, M. H. (1995) J. Cell Biol. 131, 807– 816. 17. Hatakeyama, M., Brill, J. A., Fink, G. R., and Weinberg, R. A. (1994) Genes Dev. 8, 1759 –1771. 18. Mumberg, D., Muller, R., and Funk, M. (1994) Nucleic Acids Res. 22, 5767–5768. 19. Shinobu, N., Maeda, T., Aso, T., Ito, T., Kondo, T., Koike, K., and Hatakeyama, M. (1999) J. Biol. Chem. 274, 17003–17010. 20. Kitagawa, M., Higashi, H., Suzuki-Takahashi, I., Segawa, K., Hanks, S. K., Taya, Y., Nishimura, S., and Okuyama, A. (1995) Oncogene 10, 229 –236. 21. Lew, D. J., Dulic, V., and Reed, S. I. (1991) Cell 66, 1197–1206. 22. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., and Vogelstein, B. (1993) Cell 75, 817– 825. 23. Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K., and Elledge, S. J. (1993) Cell 75, 805– 816.
24. Xiong, Y., Hannon, G. J., Zhang, H., Casso, D., Kobayashi, R., and Beach, D. (1993) Nature 366, 701–704. 25. Dawson, I. A., Roth, S., and Artavanis-Tsakonas, S. (1995) J. Cell Biol. 129, 725–737. 26. Weinstein, J., Jacobsen, F. W., Hsu-chen, J., Wu, T., and Baum, L. G. (1994) Mol. Cell. Biol. 14, 3350 –3363. 27. Adams, P. D., Sellers, W. R., Sharma, S. K., Wu, A. D., Nalin, C. M. and Kaelin, W. G., Jr. (1996) Mol. Cell. Biol. 16, 6623– 6633. 28. Castano, E., Klyner, Y., and Dynlacht, B. D. (1998) Mol. Cell. Biol. 18, 5380 –5391. 29. Shulman, B., A., Lindstrom, D. L., and Harlow, E. (1998) Proc. Natl. Acad. Sci. USA 95, 10453–10458. 30. Vlach, J., Hennecke, S., and Amati, B. (1997) EMBO J. 16, 5334 –5344. 31. Woo M. S.-A., Sanchez, I., and Dynlacht, B. D. (1997) Mol. Cell. Biol. 16, 4632– 4638. 32. Funakoshi, M., Sikder, H., Ebihara, H., Irie, K., Sugimoto, K., Matsumoto, K., Hunt, T., Nishimoto, T., and Kobayashi, H. (1997) Genes Cells 2, 329 –343. 33. Chew, Y. P., Ellis, M., Wilkie, S., and Mittenacht, S. (1998) Oncogene 17, 2177–2186. 34. Knudsen, E. S., Buckmaster, C., Chen, T.-T., Feramisco, J. R., and Reed, S. I. (1998) Genes Dev. 12, 2278 –2292. 35. Ikeda, M.-A., Jakoi, L., and Nevins, J. R. (1996) Proc. Natl. Acad. Sci. USA 93, 3215–3220. 36. Moberg, K., Starz, M. A., and Lees, J. A. (1996) Mol. Cell. Biol. 16, 1436 –1449. 37. Kashiwagi, H., Schwartz, M. A., Eigenthaler, M., Davis, K. A., Ginsberg, M. H., and Shattil, S. J. (1997) J. Cell Biol. 137, 1433–1443. 38. Eigenthaler, M., Hoferer, L., Shattil, S. J., and Ginsberg, M. H. A (1997) J. Biol. Chem. 272, 7693–7698.
952