Auto-activation of the transforming viral interferon regulatory factor encoded by Kaposi's sarcoma-associated herpesvirus (human herpesvirus-8)

Kaposi's sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus-8, is the most recently identified human herpesvirus (Chang et al., 1994). A large number of studies suggest that KSHV is aetiologically associated with Kaposi's sarcoma and several other lymphoproliferative diseases, including primary effusion lymphoma and a subset of multicentric Castleman's disease (Boshoff et al., 1995; Carbone et al., 1996; Cesarman et al., 1995a; Chang et al., 1994, 1996; Dupin et al., 1995; Gessain et al., 1996; Karcher & Alkan, 1995; Moore & Chang, 1995; Nador et al., 1995; Pastore et al., 1995; Schalling et al., 1995; Soulier et al., 1995).

Similar to other tumour viruses, KSHV has evolved specific mechanisms to evade host antiviral defences. KSHV encodes a unique set of non-structural genes that target specific cellular signal pathways, contributing to the pathogenesis of KSHV-related diseases (Russo et al., 1996). One of these non-structural genes is the open reading frame (ORF) K9, which encodes the viral interferon regulatory factor (vIRF), a homologue of cellular IRFs (Gao et al., 1997). IRFs are a family of transcriptional factors that regulate interferon signal transduction through binding to interferon-stimulated response elements (ISREs) in the promoters of interferon-responsive genes (Fujita et al., 1989; Harada et al., 1989; Imam et al., 1990; Nguyen et al., 1995; Pine et al., 1990; Taniguchi, 1995; Taniguchi et al., 1998). Similar to IRF2, overexpression of vIRF transforms NIH3T3 and Rat-1 cells (Gao et al., 1997; Zimring et al., 1998). Early reports have demonstrated that vIRF represses interferon signal transduction through direct binding to IRFs, and p300 and CREB-binding protein (CBP) transcriptional coadaptors (Burysek et al., 1999; Seo et al., 2000). Through interaction with p300, vIRF displaces p300/CBP-associated factor from the transcriptional complexes, inhibits the histone acetyltransferase activity of p300 and blocks IRF3 recruitment of p300/CBP (Li et al., 2000; Lin et al., 2001). vIRF is also a transcriptional activator and regulates the expression of other KSHV genes. More recently, vIRF has been shown to interact with p53 tumour suppressor protein and inhibits p53-mediated transcriptional regulation and apoptosis (Nakamura et al., 2001; Seo et al., 2001).

Given the likely important function of vIRF in KSHV-related pathogenesis, elucidation of the molecular mechanisms controlling vIRF gene expression could lead to an understanding of its precise role in regulating the expression of cellular and KSHV genes. In previous studies, we have mapped the vIRF core promoter region and a transcriptional silencer, Tis, and defined a 12-O-tetradecanoylphorbol 13-acetate (TPA)-responsive region in the upstream regulatory sequence of the vIRF gene (Wang et al., 2001, 2002). Here, we have shown that vIRF also auto-activates its own promoter through two unidentified cis elements that are unresponsive to interferons.

Cell culture.
KSHV-positive cells BC-1 and BCBL-1 (Cesarman et al., 1995b; Renne et al., 1996) and KSHV-negative BJAB cells were grown in RPMI 1640 medium (Sigma), supplemented with 10 % foetal bovine serum (Sigma), 10 µg gentamicin ml^-1 and 2 mM L-glutamine (Sigma). HeLa, COS7 and 293 cells were obtained from ATCC and grown in Dulbecco's modified Eagle's medium (DMEM; Sigma), supplemented with 10 % foetal bovine serum, 10 µg gentamicin ml^-1 and 2 mM L-glutamine.

Construction of plasmids.
vIRF expression plasmid pCMV-Tag2/vIRF was generated by inserting the PCR-amplified ORF K9 gene DNA fragment into the EcoRI/XhoI site of pCMV-Tag2B (Stratagene).

vIRF promoter reporter plasmids for mapping the region responsive to vIRF auto-activation have been described previously (Wang et al., 2001). Briefly, a 1·052 kbp DNA fragment spanning the region -991 to +62 relative to the vIRF gene transcriptional start site (+1) was inserted into the HindIII/XbaI sites of a promoterless and enhancerless chloramphenicol acetyltransferase (CAT) vector, pCAT-Basic, to generate the reporter construct pCAT-991. The 5' end sequence of pCAT-991 was then sequentially deleted to generate constructs pCAT-499 (-499 to +62), pCAT-337 (-337 to +62), pCAT-125 (-125 to +62) and pCAT-56 (-56 to +62) (see Fig. 3A).

(21K): Fig. 3. vIRF activates its own promoter by targeting two cis elements. (A) Constructs of the vIRF promoter used for mapping the regions involved in vIRF auto-activation. The constructs were generated by inserting different PCR-amplified DNA fragments upstream of the vIRF-initiating ATG into the pCAT-Basic vector, a promoterless and enhancerless vector, and named according to the 5' end nucleotide number upstream of the transcription start site (+1). (B) Ectopic expression of vIRF exhibits different effects on the vIRF promoter constructs. vIRF promoter reporter plasmids pCAT-56, pCAT-125, pCAT-337 or pCAT-499 were cotransfected with vIRF expression plasmid pCMV-Tag2/vIRF into 293 cells. The cells were harvested 48 h post-transfection and the CAT activity was determined as in Fig. 1. The results are shown as the averages of relative CAT activities with standard deviations from three independent experiments. (CE) vIRF auto-activation via Vac1 and Vac2 in BJAB (C), COS7 (D) and HeLa (E) cells.

Transient transfection and CAT assay.
All plasmids used for transient transfection were prepared with the QIAfilter Plasmid Maxi Kit (Qiagen). For HeLa, 293 and COS7 cells, transfection experiments were performed with Lipofectamine 2000 reagent according to the instructions of the manufacturer (Invitrogen). For BC-1, BCBL-1 and BJAB cells, transfection experiments were performed as described previously (Wang et al., 2002). The CAT assay was performed as previously reported (Gao et al., 1997). Transfection efficiencies were normalized by cotransfection with a reporter plasmid, pSV-β-galactosidase, followed by determination of β-galactosidase activity according to the instructions of the manufacturer (Promega). The conversion rate of the modified ¹⁴C-labelled chloramphenicol was measured with a GS525 Molecular Imager (Bio-Rad) and calculated with the Multi-Analysis Program (Bio-Rad).

Treatment of cells with interferons.
To determine the effect of interferons on vIRF promoter activity, cells were either transfected alone with CAT reporter constructs pCAT-125, pCAT-337, pCAT-499 and pCAT-991 or cotransfected with pCMV-Tag2/vIRF and were then treated for 24 h with either 1000 U interferon-β or -γ ml^-1 at 24 h post-transfection. The cells were then harvested for CAT assay.

Western blot analysis.
Ectopic expression of pCMV-Tag2/vIRF plasmids transfected in 293, HeLa and COS7 cells was detected by Western blot analysis. Briefly, cells (1x10⁷) were harvested, washed twice with PBS, pelleted and resuspended in 200 µl SDS sample buffer. The samples were then subjected to SDS-PAGE before proteins were transferred to nitrocellulose membrane. The membrane was probed with vIRF-specific monoclonal antibody 2H5, followed by probing with a 1 : 5000 dilution of the rabbit anti-mouse immunoglobulin alkaline phosphatase conjugate (Sigma) and developed using NBT/BCIP as substrates.

RNA isolation and Northern blot analysis.
Untreated BC-1, BCBL-1 and BJAB cells and cells either treated with TPA (Sigma) or transfected with vIRF expression plasmids (pCMV-Tag2/vIRF) were collected for total RNA isolation using TRI AGENT according to the instructions of the manufacturer (Sigma). Northern blot analysis was performed as previously reported (Wang et al., 2001). Briefly, the isolated total RNA was treated with DNase I (Promega) at 37 °C for 30 min, fractionated on agarose gel containing 1 % formaldehyde and transferred to Hybond-N+ nylon membrane (Amersham Pharmacia Biotech). The vIRF probe was prepared by labelling the vIRF DNA with [α-³²P]dCTP (NEN Life Science Products) using the Rediprime II kit (Amersham Pharmacia Biotech). Hybridization of the probe to the nylon membrane was carried out at 68 °C for 3 h with the PerfectHyb Plus Hybridization Buffer (Sigma). β-Actin probe was used to calibrate RNA loading quantity. The specific hybridization signals on the membrane were captured and analysed with a Molecular Imager.

vIRF activates its own promoter
Early reports have demonstrated that vIRF represses interferon signal transduction through interaction with IRFs, CBP and p300, transcriptional factors that form transcriptional complexes on the ISRE in the promoters of interferon-stimulated genes. vIRF has also been shown to have transactivation activity in a mammalian one-hybrid assay and regulates the expression of other KSHV genes (Li et al., 1998; Zimring et al., 1998). Many herpesviruses transactivators have an auto-activation function, contributing to the transcriptional regulation of their own genes (Deng et al., 2000; Flemington & Speck, 1990; Ragoczy & Miller, 2001). To determine whether vIRF activates its own promoter, we cotransfected a vIRF mammalian expression plasmid, pCMV-Tag2/vIRF, with a reporter plasmid, pCAT-991 (Wang et al., 2001), containing a 1·052 kbp sequence from the 5'-flanking region of the vIRF promoter into 293 cells. Two µg pCAT-991 DNA was cotransfected with 2 µg pCMV-Tag2/vIRF DNA into 293 cells by lipofection. The cells were harvested 48 h post-transfection and CAT activity was determined as described previously (Gao et al., 1997). Compared with the control cells transfected with pCAT-991, cotransfection with vIRF expression plasmid pCMV-Tag2/vIRF increased the CAT activities of pCAT-991 by 11-fold (Fig. 1A). Western blot analysis with vIRF-specific monoclonal antibody 2H5 showed that vIRF protein was expressed in cells transfected with pCMV-Tag2/vIRF but not in control cells (Fig. 1B). These results suggest that vIRF can activate its own promoter.

(18K): Fig. 1. vIRF activates its own promoter. (A) Ectopic expression of vIRF activates its own promoter by 11-fold. Two µg of expression plasmid pCMV-Tag2/vIRF was cotransfected with an equal amount of pCAT-991 into 293 cells. The cells were harvested for CAT assay 48 h post-transfection. CAT activity was measured with a Molecular Imager and calculated as the conversion rate of the modified 14C-labelled chloramphenicol. The total amount of transfected DNA in each experiment was calibrated by adding DNA of the blank plasmid, pCMV-Tag2. Transfection efficiencies were normalized by cotransfection with a reporter plasmid, pSV-β-galactosidase, followed by determination of β-galactosidase activity. CAT activity of cells transfected with reporter construct pCAT-991 alone was normalized to a value of 1. The results are shown as the average fold activation of CAT activity with standard deviations from three independent experiments. (B) Representative illustration of vIRF protein expression in cells transfected with plasmid pCMV-Tag2/vIRF. Transfected cells were analysed by Western blot with vIRF-specific monoclonal antibody 2H5. Lane 1: cells cotransfected with pCAT-991 and pCMV-Tag2/vIRF. Lane 2: cells transfected with pCAT-991. The loading of total cellular proteins in each lane was monitored by the detection of tubulin expression using an anti-tubulin monoclonal antibody.

vIRF activation of its own promoter is cell line-independent
To extend the above observation, similar cotransfection experiments were performed in three other cell lines, BJAB, COS7 and HeLa. Transfection of COS7 and HeLa cells was carried out as described above for 293 cells. BJAB cells (1x10⁷) were cotransfected with 15 µg pCAT-991 and 5 µg pCMV-Tag2/vIRF by electroporation. As shown in Fig. 2, vIRF also activated its own promoter in BJAB, COS7 and HeLa cells, increasing the promoter activity of pCAT-991 by 9-, 5·1- and 3·7-fold, respectively. These results indicate that vIRF activation of its own promoter is cell line-independent.

(12K): Fig. 2. vIRF activation of its own promoter is cell line-independent. vIRF activated its own promoter in BJAB (A), COS7 (B) and HeLa (C) cells. BJAB cells (1x107) were cotransfected with 15 µg CAT reporter construct pCAT-991 and 5 µg vIRF expression plasmid pCMV-Tag2/vIRF by electroporation. COS7 and HeLa cells were cotransfected with 2 µg pCAT-991 and an equal amount of pCMV-Tag2/vIRF by lipofection. CAT activity was measured and calculated as in Fig. 1. The results are shown as the average fold activation of CAT activity with standard deviations from three independent experiments.

vIRF activates its own promoter via two cis elements
To map the regions involved in vIRF activation of its own promoter, we performed deletion analysis on the upstream regulatory sequence of the vIRF promoter. vIRF expression plasmid pCMV-Tag2/vIRF was cotransfected with four reporter constructs, pCAT-56, pCAT-125, pCAT-337 and pCAT-449, containing different DNA fragments from the upstream regulatory sequence of the vIRF promoter (Fig. 3A), into 293 cells. As shown in Fig. 3(B), ectopic expression of vIRF activated CAT activities of pCAT-56 and pCAT-125 by 3·2- and 2·7-fold, respectively. No activation by vIRF was observed in cells transfected with pCAT-337 and pCAT-499. We have previously mapped the vIRF core promoter within the region from -56 to +1, the transcriptional silencer Tis within the region from -241 to -219 and a TPA-responsive element (TRR) within the region from -499 to -991 in the upstream regulatory sequence of the vIRF promoter (Wang et al., 2001, 2002). Tis has a strong repression effect on the vIRF promoter activity, which is relieved by TPA induction in KSHV-infected cell lines but not KSHV-negative cell lines (Wang et al., 2001, 2002). The above results indicated that vIRF activates its own promoter via two cis elements: one is the core promoter region from -56 to +1 and one is in the region that also contains the TRR from -499 to -991. We named these two cis elements Vac1 and Vac2, respectively. As shown in Fig. 3(B), vIRF activation of its own promoter via Vac1 was repressed by Tis, which was further reversed by vIRF activation via Vac2. vIRF activation of its own promoter via Vac1 and Vac2 could also be observed in BJAB (Fig. 3C), COS7 (Fig. 3D) and HeLa (Fig. 3E) cells.

vIRF activation of its own promoter is dose-dependent
To determine further the specificity of vIRF activation of its own promoter, we assayed the dose-responsiveness of Vac1 and Vac2 to ectopic expression of vIRF protein. vIRF promoter reporter constructs were cotransfected with different doses (0·1, 0·5, 0·75, 1·0, 1·5, 2·0, 2·5, 3·0 and 3·5 µg) of vIRF expression plasmid pCMV-Tag2/vIRF into 293 cells. As shown in Fig. 4(A) and (B), the expression levels of vIRF protein detected by monoclonal antibody 2H5 increased in a step-wise fashion following the increasing amount of vIRF plasmid DNA. As expected, the CAT activities of cells transfected with pCAT-56 and pCAT-125 containing Vac1 and pCAT-991 containing Vac1 and Vac2 also increased in a dose-dependent fashion, ranging from 2·3- to 11-fold (Fig. 4C). Again, the increase in vIRF protein expression had minimal effect on the reporter constructs pCAT-337 and pCAT-499.

(30K): Fig. 4. vIRF activation of its own promoter is dose-dependent. (A) Results from a representative Western blot assay showing that vIRF expression increased in a step-wise fashion correlating with increasing amounts (0·1, 0·5, 0·75, 1·0, 1·5, 2·0, 2·5, 3·0 and 3·5 µg) of vIRF plasmid pCMV-Tag2/vIRF (upper panel). vIRF protein was detected with monoclonal antibody 2H5. Anti-tubulin monoclonal antibody was used to detect tubulin and monitor quantities of loaded proteins (lower panel). (B) The vIRF protein expression level was quantified after calibration based on tubulin expression levels. (C) Dose-responsiveness of different vIRF promoter reporter constructs to vIRF activation. Different amounts of pCMV-Tag2/vIRF (0, 0·2, 0·5, 1·0, 1·5, 2·0, 2·5 and 3·0 µg; bars, left to right) were cotransfected with CAT reporter constructs pCAT-56, pCAT-125, pCAT-337, pCAT-499 and pCAT-991. Cells were harvested 48 h post-transfection and used to confirm vIRF protein expression and determine CAT activities as described in Fig. 1. The results are the averages of relative CAT activities with standard deviations from three independent experiments.

Ectopic expression of vIRF activates endogenous vIRF gene expression from viral genomes in KSHV-infected cell lines
The above results demonstrated that vIRF activates its own promoter in an engineered reporter system. To examine further whether vIRF could activate the expression of its own gene in endogenous viral genomes in KSHV-infected cells, 20 µg vIRF expression plasmid pCMV-Tag2/vIRF was transiently transfected into 1x10⁷ KSHV-positive BCBL-1 and BC-1 cells. Two days after transfection, the total RNA was isolated and Northern blot hybridization was performed to detect the expression of both ectopic and endogenous vIRF transcripts. The vIRF gene encodes an endogenous transcript of 1·7 kb in BCBL-1 and BC-1 cells (Wang et al., 2001), while pCMV-Tag2/vIRF encodes a vIRF transcript of 1·8 kb (Fig. 5A). As shown in Fig. 5, the ectopic expression of vIRF pCMV-Tag2/vIRF activated the expression of endogenous vIRF transcripts. The level of endogenous vIRF transcripts was increased 12-fold in BCBL-1 cells and 1830-fold in BC-1 cells (Fig. 5B). These results indicated that vIRF is capable of auto-activating the expression of its own gene from endogenous viral genomes in KSHV-infected cells.

(21K): Fig. 5. Exogenous vIRF expression activates endogenous vIRF gene expression in KSHV-positive cell lines. (A) vIRF expression plasmid pCMV-Tag2/vIRF was transfected into KSHV-positive BCBL-1 and BC-1 cells by electroporation. Total RNA was isolated from transfected cells 48 h post-transfection and used to detect different vIRF transcripts by Northern blot hybridization. BJAB cells transfected with pCMV-Tag2/vIRF (lane 1) were used as an exogenous vIRF transcript control (1·8 kb). TPA-treated BCBL-1 (lane 3) and BC-1 (lanes 6 and 9) cells were used as endogenous vIRF transcript controls (1·7 kb). Untreated BCBL-1 (lane 2) and BC-1 (lanes 5 and 8) cells had minimal expression of endogenous vIRF transcript while TPA treatment significantly induced expression of the endogenous vIRF transcript. Ectopic expression of pCMV-Tag2/vIRF also strongly activated the expression of endogenous vIRF transcripts from latent KSHV genomes in BCBL-1 (lane 4) and BC-1 (lanes 7 and 10) cells. The relative RNA loading was monitored by hybridization with a β-actin gene probe after stripping of the membrane (lower panel). (B) Activation of endogenous vIRF transcripts by TPA treatment and exogenous vIRF expression (pCMV-Tag2/vIRF). The expression level of endogenous vIRF transcripts was calibrated with the β-actin gene. The cell sample treated with TPA was taken as 100. The experiments were performed five times with similar results. Results from one experiment are presented.

Vac1 and Vac2 are unresponsive to induction with interferons
Since vIRF targets cellular genes whose promoters contain ISREs through interaction with IRFs, p300 and CBP, it is possible that vIRF auto-activation via Vac1 and Vac2 is responsive to induction with interferons. However, examination of the upstream regulatory sequences in the vIRF promoter has failed to identify any ISRE- and GAS-like elements that are known to be responsive to interferons, suggesting that Vac1 and Vac2 are unlikely to be responsive to induction with interferons. To confirm this, CAT reporter constructs pCAT-125, pCAT-337, pCAT-499 and pCAT-991 were transfected either alone or cotransfected together with pCMV-Tag2/vIRF into 293 cells. After 24 h of transfection, the cells were treated with either 1000 U interferon-β or -γ ml^-1 for 24 h. As shown in Fig. 6(A), the CAT activities of the cells transfected with all reporter constructs had minimal changes after treatments with either interferon-β or -γ compared with untreated controls. Treatment with interferon-β or -γ also had no effect on vIRF auto-activation via Vac1 and Vac2 (Fig. 6B). These results demonstrated the unresponsiveness of Vac1 and Vac2 cis elements alone or after activation by vIRF to induction with interferons.

(29K): Fig. 6. vIRF auto-activation is independent of interferon signal pathways. Different vIRF promoter reporter constructs were transfected either alone (A) or together with 2·0 µg pCMV-Tag2/vIRF plasmid (B) into 293 cells and cultured for 24 h before being treated with either interferon-β or -γ for 24 h. CAT activities were determined as described in Fig. 1. The results are shown as the averages of relative CAT activities with standard deviations from three independent experiments.

vIRF is a transforming protein that causes cellular transformation when expressed in NIH3T3 and Rat-1 cells (Gao et al., 1997; Zimring et al., 1998). vIRF inhibits interferon-mediated signal transduction via direct binding to IRF-1, IRF-3, p300 and CBP (Burysek et al., 1999; Seo et al., 2000) and inhibits p53-mediated transcriptional regulation through interaction with p53 (Nakamura et al., 2001; Seo et al., 2001). vIRF also behaves as a transcriptional activator when directed to DNA by the GAL4 DNA-binding domain in a mammalian one-hybrid assay and transactivates the c-myc proto-oncogene through an ISRE-like sequence called the plasmacytoma repressor factor element in the c-myc promoter (Jayachandra et al., 1999; Roan et al., 1999). In this study, we have examined vIRF regulation of its own expression. Our results have demonstrated that ectopic expression of vIRF transactivates its own promoter in a reporter assay and the expression of the vIRF gene in endogenous viral genomes of KSHV latently infected cells. This auto-amplification mechanism could ensure the rapid and efficient expression of the vIRF gene following induction by endogenous or exogenous factors during KSHV infection.

vIRF activates its own promoter through two cis elements, Vac1 and Vac2. Vac1 is located between -56 and +1, a region that contains the core promoter of the vIRF gene (Wang et al., 2001). Vac2 is located between -499 and -991, a region that contains the TRR (Wang et al., 2001). vIRF auto-activation through Vac1 is suppressed by Tis (-241 to -219) (Wang et al., 2002); however, vIRF auto-activation of its own promoter through Vac2 overcomes the suppression effect of Tis. It is likely that these two vIRF-responsive cis elements cooperate with each other to produce a synergistic effect to ensure efficient vIRF expression during KSHV lytic replication. The fact that Vac2 is located within a region containing the TRR suggests that this region is most likely a common target site of viral or cellular transactivators. Further delineation of this region could shed light on the transcriptional regulation of other KSHV lytic genes during viral lytic replication.

All the cellular IRFs bind to ISREs to regulate the expression of interferon-stimulated genes (Darnell et al., 1994). Although vIRF does not directly bind to the ISRE sequence, it interacts with IRF1, IRF3 and IRF7 to inhibit interferon signalling and regulate the expression of interferon-stimulated genes (Burysek et al., 1999; Gao et al., 1997; Seo et al., 2000). Unlike other vIRF-responsive genes identified so far, the expression of the vIRF gene itself is not responsive to treatments with interferon-β and -γ. Indeed, both Vac1 and Vac2, regardless of activation by vIRF, are not responsive to treatments with interferon-β and -γ, indicating that vIRF also targets genes whose promoters do not contain ISRE-like sequences. Further mapping of Vac1 and Vac2 will most likely provide insight into the new function of vIRF and transcriptional regulation of viral and cellular genes by vIRF.

This work was supported by NIH grant HL60604 (S.-J. G.) and a grant from the Association for International Cancer Research (S.-J. G.).