Antiviral activity of CYC202 in HIV-1-infected cells.

There are currently 40 million individuals in the world infected with human immunodeficiency virus (HIV). The introduction of highly active antiretroviral therapy (HAART) has led to a significant reduction in AIDS-related morbidity and mortality. Unfortunately, up to 25% of patients discontinue their initial HAART regimen. Current HIV-1 inhibitors target the fusion of the virus to the cell and two viral proteins, reverse transcriptase and protease. Here, we examined whether other targets, such as an activated transcription factor, could be targeted to block HIV-1 replication. We specifically asked whether we could target a cellular kinase needed for HIV-1 transcription using CYC202 (R-roscovitine), a pharmacological cyclin-dependent kinase inhibitor. We targeted the cdk2-cyclin E complex in HIV-1-infected cells because both cdk2 and cyclin E are nonessential during mammalian development and are likely replaced by other kinases. We found that CYC202 effectively inhibits wild type and resistant HIV-1 mutants in T-cells, monocytes, and peripheral blood mononuclear cells at a low IC(50) and sensitizes these cells to enhanced apoptosis resulting in a dramatic drop in viral titers. Interestingly, the effect of CYC202 is independent of cell cycle stage and more specific for the cdk2-cyclin E complex. Finally, we show that cdk2-cyclin E is loaded onto the HIV-1 genome in vivo and that CYC202 is able to inhibit the uploading of this cdk-cyclin complex onto HIV-1 DNA. Therefore, targeting cellular enzymes necessary for HIV-1 transcription, which are not needed for cell survival, is a compelling strategy to inhibit wild type and mutant HIV-1 strains.

As of the end of 2001, there were an estimated 40 million people living with human immunodeficiency virus type 1 (HIV-1) 1 globally. The resurgence and development of an additional 16,000 new HIV-1 infections were calculated to take place daily, with 95% of these cases occurring in developing countries and 50% in women (1,2). Only two viral proteins, reverse transcriptase and protease, are currently targets for HIV-1 inhibitors. However, virologic failure of antiretroviral therapy is common and primarily associated with the emergence of drug resistance (3-6). The emergence of resistance when viral loads are low during HAART or other antiviral treatments is explained by the selection for those viral variants that are less fit for replication as evident in the quasi-species theory, where viral populations are a mixture of distinct but related variants; the dominant wild type strain and mutants with varying degrees of fitness, i.e. mutants that are "over" and "under" fit (7). Therefore, it is widely believed that the success in HIV-1 treatment will require targeting of other HIV and/or host cellular proteins. To that end, we selected to focus on HIV-1 transcription events postintegration and its functional and physical interaction with the host cell cycle machinery.
Although HIV-1 entry, an early stage of infection, occurs regardless of cell cycle status, active high titer virion production of either M-or T-tropic viruses requires the host cell to be at the late G 1 or S phase of the cell cycle (8 -28). Therefore, activation (a process where cells go from G 0 to the G 1 /S stage of the cell cycle) of latently infected cells by various stimuli leads to viral expression followed by progeny formation and cell death. Establishment of the latent infection is mainly the result of the absence of necessary host factors that are present only in induced cells, such as activated nuclear factor-B (29 -31), AP-1 (32), and CBP/p300 (33), and therefore minimal transcription from the HIV-1 long terminal repeat (LTR) occurs in G 0 quiescent cells.
Tat has been shown to interact with a number of cellular factors, which may or may not be important for HIV-1-specific activated transcription (33). However, one complex, Tat-cdk9cyclin T1 (p-TEFb), has emerged as a significant binding partner for elongation of transcription from the LTR. This complex has been shown to associate with and phosphorylate the carboxyl-terminal domain (CTD) of RNA pol II, thereby enhancing elongation of transcription (34 -39). In vivo treatment of peripheral blood lymphocytes and purified resting CD4 ϩ T-lymphocytes with phorbol 12-myristate 13-acetate, phytohemagglutinin, interleukin-2, interleukin-6, or TNF-␣ also results in an increase in cdk9 and cyclin T1 protein levels and an increase in TAK (Tat-associated kinase) enzymatic activity (38). Furthermore, cdk9 expression during the cell cycle is periodic and peaks at the G 1 /S border (40). * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Recently, a growing body of evidence has indicated the role of yet another cdk-cyclin complex, namely cdk2-cyclin E, in Tatactivated transcription. Cdk2-cyclin E is the major cdk-cyclin complex whose maximal activity is observed at the late G 1 /S boundary regulating the release of Rb sequestered factors, including E2F (41). We and others have shown previously that the increased kinase activity of cdk2-cyclin E complexes in HIV-1 latently infected cells is the result of the loss of p21/waf1 (42). Cdk2-cyclin E isolated from phytohemagglutinin-treated human primary T-cells phosphorylated RNA pol II CTD (43). Studies revealed that maximal activity of the kinase was during the late G 1 phase of the cell cycle, which coincides with the maximal activity of the cdk2-cyclin E complex (43,44) and that cdk2-cyclin E has been shown to be necessary for Tat-dependent transcription in vitro and in reconstituted transcription assays (45). Furthermore, it has been shown that during late G 1 , Tat transactivation is transactivation response-dependent, whereas at G 2 , Tat transactivation is transactivation responseindependent (46). Here, we asked whether we could target a cellular kinase needed for HIV-1 transcription using CYC202 (R-roscovitine), a pharmacological cyclin-dependent kinase inhibitor (PCI). We specifically decided to target the cdk2-cyclin E complex in HIV-1-infected cells because both cdk2 and cyclin E are not essential for viable mammalian development, as demonstrated by mouse knock-out models (47)(48)(49)(50). Therefore, targeting cellular enzymes necessary for HIV-1 transcription, which are not essential for the survival of the vast majority of uninfected cells, is a compelling strategy to inhibit wild type and mutant HIV-1 strains. We found that CYC202 effectively inhibits wild type and resistant HIV-1 mutants in T-cells, monocytes, and PBMCs at a low IC 50 . Interestingly, the effect of CYC202 is independent of cell cycle stage and more specific for the cdk2-cyclin E complex. Finally, we show that cdk2cyclin E is in fact loaded onto the HIV-1 genome in vivo and that CYC202 is able to inhibit the uploading of this cdk-cyclin complex onto HIV-1 DNA.

EXPERIMENTAL PROCEDURES
Cells and Reagents-OM10.1, ACH2, and U1 cells are HIV-1-infected lymphocytic and monocytic cells, respectively. OM10.1 and ACH2 contain a single integrated copy of a wild type genome in parental CEM (12D7) cells, whereas U1 cells harbor two copies (one wild type and one mutant) of the viral genome in parental U973 cells (46). HLM1 cells (HIV-1 ϩ /Tat Ϫ (51)) contain a single copy of full-length HIV-1 provirus with a triple termination codon at the first AUG of the Tat gene. These cells can be made to produce HIV-1 virion in the presence of wild type Tat, TNF-␣, or sodium butyrate. MT-2 cells are infected with several copies of human T-cell leukemia virus type 1 (HTLV-1) and produce full-length viral particles. CEM and H9 (T-lymphocytes), HL-60 (promyelocytic), and THP-1 (monocytic) cells are uninfected and have been described previously (46). All cells were cultured at 37°C with up to 10 5 cells/ml in RPMI 1640 medium (except HLM-1 in Dulbecco's modified Eagle's medium) containing 10% fetal bovine serum and treated with a mixture of 1% streptomycin and penicillin antibiotics and 1% L-glutamine (Invitrogen).
Tat protein (both wild type and K41A mutant) was produced in Escherichia coli and purified using Sephacryl S-200 followed by C18 reversed phase high pressure liquid chromatography. The purified Tat proteins were then dried and resuspended in phosphate-buffered saline (PBS) containing 1 mM dithiothreitol and 0.01% bovine serum albumin prior to use. HLM1 cells (HIV-1 ϩ /Tat Ϫ ) were blocked in complete medium with low serum (0.1% fetal bovine serum) for 3 days. Cells were subsequently released in the presence of complete medium (10% fetal bovine serum) and transfected with 1 g of either purified wild type or mutant Tat protein. Next, cells were plated with 0.5 M CYC202 or m-CYC202, 100 nM wortmannin, 10 M Lys 294002, 5 g of wild type Tat peptide (CTKALGC-GG-YGRKKRRQRRR), or 5 g of a dominant negative Tat peptide with mutations at positions 41 and 44 (CTAALSC-GG-YGRKKRRQRRR). The cyclized Tat peptides contain two cysteines at amino and carboxyl termini of each peptide, followed by the addition of two glycine residues for flexibility, and the Tat basic nuclear trans-location signal sequence. Cells were kept at 37°C for 12 days, and supernatants were collected for p24 ELISA.
Kinase Assays-Kinase assays were performed with immunoprecipitates from both infected and uninfected cells using anti-cdk2-cyclin E, anti-cdk4-cyclin D, anti-cdk7-cyclin H, and anti-cdk9-cyclin T antibodies in a 20-l reaction volume containing 50 M ATP, 1 Ci of [␥-32 P]ATP, and 100 ng of GST-CTD or 200 ng of histone H1 in TTK kinase buffer containing 50 mM HEPES (pH 7.9), 10 mM MgCl 2 , 6 mM EGTA, and 2.5 mM dithiothreitol for 30 min at 30°C. Phosphorylated GST-CTD or histone H1 was resolved on 4 -20% SDS-PAGE and subjected to autoradiography and quantitation with a PhosphorImager.
Phosphorylation of p53 by GSK3␤ was carried out according to a procedure published previously (53). Phosphorylation of serine 37 of p53 by DNA-PK has been shown to create a site for GSK3␤ phosphorylation at serine 33 in vitro. Therefore, 3 g of wild type GST-p53 (42) was incubated with DNA-PK (20 units, Promega Corp., Madison, Wl) for 5 h at 30°C in the following buffers (40 l final volume): 25 mM HEPES (pH 7.5), 150 mM KCI, 10 mM MgCl 2 , 20% glycerol, 0.1% Nonidet P-40, 20 M ZnCl 2 , 1 mM dithiothreitol, 250 ng of DNA, 4.2 mM spermidine, 10 M ATP. Samples were then incubated at 65°C for 20 min to inactivate DNA-PK and the reaction subsequently incubated for 30 min with GSK3␤ in GSK3 kinase buffer (KGB buffer: 8 mM MOPS (pH 7.2), 10 mM MgCl 2 , 0.2 mM EDTA, 5 M ATP, 10 Ci of [ 32 P]ATP) in a final volume of 40 l. Reactions were terminated by the addition of SDS sample buffer and GST-p53 phosphorylation was detected by SDS-PAGE and autoradiography.
For the peptide kinase assays, we used our previously published procedure (54). Whole cell lysates were prepared from OM10.1 cells and PBMCs in IP buffer. Lysates (2 mg) were treated with protein A-Sepharose CL-4B (Sigma) to avoid nonspecific binding and were centrifuged. The supernatants were incubated with antibodies against cdk2 and cdk4 and then with protein A-and G-agarose beads. After centrifugation, the immunoprecipitates were washed five times with IP buffer. The immunopurified cyclins and substrates were incubated at 30°C for 30 min in R buffer (20 mM Tris-HCl (pH 7.4), 10 mM MgCl 2 , 4.5 mM 2-mercaptoethanol, 1 mM EGTA) that contained 50 M ATP and 10 mCi of [␥-32 P]ATP (6,000 Ci/mmol; Amersham Biosciences) in a final volume of 25 l with 5 g of Rb peptides (J and G1), incubated for 30 min, and run on an 18% SDS-PAGE, dried, and exposed to a PhosphorImager cassette. The Rb peptide J sequence was: 818 GLPTPTKMTPRSRIL, and peptide G1 was 775 RPPTLSPIPHIPR.
Immunoblotting-Cells were pelleted by centrifugation, washed with PBS without Ca 2ϩ and Mg 2ϩ , and lysed with lysis buffer (55). The lysate was incubated on ice for 15 min and centrifuged at 4°C for 10 min. Total cellular protein was separated on 4 -20% Tris-glycine gels (Invitrogen) and transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore Corp.) overnight at 0.08 A. After transfer, the blots were blocked with 5% nonfat dry milk in TNE 50 (100 mM Tris-HCl (pH 8.0), 50 mM NaCl, 1 mM EDTA) plus 0.1% Nonidet P-40. Membranes were probed with a 1:200 to 1:1,000 dilution of antibodies at 4°C overnight, followed by three washes with TNE 50 plus 0.1% Nonidet P-40. The next day, the blots were incubated with 10 ml of 125 I-labeled protein G (Amersham Biosciences; 50-l/10 ml solution) in TNE 50 plus 0.1% Nonidet P-40 for 2 h at 4°C. Finally, the blots were washed twice in TNE 50 plus 0.1% Nonidet P-40 and placed on a PhosphorImager cassette for further analysis.
MTT Viability Assay-The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) viability assay has been described previously (56). Briefly, cells were counted, and 10 5 cells/well in flat bottom 96-well tissue culture plates were used for each MTT assay. Medium containing 75 g/ml MTT was added, and the cells were incubated for 1.5 h at 37°C and 5% CO 2 . The resulting formazan crystals were solubilized in 115 l of 0.04 N HCl in isopropyl alcohol. The optical density of the solubilized formazan was read at 575 and 650 nm with a SpectraMax 340 (Molecular Devices, Sunnyvale, CA) plate reader. The reduction in optical density was used as a measurement of cell viability, normalized to cells incubated in control medium, which were considered 100% viable.
Flow Cytometry-For cell cycle analysis, cells treated with or without drugs were collected by low speed centrifugation and washed with PBS without Ca 2ϩ and Mg 2ϩ and then fixed with 70% ethanol. For fluorescence-activated cell sorting (FACS) analysis, cells were stained with a mixture of propidium iodide buffer (PBS with Ca 2ϩ and Mg 2ϩ , 10 g/ml RNase A, 0.1% Nonidet P-40, and 50 g/ml propidium iodide) followed by cell sorting analysis. The acquired FACS data were analyzed by ModFit LT software (Verity Software House, Inc.). Cells were washed twice with cold PBS without Ca 2ϩ and Mg 2ϩ , resuspended in 1ϫ binding buffer (10 mM HEPES-NaOH (pH 7.4), 140 mM NaCl, 2.5 mM CaCl 2 ) and 5 l of propidium iodide/10 5 cells, and incubated at room temperature for 15 min. Cells were acquired and analyzed using CELLQuest software (BD Biosciences).
DNA Fragmentation Assay-CYC202-treated and untreated cells were lysed with buffer containing 10 mM Tris-HCl (pH 7.6), 1 mM EDTA, and 0.2% Triton X-100, and the fragmented DNA in the lysate was separated from the unfragmented chromosomal DNA by precipitation at 12,000 ϫ g for 30 min. Fragmented DNA in the supernatant was then digested with 100 ng/ml ribonuclease A (Invitrogen), 20 ng/ml protease K (Invitrogen), and 1% SDS at 37°C for 45 min, purified by phenol/chloroform extraction, and precipitated with ethanol/sodium deoxycholate. The DNA was then electrophoresed on a 1.6% agarose gel containing 0.5 g/ml ethidium bromide.
PBMC Infection-Phytohemagglutinin-activated PBMCs were kept in culture for 2 days prior to each infection. Isolation and treatment of PBMCs were performed by following the guidelines of the Centers for Disease Control (55). Approximately 5 ϫ 10 6 PBMCs were infected with either an SI (UG/92/029 Uganda strain, subtype A envelope, 5 ng of p24 gag antigen) or an NSI (THA/92/001, Thailand strain, subtype E envelope, 5 ng of p24 gag antigen) strain of HIV-1. Other HIV-1 mutant viruses (AZT, 3TC, TIBO, and protease) were also used for PBMC infections (50 ng of p24 gag antigen/virus). All viral isolates were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program. After 8 h of infection, cells were washed and fresh medium was added. Drug treatment was performed (only once) immediately after the addition of fresh medium. Samples were collected every 6th day (for SI and NSI viruses) and stored at 20°C for p24 gag ELISA. For HIV-1 p24 ELISA, media from infected cell lines were centrifuged to pellet the cells, and supernatants were collected and diluted to 1:100 to 1:1,000 in RPMI 1640 prior to analysis. Supernatants from the infected PBMCs were collected and used directly for the p24 antigen assay. The p24 gag antigen level was analyzed using the HIVAG-1 monoclonal antibody kit (Abbott Laboratories, Diagnostics Division).
siRNA Analysis-siRNA sequences were designed using the Oligoengine Work station (www.oligoengine.com) and were purchased from Qiagen-Xeragon. Candidate sequences were chosen based on general siRNA design criteria, including a % GC content between 45 and 55% and avoiding more than three consecutive guanosines. Selected target sequences were also BLASTed (www.ncbi.nlm.nih.gov/BLAST/) with a standard nucleotide-nucleotide BLAST to ensure they were not homologous to other genes. Each candidate siRNA was generated from the 5Ј-end and consisted of 19 nucleotides with a d(TT) overhang. The following genes were chosen for siRNA analysis with the GenBank accession numbers in parentheses: cdk2 (AF512553), cdk4 (U37022), cdk7 (BC001968), and cdk9 (AF517840). Two candidate siRNAs were chosen for each of the four genes to ensure silencing of protein expression. For each duplex siRNA, the first sequence represents the sense sequence, and the second, the antisense sequence.
OM10.1 (5 ϫ 10 6 cells at mid-log phase of growth) cells were treated with TNF-␣ for 2 h, washed, and subsequently electroporated with siRNA against cdk2, 4, 7, and 9 (10 g each). 48 h later, cells were cross-linked and processed for ChIP assay, and supernatants saved for p24 ELISA. PCRs were performed for both the promoter (5Ј-LTR) and Env gene (3Ј-most open reading frame).

Inhibition of HIV-1 in T-cells, Monocytes, and PBMCs in the
Presence of CYC202-We reported previously that in latently HIV-1-infected cells, the expression of a number of cdk-cyclin complexes can be targeted to inhibit viral progeny formation (55). We chose several ATP analogs, including roscovitine, which was a mixture of two R-and S-substituted purine enantiomers. In this report, we specifically used R-roscovitine (CYC202) and the acid metabolite of CYC202 known to have reduced kinase activity compared with the parent compound (m-CYC202) to test further the effect of this cdk inhibitor in HIV-1-infected cells. To test the effect of CYC202, we initially used HIV-1 latently infected OM10.1 and U1 cells. OM10.1 cells are T-cells with a single integrated copy of wild type HIV-1 genome, and U1 is a monocytic cell line containing one integrated copy of wild type and one integrated copy of a mutant HIV-1 genome. Both cell lines can be induced with various mitogens including TNF-␣, interleukin-2, and phytohemagglutinin to produce HIV-1 virions. We first induced OM10.1 and U1 cells with TNF-␣ for 2 h, washed, and subsequently added complete medium with either CYC202 (0.1, 1, 10 M) or m-CYC202 (0.1, 1, 10 M). Cells were grown for 5 days, and supernatants were collected and analyzed for p24 HIV-1 gag using ELISA. TNF-␣ induced HIV-1 production in OM10.1 cells (Fig. 1A, compare lanes 1 and 2) as well as in U1 cells (Fig. 1A, compare lanes 9 and 10). CYC202 suppressed viral production in a dose-dependent manner in both cell types (Fig. 1A, lanes 3-5 and lanes [11][12][13]. In contrast, m-CYC202 did not suppress HIV-1 production at any of the concentrations examined ( and lanes 14 -16). It has been well established that HIV-1 replicates at lower titers in monocytes during a short infection period, which may partly explain why these cells do not produce robust levels of virus after 5 days. Interestingly, a dual tropic virus, 89.6, was able to infect both CEM and THP cells (Fig. 1B, lanes 18 and 22), and CYC202 inhibited viral production in both cell types (Fig. 1B, lanes 19 -21 and 23-25). To determine the effect of CYC202 on HIV-1 infection in primary cells, PBMCs were infected with two field isolates of virus, one that preferentially induced syncytium (SI, THA/92/00) and one that was a non-syncytium inducing (NSI, UG/92/00) virus (55). Activated PBMCs were efficiently infected with either SI or NSI strains (Fig. 1C), and CYC202 but not m-CYC202 effectively inhibited SI or NSI viral production (Fig. 1C). The kinetics of viral replication in PBMC showed maximal viral progeny at day 18. Both 1 and 10 M concentrations of CYC202 inhibited NSI virus production, but only 10 M concentrations of CYC202 inhibited SI virus production. No inhibition was seen with these viral strains at 0.1 M CYC202 (data not shown). Collectively, these data indicate that CYC202 is a specific inhibitor of HIV-1 during latent and acute viral infection in cultured cell lines and in primary cells.
Titration of CYC202 and Determination of IC 50 in Infected Cells-Results presented in the previous section demonstrated that CYC202 inhibited various HIV-1 strains. To determine accurately the IC 50 of CYC202, we performed a closer titration curve for CYC202 in activated OM10.1 and U1 cells that were treated with various concentrations of either CYC202 or m-CYC202 (0.1, 0.5, 1, 5, 10, 15, 20, 40, 60, 80, and 100 M). Cells were cultured for 5 days in the presence of the drug, and supernatants were collected and examined for HIV replication using the p24 ELISA. The IC 50 of CYC202 in T-cells (OM10.1) was determined to be 0.36 M, whereas in monocytes (U1) it was 1.8 M. The IC 50 for m-CYC202 was estimated to be greater than 1 mM because no inhibition was observed ( Fig.  2A). These results indicate that CYC202 may be a better inhibitor in cells harboring a single copy of HIV-1 (OM10.1) as opposed to two genome copies (U1). We do not believe that this discrepancy is the result of cell type differences because both cell types had similar kinase activity in an in vitro kinase assay for cdk2, cdk7, and cdk9 enzymes (data not shown). Finally, to determine the cellular cytotoxicity of CYC202 and m-CYC202 in both T-cells and monocytes, we performed similar titrations in uninfected CEM, THP1 cells, and two PBMCs from two separate donors. The IC 50 for CYC202 was determined to be at 28 M for T-cells (CEM), 35 M for monocytes (THP1), and 22 and 23 M for activated PBMCs using MTT assays (Fig. 2, B and C). In contrast, the IC 50 for m-CYC202 was 402 M for T-cells (CEM), 690 M for monocytes (THP1), and 139 and 97 M for activated PBMCs (Fig. 2, B and C). Collectively, these results indicate that viral inhibition was significantly more sensitive to CYC202 than inhibition of cell growth, requiring 20-fold and 80-fold less drug in T-cells and monocytes, respectively.

CYC202 Inhibits HIV-1 Transcription in Cells Undergoing
Apoptosis-We next examined the effect of CYC202 on apoptosis of both HIV-1-infected and uninfected cells. We designed two sets of experiments, one with OM10.1 cells induced with TNF-␣ and subsequently treated with CYC202 at an IC 50 of 0.36 M, and the other with PBMCs (from donor 1, Fig. 2C) infected with the NSI strain and treated with CYC202 at an IC 50 of 1.0 M. In both systems, cells were carried out for an extended period of time. Results are shown in Fig. 3. When OM10.1 cells were treated with TNF-␣ to induce HIV-1 transcription, increased levels of HIV-1 Env RNA were seen 3 days postinduction (Fig. 3A, lane 2, RT/PCR/Env). At the same time low levels of caspase-3 were also observed. By day 6 there was a high level of PARP degradation (indicator of apoptosis), and no change in actin protein or RNA was observed. These results indicate that HIV-1 alone causes apoptosis in infected cells after reactivation. We next examined the effect of CYC202 in TNF-␣-treated OM10.1 cells. As expected, we saw a dramatic decrease in the level of HIV-1 RNA in these cells after CYC202 treatment (Fig. 3A). There was no change in actin RNA or protein levels. As expected, there was activation of caspase-3 and PARP cleavage in these cells. Therefore, the resulting apoptosis in these cells was the result of activation of the virus and not treatment with CYC202. Finally, we performed the same experiment in CEM uninfected cells treated with TNF-␣ and carried the experiment up to 6 days using a similar CYC202 concentration (IC 50 of 0.36 M). We observed no PARP cleavage, caspase activation, or actin degradation in these cells. These results imply that CEM cells treated with TNF-␣ and CYC202 (at IC 50 of 0.36 M) do not cause any apparent apoptosis.
We next performed a similar experiment in PBMC cells. Donor 1 (Fig. 2C) was used for infection with the THA/92/00 NSI strain (similar to Fig. 1C), subsequently treated with CYC202 (IC 50 of 1 M), and samples were taken at 6 and 12 days. Similar to OM10.1 results, PBMC cells infected with the NSI strain showed Env expression and PARP degradation by day 6 (Fig. 3B), indicating that HIV-1 infection caused apoptosis in these cells. When infected PBMCs were treated with CYC202, there was a complete drop in Env expression by day 6 and an 80% drop by day 12. We then performed the control experiment with CYC202 using the same PBMCs. As expected we observed no toxicity in these cells when using 1 M CYC202.
To determine further the level of apoptosis in both OM10.1 and  0 M (B). A, OM10.1 cells treated with TNF-␣ were carried out to day 6. Samples were collected for days 0, 3, and 6 and processed for Western blots (100 g of total protein extract) or reverse transcription-PCR PBMCs, we performed a number of assays including FACS and DNA laddering in both infected and uninfected cells. Fig. 3C shows results of all treatments at day 6 for OM10.1 and CEM as well as day 12 for PBMC infected and uninfected cells. OM10.1 cells induced with TNF-␣ showed an apoptotic population (16.03%), which was slightly elevated (27.19%) in CYC202-treated cells. No change in apoptotic cell populations was observed in CEM cells. Similarly, PBMCs infected with the NSI strain of HIV showed an enhanced apoptotic cell population (29.81%) compared with uninfected cells (3.08%). CYC202 treatment showed similar results in infected cells and no change in uninfected cells. Finally, we extracted nuclear DNA from these cells and ran it on a 1.6% agarose gel. As seen in Fig.  3D, OM10.1, OM10.1 ϩ CYC202, as well as infected and drug-treated PBMCs showed the typical laddering pattern indicative of apoptosis. However, no DNA laddering was observed in control, uninfected cells. Collectively, these results indicate that CYC202 treatment at anywhere from 0.3 to 1.0 M doses inhibit HIV-1 transcription but has no effect in uninfected cell lines or PBMCs.
Effect of CYC202 in Cells Infected with Resistant HIV-1 Viruses-HIV-1 rapidly mutates in vivo and during culturing in vitro. Selection of mutant strains usually occurs either as a result of immune mediated positive selection or treatment with various HIV-1 inhibitors. To date, there are many forms of escape mutants related to all of the HIV-1 open reading frames; however, much attention has been given to clinically relevant strains harboring mutations in reverse transcriptase and protease. We therefore decided to test the effect of CYC202 on mutants resistant to nucleoside (AZT, 3TC) and nonnucleoside (TIBO) inhibitors as well as a protease-resistant virus. The resistant viral strains used in this study were initially grown in the presence of appropriate inhibitors (i.e. AZT, 3TC), and their resistance status has already been determined (3, 58 -61). Activated PBMCs were infected with TIBO, 3TC, or proteaseresistant viruses, and MT-2 cells were infected with an AZTresistant virus. After infection, cells were treated with either CYC202 or m-CYC202 (0.1, 0.5, 1, 5 and 10 M). Infected PBMCs were cultured for up to 12 days and MT-2 infected cells were cultured for up to 5 days. MT-2 cells (HTLV-1-infected) were cultured for a shorter period because they also contained the Tax transactivator, which induces nuclear factor-B and CBP/p300, and thus additionally enhances HIV-1 transcription. Treatment with CYC202, but not with m-CYC202, consistently inhibited all mutant viruses tested (Fig. 4, A-D). The IC 50 values were estimated to be 0.5 M (protease-resistant strain; Fig. 4D), 1.0 M (TIBO-and 3TC-resistant strains, Fig.  4, B and C, respectively), and 5 M (MT-2 cells with the AZTresistant virus, Fig. 4A). The higher IC 50 observed in MT-2 cells (HTLV-1-positive) could be a result of Tax-mediated induction of NF-B and CBP/p300, which allows a higher titer of virus (ng/ml versus pg/ml p24 levels, Fig. 4A), and thus more drug is needed to inhibit the transcription of HIV-1. This result also indicates that other viral coinfections may require higher levels of CYC202 to treat HIV-1 infection. Collectively, these data suggest that CYC202 is an effective inhibitor of HIV-1 mutant viruses regardless of whether the mutations are in reverse transcriptase or protease.
Effect of CYC202 Is Independent of Cell Cycle Stage-We have shown previously that TAR-specific activated transcription initially takes place at the G 1 /S border (46) and that cells blocked at G 1 /S with hydroxyurea have virtually undetectable titers compared with cells that have passed the checkpoint and have reached the G 2 /M phase (62). Therefore, we asked whether CYC202 could inhibit HIV-1 production in cells prior to, or after the restriction point in the mid-G 1 phase, where Rb (5 g of total RNA) for both HIV-1 envelope (Env) and actin. A shows the results from OM10.1 treated cells alone (left column) or OM10.1 cells treated with CYC202. Protein extracts were processed for PARP, caspase-3, and actin Western blots. The right panel shows similar experiments with TNF-␣-treated CEM either alone or with CYC202. The percents of trypan blue-positive cells for each treatment were as follows: OM10.1: 3%, 12%, and 74% for days 0, 3, and 6. OM10.1 ϩ CYC202: 1%, 29%, and 68% for days 0, 3, and 6. CEM: 1%, 3%, 5%. CEM ϩ CYC202: 0%, 3%, 4%. Two independent counts were performed for each treatment. Primers for reverse transcription-PCR for Env and active have been described previously (95). B, activated PBMCs from donor 1 (Fig. 2C) were used for infection with the THA/92/00 NSI strain (similar to Fig. 1C) and subsequently treated with CYC202 at an IC 50 of 1 M and carried out to day 12. Note that there were no TNF-␣ treatments with PBMCs. Cells were processed, and Western blots and reverse transcription-PCR were done as outlined in A. The percents of trypan blue positive cells were as follows: PBMC ϩ NSI: 6%, 36%, and 58% for days 0, 6, and 12. PBMC: 4%, 7%, and 12% for days 0, 6, and 12. PBMC ϩ CYC202: 1%, 4%, 3% for days 0, 6, and 12. Two independent counts were performed for each treatment. C, FACS analysis of infected and treated cells. Cells were collected by low speed centrifugation and washed with PBS without Ca 2ϩ and Mg 2ϩ and then fixed with 70% ethanol. Subsequently, cells were stained with a mixture of propidium iodide buffer followed by cell sorting analysis. FACS data acquired were analyzed by ModFit LT software. The panel shows results of all treatment at day 6 for OM10.1 and CEM as well as day 12 for infected and uninfected PBMCs. D, a DNA fragmentation assay was performed according to a previously published procedure (96). CYC202-treated and untreated cells at day 6 for OM10.1 and day 12 for PBMCs were lysed with buffer, and the fragmented DNA in the lysate was separated from the unfragmented chromosomal DNA, purified, and electrophoresed on a 1.6% agarose gel containing ethidium bromide.

FIG. 3-continued
protein is hyperphosphorylated. To do this, we used HLM-1 cells, which contain a single integrated copy of HIV-1 with a mutation in the Tat gene (triple termination codon). The advantage of using these cells is that they can easily be blocked at any phase of the cell cycle, released, and virus could be produced by introduction of the Tat gene in trans. Therefore, to test the effect of CYC202 at various phases of the cell cycle, we designed two sets of experiments. First HLM-1 cells were serum starved for 3 days to induce cell stasis at the G 0 phase and subsequently treated with 1 g of purified Tat protein. To introduce inhibitors in the mid-G 1 phase or past the G 1 /S checkpoint, cells were either immediately treated with CYC202, DNA-PK inhibitor, wild type Tat peptide, or a mutant dominant negative Tat peptide (Fig. 5A) or treated 16 h postinduction with Tat (Fig. 5C). Treated cells were cultured for 12 days, and p24 ELISAs were then performed on the supernatants. As expected, wild type Tat protein but not a mutant form (Tat, K41A) induced viral production in HLM-1 cells (Fig. 5, B  and D, lanes 2 and 3). The addition of CYC202, but not m-CYC202, inhibited viral production when added during the mid-G 1 phase (Fig. 5B, lanes 4 and 5). Two different DNA-PK inhibitors, as well as the dominant negative Tat peptide, reduced viral production only when added during the mid-G 1 phase (Fig. 5B, lanes 6, 7, and 9). When cells were allowed to pass the G 1 /S border (16 h postrelease), neither the DNA-PK inhibitors nor the Tat peptide inhibitors affected viral production (Fig. 5D, lanes 6, 7, and 9).
To ensure that cells were in fact at the G 0 phase of the cell cycle and subsequently entered the G 1 /S after the addition of serum, we performed two independent experiments. To verify that the majority of cells were arrested in G 0 , transcription factor binding to the cyclin A promoter was analyzed. Takahashi et al. (63), utilizing ChIP, examined the cyclin A promoter at various stages of the cell cycle and demonstrated that at G 0 and early G1, the cyclin A gene was repressed by having two repressor factors, namely E2F4 and p130, bound to the promoter. We have also previously utilized a similar approach for detecting G 0 cells in HTLV-1-infected cells (64). In contrast, cells that are at late G 1

FIG. 4. Effect of CYC202 in cells infected with resistant HIV-1 viruses.
Effects of CYC202 and m-CYC202 were tested in activated PBMCs infected with nucleoside (AZT, 3TC), nonnucleoside (TIBO), and protease-resistant mutant viruses. A-D, approximately 5 ϫ 10 6 PBMCs were infected with HIV-1 mutant viruses (50 ng of p24 gag antigen/virus). All viral isolates were obtained from the NIH AIDS Research and Reference Reagent Program. After 8 h of infection, cells were washed, and fresh media were added. Activated PBMCs were infected with TIBO, 3TC, or protease-resistant viruses, and MT-2 cells (HTLV-1-and Tax-positive) were infected with an AZT-resistant virus. Treatments with CYC202 or m-CYC202 (0.1, 0.5, 1, 5 and 10 M) were performed immediately after removal of the unadsorbed virus and the addition of fresh media. Infected PBMCs were cultured for 12 days and infected MT-2 cells were cultured for 5 days. Note that the p24 results for the AZT mutant in MT-2 cells are in ng/ml as opposed to pg/ml for the other viral strains. and S phases do not have E2F4 or p130 present at the cyclin A promoter. Therefore, ChIP assays were performed as a control to verify that the majority of the cells had been arrested in the G 0 phase (Fig. 5E). Chromatin from HLM-1 cells at both G 0 and G 1 /S were incubated with control IgG, anti-E2F4, anti-p130, and anti-p300 antibodies, and primers for the cyclin A promoter (marker  (46). Cells were kept at 37°C for 12 days, and supernatants were collected for p24 ELISA as shown in B and D. E, ChIP assays were performed using cells at G 0 (0 h) and G 1 /S (16 h), as described previously (64). Cyclin A primers, specific for cyclin A promoter positions Ϫ135 to Ϫ113 and ϩ13 to ϩ33, were used to amplify DNA obtained from immunoprecipitations using antibodies for E2F4, p300, and p130. PCR products were run on a 1% agarose gel and visualized by ethidium bromide staining. Lane 1 is input DNA prior to ChIP, and lane 2 is control IgG. F, immunoprecipitation with anti-cyclin D1 (50 l of polyclonal rabbit) was performed from cells at G 0 and G 1 /S (1 mg of total protein), washed, and used for a kinase assay with GST-Rb(773-928) (lanes 1 and 2) as described previously (64). 200 ng of purified cyclin D1/cdk4 (generous gift of C. Lee, UMDNJ) were used in a kinase assay as a positive control (lane 3).
for late G 1 /S transcription) were used for PCR. We observed the presence of E2F4 and p130 (G 0 phase markers) at the cyclin A promoter in HLM-1 cells serum starved for 3 days (Fig. 5E, lanes  3 and 4). In contrast, an activator of transcription, p300, was not detected at the cyclin A promoter in the G 0 phase. HLM-1 cells in G 1 /S, in contrast, had no p130, and a decreased amount of E2F4 at the cyclin A promoter was observed. In addition, p300 was recruited to the cyclin A promoter at the G 1 /S boundary. These results indicate that HLM-1 cells were properly arrested at the G 0 phase and subsequently released into G 1 /S by the addition of complete medium. Finally, as a further proof that serum-starved cells were indeed at the G 0 phase, we performed an immunoprecipitation with anti-cyclin D1 antibody from both G 0 and G 1 /S cells followed by a kinase reaction using GST-Rb as a substrate (64). Results are presented in Fig. 5F, where IP with cyclin D1 antibodies showed no phosphorylation of Rb from G 0 cells (lane 1) but only from G 1 /S cells released after 16 h in complete medium (lane 2). Control cdk4-cyclin D1 purified protein complex showed abundant phosphorylation of GST-Rb (lane 3). Collectively, these cell cycle results indicate that in addition to inhibition of Tatactivated transcription, CYC202 may also affect post-transcriptional events including RNA processing, translation, packaging, or virion release. These results also indicate that inhibition of HIV-1 by CYC202 was not cell cycle phase-specific but more likely related to cdk inhibition.
Effect of CYC202 on Cdk2-Cyclin E in HIV-1-infected Cells-It has been shown previously that roscovitine exhibited the highest potency and selectivity for cdk2-cyclin E with an apparent IC 50 of 0.1 M, followed by cdk7-cyclin H with an IC 50 of 0.49 M (55). The potency of CYC202 appeared to be 7-fold greater for cdk2-cyclin E than reported previously for roscovitine, even in the presence of a 100 M excess of free ATP. The improved potency of CYC202 was attributed to the higher potency of the R form of the enantiomer compared with the mixture of the R and S forms present in roscovitine (52). We therefore asked which of the various cdk-cyclin complexes in HIV-1-infected cells were most sensitive to CYC202. A typical kinase assay from OM10.1 cells is shown in Fig. 6. Induced OM10.1 cells were used for immunoprecipitation with various antibodies, washed, and increasing concentrations of CYC202 (0.0, 0.5, 1, 5, and 25 M; i.e. lanes 1-5, respectively) were added to kinase reactions containing histone H1 or GST-Rb as substrates. As seen in Fig. 6A, 0.5 M CYC202 completely inhibited the cdk2 activity on histone H1 (lane 2), whereas 5 M CYC202 inhibited the cdk1 activity (lane 9). We also performed similar experiments for cdk4 activity and observed a complete inhibition at 25 M CYC202 (Fig. 6B, lane 5) using GST-Rb as a substrate.
We next performed a more comprehensive set of IP/kinase assays using various cdk-cyclin complexes and other more classical enzymes as controls. Kinase assays were performed using immunoprecipitates of cdk1, 2, 4, 7, 9, and GSK3␤, followed by a wash, and incubation with various substrates. Kinase reactions were performed with recombinant RNA pol II CTD (GST-CTD) for cdk2, cdk7, and cdk9, GST-Rb(773-928) for cdk4cyclin D1, and purified GST-p53 for GSK3␤. We have shown previously that cdk2-cyclin E can phosphorylate pol II CTD (69). PKC and PKA were also assayed by using commercial assay kits (Amersham Biosciences Biotrak PKC and Upstate PKA assay kits). Both infected lysates from induced latent (OM10.1 and ACH2) and uninfected (CEM and HL-60) cell lysates were used for immunoprecipitation followed by kinase assays in vitro in the presence of various concentrations of CYC202 (0.01, 0.05, 0.1, 0.2, 0.4, 0.6, 0.8 and 1.0 M). Results in Table I indicate that the cdk2-cyclin E complex was more sensitive to CYC202 (IC 50  . Similar levels of cdk-cyclin complexes were immunoprecipitated using these four cdk antibodies as observed in IP/Western blots (data not shown). Finally, to validate further the effect of CYC202 and its specificity on the cdk2-cyclin E complex, we performed an additional experiment with infected OM10.1 and PBMCs. A similar experiment has been described in Fig. 3, A and B, where OM10.1 cells were treated with TNF-␣ and subsequently with 0.36 M CYC202, and PBMCs were infected with the NSI strain and treated with 1 M CYC202. Next, cells were processed for immunoprecipitation with anti-cdk2 or anti-cdk4 and used for an in vitro kinase assay using purified Rb peptides. Two sets of Rb peptides specific for cdk2 (Ser-612) or cdk4 (Thr-826) have been used previously by others and by us (54,65). Results in Fig. 6C indicate that cdk2 activity is reduced dramatically by day 6 and 12 in both OM10.1 and PBMC infected cells. However, virtually  Fig. 3, A and B, were performed with OM10.1 cells induced with TNF-␣ and subsequently treated with CYC202 at an IC 50 of 0.36 M, or PBMCs (from donor 1) were infected with the NSI strain and treated with CYC202 at an IC 50 of 1.0 M. Samples were collected postinduction of infection, and lysates were processed for immunoprecipitations with anti-cdk2 and anti-cdk4 antibodies. To avoid nonspecific binding, whole cell lysates were treated with protein A-Sepharose CL-4B and subsequently used for immunoprecipitations. The immunopurified cyclins and substrates were incubated at 30°C for 30 min in R buffer that contained [␥-32 P]ATP and 5 g of Rb peptides (J and G1), and separated on an 18% SDS-PAGE, dried, and exposed to PhosphorImager cassette. The Rb peptide J sequence was 818 GLPTPTKMTPRSRIL, and peptide G1 was 775 RPPTLSPIPHIPR. no reduction in activity is seen for cdk4 in CYC202-treated infected cells. Collectively, these results clearly indicate that CYC202 preferentially inhibits the cdk2-cyclin E complex with an IC 50 of at least 4-fold lower than the IC 50 for cdk7-cyclin H and 10-fold lower than the IC 50 for cdk9-cyclin T1, all of which are required for HIV-1 transcription.
CYC202 Prevents the Recruitment of the Cdk2-Cyclin E Complex to the HIV-1 Genome-The kinase experiments presented above indicate that CYC202 preferentially inhibits cdk2-cyclin E. To address whether cdk2-cyclin E directly participates in HIV-1 transcription, we determined whether cdk2-cyclin E was recruited to the HIV-1 DNA in vivo using ChIP. We also selectively inhibited expression of various cdk-cyclin complexes by siRNA interference followed by ChIP to measure the binding of various cdk-cyclin complexes to the HIV-1 DNA in vivo. These experiments served as a proof of principle to determine whether cdk-cyclin complexes could indeed be found on the HIV-1 promoter and possibly on the downstream sequences such as the Env gene. OM10.1 cells were induced by TNF-␣ and subsequently transfected with siRNAs against cdk2, 4, 7, or 9. 48 h post-transfection, cells were cross-linked and processed for ChIP assay. PCRs of immunoprecipitated cdks were performed for the promoter (5Ј-LTR) and the Env gene (3Ј-most open reading frame). Treatment with the corresponding siRNA reduced the level of cdk2, cdk4, and cdk7 and cdk9 but not actin (Fig. 7B) and prevented recruitment of these kinases to the HIV-1 LTR (Fig. 7A, lanes 4, 8, and 12, LTR panel). RNAimediated knock-down of cdk2 and cdk9 correspondingly reduced the presence of these enzymes at the Env region (Fig. 7A,  lanes 4 and 12, Env panel). No cdk7 or cdk4 was found on the Env region in the controls, in the absence of RNAi treatment (Fig. 7A, lanes 6 and 14, Env panel), indicating that these two kinases were not participating in HIV-1 transcription elongation.
We next asked whether CYC202 could inhibit recruitment of cdk2 or cdk9 to the HIV-1 LTR or Env region. Treatment with a low concentration of CYC202 had a slight effect on cdk2 loading onto the LTR (30% drop), but CYC202 dramatically reduced the presence of cdk2 on the 3Ј-Env region (Fig. 7C,  lane 2). At the same time, a low concentration of CYC202 had no effect of cdk9 on the LTR and partially reduced (60%) its presence in the Env region (Fig. 7C, lane 6). In contrast, higher concentrations of CYC202 inhibited loading of both cdk2 and cdk9 onto the LTR and Env regions (Fig. 7C, lanes 10 and 14). These results demonstrate that inhibition of cdk activity by a specific inhibitor eliminates its presence on the HIV-1 DNA. Moreover, cdk2 recruitment is more sensitive to inhibition by CYC202 as the RNA pol II travels downstream toward the 3Ј-end of the HIV-1 genome. These results further argue that both cdk2 and cdk9 are present on the HIV-1 genome (throughout the entire 9.5 kb) and that the recruitment of both is targeted by CYC202. DISCUSSION Anticancer drugs such as cdkI, topoisomerase 1 enzyme (top 1) inhibitors, nonnucleoside antimetabolites, and estrogen receptor ligands have recently been shown to be promising candidates for viral inhibition studies. At high doses, these drugs are used for cancer therapy; however, at lower concentrations they exhibit antiviral activities in cultured cells and promote infected cells to apoptose (66).
During the past few years, two laboratories have contributed significantly to our understanding and scope of cdk inhibitors in virally infected cells. The Schang and Schaffer laboratories have demonstrated the effect of roscovitine in viral inhibition and have defined the mechanism of inhibition. Initially, their results suggested the involvement of a cellular cdk(s) in the replication of a DNA virus. They observed the involvement of cellular cdk in the transcription of viral genes. When infected cells were treated with ATP analogs, viral replication was inhibited, and no resistant mutants were observed after extensive passage (67). Roscovitine inhibited HSV replication, transcription of IE and E genes, and viral DNA synthesis when added postinfection or after release from a cycloheximide block (68). Other related experiments showed that roscovitine inhibited HSV replication after release from a block in viral DNA synthesis induced by phosphonoacetic acid or following a shift down of cells infected with HSV-1 DNA2 ts mutants from the nonpermissive to the permissive temperature (68). To show further the range of specificity, it was shown that PCIs (i) inhibit replication of wild type strains of HSV-1 and -2 and HIV-1, but not vaccinia virus or lymphocytic choriomeningitis virus; (ii) inhibit replication of strains of HSV-1 and HIV-1 resistant to conventional antiviral drugs that target thymidine kinase or DNA polymerase (HSV-1), or reverse transcriptase or protease (HIV-1); and (iii) bind to the same subset of proteins in mock-and HSV-infected cells (69,70). Along these lines, HSV-1 has been shown to encode a protein, DNA polymerase processivity factor encoded by the UL42 gene, which behaves much like a cyclin and binds to cdk1. This new complex is active and can be inhibited by roscovitine (71). Finally, readers are referred to a publication with a comprehensive list of all RNA and DNA viruses that utilize cellular kinases for their latency, transcription and replication (72).
Studies related to cdks and HCMV, Epstein-Barr virus, and VZV inhibition are also worth noting here. In HCMV-infected cells cdk2-cyclin E activity is induced and can be inhibited with roscovitine at an IC 50   cdk activity is also involved in the progression from G 1 to the S phase, and S phase cdk activity appears to be essential for the expression of viral immediate early and early lytic proteins and hence for lytic viral replication. Further, it was observed that both purvalanol A and roscovitine block Epstein-Barr virus lytic replication, whereas well characterized inhibitors of other protein kinases, such as mitogen-activated protein kinase, phosphatidylinositol 3-kinase, and PKC, did not. The effects of purvalanol A was shown to be the result of inhibition of expression of viral immediate-early and early proteins, including the replication proteins encoded by the BALF2, BALF5, BMRF1, and BBLF2/3 genes (75). Finally, the effect of roscovitine has been tested in VZV infection. Roscovitine prevented VZV replication at levels that were not cytotoxic, did not induce apoptosis in the host cells, and were lower than that needed to block HSV-1 (76). Control cells treated with 25 M roscovitine did not appear to undergo apoptosis; however, infected cells with VZV treated with 25 M roscovitine or 10 M purvalanol A caused a 10-fold reduction in VZV yield after 24 h (77). Retroviral infections have also been shown to utilize cellular cdk activity and hence may be a good target of inhibition by PCIs. For instance, we have shown previously that purvalanol A can inhibit HTLV-1-activated transcription at an IC 50 of 0.035 M. The target of inhibition was shown to be cdk2-cyclin E in HTLV-1-infected cells, and RNA pol II CTD phosphorylation was inhibited in drug-treated cells (78). The effect of PCIs had also been investigated on other retroviral infections including HIV-1. Roscovitine inhibited HIV-1-activated transcription by inhibiting cdk2, 7, and 9 (62,78,79). Furthermore, doseresponse analysis had shown that both flavopiridol and roscovitine reversibly suppressed HIV-1 transcription in podocytes at an IC 50 of 25 and 3 M, respectively. Interestingly, despite equivalent suppression of HIV-1 transcription, roscovitine was a more effective inhibitor of podocyte proliferation than flavopiridol, and suppression of HIV-1 transcription by flavopiridol or roscovitine was marked by reexpression of the podocyte differentiation markers (80). Finally, the effect of roscovitine has also been shown on drug-resistant strains of HIV-1 (81). Collectively, these studies imply that cdk inhibitors may be a new class of drugs that could either be used alone or in combination with HAART in HIV-1-infected patients.
Although HAART has been shown to be effective for drugnaive patients, as a long term solution, HAART might ultimately result in the development of multi-drug-resistant viruses. Alternative therapies, including topical microbicides (82) and vaccines (83), for prevention of HIV infection are currently being explored; yet many challenges still remain. We believe that because of the emergence of HIV-1 drug-resistant viruses, the current lack of HIV-1 Tat transcription inhibitors on the market, and perhaps most importantly, the fact that multiple transcription factors or kinases are used for the optimal function of the HIV-1 LTR promoter, the use of cdk inhibitors (PCI) FIG. 7. Physical presence of cdk2-cyclin E, cdk7-cyclin H, and cdk9-cyclin T complexes on the HIV-1 genome and inhibition by CYC202. A, presence of various cdk-cyclin complexes on the HIV-1 genome using RNAi followed by ChIP and PCR. OM10.1 cells were treated with TNF-␣ for 2 h, washed, and subsequently electroporated with siRNA against cdk2, 4, 7, and 9 (10 g each). 48 h later cells were cross-linked and processed for ChIP. Supernatants were also processed for p24 ELISA after 48 h (untreated, 8.3 ng/ml; cdk2 RNAi, 1.6 ng/ml; cdk4 RNAi, 12.8 ng/ml; cdk7 RNAi, 3.1 ng/ml; and cdk9 RNAi, 1.9 ng/ml). Specific DNA sequences in the immunoprecipitates were detected by PCR by using primers specific for the HIV-1 LTR (forward primer, 5Ј-ACTTTTCCGGGGAGGCGCGATC-3Ј; reverse primer, 5Ј-GCCACTGCTAGAGATTTCCACACTG-3Ј) or Env region (forward primer, 5Ј-CCTTG(T)GAGCCAATTCCCATA-3Ј; reverse primer, 5Ј-TAACAAATGCTCTCCCTGGTC-3Ј). siRNA sequences were designed using the Oligoengine Work station (www.oligoengine.com). (Ϫ)RNAi indicates no RNA added. I indicates input DNA isolated without immunoprecipitation (for RNAi sequences, see "Experimental Procedures"). B, Western blot of RNAi-transfected cells in A. Whole cell lysates (100 g) were processed after transfection (A, 48 h) and Western blotted against cdk2, 4, 7, 9, and actin antibodies. C, inhibition of loading of cdk2 and cdk9 on the HIV-1 genome using CYC202. After induction of OM10.1 with TNF-␣, cells were treated with two concentrations of CYC202 (1.5 and 12 M) or m-CYC202. 48 h later, samples were collected and treated for ChIP analysis and PCR using LTR and Env primers as described in A.
to inhibit HIV-1 transcription may be a viable form of treatment. CYC202 inhibits cellular kinases that are not essential for cell survival (i.e. cdk2-cyclin E) (52) but are important for HIV-1 progeny formation. Consistent with this notion, several laboratories have recently shown that targeting cellular proteins (using siRNA) can reduce HIV-1 infection and progeny formation (84 -93).
We determined previously that viral replication is cell cycledependent and requires cellular cdk-cyclin complexes and that Tat-activated transcription predominantly occurs at the G 1 /S phase of the cell cycle (46). We confirmed these results in vivo by transfecting Tat into HLM1 cells (HIV-1 ϩ /Tat Ϫ ), which contain a single full-length HIV-1 provirus with a triple termination codon at the first AUG of the Tat gene, and chemically blocking the cells with either hydroxyurea (a general G 1 /S blocker) or nocodazole (a general M phase blocker). HIV-1 attained peak viral replication in cells blocked with nocodazole, whereas G 1 /S blockage by hydroxyurea resulted in the dramatic inhibition of virion production (62). These results were consistent with the notion that G 1 /S kinases, such as cdk2cyclin E, could be targeted to inhibit HIV-1 replication.
In 2003, the Barbacid laboratory first reported that embryonic fibroblasts lacking cdk2 proliferate normally and become immortal after continuous passage in culture. Additionally, cdk2 Ϫ/Ϫ mice were viable and survived for up to 2 years, indicating that cdk2 is also dispensable for the proliferation and survival of most cell types (50). Others have reported similar findings demonstrating that the E type cyclins were also largely dispensable for mouse development. Cyclin E-deficient cells proliferated actively under conditions of continuous cell cycling but were unable to reenter the cell cycle from the quiescent G 0 state (47)(48)(49). Therefore, cdk2 and cyclin E are not essential for the growth of normal, noncancerous cells, and use of inhibitors against the cdk2-cyclin E complex may pose a viable option to inhibit HIV-1 in infected cells, perhaps without compromising uninfected cells of the host.
The HIV-1 Tat protein activates viral gene expression by promoting transcriptional elongation via phosphorylation of the RNA pol II CTD. To date, three cdk-cyclin complexes, cdk7-cyclin H (subunits of transcription factor TFIIH), cdk9cyclin T, and cdk2-cyclin E, have been shown to phosphorylate the CTD on the HIV-1 promoter. We have recently reported that cdk2-cyclin E could phosphorylate the CTD (78), was associated with RNA pol II, and was found in elongation complexes assembled on the promoter (44,94). Recombinant cdk2cyclin E stimulated Tat-dependent HIV-1 transcription in a reconstituted transcription assay, and immunodepletion of cdk2-cyclin E blocked Tat-dependent transcription.
Here, we examined the effect of the cdk inhibitor CYC202 (R-roscovitine, an ATP analog) on HIV-1 latent and infected PBMCs. CYC202 was an effective inhibitor of latently infected cells with an IC 50 between 0.36 and 1.8 M. CYC202 also inhibited both SI and NSI strains effectively in PBMCs. Observing a range of IC 50 , rather than absolute numbers in infected cells, is expected because of drastic variations in ATP levels in cells, variability of cell type, and differences in the activity of a specific kinase in question, the number of copies of the viruses, the stage of the host cell cycle, or the presence of mixed infections. This is in contrast to in vitro kinase assays where the levels of ATP and other components are carefully controlled (Table I). Therefore, in vivo results can be considered as more reliable in assessing the efficacy of a drug against HIV-1. CYC202 also selectively induced apoptosis in HIV-1infected cells without virion release (Fig. 3B), which agrees with previous data using a mixture of the two isomeric forms of roscovitine (55,62,69,72). This is in contrast to the killing of HIV-1-infected cells with agents such as DNA-damaging agents (␥-irradiation), high concentrations of TNF-␣, or sodium butyrate, where apoptosis is followed by massive virion release (42). We also consistently observed a dose-dependent response with four different resistant viral strains (Fig. 4). This was also expected because the initial mutant viruses were developed against reverse transcriptase or protease and not against the LTR or Tat. To date, the wild type virus in OM10.1 cells has not shown any resistant escape mutants (in the LTR, gag, or Env; data not shown) after ongoing culture in the presence of CYC202 for up to a year. Therefore, treatment with cdk inhibitors may avoid the generation of escape mutants because the inhibitory target is a cellular protein and not a viral protein.
An interesting and provocative finding in our studies was observed after CYC202 treatment of infected cells prior to or after the G 1 checkpoint (Fig. 5). To our surprise, only CYC202, among several inhibitors tested, was able to inhibit HIV-1 production regardless of whether cells were treated before or after the G 1 /S checkpoint. This indicates that CYC202 is able to inhibit Tat-activated transcription through a specific complex such as cdk2-cyclin E (Table I) and is consistent with previously published results demonstrating that CYC202 inhibits the cdk2-cyclin E complex (52). We are currently trying to define the mechanism of this inhibition and are studying the effect of cdk2-cyclin E in transcriptional elongation, capping of HIV-1 RNA, the presence of either singly spliced or genomic RNA after CYC202 treatment, and possible involvement in transport of RNA from the nucleus to the cytoplasm, translation or packaging of virions. Our preliminary data indicate that CYC202 can inhibit transport of singly spliced and genomic RNA. 2 Our in vitro kinase assays clearly indicate that the mode of action of CYC202 is inhibition of cdk2-cyclin E (Table I). To determine whether cdk2-cyclin E is loaded onto the HIV-1 genome, we performed ChIP assays and demonstrated that CYC202 treatment inhibited loading of cdk2 onto the HIV-1 genome (Fig. 7C). This is an interesting result and indicates that inhibition of cdk activity eliminates the presence of cdk2 from the HIV-1 DNA and that cdk2 physically becomes more sensitive to CYC202 during the transcriptional elongation process. This may suggest that after CYC202 treatment, cdk2 is either free in the nucleus or cytoplasm without its cyclin partner or simply has a shorter half-life and gets degraded. Along these lines, we have observed previously that the Tat peptide inhibitor (41/44), which targets the cdk-cyclin complexes, physically removes the cdk from the complex and leaves behind the cyclin partner on the HIV-1 DNA. 3 Studies are in progress to address whether CYC202 has similar functions in infected cells.
PCIs have emerged as the most promising antiviral agents within the past few years. We have demonstrated that PCIs are potent in inhibiting cdk-cyclin complexes, which are important for viral activation, without the emergence of resistant viruses, making PCIs ideal candidates as broad range antiviral drugs. Further studies dealing with the timing involved in the initiation of treatment, possible combination with current HAARTs, and the effectiveness of these drugs in coinfection, will allow a more complete assessment of the therapeutic promise of PCIs in the treatment of HIV-infected individuals.