Molecular mechanism of cell cycle progression induced by ... - Nature

Oncogene (2001) 20, 2055 ± 2067 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

ORIGINAL PAPERS

Molecular mechanism of cell cycle progression induced by the oncogene product Tax of human T-cell leukemia virus type I Ritsuko Iwanaga1, Kiyoshi Ohtani*,1, Takeshi Hayashi1 and Masataka Nakamura1 1

Human Gene Sciences Center, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan

The trans-activator protein Tax of human T-cell leukemia virus type I (HTLV-I) plays an important role in the development of adult T-cell leukemia through, at least in part, its ability to stimulate cell growth. We previously reported that Tax induced cell cycle progression from G0/G1 phase to S and G2/M phases in human T-cell line Kit 225 cells. To elucidate molecular mechanism of Tax-induced cell cycle progression, we systematically examined the eects of Tax on biochemical events associated with cell cycle progression. Introduction of Tax into resting Kit 225 cells induced activation of the G1/S transition regulation cascade consisting of activation of cyclin dependent kinase 2 (CDK2) and CDK4, phosphorylation of the Rb family proteins and an increase in free E2F. The kinase activation was found to result from Tax-induced expression of genes for cell cycle regulatory molecules including cyclin D2, cyclin E, E2F1, CDK2, CDK4 and CDK6, and Tax-induced reduction of CDK inhibitors p19INK4d and p27Kip1. These modulations by Tax always paralleled the ability of Tax to activate the NF-kB transcription pathway. These results indicate the important role of Tax-mediated trans-activation of the genes for cell cycle regulatory molecules in Tax-induced cell cycle progression. Oncogene (2001) 20, 2055 ± 2067. Keywords: cell cycle; E2F; HTLV-I; NF-kB; Tax; cyclin D2 Introduction Human T-cell leukemia virus type I (HTLV-I) is the etiological agent of adult T-cell leukemia (ATL) (Hinuma et al., 1981; Poiesz et al., 1980; Yoshida et al., 1982), and HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) (Gessain et al., 1985; Osame et al., 1986). Accumulating evidence displays that Tax encoded by HTLV-I plays critical roles in leukemogenesis. Introduction of the Tax gene induces phenotypic transformation in ®broblast cell lines (Tanaka et al., 1990), persistent growth of primary T-

*Correspondence: K Ohtani Received 12 November 2000; revised 18 January 2001; accepted 24 January 2001

cells in vitro in the presence of interleukin 2 (IL-2) (Akagi et al., 1995; Grassmann et al., 1989), and development of tumors and leukemia in mice in vivo (Grossman et al., 1995; Nerenberg et al., 1987). These results indicate that Tax has growth stimulatory properties. Tax was initially identi®ed as a trans-acting transcriptional activator of the HTLV-I promoter in the long terminal repeat (LTR) (Felber et al., 1985; Seiki et al., 1986; Sodroski et al., 1985). Tax was subsequently demonstrated to trans-activate transcription of various cellular genes involved in cell growth signaling, such as genes for growth factors, growth factor receptors, cytoplasmic signal transmitters, and nuclear transcription factors (Yoshida, 1995). Trans-activation of the growth related cellular genes by Tax is thought to make cellular circumstances advantageous for cell growth. Tax is considered to activate cellular genes through modulation of three transcription factors, cAMP responsive element binding factor (CREB), serum responsive factor (SRF) and nuclear factor (NF)-kB (Yoshida, 1995). Recent ®ndings that Tax can interact with molecules involved in regulation of the cell cycle suggest that Tax may be directly implicated in cell cycle control. Tax binds and inactivates p16INK4a, and suppresses expression of p18INK4c, both of which are inhibitors of cyclin D-dependent kinases, suggesting the possibility that Tax lowers the threshold of activation of G1 cyclin dependent kinases (CDKs) (Low et al., 1997; Suzuki et al., 1996, 1999). Collectively, Tax is a potential oncoprotein which induces leukemogenesis presumably through functional modulation of molecules and genes related to growth signaling and cell cycle control. It is still unclear how Tax-induced modulation is involved in the promotion of cell proliferation. Cell proliferation is strictly controlled by the cell cycle. Progression through G1 phase into S phase is controlled by CDKs which associate with cyclin Ds and cyclin E. One of the roles of these kinases in G1/S progression is phosphorylation of members of the retinoblastoma tumor suppressor protein (Rb) family, pRb, p130 and p107, consequently activating the transcription factor E2F. E2F plays crucial roles in G1/S progression by regulating various genes which encode molecules involved in cell cycle progression and DNA replication. It is important to use human T-cells arrested at G0/G1 phase to exactly evaluate eects of Tax on the cell cycle. Kit 225 human T-cell line is

Mechanism of cell cycle progression by HTLV-I Tax R Iwanaga et al

2056

Figure 1 Kinetics of cell cycle progression induced by Tax. (A) Cell cycle progression of Kit 225 cells infected with the Tax recombinant viruses. Kit 225 cells starved of IL-2 for 48 h (0 h in ®gure) were infected with indicated recombinant adenoviruses for Tax expression; AxCAIY-Tax, AxCAIY-Taxd17/5, AxCAIY-Taxd3 and AxCAIY-TaxM22. A control virus (Ad-Con) was also used. Cells were further cultured in the absence of IL-2 for 24, 48 and 72 h. As controls, mock infected cells were either cultured in the absence of IL-2 or restimulated with IL-2. Cells were harvested for staining DNA with propidium iodide, followed by ¯ow Oncogene


suitable for this purpose, because it can be made quiescent by deprivation of IL-2 (Hori et al., 1987). We have reported that the expression of Tax in quiescent Kit 225 cells induced activation of endogenous E2F and cell cycle progression to S and G2/M phases (Ohtani et al., 2000). The same results were obtained with phytohemagglutinin-stimulated peripheral blood lymphocytes, which also grow in the presence of IL-2, indicating that the ability of Tax to stimulate cell cycle progression is not speci®c to Kit 225 cells. In this study, we systematically examined Tax-induced biochemical events leading to the activation of E2F with IL-2-starved Kit 225 cells. Our results demonstrate that Tax induces the expression of a group of cellular genes encoding molecules for cell cycle control including genes for cyclin D2, cyclin E, CDK2 and CDK4. As a consequence, CDK2 and CDK4 are activated, leading to phosphorylation of the Rb family proteins and augmentation of free E2F binding activity. The ability of Tax mutants to induce these events always parallels the activation of NF-kB transcription pathway and promotion of the cell cycle. This implies that Tax may play an important role in cell cycle progression through trans-activating cell cycle regulatory genes.

Results Kinetics of cell cycle progression induced by infection with the Tax recombinant virus We previously reported that expression of HTLV-I Tax in the human T-cell line Kit 225, which was made quiescent by deprivation of IL-2, caused cell cycle progression at 48 h after infection with the Taxexpressing recombinant adenovirus (Ohtani et al., 2000). To examine biochemical events associated with Tax-induced cell cycle progression, we ®rst compared the kinetics of cell cycle progression induced by Tax with that by IL-2. Kit 225 cells were IL-2-starved for 48 h and Tax was introduced by infection with the Tax recombinant virus (AxCAIY-Tax). In parallel, mock infected cells were either kept IL-2-starved or restimulated with IL-2. Cells were then examined for cell cycle properties. Restimulation with IL-2 for 24 h dramatically increased population of cells in S and G2/M phases (from 14.9 to 63.7%), which declined to 47.9% at 48 h and 36.7% at 72 h post stimulation (Figure 1A). Cells kept IL-2-starved for 72 h showed a small population (down to 11%) in S and G2/M phases. Infection with the Tax recombinant virus did not change the population of cells in S and G2/M phases until 24 h after infection. Afterwards, the population of cells in S and G2/M phases increased with time; up to 19.7 and 24.3% at 48 h and 72 h after infection, respectively, while infection

with a control virus did not show any appreciable change (Figure 1A). These results indicate that cell cycle progression induced by Tax is a gradual process as compared to that induced by IL-2. To address the reason of the slow cell cycle progression, we examined the kinetics of Tax expression in the virus infected cells by ¯uorescence-activated cell sorter (FACS) analysis. A small portion (20.5%) of Kit 225 cells was positive for Tax at 24 h after infection (Figure 1B). Tax positive cells increased to 42% at 48 h and 49% at 72 h. These results indicate that more than 24 h was required for maximum expression of Tax when Tax was introduced by adenovirus infection. It should be noted that even 72 h after infection only half the population of Kit 225 cells were positive for Tax (Figure 1B), presumably re¯ecting infection eciency. The population of cells in S and G2/ M phases did not further increase at later time points (data not shown). When Tax positive cells were gated, the population of cells in S and G2/M phases was 24.5% at 24 h after infection and increased up to 35% at 48 h and 40.3% at 72 h, indicating a time lag in the progression of the cell cycle after Tax expression. This suggests that a cascade of events is induced by Tax for the cell cycle to be promoted. Based on these results, biochemical events associated with the regulation of the cell cycle were examined 24 h, 48 h and 72 h after infection with Tax recombinant viruses. As reported previously (Ohtani et al., 2000), infection with a mutant Tax recombinant adenovirus (AxCAIY-Taxd3) capable of activating the NF-kB and SRF pathways but not CREB pathway promoted cell cycle progression (Figure 1A). Whereas another Tax mutant (AxCAIY-Taxd17/5) incapable of activating any of these pathways was ineective. To determine whether the NF-kB pathway or SRF pathway is responsible for Tax-mediated cell cycle progression, we examined the ability of another recombinant adenovirus (AxCAIY-TaxM22) carrying a mutant Tax gene capable of activating the CREB and SRF pathways but not NF-kB pathway. No signi®cant increase in the population of cells in S and G2/M phases was seen in cells infected with AxCAIYTaxM22 virus (Figure 1A). Western blot analysis showed amounts of Tax and mutant Tax proteins were comparable among cells infected with these viruses (data not shown). These results demonstrate that the ability of Tax to induce cell cycle progression coincides with the ability of Tax to activate the NF-kB pathway as far as examined.

2057

Increase in free E2F by Tax Since Tax activated endogenous E2F activity in IL-2starved Kit 225 cells (Ohtani et al., 2000), we addressed the question of how Tax aected expression of the E2F

cytometric examination. Populations of cells in S and G2/M phases were indicated. (B) Cell cycle progression in Tax-expressing cells. Kit 225 cells infected with AxCAIY-Tax as in (A) were stained for Tax and DNA with anti-Tax monoclonal antibody, Lt-4, and propidium iodide, respectively. Cells expressing (Tax +) and not expressing Tax (Tax 7) were gated and examined for DNA content with a ¯ow cytometer. Populations of cells positive and negative for Tax and cells in S and G2/M phases were indicated Oncogene


2058

family members. E2F1 mRNA levels increased 3.5-fold in cells infected with the wild type Tax recombinant virus at 24 h and further increased 4.5- and 3.6-folds at 48 h and 72 h, respectively (Figure 2A). The levels of mRNAs were normalized by the expression of GAPDH mRNA as stimulation with IL-2 or Tax also increased GAPDH mRNA. A slight increase (up to 1.5-fold) was observed with E2F4 mRNA at 48 h. In contrast, E2F2 mRNA levels clearly decreased in cells infected with the Tax recombinant virus to 24 and 19% at 48 h and 72 h, respectively. E2F3 mRNA levels did not change signi®cantly. Western blot analysis showed increase in E2F1 protein 48 h and 72 h after infection (Figure 2B). The status of endogenous E2F in cells infected with the Tax recombinant viruses was further examined using gel mobility shift assay with a probe of the typical E2F binding site from the DHFR gene (Ikeda et al., 1996). A major complex, in cell lysate from IL-2-starved Kit 225 cells, was con®rmed by supershift with antibodies to be the p130 complex which is known to be dominant in quiescent cells (Figure 3A,B, left panels) (Smith et al., 1996). Free E2F was scarcely detected in quiescent Kit 225 cells. Importantly, a complex containing free E2F4

was evident with lysates from cells 48 h and 72 h after infection with the wild type Tax recombinant virus. Tax induced decrease in p130 complex and formation of the slowest migration complex (S phase complex), which was either supershifted or abolished by addition of antibodies against p130, p107, E2F4, CDK2 and cyclin A, indicating that the complex contained p130, p107, E2F4, CDK2 and cyclin A. No appreciable change was observed in cells 24 h after infection with the wild type Tax recombinant virus. Similar complex formation was seen with cells stimulated with IL-2 with advanced kinetics (Figure 3A left panel and B middle panel). Since it was found that among the E2F family only E2F4 was detected as a free molecule, we tried to detect binding activity of the other members of the E2F family. To release E2F from the complexes with Rb family proteins, cell lysates were pretreated with deoxycholic acid (DOC). Even after treatment with DOC, binding activities of E2F1, E2F2 and E2F3 were only detected as supershifts with antibodies against those E2F molecules (Figure 3B right panel). Under the same conditions, E2F4 binding activity increased with time after infection with the wild type Tax

Figure 2 Eects of Tax on expression of E2F. (A) Eects of Tax on the E2F genes. mRNAs (4 mg) prepared from Kit 225 cells infected with indicated Tax recombinant adenoviruses and mock infected cells left IL-2-starved or restimulated with IL-2 as in Figure 1A were analysed by Northern blotting. As a control, mRNA from cells IL-2-starved for 48 h (0 h in ®gure) was also used. (B) Western blot analysis of E2F1 expression. Cell lysates were prepared from Kit 225 cells infected with indicated recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A Oncogene


recombinant virus. These results indicate that the expression of Tax leads to an increase in total and free E2F4 binding activity. All of these changes were also induced by infection with the mutant Taxd3 recombinant virus.

2059

Phosphorylation of Rb family proteins induced by Tax The observation that Tax enhanced E2F activity also prompted us to examine the phosphorylation status of Rb family members, pRb, p130 and p107. Western blot

Figure 3 Eects of Tax on complex formation of endogenous E2F. (A) Eects of Tax on E2F binding activity. Whole cell lysates were prepared from Kit 225 cells infected with indicated Tax recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A. E2F binding activity in the lysates were examined by gel mobility shift assay using typical E2F sites from DHFR as a probe, without (DOC 7) or with (DOC +) DOC treatment. The same probe was used as a competitor in the IL-2stimulated lane (IL-2+, Comp+). S phase and p130 complexes, and free E2F4 were indicated. (B) Supershift analysis of E2F complexes and free E2Fs. Antibodies indicated were added to the reactions 10 min before the addition of the probe. Lysates were prepared from IL-2-starved Kit 225 cells (left panel), cells stimulated with IL-2 for 24 h (middle upper panel) or cells 72 h after infection with the Tax recombinant virus (middle lower panel). Free E2Fs were released by DOC treatment (DOC +) of lysates from cells stimulated with IL-2 for 24 h (right upper panel) or by Tax for 72 h (right lower panel) Oncogene


2060

analysis with antibodies for pRb and p130 detected two close but distinct bands; the upper and lower bands represent hyperphosphorylated and unphosphorylated/hypophosphorylated forms, respectively. Infection with the Tax recombinant virus increased the ratio of phosphorylated forms to unphosphorylated ones of both Rb family members similar to stimulation with IL-2 which phosphorylated the molecules more eciently than Tax with advanced kinetics (Figure 4). It should be considered that eect of Tax appeared only in a half of the population of cells because of the infection eciency. Similar increases in phosphorylation of the Rb family members were seen with cells infected with the mutant Taxd3 recombinant virus that could also induce E2F activation and cell cycle progression. Other mutants, Taxd17/5 and TaxM22, were not eective. Amounts of p107 increased with a slight shift in mobility after expression of either Tax or Taxd3 like stimulation with IL-2. These results indicate that Tax has the ability to induce phosphorylation of the Rb family members coincided with cell cycle progression. Based on these observations, we speculated that Tax induced activation of G1 CDKs. Activation of G1 cyclin dependent kinases by Tax D-type CDKs (CDK4 and CDK6) are responsible for phosphorylation of the Rb family members. In addition cyclin E expressed at G1/S boundary is also involved in phosphorylation of the Rb family members through activation of CDK2. We thus examined whether expression of Tax induced activation of CDK2 and CDK4 in IL-2-starved Kit 225 cells. Cell lysates prepared from Kit 225 cells 24 h, 48 h and 72 h after infection with the Tax recombinant virus were subjected to immunoprecipitation with antibodies for CDK2 and CDK4. The immunoprecipitates were examined for their kinase activities using GST-Rb and histone H1 as substrates. As controls, cells either kept IL-2-starved or restimulated with IL-2 were used. Any anti-CDK4 antibody-mediated immunoprecipi-

tates from cells 24 h post infection showed the same background levels of kinase activities as that from IL-2starved cells. Further culture of cells for 48 h and 72 h after infection induced CDK4 kinase activity (Figure 5). Similar activation was seen with cells infected with the mutant Taxd3 recombinant virus but not other mutant Tax (Taxd17/5 and TaxM22) recombinant viruses. Rb kinase activity associated with anti-CDK6 antibody-mediated immunoprecipitates could not be detected under similar conditions (data not shown). Kinase activity associated with anti-CDK2-mediated immunoprecipitates from cells infected with the Tax and mutant Taxd3 recombinant viruses phosphorylated GST-Rb and histone H1 with the same time course as that with CDK4. CDK4 and CDK2 were activated by IL-2 stimulation with advanced kinetics. These results indicate that Tax induces the activation of CDK4 and CDK2 with kinetics coinciding with that of cell cycle progression induced by Tax. Effect of Tax on CDK inhibitors CDK inhibitors are regulatory molecules for kinase activity of CDK complexes. Kit 225 cells have been found to express little or no detectable level of p16INK4a, indicating that activation of G1 CDKs by Tax is independent of p16INK4a (Ohtani et al., 2000). Thus we examined whether Tax aected the expression of CDK inhibitors p15INK4b, p18INK4c, p19INK4d, p21Cip1 and p27Kip1 at mRNA and protein levels. Tax has been shown to repress promoter activity of the p18INK4c gene (Suzuki et al., 1999). Northern blot hybridization with a probe speci®c for the p18INK4c gene detected three distinct bands (Figure 6A). The middle band appeared after introduction of Tax and increased in density with time. High molecular weight species of mRNA for p18INK4c also increased by Tax as well as IL2. These results clearly demonstrate that Tax rather induces the expression of p18INK4c at transcriptional level. No signi®cant change in amount of p18INK4c protein was seen by Tax introduction and IL-2 treatment (Figure 6B).

Figure 4 Phosphorylation of Rb family proteins by Tax. Cell lysates were prepared from Kit 225 cells infected with indicated recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A. As a control, cells starved of IL-2 for 48 h (0 h in ®gure) were also used. Western blots were carried out with 40 mg of sample proteins and pRb, p130 and p107 were detected with corresponding antibodies Oncogene


2061

Figure 5 Activation of CDK4 and CDK2 by Tax. Cell lysates were prepared from Kit 225 cells infected with indicated recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A. As a control, cells starved of IL-2 for 48 h (0 h in ®gure) were also used. Cell lysates (100 mg) were immunoprecipitated with anti-CDK4 or anti-CDK2 antibodies and kinase activities of the immunoprecipitates were measured by in vitro kinase assay using GST-Rb or histone H1 as substrates

Tax gradually reduced the expression of the p19INK4d gene down to 67% 72 h after infection (Figure 6A). A profound decrease in p19INK4d protein was detected 48 h after infection, presumably indicating that Tax may aect p19INK4d post-transcriptionally (Figure 6B). No appreciable change in the expression of p15INK4b was seen after introduction of Tax at both mRNA and protein levels. Expression of the p27Kip1 gene was signi®cantly decreased down to 50% at 48 h and 72 h after infection in association with decrease in the amount of p27Kip1 protein. These decreases were seen with cells either treated with IL-2 or infected with the mutant Taxd3 recombinant virus but not other mutant viruses. Interestingly, the p21Cip1 gene was not expressed in quiescent cells and induced by Tax introduction and protein levels increased with time after infection. Tax-mediated modulation of gene expression involved in CDK activity Modulation of cyclin and CDK expression by Tax could be one of mechanisms by which CDK activity was enhanced in Tax-expressing cells. We then examined the eects of Tax on expression of cyclins and CDKs by Northern and Western blot analyses. Cyclin D2 mRNA increased fourfold 24 h after infection, even when only a marginal progression of the cell cycle was induced by Tax (Figure 7A). Elevation of cyclin D2 mRNA levels by Tax remained until 72 h. Cyclin D2 protein was detected 48 h and 72 h after infection with the Tax and mutant Taxd3 recombinant viruses by Western blot analysis (Figure 7B). Except for these two points, little, if any, or no expression of cyclin D2 protein was seen, even though mRNA for cyclin D2 was detected. These results indicate that Tax induces endogenous cyclin D2 through activation of the cyclin D2 gene expression prior to progression of cell cycle by Tax. Expression of cyclin D3 was not signi®cantly aected by Tax at both mRNA and protein levels. A twofold increase in cyclin E mRNA was observed in cells

infected with the Tax recombinant virus at 48 h, but Western blot analysis showed no signi®cant change in cyclin E protein. Quiescent Kit 225 cells showed undetectable or background level of mRNA for cyclin A. Tax clearly induced expression of mRNA for cyclin A at 48 h after infection, and the induced levels of the expression remained until 72 h. Levels of mRNA for CDK2, CDK4 and CDK6 were all elevated in cells with wild type Tax and mutant Taxd3 with slightly dierent kinetics. cdk2 mRNA levels were induced three and eightfolds at 48 h and 72 h after infection, respectively. cdk4 mRNA similarly increased by two and fourfolds. Twofold increase in cdk6 mRNA by Tax was detected 24 h after infection and was kept at the same level until 72 h. Increase in CDK2 and CDK4 proteins was also detected at 72 h as shown by Western blot analysis and increase in CDK6 protein by Tax was observed at 24 h through 72 h. Kinetics of induction of cdc2 mRNA by Tax were similar to that of cyclin A. Many of the changes induced by Tax were also induced by mutant Taxd3 but not by other mutants, TaxM22 and Taxd17/5, with the exception of cdk6, indicating that ability to transactivate the NF-kB pathway is important for modulation of expression these genes by Tax. TaxM22 induced cdk6 expression, indicating that CREB and/or SRF transcription pathway is sucient for induction of cdk6. Discussion The present study clearly demonstrates that Tax activates G1 CDKs, which are shown to be the key regulators of the G1 to S phase transition. The activation presumably results from modulation of the expression of a set of genes for cell cycle regulatory molecules. Our observations underscore the importance of the transactivating function of Tax in cell growth stimulation in that not only inducing growth signaling molecules but also inducing cell cycle regulatory molecules. Oncogene


2062

Figure 6 Eects of Tax on expression of CDK inhibitors. (A) Eects of Tax on mRNA expression of CDK inhibitors. mRNAs (4 mg) prepared from Kit 225 cells infected with indicated Tax recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A were analysed by Northern blotting. (B) Western blot analysis. Cell lysates were prepared from Kit 225 cells infected with indicated recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A

Initiation of cell cycle progression in cells infected with the Tax recombinant adenoviruses was signi®cantly later than by stimulation with IL-2. One of the reasons may be attributable to the time lag in the expression of Tax from the recombinant virus as shown by FACS analysis (Figure 1B). Cells require 48 ± 72 h after infection to accumulate enough Tax for cell cycle progression. Another reason is that only half population of the cells infected with the virus expressed Tax even at 72 h. Thus the eects of Tax appeared only in a half of the population when the whole population of cells was analysed. We previously demonstrated that Tax activated endogenous E2F, resulting in activation of the Oncogene

promoters of a series of genes with the E2F binding sites (Ohtani et al., 2000). The present study further demonstrates that Tax releases active E2F4 from the complex with the Rb family members. Tax has not been known to directly associate with the Rb family members, thus it is less likely that Tax directly modi®es the pRb-E2F complex unlike oncogene products of DNA tumor viruses such as the T antigen of SV40 and Ela of adenovirus (Nevins, 1992). These observations imply that the generation of active E2F4 is a result of indirect eects of Tax, presumably due to phosphorylation of the Rb family members, in particular p130. Northern and Western blot analyses indicated that expression of E2F1 was enhanced by Tax at mRNA


2063

Figure 7 Eects of Tax on expression of cyclins and CDKs. (A) Northern blot analysis. (B) Western blot analysis. Cell lysates were prepared from Kit 225 cells infected with indicated recombinant adenoviruses and cells left IL-2-starved or restimulated with IL-2 as in Figure 1A. Western blots were carried out with 40 mg of sample proteins. Cell lysates from MT-2 and MOLT-4 cells were used as controls for cyclin D2 and cyclin D3, respectively. Cyclin D2 molecules are indicated by arrows

and protein levels. Although the activity of E2F1 to bind to the DNA element was not detected in our gel

shift assay, it may contribute to activation of endogenous E2F by Tax. Oncogene


2064

Oncogene

Kinase activities responsible for phosphorylation of the Rb family members were monitored in cells after Tax expression. In vitro kinase assay showed that Tax induced activation of CDK4 and CDK2 in quiescent Kit 225. The same assay failed to detect kinase activity in the immunoprecipitates mediated by anti-cyclin D2 and anti-cyclin D3 antibodies (data not shown), thus it is not known which cyclin is involved in Tax-induced activation of CDK4. We demonstrated that Tax transactivated the cyclin D2 gene, but not cyclin D3 gene, resulting in elevation of the cyclin D2 protein level. It may be possible that cyclin D2 is involved, at least in part, in the induction of activation of CDK4 in Taxexpressing cells. cdk4 was tanscriptionally up-regulated by Tax. It is therefore likely that there could be a close link between Tax-induced increases in amount and activity of CDK4. Tax profoundly decreased the amount of p19INK4d protein, which could also contribute to activation of CDK4. The level of cyclin E, which binds and regulates CDK2, did not ¯uctuate between quiescent and Taxexpressing cells, although its mRNA level was induced twofolds by Tax. Tax thus seemed to induce activity of CDK2 in a cyclin E-independent manner. In this context, it should be noted that the expression of CDK2 was elevated by Tax at both mRNA and protein levels. Tax-dependent increase in the amount of CDK2 may be the main cause of induction of CDK2 activity. Another possibility is sequestration of p27Kip1 by cyclin D2 and CDK4 increased by Tax, leading to activation of the cyclin E/CDK2 complex that may phosphorylate p27Kip1. Consequently level of p27Kip1 is reduced by ubiquitin-mediated degradation. In addition, induction of cyclin A by Tax may also contribute to activation of CDK2. The integration of these modulations induced by Tax probably leads to the activation of CDKs. Kinase activities of CDK4 and CDK2 in cells infected with the Tax recombinant virus were activated at 48 h and were most evident at 72 h in contrast to those induced by IL-2 which were most evident at 24 h. All of the events downstream of activation of G1 CDKs such as phosphorylation of Rb family proteins, down regulation of p27Kip1 and induction of free E2F4 were induced at 24 h by stimulation with IL-2 and at 72 h by Tax. These results suggest that Tax-induced cell cycle progression seems to be associated with activation of G1 CDKs. Unlike other cell cycle regulatory events which are activated at 48 h through 72 h by Tax, the cyclin D2, E2F1 and cdk6 genes were activated by Tax as early as 24 h when activation of CDKs and progression to S and G2/M phases were not yet observed. These observations suggest that the induction of the three genes is the primary eect mediated by Tax. It may be expected that Tax induces the same group of genes to the same extent as stimulation with IL-2. It is intriguing to note that, there are dierences in induction of cell cycle regulatory genes between Tax and IL-2. Tax stimulated the expression of the cyclin E, p21Cip1 and cdk6 genes that were not signi®cantly

aected by IL-2. Induction of E2F1, cdk2 and cdk4 genes by Tax was much stronger than that with stimulation with IL-2. Moreover, IL-2 induced the upper band of p18INK4c, whereas Tax induced the middle band. These results imply that Tax does have the distinct eects on cell cycle regulatory genes from IL-2, indicating that Tax does not simply stimulate growth signaling pathway induced by IL-2. These results support the notion that Tax directly mediates induction of cell cycle regulatory genes rather than through growth signaling pathway. Tax greatly induced expression of p21Cip1, at both mRNA and protein levels. High expression of p21Cip1 mRNA has been reported in HTLV-I transformed cell lines (Akagi et al., 1996), suggesting that induction of p21Cip1 by Tax is not speci®c to Kit 225 cells. The biological signi®cance of induction of p21Cip1 by Tax is not known at present. p21Cip1 may facilitate assembly of cyclin D2 and cyclin D3 with CDK4 and CDK6 (LaBaer et al., 1997), contributing in activation of Dtype CDKs. Further studies are needed for elucidation of role(s) of p21Cip1 induced by Tax. One of the striking results in the present study is that the ability of Tax to activate the NF-kB pathway is closely associated with activation of a set of endogenous cell cycle regulatory genes. As shown in our previous study (Ohtani et al., 2000), exogenous introduction of active NF-kB complex is not eective in activation of E2F. Tax may activate yet unidenti®ed pathways that are also necessary for cell cycle progression. The speculation may be enforced by our previous observation that Tax expressed in rat ®broblast trans-activates the NF-kB sites but does not activate E2F (Ohtani et al., 2000). As a consequence of activation of G1 CDKs by Tax, phosphorylation of the Rb family members was induced as shown by Western blot analyses (Figure 4). Consistent with phosphorylation of p130, the p130/ E2F4 complex disappeared and free E2F4 binding activity and the S phase complex were induced in Tax expressing cells (Figure 3A). Recently a murine pro-B cell line, Ba/F3, has been shown to have p130 as a regulator of cell growth inhibition rather than pRb (Hoshikawa et al., 1998). In addition, the growth inhibitory eect of p130 was abrogated by expression of E2F4. Thus, p130 and E2F4 are thought to be major determinants of cell growth of lymphoid lineage. Much knowledge concerning regulation of Rb/E2F complex has been obtained by investigations with ®broblast cells (Dyson, 1998). These studies have revealed that a shift from the p130/E2F complex at G0 phase to p107/E2F complex at S phase is observed in ®broblasts. In Kit 225 cells, however, this shift is not dramatic as in ®broblasts and the p130/E2F complex remains at S phase. The dierence may re¯ect dierential usage of Rb family members in dierent tissues. In summary, our results demonstrate that the transactivating ability of Tax is important in Tax-induced cell cycle promotion with the identi®cation of a group of cell cycle regulatory genes as new targets of Tax.


Elucidation of yet unknown targets of Tax-mediated trans-activation should provide further insights into understanding cell growth promotion and eventually leukemogenesis induced by Tax.

Materials and methods Cell culture IL-2-dependent human T-cell line Kit 225 (Hori et al., 1987) was maintained in RPMI 1640 medium containing 10% fetal calf serum (FCS) and 0.5 nM IL-2 (Ajinomoto, Yokohama, Japan). HTLV-I infected human T-cell line MT-2 (Miyoshi et al., 1981) and HTLV-I unrelated MOLT-4 (Minowada et al., 1972) were maintained in RPMI 1640 medium containing 10% FCS. Human kidney epithelial cell line 293 (Graham et al., 1977) was maintained in Dulbecco's modi®ed Eagle's medium containing 10% FCS. Infection with recombinant adenoviruses The recombinant adenovirus for expression of Tax (AxCAIY-Tax) and its mutants Taxd3 (AxCAIY-Taxd3), Taxd17/5 (AxCAIY-Taxd17/5) and TaxM22 (AxCAIYTaxM22), were kindly provided by Dr M Yoshida and described previously (Hirai et al., 1992; Ohtani et al., 2000). Taxd3, TaxM22 and Taxd17/5 have defect in activation of CREB, NF-kB and all of the transcription pathways, respectively. The control virus, Ad-Con, has been described previously (Schwarz et al., 1995). For all assays, Kit 225 cells were starved of IL-2 for 48 h and infected with the recombinant viruses at multiples of 100 plaque forming units/cell at 107 cells/ml in RPMI 1640 for 1 h at 378C. Cells were further cultured in the absence of IL-2 for 24, 48 and 72 h and harvested for analyses. Mock infected cells cultured in the absence and presence of IL-2 were used as negative and positive controls, respectively. Tax expression and DNA content analysis Tax was stained with an anti-Tax monoclonal antibody (Lt-4) (Lee et al., 1989) and DNA was stained with propidium iodide. Cells were examined by FACS analysis as described previously (Ohtani et al., 2000). In vitro kinase assay Kinase assays with pRb and histone H1 were performed as described previously with minor modi®cations (DeGregori et al., 1995). Brie¯y, cells were resuspended at 107 cells/ml in 16lysis buer [50 mM N-(2-hydroxyethyl) piperazine-N'-(2ethanesulfonic acid) (HEPES) (pH 7.5), 150 mM NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 2 mM ethylene glycol-bis(b-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), 0.1% Tween 20, 10% glycerol, 10 mM bglycerophosphate, 1 mM NaF, 1 mM dithiothreitol (DTT), 0.1 mM phenylmethylsulfonyl ¯uoride (PMSF), 10 mg/ml leupeptin, 0.1 mM sodium orthovanadate, 5 mg/ml aprotinin and 1 mg/ml pepstatin A] and were frozen in liquid nitrogen. Cells were thawed, incubated on ice for 2 h and then microcentrifuged for 10 min at 13 000 r.p.m. Protein concentrations of cell lysates were measured by the Bradford assay. For kinase assays, cell lysates (400 mg for pRb and 100 mg for histone H1) were precleared with 20 ml of 50% protein G-sepharose beads for 1 h. Antibodies (5 mg) against

CDK2 (sc-163-G), CDK4 (sc-601-G) or CDK6 (sc-177-G) which were purchased from Santa Cruz Biotechnology were added to the clari®ed lysates and incubated for 1 h on ice, followed by addition of 10 ml of 50% protein G-sepharose beads and incubation for 2 h at 48C with rocking. Immune complexes on the beads were washed three times with 16 lysis buer and twice with kinase buer [50 mM HEPES (pH 7.5), 10 mM MgCl2 and 1 mM DTT]. The beads containing the immune complexes were then suspended in 50 ml of kinase buer either containing 5 mg of GST-Rb (Ewen et al., 1993), 2.5 mM EGTA, 10 mM b-glycerophosphate, 0.1 mM sodium orthovanadate, 1 mM NaF, 20 mM ATP and 5 mCi of [g-32P]ATP or containing 5 mg of histone H1, 1 mM ATP and 5 mCi of [g-32P]ATP, and incubated for 30 min at 308C with occasional mixing. The reaction was stopped by addition of 50 ml of 26 SDS sample buer and then the samples were boiled for 10 min. Samples (30 ml) were electrophoresed through either a 12 or 8% SDSpolyacrylamide gel. The gels were dried and exposed to an imaging plate 2040S (Fuji Film), followed by analysis with an image analyser BAS 1500 (Fuji Film).

2065

Gel mobility shift assay Preparation of whole cell extract and gel mobility shift assay were performed as described previously (Ikeda et al., 1996). The typical E2F site from the DHFR promoter was used for a probe and a competitor at 100-fold molar excess. For supershift assays, reaction mixtures without radio-labeled probe were preincubated in the presence of indicated antibodies for 10 min. The radio-labeled probe was then added and further incubated for 20 min. Antibodies against p130 (sc-317X), p107 (sc-250X), cyclin A (sc-751), CDK2 (sc163-G), E2F1 (sc-251X), E2F2 (sc-633X), E2F3 (sc-878X) and E2F4 (sc-512X) were from Santa Cruz Biotechnology. C36 monoclonal antibody against pRb (14031A) was from Pharmingen. Northern (RNA) blot assay Total RNA extraction and polyA RNA puri®cation were carried out using Isogen (Nippon Gene) and PolyA Tract (Promega), respectively, according to the protocol recommended by the manufacturers. Gel electrophoresis, transfer onto nylon membranes, and hybridization were performed as described previously (Johnson et al., 1994). The probes used were cDNA fragments from the following plasmids; cyclin D2: HindIII-XbaI fragment from pRc/RSV-cyc D2 (Quelle et al., 1993), cyclin D3: SpeI-XbaI fragment from pRc/RSV-cyc D3 (Quelle et al., 1993), cyclin E: Asp718-XbaI fragment from pRc/cyc E (Hinds et al., 1992), cyclin E2: BamHI fragment from pKS-HCN1 (Lauper et al., 1998), cyclin A: HindIII-BamHI fragment from pCMX-A, cdk2, cdk4, cdk6, cdc2: BamHI fragment from pCMVcdk2WT-HA, pCMVcdk4WT-HA, pCMVcdk6WT-HA, pCMVcdc2WT, respectively (van den Heuvel and Harlow, 1993), E2F1: BamHI fragment from pDCE2F (Johnson et al., 1993), E2F2, E2F3: BamHI fragment from pCMVE2F-2(85-437), pCMVE2F-3(132-425), respectively (Lees et al., 1993), E2F4: BamHI-EcoRI fragment from pCMV-E2F-4 (Ginsberg et al., 1994), p15INK4b: EcoRI-HindIII fragment from pBSp15, p18INK4c: ApaI-SacI fragment from pTAp18, p19INK4d: BamHI-HindIII fragment from pBSp19, p21Cip1: ClaI-XbaI fragment from pBS-p21 (DeGregori et al., 1995), p27Kip1: HindIII-XbaI fragment from pBS-p27 (DeGregori et al., 1995). pBSp15 and pTAp18 were kindly provided by Dr G Peters. pBSp19 was made by cloning of nt 210 ± 796 of p19 Oncogene


2066

cDNA into the EcoRV site of pBSK- using RT ± PCR with MOLT-4 mRNA as a template. GAPDH cDNA was used as a control probe. The blot was exposed to an imaging plate 2040S and analysed with an image analyser BAS 1500. Antibodies and immunoblotting Western blot analysis was performed as described previously with minor modi®cations (Iwanaga et al., 1999). Brie¯y, cell lysates prepared from Kit 225 cells infected with the recombinant adenoviruses were resolved by electrophoresis on a 6% (for pRb, p107 and p130), 10% (for CDK2, CDK4, CDK6, cyclin D2, cyclin D3, cyclin E, p27Kip1 and E2F1) or 15% (for p18INK4c, p19INK4d and p21Cip1) SDS-polyacrylamide gel and transferred to polyvinylidene di¯uoride membranes (Atto). After blocking with Block Ace (Dainihonseiyaku) (for CDK2, CDK4, CDK6, cyclin E, E2F1, p18INK4c and p19INK4d) or 5% skim milk (GIBCO BRL) in PBS containing 0.1% Tween 20 (for pRb, p107, p130, p21Cip1, p27Kip1, cyclin D2 and cyclin D3) for 1 h at room temperature and washing with PBS, the blots were incubated with primary antibodies against p107 (sc-318), p130 (sc-317), CDK2 (sc-163-G), CDK4 (sc-601-G), CDK6 (sc-177-G), p18INK4c (sc-865), cyclin D2 (sc-452), cyclin D3 (sc-453) (which were used at 1 : 1000), pRb (sc-50), p27Kip1 (sc-1641) (which were used at 1 : 3000), p21 (sc-6246) used at 1 : 50, E2F1 (sc-251) used at 1 : 60, p19INK4d (65911A), cyclin E (14591A) (which were used at 1 : 1000), in 10% Block Ace in water containing 0.1% Tween 20 for 1 h at room temperature or 0.25% skim milk in PBS containing 0.1% Tween 20 for overnight at 48C and then washed with PBS containing 0.1% Tween 20. These

antibodies were obtained from Santa Cruz Biotechnology except for anti-p19INK4d and anti-cyclin E antibodies from Pharmingen. The blots were then incubated with peroxidaselabeled anti-mouse IgG (1 : 5000, NA9310, Amersham) (for anti-cyclin E, anti-p19INK4d, anti-p21Cip1, anti-p27Kip1 and antiE2F1), anti-rabbit IgG (1 : 5000, NA934, Amersham) (for anti-pRb, anti-p107, anti-p130, anti-p18INK4c, anti-cyclin D2 and anti-cyclin D3), or anti-goat IgG (1 : 5000, 6025-05, Southern Biotech) (for anti-CDK2, anti-CDK4 and antiCDK6), in 10% Block Ace in water containing 0.1% Tween 20 or 0.05% skim milk in PBS containing 0.1% Tween 20 for 1 h at room temperature and then washed with PBS containing 0.1% Tween 20 and ®nally with PBS alone. Proteins recognized by the antibodies were visualized using the enhanced chemiluminescence Western blotting detection system (Amersham).

Acknowledgments We thank J Hamuro for IL-2, B Amati and G Peters for plasmids and M Yoshida for recombinant adenoviruses. This work was supported in part by a Grant-in-Aid for General Scienti®c Research and Cancer Research from the Ministry of Education, Science, Sports and Culture, in part by a grant from CREST (Core Research for Evolutional Science and Technology) of the Japan Science and Technology Corporation (JST), in part by a grant from NOVARTIS Foundation (Japan) for the Promotion of Science, in part by a grant from Osaka Cancer Foundation, and in part by a grant from The Naito Foundation.

References Akagi T, Ono H and Shimotohno K. (1995). Blood, 86, 4243 ± 4249. Akagi T, Ono H and Shimotohno K. (1996). Oncogene, 12, 1645 ± 1652. DeGregori J, Leone G, Ohtani K, Miron A and Nevins JR. (1995). Genes Dev., 9, 2873 ± 2887. Dyson N. (1998). Genes Dev., 12, 2245 ± 2262. Ewen ME, Sluss HK, Sherr CJ, Matsushime H, Kato J and Livingston DM. (1993). Cell, 73, 487 ± 497. Felber BK, Paskaris H, Kleinman-Ewing C, Wong-Staal F and Pavlakis GN. (1985). Science, 229, 675 ± 679. Gessain A, Barin F, Vernant JC, Gout O, Maurs L, Calender A and de The G. (1985). Lancet, ii, 407 ± 410. Ginsberg D, Vairo G, Chittenden T, Xiao ZX, Xu G, Wydner KL, DeCaprio JA, Lawrence JB and Livingston DM. (1994). Genes Dev., 8, 2665 ± 2679. Graham FL, Smiley J, Russell WC and Nairn R. (1977). J. Gen. Virol., 36, 59 ± 74. Grassmann R, Dengler C, Muller-Fleckenstein I, Fleckenstein B, McGuire K, Dokhelar MC, Sodroski JG and Haseltine WA. (1989). Proc. Natl. Acad. Sci. USA, 86, 3351 ± 3355. Grossman WJ, Kimata JT, Wong FH, Zutter M, Ley TJ and Ratner L. (1995). Proc. Natl. Acad. Sci. USA, 92, 1057 ± 1061. Hinds PW, Mittnacht S, Dulic V, Arnold A, Reed SI and Weinberg RA. (1992). Cell, 70, 993 ± 1006. Hinuma Y, Nagata K, Hanaoka M, Nakai M, Matsumoto T, Kinoshita KI, Shirakawa S and Miyoshi I. (1981). Proc. Natl. Acad. Sci. USA, 78, 6476 ± 6480.

Oncogene

Hirai H, Fujisawa J, Suzuki T, Ueda K, Muramatsu M, Tsuboi A, Arai N and Yoshida M. (1992). Oncogene, 7, 1737 ± 1742. Hori T, Uchiyama T, Tsudo M, Umadome H, Ohno H, Fukuhara S, Kita K and Uchino H. (1987). Blood, 70, 1069 ± 1072. Hoshikawa Y, Mori A, Amimoto K, Iwabe K and Hatakeyama M. (1998). Proc. Natl. Acad. Sci. USA, 95, 8574 ± 8579. Ikeda MA, Jakoi L and Nevins JR. (1996). Proc. Natl. Acad. Sci. USA, 93, 3215 ± 3220. Iwanaga Y, Tsukahara T, Ohashi T, Tanaka Y, Arai M, Nakamura M, Ohtani K, Koya Y, Kannagi M, Yamamoto N and Fujii M. (1999). J. Virol., 73, 1271 ± 1277. Johnson DG, Ohtani K and Nevins JR. (1994). Genes Dev., 8, 1514 ± 1525. Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993). Nature, 365, 349 ± 352. LaBaer J, Garrett MD, Stevenson LF, Slingerland JM, Sandhu C, Chou HS, Fattaey A and Harlow E. (1997). Genes Dev., 11, 847 ± 862. Lauper N, Beck AR, Cariou S, Richman L, Hofmann K, Reith W, Slingerland JM and Amati B. (1998). Oncogene, 17, 2637 ± 2643. Lee B, Tanaka Y and Tozawa H. (1989). Tohoku J. Exp. Med., 157, 1 ± 11. Lees JA, Saito M, Vidal M, Valentine M, Look T, Harlow E, Dyson N and Helin K. (1993). Mol. Cell. Biol., 13, 7813 ± 7825.


Low KG, Dorner LF, Fernando DB, Grossman J, Jeang KT and Comb MJ. (1997). J. Virol., 71, 1956 ± 1962. Minowada J, Onuma T and Moore GE. (1972). J. Natl. Cancer Inst., 49, 891 ± 895. Miyoshi I, Kubonishi I, Yoshimoto S, Akagi T, Ohtsuki Y, Shiraishi Y, Nagata K and Hinuma Y. (1981). Nature, 294, 770 ± 771. Nerenberg M, Hinrichs SH, Reynolds RK, Khoury G and Jay G. (1987). Science, 237, 1324 ± 1329. Nevins JR. (1992). Science, 258, 424 ± 429. Ohtani K, Iwanaga R, Arai M, Huang Y, Matsumura Y and Nakamura M. (2000). J. Biol. Chem., 275, 11154 ± 11163. Osame M, Igata A, Usuku K, Rosales RL and Matsumoto M. (1986). Lancet, i, 1031 ± 1032. Poiesz BJ, Ruscetti FW, Gazdar AF, Bunn PA, Minna JD and Gallo RC. (1980). Proc. Natl. Acad. Sci. USA, 77, 7415 ± 7419. Quelle DE, Ashmun RA, Shurtle SA, Kato JY, Bar-Sagi D, Roussel MF and Sherr CJ. (1993). Genes Dev., 7, 1559 ± 1571. Schwarz JK, Bassing CH, Kovesdi I, Datto MB, Blazing M, George S, Wang XF and Nevins JR. (1995). Proc. Natl. Acad. Sci. USA, 92, 483 ± 487.

Seiki M, Hattori S, Hirayama Y and Yoshida M. (1986). EMBO J., 5, 561 ± 565. Smith EJ, Leone G, DeGregori J, Jakoi L and Nevins JR. (1996). Mol. Cell. Biol., 16, 6965 ± 6976. Sodroski J, Rosen C, Goh WC and Haseltine W. (1985). Science, 228, 1430 ± 1434. Suzuki T, Kitao S, Matsushime H and Yoshida M. (1996). EMBO J., 15, 1607 ± 1614. Suzuki T, Narita T, Uchida-Toita M and Yoshida M. (1999). Virology, 259, 384 ± 391. Tanaka A, Takahashi C, Yamaoka S, Nosaka T, Maki M and Hatanaka M. (1990). Proc. Natl. Acad. Sci. USA, 87, 1071 ± 1075. van den Heuvel S and Harlow E. (1993). Science, 262, 2050 ± 2054. Yoshida M. (1995). J. Cancer Res. Clin. Oncol., 121, 521 ± 528. Yoshida M, Miyoshi I and Hinuma Y. (1982). Proc. Natl. Acad. Sci. USA, 79, 2031 ± 2035.

2067

Oncogene