|Home | About | Journals | Submit | Contact Us | Français|
Kaposi's sarcoma-associated herpesvirus (KSHV) plays a significant role in the development of Kaposi's sarcoma, primary effusion lymphoma, and some forms of multicentric Castleman's disease. The KSHV open reading frame K9 encodes the viral interferon (IFN) factor 1 (vIRF1), which downregulates IFN- and IRF-mediated transcriptional activation, and leads to cellular transformation in rodent fibroblasts and induction of tumors in nude mice. Using the yeast two-hybrid assay, we identified genes associated with retinoid-IFN-induced mortality-19 (GRIM19), which interacts directly with vIRF1, both in vivo and in vitro. The N-terminal region of vIRF1 is required for binding GRIM19. Colocalization of vIRF1 and GRIM19 was observed in 293T cells. The vIRF1 protein deregulates GRIM19-induced apoptosis in the presence of IFN/all-trans-retinoic acid (RA) and inhibits IFN/RA-induced cell death. Another DNA tumor viral protein, human papillomavirus type 16 E6, also binds GRIM19, suggesting that this is a general target of viral proteins. Our results collectively indicate that vIRF1 modulates IFN/RA-cell death signals via interactions with GRIM19.
Kaposi's sarcoma (KS)-associated herpesvirus (KSHV), also called human herpesvirus 8 (HHV8), is a recently discovered human DNA tumor virus that plays a critical role in the development of KS lesions, body cavity-based primary effusion lymphoma, and a subset of multicentric Castleman's disease (5, 6, 43). KSHV belongs to the gammaherpesvirus group and displays significant genetic similarity to herpesvirus saimiri and Epstein-Barr virus (39). Viral infection triggers the activity of multiple interferons (IFNs), which participate in host immune surveillance. IFNs exhibit a wide range of biological activities, including cell growth inhibition and immune activation. Viruses have developed a variety of strategies to cope with the inhibitory effects of IFNs (35). IFN regulatory factors (IRFs) are transcription factors that serve as mediators of the IFN signal. Interestingly, KSHV contains at least three open reading frames (ORFs) encoding proteins homologous to IRF families, designated viral IRFs (vIRFs). These include ORF K9 (also known as vIRF1), vIRF2, and vIRF3 or latent-associated nuclear antigen 2 (3, 24, 29, 37).
The vIRF1 protein comprises 449 amino acids (aa) with an N-terminal region containing a conserved tryptophan-rich DNA-binding region and displaying 70% identity to the IFN consensus sequence binding protein (39). In a transient-transfection assay, vIRF1 represses IFN- and IRF1-mediated transcription (24, 48). NIH 3T3 cells that stably express vIRF1 undergo transformation and consequently display features of malignant fibrosarcoma in nude mice (14, 24). Recently, we and another group reported that vIRF1 associates with the tumor suppressor p53 protein, leading to the repression of p53-dependent transcription and apoptosis (33, 41). These observations collectively indicate that vIRF1 induces tumorigenicity. The vIRF1 protein associates with p300/CREB binding protein (CBP), leading to the inhibition of transactivation of CBP, histone acetyltransferase activity of p300 and the formation of transcriptionally active IRF3-p300/CBP complexes (4, 23, 25, 40). Earlier studies have demonstrated that vIRF1 acts as a transcriptional activator, despite its many repressive activities (38).
The importance of the IFN pathway in cell growth suppression is confirmed in a number of reports (17, 20, 44). In addition to IFNs, all-trans retinoic acid (RA) and its derivatives are important in tumor suppression (27, 45). Clinical studies show that a combination of IFN and RA inhibits cell growth in vitro and in vivo more potently than either agent alone (26, 28, 31). Hofmann et al. (18) identified the specific genes that play a role in IFN/RA-induced cell death by an antisense technical-knockout approach (8). These genes were designated according to their association with retinoid-IFN-induced mortality (i.e., GRIMs). Out of these, GRIM12 is identical to human thioredoxin reductase. Overexpression of GRIM12 causes a small amount of cell death and increases the susceptibility of cells to IFN/RA-induced cell death. GRIM12 modulates the activity of capase-8 and increases the expression of death receptors to evoke cell death (18, 30). Another GRIM gene, GRIM19, was also characterized (1). GRIM19 is a novel cell death-associated gene that is not included in any of the known death gene categories. This gene encodes a 144-aa protein that localizes predominantly to the nucleus. A sequence search revealed that other species, specifically, mice, Arabidopsis thaliana, and Staphylococcus aureus, as well as humans, contain the GRIM19 gene, suggesting that the protein is ubiquitous in most organisms. Overexpression of GRIM19 enhances caspase-9 activity and apoptotic cell death in response to IFN/RA treatment (1). GRIM19 is located in the 19p13.2 region of the human chromosome essential for prostate tumor suppression, signifying that the protein may be a novel tumor suppressor (7).
To transform cells and induce tumors, various tumor suppressors need to be inactivated, which may occur in a number of ways. Since vIRF1 triggers cellular transformation without the aid of any other oncogenic protein, we reasoned that, in addition to p53 and p300/CBP, the viral protein may target factors participating in other cell death pathways. Using the yeast two-hybrid assay with vIRF1 as bait, we identified the newly characterized cell death regulatory protein, GRIM19, as a novel interaction partner. This interaction specifically suppresses the cell death activity of GRIM19 in response to IFN/RA.
The pcDNA3-vIRF1 construct was described previously (40). To generate glutathione S-transferase (GST)-tagged vIRF1 and its deletion mutants, the corresponding cDNA sequences from pcDNA3-vIRF1 were subcloned into pGEX4T-1 (Amersham Pharmacia Biotech, Uppsala, Sweden). The pGBKT7-vIRF1ΔC plasmid, which was used as bait for the yeast two-hybrid screen, was generated by cloning into the EcoRI/SalI sites of pGBKT7. To obtain FLAG-tagged vIRF1 and its N-terminal deletion mutants, the corresponding cDNA sequences were cloned into the HindIII/BglII sites of pFIN, thus generating pFIN-vIRF1 and pFIN-vIRF1ΔN, respectively. GAL4 DNA-binding domain (DBD) fused to vIRF1 was constructed by inserting the EcoRI/NotI fragment from pcDNA3-vIRF1 into the corresponding restriction sites of pcDNA3-GAL4 (including the cytomegalovirus promoter) to generate a chimeric protein with GAL4 DBD fused to the N terminus of the viral factor. GRIM19 cDNA was amplified by PCR with pCXN2-GRIM19 plasmid as a template, as described previously (1). Amplified products were digested with appropriate restriction enzymes and cloned into the EcoRI/XhoI sites of pcDNA3, pcDNA3-HA, and pGADT7, respectively. The EcoRI/XhoI fragment of pcDNA3-GRIM19 was cloned into the EcoRI/SalI sites of pGEX4T-1 and pVP16 (Clontech, Palo Alto, Calif.). To generate a GST-GRIM19 expression plasmid, the BamHI/NotI fragment from pGEX4T-1-GRIM19 was inserted into the corresponding restriction sites of the pEBG vector (including the EF-1α promoter) to generate a chimeric protein with GST fused to the N terminus of GRIM19. The pEGFP-GRIM19 construct expressing green fluorescent protein (GFP) fused to vIRF1 was generated by subcloning into the BglII/SalI sites of pEGFP-C1 (Clontech). pcDNA3-E1A and pcDNA3-E6 were generated by inserting the required sequences into the EcoRI/XhoI sites of pcDNA3. The pSG5-simian virus 40 (SV40) large T (LT) antigen was a generous gift from T. M. Roberts (Dana-Farber Cancer Institute). The pFR-Luc plasmid was described previously (22, 40).
Yeast two-hybrid screening was performed by using the Matchmaker GAL4 Two-Hybrid System 3 (Clontech). The yeast strain, AH109 (19), containing pGBKT7-vIRF1ΔC was transformed with a human brain cDNA library cloned into a GAL4 activation domain vector, pACT2 (Clontech). Transformants were selected on plates containing synthetic dextrose (SD)-glucose-20 mM 3-amino-1,2,4-triazole (3-AT) lacking tryptophan (−Trp), leucine (−Leu), and histidine (−His) or on plates containing SD-glucose but lacking tryptophan, leucine, histidine, and adenine (−Ade). Library plasmids of positive clones were isolated, subcloned into the pcDNA3 vector, and sequenced with a T7 promoter primer.
293T and HeLa cells were grown in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum. BJAB, BCBL-1, and MCF-7 cells were maintained in RPMI 1640 containing 10% fetal bovine serum. Transfections were performed by the calcium phosphate method (16) in 293T cells. HeLa cells were transfected with Superfect transfection reagent (Qiagen, Hilden, Germany) according to the manufacturer's recommendations. To generate stable cells expressing FLAG-vIRF1 and FLAG-vIRF1ΔN, MCF-7 cells were transfected with pFIN-vIRF1 and pFIN-vIRF1ΔC by using Lipofectamine Plus reagent (Gibco-BRL, Gaithersburg, Md.) according to the manufacturer's instructions. Transfected cells were selected for 3 weeks in complete medium supplemented with 1 mg of G418 (Gibco-BRL)/ml.
293T cells were transiently transfected either with GST or GST-GRIM19 in combination with FLAG-vIRF1, or with control plasmid (pFIN), FLAG-vIRF1, in conjunction with hemagglutinin (HA)-GRIM19. After 48 h of transfection, cells were lysed in EBC buffer (50 mM Tris-HCl, pH 7.5; 120 mM NaCl; 0.5% Nonidet P-40; 50 mM NaF; 200 μM sodium orthovanadate; 1 mM phenylmethylsulfonyl fluoride) and incubated with glutathione-Sepharose 4B beads or protein G-Sepharose beads (preincubated with anti-FLAG monoclonal antibody [Sigma, St. Louis, Mo.]) at 4°C for 2 h with rocking. Bound protein complexes were washed three times with EBC buffer and then resuspended in loading buffer (10 μl). Samples were immunoblotted with anti-FLAG, anti-GST, or anti-HA antibody, respectively. BJAB and TPA (12-O-tetradecanoylphorbol-13-acetate)-treated BCBL-1 cells were lysed with EBC buffer and immunoprecipitated with anti-GRIM19 mouse monoclonal antibody, as described above. Samples were immunoblotted with anti-vIRF1 rabbit polyclonal antibody.
Wild-type GST and GST fusion proteins were prepared by induction of Escherichia coli containing the fusion vector with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). After lysis by sonication, GST fusion proteins were bound to glutathione-Sepharose 4B beads, washed with phosphate-buffered saline (PBS), and eluted with buffer containing 25 mM glutathione. 35S-labeled proteins were synthesized in vitro by using the TNT-Coupled Transcription-Translation System (Promega, Madison, Wis.) as described by the manufacturer. Purified or Sepharose 4B-bound GST fusion protein was incubated with 35S-labeled proteins in 500 μl of binding buffer (50 mM Tris-HCl, pH 7.5; 150 mM NaCl; 5 mM EDTA, pH 8.0; 2.5 mM dithiothreitol; 0.7 mg of bovine serum albumin [BSA]/ml; 0.1% Nonidet P-40) at 4°C for 90 min. The precipitated protein complexes were washed five times with binding buffer. Bound proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography.
Cells were fixed with 3.7% formaldehyde for 30 min, permeabilized with PBS containing 0.2% Triton X-100 (PBST) on ice, blocked with 1% BSA, incubated with primary antibody in PBST containing 1% BSA for 1 h at room temperature, and washed three times with PBST. Cells were incubated with secondary antibody in PBST containing 1% BSA for 1 h and washed three times with PBST. Analyses were performed with a Zeiss (Oberkochen, Germany) confocal microscope with fluorescein isothiocyanate (FITC) and TRITC (tetramethyl rhodamine isothiocyanate) filter sets. The Nucleus of cell was stained with DAPI (4′,6′-diamidino-2-phenylindole; Sigma, St. Louis, Mo.).
HeLa cells grown on coverslips in 35-mm-diameter dishes were transfected with Superfect reagent. Cells were treated with IFN-β (3,000 U/ml; Calbiochem, San Diego, Calif.), together with RA (5 μM; Sigma), for 72 h and fixed. The TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling) reaction (15) was performed with an in situ cell detection kit, as specified by the manufacturer (Roche Molecular Biochemicals, Mannheim, Germany). Analyses were performed with a Zeiss confocal microscope by using an FITC filter set. The percentage of apoptosis was determined by counting the green apoptotic cells from multiple fields.
Cell growth was measured with the Sulforhodamine-B (SRB) method (42) at 3, 5, and 7 days after IFN/RA treatment, respectively. Each group comprised eight replicates. Cells (3,000 cells/well) were plated in 96-well plates 6 h before IFN/RA treatment. Next, cells were fixed with 10% trichloroacetic acid at 4°C for 1 h, washed in water four times, stained with 0.4% SRB (dissolved in 1% acetic acid), and rinsed four times with 1% acetic acid. Bound dye was eluted with 100 μl of 10 mM Tris-HCl (pH 10.5), and each absorbance value was measured at 570 nm. One plate was fixed at 6 h after the cells were plated, and the absorbance value obtained was taken as 0% growth. The absorbance values obtained from untreated cells were taken as 100% growth. The changes in absorbance values in experimental wells relative to the initial value were measured as cell growth and death, respectively. When plotted as a percentage of untreated control growth and death, values measured were on the positive and negative scales of the y axis, respectively.
A yeast two-hybrid assay was performed to identify cellular proteins that interact with vIRF1. Since vIRF1 is a strong transactivator (38, 40), the C-terminal deletion construct, vIRF1ΔC (aa 1 to 360) fused to the DBD of GAL4, was used as bait (Fig. (Fig.1A).1A). GAL4-fused vIRF1ΔC did not activate reporter gene expression independently in yeast (Fig. (Fig.1B).1B). We screened a human brain cDNA library expressing proteins fused to the GAL4 transcriptional activation domain. From 106 yeast transformants we isolated 38 clones that activated a reporter gene in the presence of GAL4-fused vIRF1ΔC. Among these, three were identified as GRIM19, a novel cell death-related gene (1). The remaining 35 clones encode eight novel and two known proteins. To confirm interactions between vIRF1 and GRIM19, we cotransformed pGBKT7-Lamin C, pGBKT7-vIRF1ΔC, pGADT7, and pGBKT7-GRIM19 into the AH109 strain. Since human lamin C neither forms complexes nor interacts with most other unrelated proteins, this was used as a negative control (47). Only yeast cells containing pGBKT7-vIRF1ΔC and pGADT7-GRIM19 were grown on SD-glucose-(−Trp, −Leu, −His)-20 mM 3-AT and SD-glucose-(−Trp, −Leu, −His, −Ade) plates, respectively, indicating that vIRF1 specifically interacts with GRIM19 (Fig. (Fig.1B).1B). To suppress leaky expression of HIS3, SD-glucose-(−Trp, −Leu, −His) plates were supplemented with 3-AT, a competitive inhibitor of the yeast HIS3 protein (9).
To verify the specific interactions between vIRF1 and GRIM19, a mammalian two-hybrid assay was performed (22). Two chimeric proteins were created by fusing full-length vIRF1 to DBD of GAL4 (GAL4-vIRF1) and full-length GRIM19 to the transcriptional activation domain of VP16 (VP16-GRIM19). Transient-cotransfection experiments were performed in 293T cells with a luciferase reporter (pFR-Luc) and GAL4-vIRF1, together with either VP16-GRIM19 or VP16. As shown in Fig. Fig.2B,2B, VP16-GRIM19 activated luciferase transcription via GAL4-vIRF1 in a dose-dependent manner, whereas VP16 did not. GAL4-vIRF1 expression was monitored by a Western blot assay, which revealed that levels of the chimeric protein remained unchanged in the presence of both VP16-GRIM19 and VP16. These data indicate that vIRF1 physically interacts with GRIM19 in mammalian cells.
We next determined whether vIRF1 binds GRIM19 in vivo. 293T cells were cotransfected with GST, GST-GRIM19, and FLAG-vIRF1 expression plasmids (pEBG, pEBG-GRIM19, and pFIN-vIRF1, respectively). At 48 h after transfection, cells were lysed and incubated with glutathione-Sepharose 4B beads for 2 h to precipitate GST and GST-GRIM19 proteins. Precipitated proteins were immunoblotted with an anti-FLAG antibody. The FLAG-vIRF1 protein coprecipitated with GST-GRIM19 but not with GST alone (Fig. (Fig.3A,3A, top panel). To establish that the FLAG-vIRF1, GST, and GST-GRIM19 proteins are expressed properly in transfected cells, we performed a Western blot assay by using total cell extracts with anti-FLAG and anti-GST antibodies (Fig. (Fig.3A,3A, top and bottom panels, respectively). In a reciprocal experiment, 293T cells were cotransfected with FLAG-vIRF1 (pFIN-vIRF1) and HA-GRIM19 (pcDNA3-HA-GRIM19), lysed, and immunoprecipitated with an anti-FLAG antibody. The precipitated complex was immunoblotted with anti-HA antibody. HA-GRIM19 coimmunoprecipitated with FLAG-vIRF1, but not FLAG alone (Fig. (Fig.3B,3B, top panel). HA-GRIM19 and FLAG-vIRF1 proteins were properly expressed from total cell extracts (Fig. (Fig.3B,3B, top and bottom panels, respectively). To confirm the interactions between vIRF1 and GRIM19, we performed coimmunoprecipitation assays in KSHV-infected BCBL-1 cells. Since vIRF1 is a lytic protein, vIRF1 expression was induced with TPA (36) (Fig. (Fig.3C,3C, left panel). Total lysates of BJAB and BCBL-1 cells were immunoblotted with anti-vIRF1 polyclonal antibody and anti-GRIM19 monoclonal antibody (Fig. (Fig.3C,3C, right panel, lanes 1 and 2). Cells were lysed and immunoprecipitated with either anti-GRIM19 or anti-HA antibody. The vIRF1 protein coimmunoprecipitated with GRIM19 in KSHV-infected BCBL-1 cells (Fig. (Fig.3C,3C, right panel, lane 4). However, vIRF1 was not detected in KSHV-negative BJAB cell extracts or by immunoprecipitation with anti-HA antibody (Fig. (Fig.3C,3C, right panel, lanes 3 and 5). Our data confirm that vIRF1 interacts with GRIM19 in lymphocytes.
Since vIRF1 binds to CBP/p300 (4, 23, 40), it is possible that vIRF1-GRIM19 interaction is originated from bridging through CBP/p300. To exclude this bridge effect, we examined the in vivo interaction between GRIM19 and CBP. 293T cells were cotransfected with GST, GST-GRIM19, GST-vIRF1, and HA-CBP expression plasmids, lysed, and precipitated with glutathione-Sepharose 4B beads. The precipitated complex was immunoblotted with anti-HA monoclonal antibody. The GST-vIRF1 coprecipitated with CBP but not GST-GRIM19 (Fig. (Fig.3D,3D, top panel). These data show that GRIM19 dose not interact with CBP, and therefore vIRF1-GRIM19 interaction is not caused by bridging CBP.
To determine whether vIRF1 and GRIM19 complexes colocalize, we examined their subcellular localization by immunofluorescence confocal microscopy. 293T cells were transfected with GFP-GRIM19 (pEGFPC1-GRIM19) and FLAG-vIRF1 (pFIN-vIRF1) expression plasmids. As shown previously, GFP-GRIM19 was located within the nucleus, whereas GFP alone displayed a diffuse pattern throughout the cytoplasm and nucleus (1). The vIRF1 protein was observed throughout the cytoplasm and nucleus, although the large proportion was localized in the nucleus (Fig. (Fig.4A).4A). In 293T cells coexpressing FLAG-vIRF1 and GFP-GRIM19, vIRF1 was colocalized with GFP-GRIM19 in the nucleus (Fig. (Fig.4B).4B). To examine the colocalization of vIRF1 and GRIM19 in physiological conditions, we performed confocal microscopy in KSHV-infected BCBL-1 cells. TPA-induced BCBL-1 cells were fixed and incubated with anti-vIRF1 and anti-GRIM19 antibodies. vIRF1 also was colocalized with GRIM19 in the nucleus of BCBL-1 cells (Fig. (Fig.4C).4C). This colocalization of vIRF1 and GRIM19 indicates specific interactions between the two proteins in vivo.
To confirm the exact subcellular localization of these two proteins, fractionated extracts from transfected cells were separated into nuclear and cytoplasmic fractions, and each fraction was analyzed by Western blotting (Fig. (Fig.4D).4D). FLAG-vIRF1 was detected mainly in nuclear fraction and a small portion of FLAG-vIRF1 was detected in cytoplasmic fraction (Fig. (Fig.4D,4D, top panel). GFP was detected in both nuclear and cytoplasmic fractions, but GFP-GRIM19 was detected only in nuclear fraction (Fig. (Fig.4D,4D, bottom panel). The nuclear fractions from the cells expressing GFP, GFP-GRIM19, and FLAG-vIRF1 were immunoprecipitated with anti-GFP polyclonal antibody and immunoblotted with anti-FLAG monoclonal antibody. Nuclear FLAG-vIRF1 was coimmunoprecipitated with nuclear GFP-GRIM19 (Fig. (Fig.4E).4E). These data indicate that vIRF1 associates GRIM19 in nucleus.
To identify the region of vIRF1 that specifically binds GRIM19, we performed a GST pull-down assay. A series of GST-vIRF1 N-terminal and C-terminal deletion mutants were constructed, specifically, GST-vIRF1ΔC (aa 1 to 360), GST-vIRF1ΔC1 (aa 1 to 270), GST-vIRF1ΔC2 (aa 1 to 230), GST-vIRF1N (aa 1 to 152), GST-vIRF1ΔN (aa 152 to 449), and vIRF1C (aa 289 to 449) (Fig. (Fig.5A).5A). In vitro-translated GRIM19 interacted with GST-vIRF1, GST-vIRF1ΔC, GST-vIRF1ΔC1, GST-vIRF1ΔC2, and GST-vIRF1N proteins, but not with GST-vIRF1ΔN and GST-vIRF1C (Fig. (Fig.5B).5B). From these data, we conclude that the N-terminal region of vIRF1 is necessary for binding GRIM19.
Since GRIM19 induces apoptotic cell death in response to the IFN/RA signal (1), we investigated whether vIRF1 regulates this activity. GRIM19 induced apoptosis that was strongly enhanced upon IFN/RA treatment, suggesting that the protein undergoes posttranslational modifications or interacts with other factors in the presence of IFN/RA (1). Consequently, we investigated cell apoptosis induced by GRIM19 in the presence of IFN/RA in this study. To examine the effects of vIRF1 on GRIM19-induced apoptosis by IFN/RA, a TUNEL assay was performed (15) in HeLa cells. Transiently transfected HeLa cells were treated with a IFN/RA combination for 72 h, fixed, and subjected to TUNEL analysis. Since the TUNEL reaction used FITC as a labeling reagent, TUNEL-positive (apoptotic) cells contained green nuclei. Transfection of the GRIM19 expression plasmid increased the number of TUNEL-positive cells in the presence of IFN/RA, whereas cotransfection of vIRF1 (pFIN-vIRF1) and GRIM19 expression plasmid (pCXN2-GRIM19) decreased the number of TUNEL-positive cells. However, cotransfection with vIRF1ΔN (pFIN-vIRF1ΔN) did not alter apoptosis by GRIM19 (Fig. (Fig.6A).6A). The vIRF1ΔN protein did not interact with GRIM19 in a GST pull-down assay (Fig. (Fig.5B).5B). Therefore, protein-protein interactions between vIRF1 and GRIM19 are required for the inhibition of GRIM19-mediated apoptosis. Apoptotic cells in the presence of IFN/RA represented 19.1, 58.7, 16.8, 54.1, and 8.9% of the total cell population during expression of no protein, GRIM19, GRIM19 in combination with vIRF1, GRIM19 plus vIRF1ΔN, and vIRF1, respectively (Fig. (Fig.6B).6B). These results suggest that vIRF1 inhibits the apoptotic activity of GRIM19.
Since GRIM19 plays a critical role in IFN/RA cell death (1), we investigated whether vIRF1 deregulates this process. MCF-7 cells were stably transfected with vector (pFIN), FLAG-vIRF1 (pFIN-vIRF1), and the N-terminal 151-aa deletion mutant (pFIN-vIRF1ΔN) that does not interact with GRIM19 (Fig. (Fig.5B).5B). Successfully transfected cells were selected in medium containing G418 (1 mg/ml) for 3 weeks. To determine whether cells expressing vIRF1 are resistant to the apoptotic effect of IFN/RA compared to those expressing vector alone, we monitored the growth of MCF-7 stable cells in the presence of IFN/RA by using the SRB assay (42). Cell growth was measured after 3, 5, and 7 days and is presented as a percentage of IFN/RA-untreated cells. The results demonstrate that MCF-7 cells expressing vIRF1 are more resistant to IFN/RA-induced cell death than those expressing vIRF1ΔN or vector alone (Fig. (Fig.7A).7A). vIRF1ΔN does not associate with GRIM19 (Fig. (Fig.5B),5B), indicating that binding between these two proteins is important for inhibiting IFN/RA-induced cell death. Expression of vIRF1 and vIRF1ΔN was detected by Western blotting with anti-FLAG (Fig. (Fig.7B)7B) antibody. The data suggest that vIRF1 inhibits the cell death-promoting activity of IFN/RA.
To investigate the possibility that DNA tumor virus proteins other than KSHV vIRF1 also target GRIM19, we examined the interaction between HPV-16 E6 and GRIM19. 293T cells were cotransfected with GST, GST-GRIM19, and FLAG-16 E6 expression plasmids (pEBG, pEBG-GRIM19, and pME18S-16 E6, respectively). At 48 h after transfection, cells were lysed and incubated with glutathione-Sepharose 4B beads for 2 h to precipitate GST and GST-GRIM19 proteins. Precipitated proteins were immunoblotted with an anti-FLAG antibody. The FLAG-16 E6 protein coprecipitated with GST-GRIM19 but not with GST alone (Fig. (Fig.8A,8A, top panel). These data show that HPV-16 E6 interacts with GRIM19 in vivo.
HPV E6 proteins of the low-risk group associate with p53 with lower affinity than HPV E6 proteins of the high-risk group (46). To determine whether GRIM19 binding also differs between the high- and low-risk groups of HPV E6, we performed GST pull-down assays with in vitro-translated HPV-11 E6 (low risk) and HPV-16 E6 (high risk). GST-GRIM19 was immobilized on glutathione-Sepharose 4B beads, and binding of in vitro-translated 11 E6, 16 E6, and luciferase to the fusion protein was observed. Luciferase (used as a negative control) did not interact with GST-GRIM19. HPV-16 E6 specifically associated with GRIM19 more strongly than HPV-11 E6, suggesting that the difference between high-risk and low-risk E6-GRIM19 interaction correlates with the transforming activity of these different papillomaviruses (Fig. (Fig.8B).8B). Interestingly, both vIRF1 and HPV-16 E6 interact with p53, p300/CBP, and GRIM19, suggesting that the two proteins share cellular interaction partners. We also examined the interaction of SV40 LT antigen and adenovirus E1A to GRIM19. In GST pull-down assays, SV40 LT antigen also interacted with GRIM19, but E1A did not (data not shown). The finding that DNA tumor viral proteins other than vIRF1 interact with GRIM19 suggests that GRIM19 may be a general target for tumor viral proteins.
Sequence analyses reveal that KSHV ORFs encode a number of accessory proteins that display significant homology to cellular regulatory proteins and an unusual degree of genetic piracy compared to other herpesviruses (13, 32, 39). Many of these viral genes encode proteins that deregulate cellular gene expression or signaling, including vCyclin (ORF72), vBCL-2 (ORF16), vGPCR (ORF74), vIRF1 (ORFK9), vIL-6 (ORFK2), and vFLIP (ORFK13) (32). Among these, vIRF1 triggers the transformation of NIH 3T3 cells, leading to malignant tumors in nude mice (14, 24). The vIRF1 protein effectively blocks IFN- and IRF1-mediated transcription. However, the issue of whether vIRF1 and IRF1 directly interact is currently controversial (4, 48). Since IRF1 is a tumor suppressor, it is hypothesized that the tumor-inducing activity of vIRF1 is partly caused by inhibition of IRF1 (48). vIRF1 additionally influences cellular regulation pathways other than the IFN pathway. Previous studies established that vIRF1 interacts with p300/CBP and inhibits the histone acetyltransferase activity of p300 (4, 23, 40). In addition, vIRF1 associates with p53, resulting in the inhibition of p53-dependent apoptosis (33, 41). These data specify the various functions of vIRF1 and partly explain the molecular mechanism of vIRF1-transforming activity. In this report, we demonstrate that GRIM19 is a cellular interaction partner of vIRF1. Moreover, vIRF1 inhibits GRIM19-induced apoptosis in HeLa cells and suppresses IFN/RA-mediated death in MCF-7 cells. Direct interactions between vIRF1 and GRIM19 are required for these activities.
GRIM19 is a newly characterized death-regulatory protein whose inactivation leads to growth advantage to cells in the presence of IFN/RA. However, mechanism of IFN/RA-mediated cell death is yet to be fully elucidated. GRIM19 is located at the 19p13.2 locus of the human chromosome essential for prostate tumor progression. This supports the possibility that the protein acts as a tumor suppressor. Interestingly, Fearnley et al. (11) reported that bovine GRIM19 is a subunit of bovine mitochondrial complex I, a finding suggesting the involvement of mitochondria in GRIM19-induced apoptotic cell death. These results collectively indicate that GRIM19 plays a critical role in cell death, although the biochemical function of GRIM19 is unclear. Angell et al. (1) showed that the C-terminal region of GRIM19 harboring a putative ATP-binding domain, IMKDVPXWKVGE, is required for cell death induction. This region is highly conserved throughout various eukaryotic GRIM19 homologs. One possible mechanism of GRIM19 inhibition involves targeting the ATP-binding domain by vIRF1, which would result in the abrogation of apoptotic activity. Although we demonstrate the importance of vIRF1 and GRIM19 binding in suppressing IFN/RA-induced cell death, we cannot exclude the possibility of interactions between vIRF1 and other GRIMs or other cellular factors that may play an important role in IFN/RA signaling.
NIH 3T3 cells expressing vIRF1 are resistant to tumor necrosis factor alpha-induced apoptosis, although the mechanism for this is currently unknown (4). Human B lymphocytes expressing vIRF1 are resistant to the antiproliferative effects of IFN-α, and vIRF1 expression inhibits p53-induced apoptosis (12, 33, 41). MCF-7 cells expressing vIRF1 are also resistant to IFN/RA-induced cell death, possibly through interactions between vIRF1 and GRIM19 (Fig. (Fig.7).7). These data strongly suggest that vIRF1 induces resistance in cells to various death signals. Cells expressing vIRF1 may be resistant to other cell death signals, such as those of Fas ligand and hypoxia. Further studies are required to confirm whether vIRF1 inhibits cell death induced by other factors. The resistance to a number of cell death and antiproliferative signals is consistent with the strong transforming activity of vIRF1.
The vIRF1 interaction partners, p300/CBP and p53, are commonly targeted by other DNA viral oncoproteins, such as SV40 LT antigen, HPV E6 and adenovirus E1A (2, 10, 21, 34, 46). In view of this observation, we hypothesize that if GRIM19 acts as tumor suppressor in multiple tissues, it may be a general target for tumor viral proteins other than vIRF1. In vitro and in vivo binding assays reveal that HPV-16 E6 specifically interacts with GRIM19 (Fig. (Fig.8).8). Interestingly, GRIM19 binds more strongly to high-risk group E6 than to low-risk group E6, suggesting the possibility that E6-GRIM19 interaction may contribute to the transforming activity of high-risk group E6. We are currently investigating whether E6 also deregulates IFN/RA-mediated cell death. During viral oncogenesis, GRIM19 may be a general target protein similar to other cellular tumor suppressors, such as p53, p300/CBP, and RB. Further studies are required to elucidate the importance of GRIM19 protein in oncogenesis of the DNA tumor virus.
This work was supported in part by grants from the National Research Laboratory Program of the Korea Institute of Science and Technology Evaluation and Planning (KISTEP), the Korea Science and Engineering Foundation (KOSEF) through the Protein Network Research Center at Yonsei University, and the BK21 Program of the Ministry of Education Korea. D.V.K. was supported by grant CA78282 from the National Cancer Institute, National Institutes of Health.