Interplay Between Cdk9 and NF-Kb Factors Determines the Level of HIV-1 Gene Transcription in Astrocytic Cells

Interplay between cdk9 and NF-kB factors determines the level of HIV-1 gene transcription in astrocytic cells

Shohreh Amini1, Anaira Clavo1, Yuri Nadraga1, Antonio Giordano2, Kamel Khalili*,1 and Bassel E Sawaya1

1Center for Neurovirology and Cancer Biology, College of Science and Technology, Temple University, 1900 North 12th Street, 015-96, Room 203, Philadelphia, Pennsylvania, PA 19122, USA; 2Sbarro Institute for Cancer Research, College of Science and Technology, Temple University, 1900 North 12th Street, Philadelphia, Pennsylvania, PA 19122, USA

Basal transcription of the HIV-1 genome is controlled by a gene transcription. In this respect, a great deal of variety of ubiquitous and inducible regulatory factors, attention has been focused on the kB DNA motif of some with the ability to associate with the viral DNA HIV-1 which is located between 7117 and 780 of the sequences within the promoter spanning the long terminal HIV-1 long terminal repeat, LTR, and its cellular repeat (LTR). In this report we demonstrate that DNA binding proteins, the NF-kB family of transcrip- activation of the HIV-1 promoter through the inducible tion factors. The two well-studied members of this DNA binding NF-kB transcription factors can be affected family are p50 and p65 whose activity can be induced by cdk9 in human astrocytic cells. Our results show that by a broad range of biological and pharmacological ectopic expression of cdk9, but not its mutant variant agents including tumor necrosis factor a (TNFa), and which lacks the domain responsible for its kinase activity, phorbol esters such as PMA (Kumar et al., 1996; Beg augments transcription of the LTR. Moreover, we et al., 1993). Treatment of the cells with these agents demonstrate that induction of the NF-kB pathway by results in degradation of IkB, a cytoplasmic partner of PMA, or overexpression of its subunits including p50/p65 p50/p65 that under unstimulated conditions, retains have a negative effect on the ability of cdk9 to stimulate these subunits in the cytoplasm. In the absence of IkB, viral gene transcription in these cells. Results from band- both p50 and p65 enter into the nuclei and by shift experiments demonstrated significant suppression of associating with their DNA target sequences stimulate p50/p65 association to its DNA target motif by cdk9. transcription of the genes containing the kB motif. Further, data from GST pull-down and combined Interestingly, infection with a variety of viruses also immunoprecipitation/Western blot analysis of the protein results in stimulation of the NF-kB regulatory path- extracts from cells expressing cdk9, p50 and p65 have way, which in turn, may lead to initiation of viral lytic revealed the interaction of cdk9 with both p50 and p65 in infection (Arenzana-Seisdedos et al., 1995). Activation the absence of DNA containing the kB motif. All of these of NF-kB upon viral infection, through stimulation of observations led us to conclude that the interaction of cdk9 various cytokines including TNFa, can indirectly affect with the NF-kB factors can determine the ability of NF-kB the severity of the disease by causing various injuries to modulate HIV-1 gene transcription. (Rostasy et al., 2000). Several laboratories have Oncogene (2002) 21, 5797 – 5803. doi:10.1038/sj.onc. identified a number of cellular and viral proteins that 1205754 through association with either subunit of NF-kB modulate their transcriptional regulatory ability Keywords: HIV-1 transcription; NF-kB; cdk9 (Bandres and Ratner, 1994; Hiscott et al., 1997). For example, the association of p50 and p65 with HTLV-1 Tax may dictate the extent of transcriptional activation of the viral promoter by Tax (Lindholm et al., 1990). Transcription of the HIV-1 promoter is regulated by a Further, communication between the HIV-1 transcrip- variety of cellular proteins with the ability to directly tion activator, Tat, and NF-kB subunits plays an or indirectly interact with the DNA sequence (Pereira important role in the level of viral gene expression in et al., 2000). These factors, some with inducible the cells. Tat is an early protein of HIV-1, which characteristics whose expression and/or activities are through interaction with the TAR RNA sequence augmented by the environmental stimuli, may commu- located in the leader of the viral transcript, augments nicate with each other, either on or off the DNA transcription initiation and elongation of the viral sequences to exert their regulatory effect upon viral genome (Berkhout et al., 1989). According to one model, interaction of Tat with TAR brings p-TEF, which is a component of cyclin T and its catalytic *Correspondence: K Khalili; E-mail: [email protected] subunit, cdk9, to the vicinity of the transcription Received 19 December 2001; revised 26 April 2002; accepted 7 initiation complex where cdk9 can phosphorylate the June 2002 carboxyl terminal domain of the large subunit of RNA Interaction of NF-kB with cdk9 controls HIV-1 S Amini et al 5798 polymerase II and potentiate its activity upon viral LTR in these cells. The mutant cyclin T1 containing gene transcription (Price, 2000; Wei et al., 1998). the residues 1 – 250, which showed no ability to Furthermore, more recent studies suggested that augment LTR transcription in unstimulated cells had elongation of RNA polymerase II can be augmented no significant effect on viral promoter activity, alone or through the interaction of NF-kB and p-TEFb together with cdk9, in cells treated with PMA (data not (Barboric et al., 2001). The ability of Tat to shown). communicate with NF-kB on one hand, and cdk9 on To examine the importance of the kB motif in the the other, prompted us to investigate the role of cdk9 observed regulatory event, a truncated LTR promoter in transcription of the LTR through its communication encompassing the sequences between 7117 to +80 with NF-kB. We chose human astrocytic cells since our was introduced into untreated and PMA-treated U- earlier studies revealed a novel pathway which includes 87MG cells in the presence of cdk9 expressing interaction of the NF-kB pathway with Tat in these plasmids. As shown in Figure 3, treatment of cells cells that potentiates the ability of Tat to stimulate with PMA decreased the level of viral transcription, viral gene transcription through the kB motif and in suggesting that the region between 7117 to +80 is the absence of TAR (Taylor et al., 1992, 1994). important for the functional communication between In the first series of experiments, we examined the cdk9 and p50/p65 (Figure 3, lane 1). Removal of the ability of cdk9 to modulate the basal transcriptional TAR region of the LTR between +3 and +80 activity of the HIV-1 LTR in the untreated and PMA- revealed that expression of cdk9 is able to down- treated human astrocytic cell line, U-87MG. As shown regulate PMA-induced activation of the viral promoter in Figure 1, ecotopic expression of cdk9, but not its in the absence of TAR (Figure 3, lane 2). Deletion of kinase deficient mutant, cdk9DN (Foskett et al., 2001; the two kB motifs which are the binding sites for p50 Garriga et al., 1996), caused an enhancement in the and p65 significantly decreased the level of LTR level of LTR transcription in untreated cells (Figure 1a, transcription in the untreated and PMA-treated cells compare lanes 2 and 3). As anticipated, treatment of and that the levels of transcription in the untreated and the cells with PMA elevated the basal activity of the PMA-treated cells expressing cdk9 remained compar- LTR promoter (compare lanes 1 in Figure 1a,b) and able with each other (Figure 3, lane 3). In contrast, that the observed induction was diminished by cdk9 (in 7117/+3 constructs with a deletion in the three GC- Figure 1a,b compare lanes 2) suggesting that PMA rich motifs remained responsive to PMA-induced treatment has a negative impact upon cdk9 transcrip- suppression of cdk9-mediated transcription of the viral tional activity. As PMA is a known inducer of the NF- promoter (Figure 3, lane 4). These results demon- kB pathway, in follow-up experiments we investigated strated that the sequence corresponding to the kB functional interaction of cdk9 and the p50 and p65 motif of the LTR is important for cdk9 and NF-kB subunits of NF-kB. Results from transfection studies communication. showed that the stimulatory effects of p50, p65, and As NF-kB subunits exert their activity through their cdk9 were decreased in cells co-expressing p50 and interaction with the kB DNA sequence, in the next cdk9 or p65 and cdk9 (Figure 1c, compare lanes 5 and series of experiments we assessed the binding ability of 6 to lane 4). A similar effect was observed in cells p50 and p65 in the presence of cdk9 protein. Results treated with PMA (Figure 1d, compare lanes 5 and 6 from band-shift experiments demonstrated that incuba- to lane 4), again supporting the notion that antag- tion of the 32P-labeled DNA oligonucleotide containing onistic communication between cdk9 and p50 or p65 the kB motif of the LTR with bacterially-produced abrogates their ability to fully up-regulate HIV-1 gene purified p50 led to the formation of a nucleoprotein transcription. Of note, co-production of p50 and p65 complex (Figure 4a). The addition of bacterially- showed no improvement in the level of transcription produced cdk9 to the binding reaction completely suppression in cells expressing cdk9 (data not shown), blocked formation of the kB : p50 complex. This suggesting that communication of cdk9 with p50 and inhibition may not be attributed to the GST-moiety p65 can occur in the presence of both NF-kB subunits. of bacterially-produced GST-cdk9 as incubation of the In the next experiment we determined whether probe with p50 (shown in Figure 4a, lane 1) was ectopic expression of cyclin T1, a known partner of carried out in the presence of GST to normalize the cdk9 can affect the observed negative interplay of NF- amounts of proteins in each reaction. The absence of kB and cdk9 upon LTR transcription. Toward this an additional complex in lane 2 suggests that cdk9 may end, cells were transfected with CMV-based plasmids not directly bind to the kB DNA probe and/or the expressing cyclin T1 and cdk9, alone or together, in the kB : p50 complex. Similar experiments were performed absence and presence of PMA-stimulation of NF-kB. with highly purified p65 in the absence and presence of As shown in Figure 2, treatment of the cells with PMA GST – cdk9. As illustrated in Figure 4b, the intensity of further enhanced the level of LTR activation by cyclin the bands corresponding to p65 DNA complex was T1 (lane 3). Co-expression of cdk9 and cyclin T1 in reduced upon incubation of the binding reaction with PMA treated cells resulted in a reduced level of LTR GST – cdk9, supporting the notion that cdk9 can transcription compared to that of untreated cells (lane diminish the interaction of p65 with the DNA probe. 4). This observation suggests that NF-kB factors are Further, cdk9 was able to drastically decrease the able to diminish the activity of cdk9 alone, but not that association of p50 : p65 with the kB probe (Figure 4c). of cyclin T1/cdk9 on the transcription of the HIV-1 Results from supershift analysis using specific anti-

Oncogene Interaction of NF-kB with cdk9 controls HIV-1 K Khalili et al 5799

Figure 1 Activation of HIV-1 promoter by cdk9 and NF-kB in U-87 MG. The human astroglial cell line, U-87 MG (56105 in 60 mm plate), was transfected with 0.5 mg of the reporter LTR-Luciferase plasmid (LTR – Luc) alone or together with 0.25 mgof a plasmid expressing cdk9 (CMV – cdk9) or its mutant (CMV – cdk9DN) (Foskett et al., 1996; Garriga et al., 1996) by calcium – phosphate precipitation (a). In parallel plates, the cells were subjected to PMA-treatment (50 ng/ml) 24 h post transfection to stimulate NF-kB pathway (b). The levels of LTR transcription were measured 48 h post transfection, the cells were harvested and subjected to Luciferase assay. (c) Transfection of U-87 MG with LTR – Luc and pCMV – cdk9 was performed in the absence or presence of 1.0 mg pCMV – p50 or pCMV – p65. All expression plasmids were created in pcDNA3 or pCEP-4-3 background plasmids and empty pcDNA or pCEP-4-3 was included in all transfection mixtures to normalize the amounts of the DNA in each reaction. (d) Transfections were performed as in b except that cells were treated with 50 ng/ml of PMA. The values shown on the top of each bar represent the fold activation over the basal HIV-1 LTR promoter activity which is set at an arbitrarily unit of one. The data represent the mean value of at least three separate transfection experiments, each conducted with various plasmid preparations. In all transfection assays, the final amount of the DNA was kept constant by using empty vector DNA. Further, transfection efficiency was closely monitored by including the plasmid DNA expressing green fluorescent protein, GFP, and the transfected cells were visualized by fluorescent microscopy. In each assay, an equal amount of the proteins were used for Luciferase assay bodies that recognize the NF-kB protein showed the enhancement in the formation of NF-kB protein presence of the p50 and p65 proteins in the observed complexes (Figure 4d, depicted by c). Transfection of nucleoprotein complex (data not shown). Again, no the cells with the cdk9 expression plasmid severely evidence for the interaction of cdk9 with kB was decreased the formation of these complexes in the observed in these studies. unstimulated cells (Figure 4d, compare lanes 1 and 3) To investigate the effect of cdk9 on the interaction of and noticeably decreased the intensity of the band NF-kB with the kB motif, protein extracts from the corresponding to the NF-kB complex in the stimulated unstimulated or PMA-stimulated cells transfected with cell extract (Figure 4d, compare lanes 2 and 3). In cdk9 were prepared and mixed with the kB DNA parallel, interaction of GC-rich binding proteins with probe. As anticipated, PMA treatment caused an the GC-rich DNA motif of HIV-1 was examined and

Oncogene Interaction of NF-kB with cdk9 controls HIV-1 S Amini et al 5800 The ability of cdk9 to inhibit the association of NF- kB proteins with the kB motif without being associated with the DNA probe provided a rationale to investigate the direct interaction of p50 and p65 with cdk9. Thus, we performed GST pull-down assay. First, we investigated the physical interaction of p50 and cdk9 using in vitro translated 35S-labeled p50 or cdk9 proteins in GST pull-down with GST – cdk9 or GST – p50, respectively. As shown in Figure 5a, a 50 kDa band corresponding to p50 was observed upon elution of the GST – cdk9, but not GST, indicating the ability of p50 to interact with cdk9. In a reciprocal approach, detection of a 38 kDa protein corresponding to cdk9 in eluates from GST – p50, but not GST, verified the direct interaction of p50 and cdk9 in this assay (Figure Figure 2 Regulation of HIV-1 LTR by cdk9 and cyclin T1 upon 5b). stimulation of NF-kB in human astroglial cells. U-87MG cells A similar strategy was adapted to assess the were transfected with LTR-Luc alone or together with CMV- interaction of p65 with cdk9. As demonstrated in cdk9 and/or CMV-cyclin T1. Cells were untreated or treated with Figure 5c, the presence of a band corresponding to p65 PMA (50 ng/ml) for 48 h. Protein extracts were prepared and Lu- upon elution of GST – cdk9 points to the ability of p65 ciferase activity was determined. The standard controls such as normalization of the plasmid DNAs with empty vector (pCMV), to associate with cdk9. Again, the reciprocal experi- measuring transfection efficiency by the level of GFP expression ments using in vitro synthesized cdk9 and GST – p65 and the use of equal amounts of proteins in Luciferase assay were confirmed the physical association of cdk9 with p65 carried out. Each transfection was repeated three times to ensure (Figure 5d). its reproducibility To further investigate the interaction of cdk9 with p50 and p65, PMA-stimulated U-87MG cells were transfected with a plasmid expressing flag-tagged cdk9 and after 48 h, protein extracts were prepared and utilized in immunoprecipitation and Western blot assay. As shown in Figure 5e, incubation of the extract with anti-p50 or anti-p65 antibodies caused precipitation of a complex which upon Western blot analysis using anti-flag antibody showed the presence of cdk9 protein in each complex (Figure 5d, see lanes 3 and 4). Direct Western blot analysis of the protein extract from the transfected cells showed the expression of Figure 3 Importance of kB motif of LTR in cdk9 stimulation of flag-tagged cdk9 and the corresponding band served as LTR. The human astroglial cell line, U-87 MG, was transfected a control in the gel (Figure 5d, lane 2). Under similar with 0.5 mg of various LTR deletion constructs as schematized conditions, this band was absent in the untransfected on the left and with 0.25 mg of the cdk9 expression plasmid by calcium – phosphate precipitation. After transfection, cells were cells verifying the specificity of this observation, and untreated or treated with PMA as described in the legend to Fig- indicating that cdk9 has the ability to physically ure 1. The values shown to the right represent the fold activation interact with p50 and, to a lesser extent, with p65. of cdk9 upon transcription of each deletion promoter construct as Thus, one may speculate that interaction of cdk9 with measured by the Luciferase assay. The cdk9 expression plasmid p50/p65 may result in the dissociation of p50/p65 from was created in pCEP-4-3 background plasmid. The empty pCEP-4-3 plasmid was added in all transfection mixtures, to nor- the kB motif and lead to a decrease in the level of malize the amounts of the DNA in each reaction. Luciferase ac- transcriptional activation of the LTR by NF-kB. tivity was determined 48 h after transfection. The data represent Indeed, this event has a reciprocal effect upon cdk9 the mean value of three separate transfection experiments. As de- function, as in the presence of p50 or p65, cdk9 loses scribed in Figure 1, the transfection efficiency was monitored by the level of GFP expression. Furthermore, equal amounts of pro- its ability to fully enhance the activity of the viral tein were utilized in each Luciferase assay. Each transfection was promoter. repeated three times to ensure its reproducibility In this study we show that cdk9, the catalytic subunit of human pTEFb, can stimulate the basal level of viral gene transcription via the kB motif. We the formation of the complex was used as a control. As also demonstrate that this activation which requires shown, while the intensity of the GC-rich associated the catalytic unit of cdk9 can be diminished once the complexes was slightly increased upon PMA treatment NF-kB subunits are overexpressed either by induction (Figure 4d, compare lanes 1 to 2 and lanes 3 to 4), with PMA or ectopic expression by plasmids leading overexpression of cdk9 showed no effect on the to speculation that the intracellular concentrations of formation of this complex, suggesting that the observed cdk9 and the NF-kB subunits are critical for the inhibition of NF-kB binding to its DNA motif by cdk9 observed regulatory events. The antagonistic effect of is not a general event. cdk9 and NF-kB was also observed in the presence

Oncogene Interaction of NF-kB with cdk9 controls HIV-1 K Khalili et al 5801

Figure 4 Interaction of p50 and/or p65 with the kB motif in the presence of cdk9. (a – c) Band shift assay was performed by using a 20-bp double-stranded synthetic DNA containing the kB motif of the LTR which was labeled at the 5’-end by 32P-g-ATP using T4 kinase. Approximately 105 c.p.m. DNA probe were incubated with 50 ng of p50 (a), 50 ng of p65 (b), and 50 ng of p50 and p65 (c) in the presence of 500 ng of GST (lane 1) or GST – cdk9 (lane 2). The p50 and p65 were bacterially produced, highly purified proteins that were kindly provided to us by Dr B Wigdhal (Penn State University, Hershey, PA, USA). The total amount of protein in each binding reaction remained constant (550 ng) by adding GST. The brackets depicted by C demonstrate complexes formed upon addition of p50, p65 or both. (d and e) Band shift assay using protein extract (10 mg) from untransfected cells or cells transfected with 15 mg of pCMV – cdk9 (56105 cells per 10 cm2 plate) which were maintained for 48 h after transfection in the absence or presence of PMA. The binding reaction was carried out upon incubation of the 32P-labeled DNA fragment containing the kB or GC- rich sequence of the HIV-1 LTR. The reaction contained 1 mg of the sheared salmon sperm DNA and 1 mg of poly(dI-dC) to pre- vent non-specific binding to the DNA probes. The positions of the NF-kB and GC-rich binding proteins are depicted by a C. The integrity of the complex was verified by supershift analysis using anti p50 antibodies (not shown). P, points to the position of the probe

Oncogene Interaction of NF-kB with cdk9 controls HIV-1 S Amini et al 5802

Figure 5 Interaction of cdk9, with p50 and p65. (a – d) GST-pull down assay was performed using in vitro translated 35S-labeled p50 (a), cdk9 (b and d), and p65 (c). The labeled proteins were mixed with either GST (lanes 2 of each panel) or GST-fusion proteins as indicated on top of lane 3 in each panel. Bound proteins were eluted and analysed by SDS – PAGE followed by autoradio- graphy. The arrows with molecular size of each protein point to the position of the speciﬁc bands that were bound to the GST- fusion proteins and eluted from each binding reaction. Lane 1 illustrates 1/10 of the in vitro translation reaction that was used in each binding reaction and represents a migration control. (e) Cell lysates were prepared from PMA-treated U-87MG cells and transfected with a plasmid expressing cdk9 fused to ﬂag epitope. Approximately 300 mg of the total protein extract were incubated with anti-p50 (lane 3) or anti-p65 (lane 4) antibodies overnight. The proteins were collected and resolved by SDS – PAGE along with 50 mg of protein extracts from untransfected cells (lane 1) or 50 mg of protein extracts from cells transfected with pCMV – cdk9-Flag plasmid (lane 2). The proteins were transferred to membrane and analysed by using anti-Flag antibody that re- cognizes cdk9-Flag protein. The position of the 38 kDa cdk9 is depicted by an arrow. Numbers on the left are molecular masses in kilodaltons

of the cdk9 partner, cyclin T1. Although ectopic through the association with IkBa (Sawaya B, study expression of cyclin T1 in the PMA treated cells led in progress). to stimulation of LTR transcription, suggesting a The association of cdk9 with cyclin T1 is important cooperativity between cyclin T1 and the NF-kB for activation of the HIV-1 promoter by the viral factors. This observation is consistent with earlier transactivator, Tat, through the TAR RNA sequence reports indicating that p-TEFb binds to NF-kBand (Wei et al., 1998; Zhu et al., 1997). It is believed that that the interaction between cyclin T1 and RelA phosphorylation of the carboxy terminal domain of augments LTR gene transcription (Barboric et al., RNA polymerase by cdk9 and cyclin T1 complex 2001). Our data from transcription assay as well as which is positioned in close proximity of the transcrip- DNA – protein and protein – protein interaction tion start site due to its association with Tat : TAR experiments suggest that cdk9, but not its mutant complex, is the underlying mechanism for Tat activa- with no kinase domain, has the ability to associate tion of the LTR (Garber et al., 1998; Zhou et al., with the 40 kDa IkBa protein (Sawaya et al., 1998). Recent studies have demonstrated that cdk9 can unpublished observations). While the signiﬁcance of associate with several other regulatory proteins includ- cdk9 interaction with IkB remains unknown, one ing TRAF (MacLachlan et al., 1998), b-myb (De Falco may speculate a cytoplasmic role for cdk9 in et al., 2000), ubiquitin (Kiernan et al., 2001), and 7SK modulating nuclear import of the NF-kB subunits RNA (Nguyen et al., 2001; Yang et al., 2001).

Oncogene Interaction of NF-kB with cdk9 controls HIV-1 K Khalili et al 5803 Moreover, several studies have indicated a functional such as cdk9, NF-kB, and others. As such, the current role for cdk9 and cyclin T1 in stimulating other models should consider the inclusion of other inducible promoters including b-myb (De Falco et al., 2000) cellular proteins that by communicating with each and TNFa (Darbinian et al., 2001). The ability of Tat other and with Tat mediate TAR-dependent and TAR- to stimulate TNFa, a cellular gene which lacks the independent regulation of viral and cellular genes. TAR RNA sequence, suggests that the TAR-indepen- Indeed, one should also consider the variations in the dent activation by Tat may utilize several pathways levels of these inducible proteins that may occur during which includes direct interaction of Tat with other the course of viral gene expression in response to regulatory proteins such as Pura (Darbinian et al., various external stimuli and phsyiologic conditions. 2001; Gallia et al., 1999; Krachmarov et al., 1996), or indirect contribution by interfering with the communication of NF-kB factors such as p65 with cdk9 as described in this report. Acknowledgments The authors wish to thank past and present members of the The picture which emerges from these observations Center for Neurovirology and Cancer Biology for their suggests that the regulatory events involved in the insightful discussion, and sharing of ideas and reagents. We basal and Tat-induced transcription of the viral and thank C Schriver for preparation of this manuscript. This cellular genes are highly complicated and are mediated work was made possible by grants awarded by NIH to K through the interplay between various cellular proteins Khalili, BE Sawaya and S Amini.

References

Arenzana-Seisdedos F, Thompson J, Rodriguez MS, Krachmarov CP, Chepenik LG, Barr-Vagell SM, Khalili K Bachelerie F, Thomas D and Hay RT. (1995). Mol. Cell and Johnson EM. (1996). Proc. Natl. Acad. Sci. USA, 93, Biol., 15, 2689 – 2696. 14112 – 14117. Bandres JC and Ratner L. (1994). J. Virol., 68, 3243 – 3249. KumarS,OrsiniMJ,LeeJC,McDonnellPC,DebouckC Barboric M, Nissen RM, Kanazawa S, Jabrane-Ferrat N and and Young PR. (1996). J. Biol. Chem., 271, 30864 – 30869. Peterlin BM. (2001). Mol. Cell, 8, 327 – 337. Lindholm PF, Marriott SJ, Gitlin SD, Bohan CA and Brady Beg AA, Finco TS, Nantermet PV and Baldwin Jr AS. JN. (1990). New Biol., 2, 1034 – 1043. (1993). Mol. Cell. Biol., 13, 3301 – 3310. MacLachlan TK, Sang N, De Luca A, Puri PL, Levrero M Berkhout B, Silverman RH and Jeang KT. (1989). Cell, 59, and Giordano A. (1998). J. Cell. Biochem., 71, 467 – 478. 273 – 282. Nguyen VT, Kiss T, Michels AA and Bensaude O. (2001). Darbinian N, Sawaya BE, Khalili K, Jaffe N, Wortman B, Nature, 414, 322 – 325. Giordano A and Amini S. (2001). J. Neuroimmunol., 121, Pereira LA, Bentley K, Peeters A, Churchill MJ and Deacon 3 – 11. NJ. (2000). Nucleic Acids Res., 28, 663 – 668. De Falco G, Bagella L, Claudio PP, De Luca A, Fu Y, Price DH. (2000). Mol. Cell. Biol., 20, 2629 – 2634. Calabretta B, Sala A and Giordano A. (2000). Oncogene, Rostasy K, Monti L, Yiannoutsos C, Wu J, Bell J, Hedreen J 19, 373 – 379. and Navia BA. (2000). J. Neurovirol., 6, 537 – 543. Foskett SM, Ghose R, Tang DN, Lewis DE and Rice AP. Taylor JP, Pomerantz R, Bagasra O, Chowdhury M, (2001). J. Virol., 75, 1220 – 1228. Rappaport J, Khalili K and Amini S. (1992). EMBO J., Gallia GL, Darbinian N, Tretiakova A, Ansari SA, 11, 3395 – 3403. Rappaport J, Brady J, Wortman MJ, Johnson EM and Taylor JP, Pomerantz R, Raj G, Kashanchi F, Brady J, Khalili K. (1999). Proc. Natl. Acad. Sci. USA, 96, 11572 – Amini S and Khalili K. (1994). J. Virol., 68, 3971 – 3981. 11577. WeiP,GarberME,FangSM,FischerWHandJonesKA. Garber ME, Wei P and Jones KA. (1998). Cold Spring Harb. (1998). Cell, 92, 451 – 462. Symp. Quant. Biol., 63, 371 – 380. Yang Z, Zhu Q, Luo K and Zhou Q. (2001). Nature, 414, Garriga J, Mayol X and Grana X. (1996). Biochem. J., 319, 317 – 322. 293 – 298. Zhou Q, Chen D, Pierstorff E and Luo K. (1998). EMBO J., Hiscott J, Beauparlant P, Crepieux P, DeLuca C, Kwon H, 17, 3681 – 3691. Lin R and Petropoulos L. (1997). J. Leukoc. Biol., 62, 82 – ZhuY,Pe’eryT,PengJ,RamanathanY,MarshallN, 92. Marshall T, Amendt B, Mathews MB and Price DH. Kiernan RE, Emiliani S, Nakayama K, Castro A, Labbe JC, (1997). Genes Dev., 11, 2622 – 2632. Lorca T, Nakayama Ki K and Benkirane M. (2001). Mol. Cell. Biol., 21, 7956 – 7970.

Oncogene