Kaposi Sarcoma-associated herpes virus (KSHV) G protein-coupled receptor (vGPCR) activates the ORF50 lytic switch promoter: A potential positive feedback loop for sustained ORF50 gene expression

Abstract

KSHV vGPCR, a lytic cycle associated protein, induces several signaling pathways leading to the activation of various transcription factors and consequently the expression of cellular and viral genes. Though the role of vGPCR in KSHV tumorigenicity has been well studied, its function related to the viral life cycle is poorly understood. Reduction in vGPCR by RNA interference also resulted in the reduction in KSHV lytic switch ORF50 gene and protein expression. Induction of vGPCR by doxycycline in BC3.14 cells also resulted in more KSHV production. When this was explored, induction of the ORF50 promoter by vGPCR expression was observed. Further examination of the molecular mechanisms by which vGPCR regulates the ORF50 promoter, using various ORF50 promoter constructs, revealed that induction of ORF50 promoter by vGPCR did not involve AP1 but was dependent on Sp1 and Sp3 transcription factors. vGPCR signaling led to an increase in Sp1 and Sp3 DNA binding activity and a decrease in histone deacetylase (HDAC) activity. These activities were pertussis toxin independent, did not involve Rho and Rac-GTPases and involved the heterotrimeric G protein subunits Gα12 and Gαq. Studies using pharmacologic inhibitors and dominant negative proteins identified phospholipase C, the novel protein kinase C (novel PKC) family and protein kinase D (PKD) as part of the signaling initiated by vGPCR leading to ORF50 promoter activation. Taken together, this study suggests a role for vGPCR in the sustained expression of ORF50 which could lead to a continued activation of lytic cycle genes and ultimately to successful viral progeny formation.

INTRODUCTION

Kaposi Sarcoma-associated herpesvirus (KSHV) or human herpesvirus-8 (HHV-8) is a γ2-human herpesvirus etiologically associated with Kaposi Sarcoma (KS), a multifocal angioproliferative disease consisting of spindle cells of lymphatic endothelial origin and infiltrating hematopoietic cells (Ganem). In addition, KSHV is also associated with two B cell malignancies, primary effusion lymphoma (PEL) and multicentric Castleman disease (MCD) (Cesarman et al., 1995; Ganem). Similar to the other members of the herpes virus family, KSHV exhibits two different life cycles namely latent or lytic cycle, in an infected host. Of its more than 90 genes, KSHV expresses only a handful of genes in the latent state such as ORF71 (coding for v-FLIP), ORF72 (coding for v-Cyclin), ORF73 (coding for LANA-1), ORF K10.5 (coding for v-IRF3 in PEL cells), kaposins A, B and C, and 12 microRNAs (Cai et al., 2005; Dourmishev et al., 2003; Ganem; Samols et al., 2005; Staskus et al., 1997; Zhong et al., 1996). In PEL cells, more than 40 copies of the viral genome are maintained as a nuclear episome and replicates with the host cell DNA. During this stage, no new viral particles are produced. About 1% to 3% of the PEL cells spontaneously enter the lytic program by an unknown mechanism, and about 10% to 25% of cells enter the lytic phase after treatment with phorbol esters or histone deacetylase inhibitors (sodium butyrate) (Ganem).

During lytic phase, a cascade of immediate early, early, and late viral genes is initiated resulting in viral DNA replication, assembly, and cell lysis / release of progeny infectious virions (Ganem). Several KSHV lytic cycle proteins are responsible for immune evasion, inhibition of apoptosis, host gene modulation, host protein expression shutoff, and modulation of signal transduction. These genes include K2 (v-interleukin-6 [vIL-6]), K4 (v-macrophage inhibitory protein II [vMIP-II]), K3 and K5 (MIR-1 and MIR-2; immunomodulatory proteins), K6 (vMIP-I), K7 (anti-apoptotic protein), K9 (v-interferon regulatory factor [vIRF]), vIRF2 (K11.1), ORF16 (vBcl-2), K13 (v-FLICE-inhibitory protein [vFLIP]), K14 (vOX-2), ORF72 (v-cyclin D), and ORF74 (v-G protein-coupled receptor) (Ganem; Sun et al., 1999). The first gene activated by chemical treatments is KSHV ORF50 coding for replication and transcription activator (RTA). It has been shown that over-expression of ORF50 (RTA) is enough to induce lytic reactivation whereas its deletion or mutation totally inhibits spontaneous as well as chemically inducted-lytic reactivation (Lukac et al., 1998; Staudt and Dittmer, 2007; Xu et al., 2005). All these observations define ORF50 (RTA) as the lytic master-switch protein. ORF50 (RTA) contains an N-terminal DNA binding domain and a C-terminal activation domain and is able to activate the transcription of several other viral genes after binding to a specific RTA-responsive element (RRE) sequence (Ganem; Staudt and Dittmer, 2007). RRE sequences were reported in several KSHV early lytic genes (e.g. tk, ORF57, K8, K2, K9, K14) as well as in the promoter of glycoprotein gB, a late gene (Ganem). Interestingly, ORF50 (RTA) can also activate its own promoter creating an amplification loop (Deng, Young, and Sun, 2000).

Due to its crucial role in lytic reactivation, ORF50 promoter regulation has been under intense investigation. All the known reactivation inducers were able to activate the ORF50 promoter by different mechanisms. TPA has been shown to induce ORF50 through the activation of the AP1 transcription factor (Wang et al., 2004) while the activation of the ORF50 promoter after sodium butyrate treatment has been shown to involve the Sp1/Sp3 transcription factors (Lu et al., 2003; Ye, Shedd, and Miller, 2005). During hypoxia, another lytic reactivation inducer, the hypoxia-inducible factor-1 (HIF-1) family of transcription factors, is able to activate the ORF50 promoter (Cai et al., 2006; Davis et al., 2001; Haque et al., 2003). ORF50 auto-regulation is also mediated through ORF50 (RTA) interaction with other transcription factors such as the RBP-Jκ transcription factor (Ganem). Whereas RBP-Jκ protein alone acts as a sequence specific transcription repressor, it becomes a potent transcription activator after binding to ORF50 (RTA) (Wang and Yuan, 2007). In addition to this mechanism, ORF50 (RTA) can also activate its own promoter through its interaction with the Oct1 transcription factor (Sakakibara et al., 2001). Several proteins have been shown to inhibit the ORF50 promoter (Lan, Kuppers, and Robertson, 2005; Lan et al., 2004; Yada et al., 2006). For example, the latency associated nuclear KSHV protein LANA-1 has been proposed to inhibit viral replication by repressing ORF50 through its interaction with ORF50 (RTA) or RBP-Jκ (Lan, Kuppers, and Robertson, 2005; Lan et al., 2004). The ORF50 promoter can also be repressed after binding of the transcriptional repressor Hey1 (Yada et al., 2006).

KSHV ORF74, an early lytic gene activated by ORF50 (RTA), codes for a seven transmembrane domain G-protein coupled receptor (vGPCR). ORF74 expression has been observed as early as ten hours in TPA treated BC3 cells and is not affected by an inhibitor of DNA replication like Cidofovir (Jenner et al., 2001; Lu et al., 2004). vGPCR possesses limited homology to human chemokine receptors such as the IL8 receptors (CXCR1 and CXCR2). In contrast to these receptors, vGPCR is constitutively active although it can still be modulated somewhat by various ligands (Rosenkilde et al., 1999). The first downstream signaling molecules activated by vGPCR are members of the heterotrimeric G protein family (Liu et al., 2004). Consequently, vGPCR can activate phosphatydylinositol 3-kinase (PI3K), phosphoinoside-dependent kinase (PDK) and AKT/protein kinase B (AKT/PKB) (Cannon, 2007). vGPCR is also known to activate MAP kinase cascades (JNK, ERK, p38) (Cannon, 2007). In addition to heterotrimeric G proteins, vGPCR can signal through small GTPases including RhoA and Rac1 (Martin et al., 2007; Montaner et al., 2004; Shepard et al., 2001). All these signaling pathways lead to the activation of key cellular transcription factors, such as activating protein-1 (AP1), nuclear factor kappa-B (NF-κB), nuclear factor activator of T cells (NFAT), cyclic AMP response element binding protein (CREB), and hypoxia-inducible factor-1 (HIF1) (Cannon, Philpott, and Cesarman, 2003; Cannon and Cesarman, 2004; Montaner et al., 2004; Sodhi et al., 2000). These transcription factors in turn regulate several cellular and viral genes.

Several studies highlight the importance of vGPCR in KS formation (Cannon, 2007; Sodhi, Montaner, and Gutkind, 2004). vGPCR can transform NIH3T3 fibroblasts in vitro and vGPCR-expressing 3T3 cells form tumors in mice. In addition, transgenic mice over-expressing vGPCR develop tumors resembling KS lesions (Montaner et al., 2003; Yang et al., 2000), and an angiogenic paracrine effect has been proposed to be involved in vGPCR tumorigenicity via the secretion of VEGF, VEGFR2 (KDR), Gro1 and IL8 (Bais et al., 2003; Martin et al., 2008). However, the role of KSHV vGPCR in viral replication is not fully understood. The function of vGPCR and other virally encoded GPCRs has been investigated in several herpesviruses. In the β-herpesvirus subfamily, the viral GPCR of murine cytomegalovirus (M78) is involved in the accumulation of several early lytic mRNAs (Oliveira and Shenk, 2001). Cells infected with a viral mutant lacking most of M78 ORF showed a reduced level of the m123 mRNA and consequently a reduced level of m54 (early) and m99 (late) messengers. In the γ-herpesvirus subfamily, the murine MHV68 viral GPCR has been shown to be involved in lytic reactivation (Lee et al., 2003; Moorman, Virgin, and Speck, 2003). In addition, KSHV vGPCR has also been shown to regulate some viral genes such as vIL6 (Cannon, Philpott, and Cesarman, 2003).

In the course of examining the vGPCR functions via the use of siRNA, we observed that si-vGPCR not only reduced the expression of vGPCR but also ORF50 expression. To follow up on this interesting observation we set out to determine the potential relationship between vGPCR expression and activation of the lytic cycle. Our studies define the molecular mechanisms by which vGPCR potentially regulates the ORF50 lytic switch promoter. Taken together, these studies suggest that vGPCR acts as part of a positive feedback mechanism activating the ORF50 promoter, and indicate a potential role for vGPCR in the KSHV lytic cycle.

RESULTS

KSHV vGPCR down-regulation reduces lytic ORF 50 gene expression in PEL cells

KSHV vGPCR has been shown to induce a wide variety of signal transduction pathways leading to the activation of various transcription factors including NF-κB, AP1, NFAT and CREB (Cannon, 2007; Ganem). Even though vGPCR function during KSHV oncogenesis has been highly investigated (Cannon, 2007; Sodhi, Montaner, and Gutkind, 2004), little is known about the role of this protein during the viral life cycle. Therefore, we set out to investigate the role of vGPCR-induced signaling in viral gene expression.

To determine the role of vGPCR in PEL cells, we used RNA interference technology to down-regulate vGPCR. We chose five different siRNA target sequences from vGPCR ORF, constructed siRNA expression cassettes driven by the U6 promoter and cloned these into the pGMT-easy plasmid (Tiscornia, Singer, and Verma, 2006a). Human embryonic kidney (HEK293T) cells were co-transfected with a vGPCR expression plasmid along with plasmids expressing the five different siRNA constructs (si-vGPCR1-5), and the level of vGPCR mRNA was measured by real-time RT-PCR. All five siRNAs reduced vGPCR mRNA levels (data not shown) and si-vGPCR-1 showed the greatest effect with about 65% reduction (Fig. 1A). The si-vGPCR-1 cassette was then sub-cloned into the lentiviral vector and HEK293T cells were transduced with the lentivectors expressing si-vGPCR-1 or control si-Luciferase (si-C). One week after transduction, the cells were transfected with a vGPCR expression plasmid and vGPCR protein level was determined by Western Blot (Fig. 1B). As expected, untransfected HEK293T cells did not express vGPCR (Fig. 1B, lane 1). Compared to HEK293T cells expressing no siRNA (Fig. 1B, lane 2) or siRNA against Luciferase (si-C) (Fig. 1B, lane 3), cells transduced with si-vGPCR-1 showed reduction in vGPCR protein (Fig. 1B, lane 4). Densitometric quantification demonstrated that the cells expressing no siRNA or si-C expressed similar levels of vGPCR (100% and 92%), whereas si-vGPCR-1 cells expressed only about 37% of vGPCR (Fig.1.B).

Figure 1.

vGPCR silencing. (A) HEK293T cells were transfected with vGPCR expressing vector together with a plasmid expressing an siRNA control against luciferase (si-C) or siRNA against vGPCR (si-vGPCR-1). RNAs were prepared, reverse transcribed into cDNA, levels of vGPCR measured by quantitative real-time RT-PCR and normalized to the levels of the house keeping gene tubulin. (B) HEK293T cells were transduced with a lentivector expressing a control siRNA (si-C) or a siRNA against vGPCR (si-vGPCR-1). The lentivector also expressed GFP under a CMV promoter thus allowing for measurement of the level of transduction. One week later, the cells were transfected with an empty expression plasmid or a plasmid expressing vGPCR. 48h later, the cells were lysed in RIPA lysis buffer and 50 µg of protein were subjected to Western blot analysis. Detection of tubulin was used as a loading control and GFP as transduction control. (C) BCBL-1 cells were transduced with the lentivector expressing the control siRNA against luciferase (si-C) or with the lentivector expressing the siRNA against vGPCR (si-vGPCR). Both lentivectors also expressed GFP under a constitutive promoter allowing for the estimation of the level of transduction by fluorescence microscopy and flow cytometry. One week after transduction, the cells were treated with TPA (20 ng/ml) for 0 or 4 days. The cells were then assayed for accumulation of vGPCR and ORF50 mRNA by quantitative real-time RT-PCR. Each reaction was done in triplicate and a representative experiment of three experiments is shown here. +/− SD; ***p<0.0005. (D) BCBL-1 cells were transduced with the lentivector expressing si-C or si-vGPCR. After one week, cells were untreated or treated with TPA (20 ng/ml) for 3 days. Total protein extracts were prepared and the level of ORF50 and vGPCR expression was assessed by western blot. Tubulin was used as loading control.

After confirming the efficiency of siRNA, we checked their effect on the TPA-induced KSHV lytic cycle. BCBL1 cells were transduced with lentivectors expressing si-vGPCR-1 or si-Luciferase at an moi of 10. The lentivectors express GFP in addition to the siRNA cassette, and in flow cytometric analysis about 80% of the cells were positive for GFP (data not shown). One week later, cells were treated with TPA for 4 days and the RNA extracted from these cells was used in real-time RT-PCR to measure the mRNA levels of vGPCR gene and the lytic switch ORF50 (RTA) gene. As shown in Figure 1C, compared to the si-Luciferase (si-C) cells, vGPCR levels were reduced considerably in si-vGPCR expressing cells (78.5% inhibition). Though the lentiviral tranduction was stable in these cells, the efficiency of vGPCR reduction appeared to decrease over time. For this reason, the experiment was carried out one week after transduction and the cells were not kept for an extended period.

Interestingly, we observed a significant reduction in ORF50 mRNA levels in si-vGPCR expressing cells with about 43.6% inhibition at day 4 (Fig. 1C), thus suggesting that vGPCR could be involved in regulation of ORF50 gene expression. To determine whether reduction in ORF50 mRNA levels in si-vGPCR expressing cells also results in the reduction of ORF50 at the protein level, BCBL-1 cells transduced with si-C or si-vGPCR were treated with TPA for 3 days and protein extracts were used for western blot analysis. TPA treatment of control si-C transduced BCBL1 cells resulted in a high induction of ORF50 protein levels with about 32-fold induction compared to uninduced cells (Fig. 1D, lane 1). In contrast, in si-vGPCR expressing cells, ORF50 expression was reduced to 12-fold with about 40% reduction (Fig.1D, upper panel, lane 2). We confirmed that vGPCR level was reduced in the knock down cells (Fig.1D, middle panel, lane 2). This further suggested that vGPCR could be involved in the regulation of ORF50 gene expression. .

vGPCR expression induces ORF50 in PEL cells

BC3.14 cells express vGPCR in a doxycycline inducible manner (Cannon et al., 2000). We utilized these cells to measure ORF50 and other related gene expression by real-time RT-PCR after induction of vGPCR by doxycycline treatment. Normalization was performed to the expression of two house keeping genes (HPRT and tubulin) which showed no changes with doxycycline (data not shown). We observed a dose-dependent up-regulation of vGPCR transcript 48h after doxycycline treatment of BC3.14 cells with about 2.9, 3.1 and 3.5- fold induction of vGPCR at 0.05, 0.1 and 0.5 µg/ml of doxycycline, respectively (Fig. 2A). Interestingly, KSHV lytic switch ORF50 mRNA was also induced under these conditions. We observed about 1.4, 1.7 and 1.7-fold induction of ORF 50 genes after 48h induction with 0.05, 0.1 and 0.5 µg/ml concentrations of doxycycline, respectively (Fig. 2A). Under the same conditions, KSHV latency associated ORF73 messenger level did not change significantly (Fig. 2A). When the parental BC3 cells were treated with doxycycline, we did not observe any induction of vGPCR or ORF50 and thus demonstrating that doxycycline treatment did not spuriously induce Rta/vGPCR expression (Fig. 2B).

Figure 2.

Induction of vGPCR and KSHV lytic gene expression. (A) BC3.14 cells were treated for 48h with the indicated doses of doxycycline (0, 0.05, 0.1 and 0.5 µg/ml). The cells were then assayed for accumulation of vGPCR, ORF50, and ORF73 mRNAs by quantitative real-time RT-PCR. Results were normalized to HPRT. Each reaction was done in triplicate and the data represent the mean of four independent experiments +/− S.E.M. (B) BC3 cells were treated for 48h with the indicated doses of doxycycline (0, 0.1 and 0.5 µg/ml) and assayed for the accumulation of vGPCR and ORF50 mRNAs by quantitative real-time RT-PCR. Results were normalized to HPRT. Each reaction was done in triplicate and the data represent the mean of three independent experiments +/− S.E.M. (C) BC3.14 cells were treated for 48h with the indicated doses of doxycycline (0, 0.05, 0.1 and 0.5 µg/ml). The cells were then assayed for the accumulation of PAN, ORF57, and ORF26 mRNAs by quantitative real-time RT-PCR. Results were normalized to HPRT. Each reaction was done in triplicate and the data represent the mean of three independent experiments +/−S.E.M. (D) BC3.14 cells were treated for 3 days with TPA (20 ng/ml) or doxycycline (1 µg/ml). Total proteins were extracted in RIPA lysis buffer and subjected to a western blot experiment using the indicated antibodies. (E) BC3 and BC3.14 cells were treated for 2 days with doxycycline (0.5 µg/ml). After centrifugation for 10 minutes at 2000 rpm, the supernatant was filtered on a 0.45 µm filter. Virus quantification was performed by real time PCR quantitation of extracted viral DNA (ORF73 gene).

The polyadenylated nuclear (PAN) promoter contains ORF50 response elements and has been shown to be highly sensitive to ORF50 (Song et al., 2001; Song et al., 2002). Over-expression of vGPCR also induced the expression of PAN mRNA, with about 2, 2.5 and 2.6- fold increase in PAN mRNA at 0.05, 0.1 and 0.5 µg/ml of doxycycline, respectively (Fig.2C). We further investigated whether the induction of ORF50 could be responsible for the induction of ORF57, another ORF50 (RTA) dependent lytic gene (Bu et al., 2008; Byun et al., 2002; Song, Deng, and Sun, 2003; Wang et al., 2003). As shown in figure 2C, we observed about 2.8 and 6.6-fold induction of ORF 57 genes after 48h induction with 0.1 and 0.5 µg/ml concentrations of doxycycline, respectively. We did not observe any induction of ORF26, a lytic gene independent of ORF50 and thus demonstrating the specificity of these observations (Fig.2C).

TPA treatment resulted in higher levels of vGPCR mRNA (2-3-fold) than by doxycycline treatment (data not shown). ORF50 mRNA level was also higher under TPA than doxycycline treatment (data not shown). These observations indicated that vGPCR could participate in ORF50 regulation but TPA induced signaling is a more potent inducer. To confirm these results at the protein level, BC3.14 cells were treated for 3 days with 1 µg/ml doxycycline or with 20 ng/ml of TPA, lysed and proteins were analyzed by western blot using anti-vGPCR, ORF50 and tubulin antibodies (Fig. 2D). As expected, TPA induced both ORF50 and vGPCR expression (23 and 12 fold, respectively) (Fig.2.D, lane 3, top and middle panel). Doxycycline treatment induced about 7-fold vGPCR over uninduced cells (Fig.2.B, lane 2, middle panel). Interestingly, about 13-fold higher ORF50 levels were also observed after doxycycline treatment (Fig.2.B, lane 2, middle panel) suggesting ORF50 induction by vGPCR. Tubulin used as loading control did not show any difference between conditions.

Next, we investigated whether the activation of ORF50 by vGPCR results in increased KSHV production. We collected the medium after 48h of doxycycline treatment of BC3 or BC3.14 cells and quantified the viral particles by ORF73 DNA PCR. As shown in figure 2E, compared to uninduced BC3.14 cells, we detected >3-fold KSHV particles in the culture supernatants of doxycycline induced BC3.14 cells. No difference was observed in BC3 cells after doxycycline treatment. These results, together with the data shown in Fig. 1 clearly suggested a role for vGPCR in the regulation of ORF50 promoter.

vGPCR up-regulates the ORF50 promoter

We next investigated the effect of vGPCR expression on the ORF50 promoter in the absence of other viral proteins. HEK293T cells were co-transfected with an increasing quantity of vGPCR expression plasmid along with the ORF50 (−2500)-Luciferase reporter gene (p2500Luc, corresponding to 2500 bases upstream of ORF50 ATG) and a RSV-β-Gal reporter gene in which the β-galactosidase gene is under a constitutive promoter (transfection control). 48h after transfection, lysates were assayed for the luciferase and β-gal activities. Consistent with previous studies (Cannon, Philpott, and Cesarman, 2003), increasing quantities of vGPCR induced the ORF50 promoter in a dose-dependant manner (Fig. 3B). Indeed, the ORF50 promoter was activated 2.6, 2.6 and 3.5-fold after transfection of HEK293T with 0.5, 1 and 2 µg of vGPCR expressing plasmid, respectively. In contrast, increasing expression of vGPCR had no effect on the minimal pGL3 plasmid promoter (Fig. 3A). In vGPCR-expressing BC3.14 cells, we observed an increase in ORF50 (RTA) activated PAN mRNA (Fig. 2C). This could be a direct effect of vGPCR signaling or through the induction of ORF50 (RTA) since the PAN promoter contains an RTA responsive element (Wang et al., 2004). The PAN promoter p1241 (spanning bp −1241 to +14. nucleotides 27426 to 28680) was induced 2.2, 2.4 and 3.6-fold after transfection of HEK293T cells with 0.5, 1 and 2 µg of vGPCR expressing plasmid respectively (Fig. 3C), suggesting that vGPCR signaling can induce PAN promoter independently of ORF50 (RTA).

Figure 3.

(A–F) Effect of vGPCR on different KSHV promoters. HEK293T cells were transfected by the calcium phosphate method with pcDNA or an increasing quantity of vGPCR expression plasmids (0, 0.5, 1 or 2 µg), and 1 µg of the indicated promoter luciferase constructs (A: pGL3, B: ORF50-Luc; C: PAN-Luc; D: LanaPc-Luc; E: LanaPi-Luc; F: K14-Luc), and 0.5 µg of the β-gal reporter construct (as transfection control). Cells were harvested 36h after transfection. The luciferase and β-galactosidase activities were measured in triplicate using Steady-glo and Beta-glo Promega kits, respectively. The luciferase activity was normalized to β-galactosidase activity as transfection control. The data represent the mean of fold induction of at least three independent experiments +/− S.E.M. ** p<0.05. ***p<0.01. (G) HEK293T cell were transfected by the calcium phosphate method with 3 µg of pcDNA or vGPCR expression plasmids. The same day, other HEK293T cells were transfected with 2 µg of the ORF50 promoter luciferase constructs and 1 µg of the β-Gal reporter construct as transfection control. The day after transfection, the medium was replaced by fresh medium and was collected 24h later. The medium isolated from pcDNA or vGPCR transfected cells was added to the cells transfected with the reporter constructs. 8h or 16h later, the luciferase and β-galactosidase activities were measured in triplicate using Steady-glo and Beta-glo Promega kits, respectively, and the luciferase activity was normalized to β-galactosidase activity. The data represent the mean of fold induction of five independent experiments +/− S.E.M.

We observed no significant changes in ORF73 mRNA levels after vGPCR induction in BC3.14 cells (Fig.2A). Two distinct promoters have been shown to activate the ORF73 gene. LanaPc is a constitutive promoter activated during latency whereas LanaPi is an inducible promoter activated by ORF50 (RTA) (Staudt and Dittmer, 2006). As shown in Figure 3E, vGPCR has no effect on the LanaPi promoter and a significant effect on the LanaPc promoter was observed only at one concentration of vGPCR plasmid (Fig. 3D). vGPCR protein is coded by a bi-cistronic messenger also coding for the K14 protein (Jeong, Papin, and Dittmer, 2001; Staudt and Dittmer, 2006). We investigated if vGPCR could also activate its own promoter (K14p) in a luciferase assay. As shown in Figure 3F, increasing quantities of vGPCR are able to significantly induce K14 promoter with about 1.9, 1.9 and 2- fold with 0.5, 1 and 2 µg of vGPCR expression plasmid, respectively. Taken together, these results indicate that vGPCR is involved in the regulation of various early lytic promoters.

The effect on the ORF50 promoter by vGPCR does not involve a paracrine mechanism

vGPCR mediates the induction of host cell intrinsic signaling pathways which leads to the secretion of a wide variety of cytokines such as Gro-α, IL-6, VEGF-C, MIP-1α and MIP-1β (Polson et al., 2002). To determine whether a paracrine effect is responsible for activation of the ORF50 promoter, HEK293T cells were transfected with an empty vector or with vGPCR. 24h later these conditioned media were applied onto HEK293T cells that were previously transfected with the ORF50 luciferase construct and the β-gal reporter control, and incubated for 8 or 16h prior to the luciferase and β-Gal assays. As shown in Fig. 3G, the medium isolated from vGPCR transfected cells was unable to induce the ORF50 promoter. This result suggests that vGPCR’s effect on the ORF50 promoter is due to an intrinsic signaling pathway rather than a paracrine mechanism.

Identification of vGPCR-responsive elements in the ORF50 promoter

Since ORF50 is the lytic switch protein, we next examined the precise mechanism of vGPCR activation of ORF50. The ORF50 promoter has been under intense investigation, highlighting the importance of several transcription factor binding sites. To determine the minimal promoter region responsive to vGPCR expression, a series of ORF50 promoter deletion constructs were assayed in a luciferase reporter experiment in HEK293T cells (Fig. 4A). vGPCR over-expression was able to activate p2500Luc, p950Luc and p134Luc at a similar level with about 3.6, 4 and 3.4-fold activation, respectively, while the p75Luc and p69Luc promoters had low basal activity (Fig. 4B). In addition, compared to pcDNA transfection, these two latter promoters were only marginally activated by vGPCR expression with about 1.4 and 1.3-fold, respectively. Together, these results suggested that the promoter region located between −134 and −75 is required for ORF50 promoter activation induction by vGPCR.

Figure 4.

(A and B) Mapping of ORF50 promoter elements involved in vGPCR induced activation. HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or vGPCR expression plasmids, 1 µg of the indicated ORF50 promoter luciferase constructs, and 0.5 µg of the β-gal reporter construct. Cells were harvested 36h after transfection, the luciferase and β-galactosidase activities measured in triplicate, and the luciferase activity was normalized to β-galactosidase activity. The data represent the mean of fold induction of four independent experiments +/− S.E.M.

AP1 site is not involved in vGPCR induction of the ORF50 promoter

The region located between −134 and −75 contains the AP1 (CGACTCA), C/EBP and Sp1 sites (Fig. 4A). TPA treatment induces activation of the KSHV lytic cycle and previously it has been shown that TPA induces the phosphorylation of cJun as well as its DNA binding activities (Yu et al., 1999). In addition, over-expression of cJun and cFos were able to activate the ORF50 promoter through the AP1 site (Wang et al., 2004). Interestingly, vGPCR can activate AP1 mediated transcription in HEK293, COS1 and PEL cells (Cannon, Philpott, and Cesarman, 2003). To test if vGPCR activation of the ORF50 promoter is mediated by AP1, an AP1 site deletion reporter construct was assayed (p115Luc) (Fig. 5A). To our surprise, p134Luc and p115Luc were both responsive to vGPCR expression with about 2.8 and 2.5-fold activation, respectively, compared to pcDNA transfection (Fig. 5A). To confirm this result, a reporter plasmid with a mutated AP1 site (p134AP1M) was tested (Fig. 5B). Compared to empty vector transfection, vGPCR transfection induced a 2.8-fold activation of p134Luc and 4.2-fold of p134LucAP1m (Fig. 5B). Taken together, these results indicate that vGPCR activation of the ORF50 promoter is not mediated through the AP1 site.

Figure 5.

Sp1, but not AP1, activates the ORF50 promoter after vGPCR expression. (A and B) HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or vGPCR expression plasmids, 1 µg of the indicated ORF50 promoter luciferase constructs (A: ORF50 promoters with deleted AP1 or Sp1 sites; B: ORF50 promoters with mutated AP1 or Sp1 sites), and 0.5 µg of the β-gal reporter construct. Cells were harvested 36h after transfection. The luciferase and β-galactosidase activities were measured in triplicate and the luciferase activity was normalized to β-galactosidase activity. The data represent the mean of fold induction of three independent experiments +/− S.E.M. (C) Effect of vGPCR induction on Sp1 and Sp3 transcription factor activation. Nuclear extracts prepared from untreated or doxycycline treated BC3.14 cells were tested for the activation of Sp1 and Sp3 transcription factors by incubating the nuclear extracts with plate-immobilized oligonucleotides containing the Sp1 family consensus binding sequence, which was followed by ELISA with antibodies to the respective transcription factors. MCF7 (human breast adenocarcinoma cells) nuclear extracts provided in the Sp1/3 ELISA kit were used as a positive control. To confirm the specificity of binding, the nuclear extracts were incubated with WT or mutated oligonucleotide prior to the plate-immobilized oligonucleotide incubation. The data represent the average and standard deviations of two independent experiments. *** p<0.001. (D) HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 promoter luciferase construct WT (p134) or mutated at the Sp1/3 site (p134Sp1M), and 0.5 µg β-galactosidase reporter, in the presence or absence of 2 µg of Sp1 or Sp3 expression plasmids. 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represent the mean fold induction of four independent experiments +/− S.E.M. ** p<0.01. *** p<0.0005. (E) HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 promoter luciferase construct WT (p2500) or mutated at the Sp1/3 site (p2500Sp1M), and 0.5 µg β-galactosidase reporter, in the presence or absence of 2 µg of Sp1 expression plasmids. 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represented as the mean fold induction of three independent experiments +/− S.E.M. (F) Mapping of PAN promoter elements involved in vGPCR-induced activation. HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or vGPCR expression plasmids, 1 µg of the indicated PAN promoter luciferase constructs, and 0.5 µg of β-gal reporter construct. Cells were harvested 36h after transfection and the luciferase and β-galactosidase activities measured in triplicate. Luciferase activity was normalized to β-galactosidase activity. The data represent the mean of fold induction of four independent experiments +/− S.E.M.

Sp1 site is involved in vGPCR induction of the ORF50 promoter

The GCCCCGCCCA Sp1 site is also present in the −134/−75 region of the ORF50 promoter that has been shown to be crucial for butyrate induced ORF50 expression (Lu et al., 2003; Ye, Shedd, and Miller, 2005). When we tested a reporter construct p95Luc with a deleted Sp1 site, as shown in Fig. 5A, the Sp1 deletion reduced the basal activity of the promoter. In addition, this deletion totally abolished the responsiveness of the ORF50 promoter to vGPCR expression. To further confirm the role of the Sp1 site, a mutational approach was used on the p134Luc construct. In the Sp1 site mutant (p134Sp1M), the GCCCCGCCCA-Sp1 site was replaced by AGGAATTCTC. As shown in Fig.5B, SP1 mutation reduced the basal activation of the ORF50 promoter. In addition, the p134Sp1M was significantly less responsive to vGPCR expression (1.7-fold). As for the control, a vGPCR expression similarly activated the construct with a mutation outside of any transcription factor’s binding site (p134ContM) with about 3.2-fold compared to pcDNA transfection. Taken together, these results suggested that the Sp1 site plays an important role in vGPCR expression mediated activation of the KSHV ORF50 promoter.

The vGPCR induced signaling pathway is known to activate several transcription factors such as NF-κB, AP1, NFAT, CREB, and HIF1 (Cannon, Philpott, and Cesarman, 2003; Cannon and Cesarman, 2004; Montaner et al., 2004; Sodhi et al., 2000). However, the effect of vGPCR on Sp1 family activation has never been reported. In order to determine whether vGPCR signaling induces the DNA binding activity of Sp1 and Sp3, nuclear extracts prepared from untreated or doxycycline treated BC3.14 cells were examined by an ELISA-based DNA binding assay using oligonucleotides containing the Sp1 family consensus binding sequence. As shown in Fig. 5C, vGPCR induction in BC3.14 cells significantly increased Sp1 as well as Sp3 DNA binding activity (3.2 and 4.2-fold, respectively). vGPCR expressing cells had the same level of Sp1 and Sp3 activation as that of the positive control MCF-7 cells (Fig. 5C). The specificity of the DNA binding activity of Sp1 and Sp3 detected in these cells was verified by the absence of reactivity when nuclear extracts were pre-incubated with WT oligonucleotide prior to the plate-immobilized oligonucleotide incubation and by the presence of reactivity when incubated with oligonucleotide with a mutated Sp1 site.

It has been previously shown that Sp1 over-expression is able to activate ORF50 in Drosophila schneider 2 cells and HKB5/B5 cells (Ye, Shedd, and Miller, 2005). Since vGPCR expression increased the Sp1 and Sp3 DNA binding activities in transfected cells (Fig. 5C), we tested the ability of Sp1 and Sp3 to induce the ORF50 promoter in HEK293T cells. As shown in Figure 5D, Sp1 and Sp3 over-expression significantly induced the p134 ORF50 promoter (31.7 and 31.4-fold induction, respectively). In contrast, when cells were transfected with the same constructs with mutated Sp1 and Sp3 binding sites (p134Sp1M), we observed about 50 % and 66% reduction in ORF50 promoter activity (Fig. 5D). Sp1 overexpression was able to activate the longer p2500 ORF50 promoter more efficiently than the same construct with a mutated Sp1 site (p2500Sp1M) (5 fold and 1.7- fold, respectively) (Fig. 5E). The p134 ORF50 promoter induction by Sp1 and Sp3 over-expression is considerably higher than after vGPCR over-expression. It is not surprising that the effect of a transcription factor over-expression is more direct and potent compared to expressing an upstream protein. Moreover, the levels of expression of the different proteins are difficult to compare as they are influenced by factors including stability of the mRNA and stability of the proteins. Nevertheless, these results further verified that Sp1 and Sp3 binding play critical roles in ORF50 promoter activity.

In addition to the ORF50 promoter, vGPCR also regulated the PAN promoter (Figure 2C and 3C). To determine the minimal promoter region responsive to vGPCR expression, a series of PAN promoter deletions were assayed in a luciferase reporter experiment in HEK293T cells (Fig. 5F). vGPCR over-expression was able to activate p261Luc, p200Luc and p122Luc at a similar level with about 3, 3.5 and 2.4-fold activation, respectively. However, the p69Luc and p38Luc promoters were only marginally activated by vGPCR expression with about 1.1 and 1.2-fold, respectively (Fig. 5F). Together, these results suggested that the promoter region located between −122 and −69 is important for the PAN promoter activation inducted by vGPCR. Interestingly, this region of the promoter contains a c/EBP site but also a Sp1 site indicating that the Sp1 family could be involved in the regulation of both ORF50 and PAN promoters. We do not exclude that c/EBP is not involved in PAN promoter activation. However, results obtained in Figure 5C indicate that vGPCR induces Sp1 and Sp3 activation. To our knowledge, the ability of Sp1 and Sp3 to induce the PAN promoter has never been demonstrated. As shown in Figure 5F, Sp1 and Sp3 over-expression significantly induced the PAN promoter constructs p261, p200 and p122 containing the Sp1/3 site (for Sp1, 5.4, 5.1, and 5.6- fold, respectively, and 3.9, 2.7, and 3.6-fold respectively for Sp3) in HEK293T cells. In contrast, the promoter construct p69 lacking the putative Sp1/3 site was not induced by Sp1 nor Sp3 over-expression (Fig. 5F). These results suggested that vGPCR could induce the PAN promoter in an Sp1/Sp3 dependent manner.

The effects of vGPCR on the ORF50 promoter are pertussis toxin insensitive

vGPCR is known to activate a variety of transcription factors and several investigations led to an understanding of the mechanism of AP1 and NF-κB activation (Cannon, 2007; Cannon, Philpott, and Cesarman, 2003; Cannon and Cesarman, 2004; Liu et al., 2004; Montaner et al., 2004). However, nothing is known about the mechanism by which vGPCR activates the Sp1 family of transcription factors. As the name of the protein indicates, vGPCR is linked to heterotrimeric G proteins (Fig. 6A). To determine which G protein subtype is involved in vGPCR-induced ORF50 promoter regulation, we investigated the effect of pertussis toxin (PTx), which prevents GDP release from the Gα subunit so that the G protein is locked in an inactive state (Burns, 1988). The G proteins Gα_i, Gα₀ and Gα_t are sensitive to this toxin. The G proteins Gα_t and Gα₀ are not expressed in our cell model. We have investigated the role of Gα_i in ORF50 promoter regulation. HEK293T cells were transfected with vGPCR expressing construct as well as the reporter plasmids (p2500Luc and RSV-β-gal) and treated overnight with different concentrations of PTx (50, 100, 200 and 1000 ng/ml). HEK293T cells were viable when exposed to PTx at experimental concentrations (data not shown). Increasing concentrations of PTx had no effect on vGPCR-induced activation of the ORF50 promoter (Figure 6B). These results indicate that if a G protein is involved in ORF50 promoter regulation by vGPCR, it is PTx insensitive.

Figure 6.

Regulation of the ORF50 promoter after vGPCR expression is partially mediated by Gαq and Gα12. (A) Schematics depicting wild-type and mutant vGPCR (m15 linked to Gαq and m8 linked to Gαi) (B) ORF50 regulation by vGPCR expression is pertussis toxin (PTx) insensitive: HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or vGPCR expression plasmids along with 1 µg of full length ORF50-luciferase reporter (p2500Luc) and 0.5 µg of the β-gal reporter plasmid (as transfection control). 24h post-transfection, the cells were treated overnight with increasing doses of PTx (0, 50, 100, 200, and 1000 ng/ml). The data represent the mean of fold induction of three independent experiments +/− S.E.M. (C) Wild-type and mutant vGPCR (m15, linked to Gαq) can activate the ORF50 promoter. HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or the indicated vGPCR expression plasmids (WT, m8, m15) along with 1µg of the full length ORF50-luciferase reporter (p2500) or its Sp1 mutated version (p2500Sp1M) and 0.5 µg of the β-gal reporter plasmid (as transfection control). 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represent the mean of fold induction of three independent experiments +/− S.E.M. In addition, total proteins were extracted in RIPA lysis buffer and subjected to a western blot experiment using vGPCR antibodies (right panel). (D) Expression of constitutively active mutants Gα12 and Gαq activate the ORF50 promoter. HEK293T cells were transfected by the calcium phosphate method with 1 µg of the indicated luciferase reporter (pGL3 or ORF50 full length promoter p2500Luc), 0.5 µg of the β-galactosidase reporter (as transfection control) and 2 µg of the indicated G protein expression plasmids. 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represent the mean of fold induction of five independent experiments +/− S.E.M. *** p<0.005. (E) The Sp1 site is involved in Gα12 and Gαq activation of the ORF50 promoter. HEK293T cells were transfected by the calcium phosphate method with 2 µg of pcDNA or Gα12QM or GαqQM expression plasmids, 1 µg of the indicated ORF50 promoter luciferase constructs, and 0.5 µg of the β-Gal reporter construct (as transfection control). Cells were harvested 36h after transfection. The luciferase and β-galactosidase activities were measured in triplicate and the luciferase activity was normalized to β-galactosidase activity. The data represent the mean of fold induction of three independent experiments +/− S.E.M.

vGPCR mutants have different effects on the ORF50 promoter

Structure function studies of vGPCR have identified residues that selectively affect its association with different subclasses of Gα proteins (Liu et al., 2004). It has been shown that while vGPCR can be coupled to Gα_i, Gα_q and Gα_12/13, some mutants are more specifically coupled to only one subunit. For example, vGPCR m8 (R322W) is predominantly Gα_i coupled whereas the m15 mutant (M325S) is predominantly Gα_q coupled (Liu et al., 2004). We tested whether these different mutants could activate the ORF50 promoter in a luciferase assay in HEK293T cells. As shown in Figure 6C, ORF50 promoter was activated by the m15 vGPCR mutant as efficiently as the wild-type protein whereas the m8 vGPCR mutants were only marginally active. We confirmed by western blot experiment that the m8 mutant was expressed at similar level as that of WT vGPCR (right panel). In addition, both WT and m15 vGPCR activated the ORF50 promoter through the Sp1 site as none of the two proteins were able to activate the mutated promoter p2500Sp1M (Fig.6C). These results indicate that the Gα_q subunit could be involved in ORF50 promoter activation induced by vGPCR.

Constitutively activated Gα₁₂ and Gα_q induce ORF50 promoter activation

The mutation of glutamine to leucine (Q to L) on the Gα subunit eliminates GTPase activity thereby inducing a constitutively active phenotype. We co-transfected the ORF50-reporter construct with plasmids expressing different Gα subunit mutants and tested their ability to modulate the ORF50 promoter (Fig. 6D). The Gα_s mutant inhibited ORF50 promoter activity significantly as well as the activity of a basal promoter (pGL3). As the PTx results predicted, the Gα_i mutant had no effect on the ORF50 promoter indicating that this subunit is not involved in the signaling pathway inducing the ORF50 promoter. Similarly, the Gα₁₃ subunit mutant did not have any effect on the ORF50 promoter. Interestingly, whereas Gα₁₂ and Gα_q have minimal effects on a basal promoter (pGL3), these two constitutively active proteins were able to induce the ORF50 promoter with 1.7 and 1.7-fold activation, respectively (Fig. 6D). Although these activations were moderate, they were statistically significant (p<0.005). These results demonstrated the role of Gα_q and/or Gα₁₂ in vGPCR mediated activation of the ORF50 promoter.

Gα_q and Gα₁₂ activation of the ORF50 promoter is mediated by the same site as vGPCR activation

To confirm that ORF50 promoter activation induced by vGPCR and G protein expression are part of the same pathway, we tested the effect of the constitutively active Gα_q and Gα₁₂ mutants on the different promoter constructs. As shown in Figure 6E, constitutively active Gα_q and Gα₁₂ expression induced the full-length promoter (1.3 and 1.7-fold, respectively) as well as the shorter version p134Luc (1.3 fold). In addition, the ORF50 promoter with an AP1 site deletion (p115Luc) was induced by the constitutively active Gα_q and Gα₁₂ expression with about 1.5 and 1.6-fold, respectively (Fig. 6E). However, the Sp1 site deleted promoter (p95Luc) or Sp1 mutated promoters (p2500Sp1M and p134Sp1M) did not respond to the expression of the constitutively active G proteins (Fig. 6E). These results confirmed that ORF50 promoter activation by vGPCR could be mediated through the activation of Gα_q and/or Gα₁₂.

Effect of Rho family proteins on vGPCR induction of the ORF50 promoter

Downstream of the G proteins, the Rho family small GTPases are involved in different signaling pathway activation. For instance, Rac1 GTPase is involved in AP1 associated NF-κB activation induced by vGPCR (Dadke et al., 2003; Montaner et al., 2004) whereas Ras is involved in the activation of ERK leading to HIF1 transcription factor activation (Sodhi et al., 2000). Finally, vGPCR has also been shown to activate the NF-κB pathway through the activation of RhoA (Shepard et al., 2001). To determine if a small G protein is involved in vGPCR activation of the ORF50 promoter, we investigated the effect of Clostridium difficile toxin B (CdTxB), which inactivates the Rho family of small GTPases through monoglucosylation of these family members. This toxin targets Rho, Rac and cdc42 but is ineffective on the Ras, Rab, Arf or Ran subfamilies as well as the heterotrimeric G proteins (Aktories, 1997). We transfected HEK293T cells with the ORF50 promoter (p2500Luc) along with pcDNA or vGPCR expressing plasmids. The cells were treated overnight with CdTxB at 200 ng/ml. As shown in Figure 7A, CdTxB had a moderate effect on the basal activity of the promoter (18.7% inhibition). In the absence of CdTxB, vGPCR expression induced ORF50 promoter by 2.4-fold and about 2-fold induction was observed in the presence of the toxin. We further investigated the effect of expression of constitutively active RhoA (RhoCA) or dominant negative RhoA (RhoDN) mutants as well as the effect of Rac mutants (RacCA and RacDN) (Fig. 7B). In the absence of vGPCR expression, we observed that RhoCA expression was able to induce a moderate ORF50 promoter activity (1.93-fold) whereas the other mutants had no significant effects. In the presence of vGPCR expression, the constitutively active mutants RhoCA and RacCA were unable to further activate the ORF50 promoter. We observed a 3-fold induction of ORF50 promoter with vGPCR alone and 3.1 and 2.8-fold when RhoCA or RacCA were co-expressed with vGPCR. In addition, the dominant negative mutants RhoDN and RacDN were unable to inhibit the effect of vGPCR expression alone since 3-fold induction for vGPCR alone was observed when compared to 3.1 and 2.7-fold with RhoDN and RacDN co-expression with vGPCR. Taken together, these results indicated that a small G protein of the Rho family is not involved in vGPCR induced ORF50 promoter activation.

Figure 7.

G-proteins Rho and Rac are not involved in ORF50 promoter regulation after vGPCR expression. (A) HEK293T cells were transfected by the calcium phosphate method with 1 µg of ORF50 full length promoter p2500Luc, 0.5 µg of the β-galactosidase reporter (as transfection control) in the presence or absence of 2 µg of the vGPCR expression construct. 24h after transfection, the cells were incubated with or without CdTxB (200 ng/ml) for 16h. Luciferase and β-galactosidase were measured in triplicate and the data represent the mean of fold induction of three independent experiments +/− S.E.M. (B) HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 full length promoter p2500Luc, 0.5 µg β-galactosidase reporter (as transfection control) and 1 µg of the indicated expression plasmids, in the presence or absence of 1 µg of the vGPCR expression construct. 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represent the mean fold induction of four independent experiments +/− S.E.M.

Effect of different signaling pathway inhibitors on vGPCR induced ORF50 promoter activation

To further investigate the signaling pathway utilized by vGPCR to activate the ORF50 promoter, a pharmacological approach was used. Since Gq coupled receptors are known to induce PLC activation (Alberts B, 2002), to determine if this pathway is involved in vGPCR mediated ORF50 regulation, we treated transfected HEK293T cells overnight with the PLC inhibitor U73122 (10 µM). As shown in Fig. 8A, PLC inhibitor clearly had an effect on basal ORF50 promoter activity (39% inhibition) as well as on vGPCR mediated ORF50 activity (48% inhibition).

Figure 8.

Effect of pharmacologic signaling pathway inhibitors. (A) Effect of PLC inhibitor. HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 full length promoter p2500Luc, 0.5 µg of the β-galactosidase reporter in the presence or absence of 2 µg of the vGPCR expression construct. 24h after transfection, the cells were incubated with or without the indicated PLC inhibitor (U73122; 10 µM) for 16h. Luciferase and β-galactosidase were measured in triplicate and the data represent the mean of fold induction of four independent experiments +/− S.E.M. p<0.05. (B) Effect of PKC inhibitors. HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 full length promoter p2500Luc, 0.5 µg of the β-galactosidase reporter in the presence or absence of 2 µg of the vGPCR expression construct. 24h after transfection, the cells were incubated with or without the indicated PKC inhibitors (Rottlerin 10 µM, GFX 10 µM, Gö6983 1 or 10 µM) for 16h. Luciferase and β-galactosidase were measured in triplicate and the data represent the mean of fold induction of three independent experiments (Gö6983) or 5 experiments (GFX and Rottlerin) +/− S.E.M. * p<0.05, ***p<0.0005. (C) BC3.14 cells were treated for 48h with or without doxycycline (0.5 µg/ml) in presence or absence of the indicated PKC inhibitors (Rottlerin 10 µM, GFX 10 µM). The cells were then assayed for accumulation of vGPCR and ORF50 mRNAs by quantitative real-time RT-PCR. Results were normalized to HPRT and the fold induction with / without doxycycline is presented. Each reaction was done in triplicate and the data represent the mean of three independent experiments.

Phospholipase Cβ activation induces the production of diacylglycerol (DAG) and inositol triphosphate (IP3) leading to an increase in cytosolic calcium. Downstream of these secondary messengers, effectors such as protein kinase C (PKC) can be activated. The PKC family can be divided into three subgroups depending on their sensitivity to activation. Classic or conventional PKC (α, β1, β2, γ) are activated by both diacylglycerol (DAG) and calcium, whereas the novel PKC (δ, ε, η and θ) are sensitive to DAG alone. The atypical PKC (ζ and λ) are insensitive to both calcium and DAG (Mackay and Twelves, 2007). Pharmacologic inhibitors were used to determine which PKC is involved in vGPCR induction of ORF50 promoter. GF109203X (GFX) inhibits both the classic and novel PKCs. As shown in Fig. 8B, GFX decreased the basal level of promoter activity as well as the vGPCR mediated ORF50 induction (39% and 52% inhibiton, respectively). Rottlerin has been shown to inhibit one novel PKC (PKCδ) (Gschwendt et al., 1994). Rottlerin had a minimal effect on basal ORF50 promoter activity and, in contrast, had a considerable inhibitory effect on vGPCR-mediated ORF50 activation (14% and 42% inhibition, respectively). Gö6983, which specifically inhibits the classic PKCs (Gschwendt et al., 1996) did not show any effect on vGPCR induced ORF50 activation at both concentrations tested (Fig. 8B). These results suggested that a novel PKC may be involved in ORF50 promoter regulation by vGPCR. Since vGPCR is also known to induce several other signaling pathways, we tested the effect of Rp-cAMP and PD98059, inhibitors of cAMP and MEK pathways, respectively. These inhibitors had no effect on basal ORF50 activity or on vGPCR induced activity (data not shown).

To confirm that the effects were significant in infected cells, we investigated the effect of these inhibitors on ORF50 expression after doxycyclin treatment of BC3.14 cells. As shown in Figure 8C, doxycycline induced about 1.5-fold induction of ORF50, whereas in the presence of the PKC inhibitors GFX and Rottlerin, activation of ORF50 induced by vGPCR was abrogated. The level of vGPCR was not inhibited by GFX or Rottlerin. Unfortunately, the effect of PLC inhibitor U73122 could not be evaluated due to the high toxicity of this drug in BC3.14 cells at the 48 hours time point used in this study (not shown).

PKD is partially involved in vGPCR induced ORF50 promoter regulation

A downstream effect of PKC activation induced by cellular G protein coupled receptors is the activation of protein kinase D (PKD, also called PKCµ). To determine if PKD is part of the vGPCR pathway involved in ORF50 promoter regulation, we investigated the effect of expression of a PKD dominant negative construct in which mutation of KW in the ATP binding site makes it a kinase dead mutant. As shown in Figure 9A, the expression of the mutant PKD had a minimal effect on basal promoter activity (1.2-fold activation). However, whereas vGPCR could activate the ORF50 promoter by 2.6- fold in the absence of the PKD construct, this activation was reduced significantly in the presence of the kinase dead protein (41% reduction). Though further confirmation that the effects are significant in infected cells BCBL-1 cells needs stable transfection with PKD dominant negative construct, nerverthless, these results suggested that PKD is at least partially involved in vGPCR regulation of the ORF50 promoter.

Figure 9.

Effect of PKD and HDAC. (A) HEK293T cells were transfected by the calcium phosphate method with 1 µg of the ORF50 full length promoter p2500Luc, and 0.5 µg β-galactosidase reporter in the presence or absence of 1 µg of PKD-KW expression plasmids as well as in the presence or absence of 1 µg of the vGPCR expression construct. 36h after transfection, luciferase and β-galactosidase were measured in triplicate and the data represent the mean fold induction of four independent experiments +/− S.E.M. ** p<0.01. (B) BC3.14 cells were treated with Doxycycline (0.5 µg/ml) for 48h. 1 µg of nuclear extracts were used in HDAC assays as described in the Material and Methods section. Briefly, nuclear extracts were incubated in the absence or presence of 250 µM NAD+ cofactor. After HDAC deacetylation of the substrate, a fluorophore is released and detected using a fluorescence plate reader (Synergy™ HT Multi-Mode Microplate Reader). Specificity of HDAC activity was checked using the HDAC inhibitor trichostatin A (1 µM). Specificity of the class III HDAC activity was analyzed in the presence of metabolite nicotinamide (5 µM). The data represent the mean fluorescent intensity of two independent experiments with the fluorescence read in triplicate ± S.D.

vGPCR expression inhibits HDAC activity

It has been shown that Sp1 and Sp3 are able to recruit HDAC proteins leading to histone hypoacetylation and transcriptional repression (Won, Yim, and Kim, 2002). In addition, in a cardiac hypertrophy model, it has been proposed that seven transmembrane receptors, such as α-adrenergic receptor or endothelin-1 receptors, are able to induce class II HDAC (HDAC4, 5, 7, 9) phosphorylation (Vega et al., 2004) while CaMK, novel PKC and PKD have been shown to phosphorylate class II HDAC (Chang et al., 2005b; Harrison et al., 2006; Vega et al., 2004). Phosphorylated HDAC binds to the protein 14-3-3 and is exported from the nucleus to the cytoplasm by a CRM1 dependent mechanism, inhibiting the transcriptional repression induced by HDACs (Grozinger and Schreiber, 2000; Li et al., 2004). The results presented above demonstrated that PKC and PKD, as well as Sp1/3, are involved in ORF50 activation induced by vGPCR. It has also been shown in a previous study that HDAC-1, 5 and 7 are associated with the ORF50 promoter in latently infected cells (Lu et al., 2003). Induction of the lytic cycle by butyrate treatment increased histone H3 and H4 acetylation due to the recruitment of the BRG1 chromatin-remodeling complex to the ORF50 promoter (Lu et al., 2003). We next investigated whether vGPCR could decrease HDAC activity. vGPCR was induced in BC3.14 cells by doxycycline treatment and the nuclear extracts were assayed for HDAC activity. 1 µg of nuclear extracts were incubated with a short peptide substrate containing an acetylated lysine residue that can be deacetylated by most HDAC. Class I, II and IV HDAC can be measured directly whereas class III HDAC, Sirt1, 2, 3, 4, 5, 6 and 7, require the addition of the NAD⁺ cofactor. Once the substrate is deacetylated by the HDAC proteins, the lysine residue reacts with the developing solution and releases a fluorophore. As shown in Fig. 9B, the general HDAC activity was similar in BC3.14 cells with or without vGPCR induction. Trichostatin A was used as an HDAC inhibitor and showed 63% and 62% reduction in non-treated and doxycycline treated BC3.14 cells, respectively. However, after addition of the NAD+ cofactor, we detected a considerable decrease in HDAC activity in vGPCR induced cells (65% reduction compared to non-induced cells). As a control, nicotinamide was used as an HDAC class III inhibitor. HDAC class III activity was reduced 78% in non-induced cells whereas this activity was reduced 64% in vGPCR induced cells. These results suggested that vGPCR signaling could decrease HDAC nuclear activity and therefore could inhibit transcriptional repression of ORF50 promoter.

DISCUSSION

Studies presented here demonstrate that KSHV vGPCR can activate the promoter of the lytic cycle switch ORF50 gene through the Sp1 family of transcription factors and suggest a role for vGPCR in the sustained expression of ORF50 (Fig.10). Similar to our studies, a role for GPCR proteins encoded by other herpesviruses in the regulation of lytic gene expression has been observed (Lee et al., 2003; Moorman, Virgin, and Speck, 2003; Oliveira and Shenk, 2001). For example, the murine CMV GPCR (M78) protein has been shown to be involved in lytic viral immediate early (M123), M54 early (M54) and late (M99) mRNA accumulation (Oliveira and Shenk, 2001).However, the mechanism behind the role of CMV GPCR in lytic cycle gene expression is not yet understood. In the γ-herpesvirus family, it has been previously shown that vGPCR encoded by murine herpesvirus 68 (MHV68) is required for efficient reactivation from latency in mice (Lee et al., 2003; Moorman, Virgin, and Speck, 2003). In addition, treatment with the CXC chemokine KC increased MHV68 replication by a mechanism that was dependent on pertussis toxin-insensitive vGPCR signaling, which indicated that the G protein Gq subunit was more likely to be involved (Lee et al., 2003). These results are similar to our in vitro observations with KSHV vGPCR (Fig. 6 and Fig. 10). KSHV vGPCR has been shown to be involved in the up-regulation of several viral promoters. Luciferase assays performed with doxycycline treated BC3.14 cells indicated that vGPCR could activate KSHV IL6 (K2), ORF50 and ORF57 promoters (Cannon, Philpott, and Cesarman, 2003). In a following study, the same group observed that vGPCR expression induced a cell cycle arrest via p21-mediated inhibition of Cdk2. In these conditions, TPA was unable to induce the lytic cycle (Cannon, Cesarman, and Boshoff, 2006). In addition, in a luciferase assay, vGPCR expression activated the ORF50 dependent PAN promoter in HeLa cells (Chiou et al., 2002). A deletion of the AP1 site, alone or in combination with a serum responsive element deletion, did not alter vGPCR activation of the PAN promoter, suggesting the involvement of another transcription factor. Our study suggests that Sp1/3 transcription factors could be involved in this activation. Our study has dissected the signaling pathway induced by vGPCR involved in ORF50 promoter activation (Fig. 10). We found that Gα₁₂ and Gα_q are probably the main proteins involved in this signaling (Fig. 6). Since neither Rac1 nor RhoA dominant-negative mutants were able to inhibit vGPCR-induced ORF50 regulation, we can not exclude the possibility that other small GTPases could be involved in vGPCR’s effect on ORF50 (Fig. 7). It has recently been shown that activated Ha-Ras, as well as activated Raf mutants, were able to induce ORF50 in BC-3 cells, through the MEK/ERK pathway (Yu et al., 2007). In addition, the authors observed that v-Ki-Ras2 expression was able to activate ORF50. Yu et al., 2007 proposed that the transcription factor Ets1 mediated Ras induced reactivation and these proteins could be interesting candidates in the evaluation of vGPCR-induction of the ORF50 promoter.

Figure 10.

Schematic representation depicting the mechanism of ORF50 promoter regulation induced by vGPCR. The constitutive activity of vGPCR through the Gαq and Gα12 proteins leads to the activation of the PLC and PKC signaling pathways, which directly increase the nuclear level of Sp1 and Sp3 transcription factors as well as their binding to their specific target sites present on the ORF50 promoter. In addition, vGPCR releases the ORF50 promoter repression induced by HDAC and the inhibition of HDAC could involve a phosphorylation step induced by PKD. Together, these activities of vGPCR could act as part of a positive feedback loop resulting in the sustained expression of ORF50, which could lead to continued activation of ORF50 dependent lytic cycle genes and to successful viral progeny formation.

One of the main signaling events downstream of a G protein is phospholipase Cβ activation and our results using the chemical inhibitor U73122 indicate that this pathway is involved in KSHV vGPCR mediated ORF50 regulation (Fig. 8). It has previously been shown that human CMV encoded US28 chemokine receptor is also able to activate the PLC pathway via a G(q/11) pathway (Casarosa et al., 2001). Activation of the phospholipase stimulates the production of inositol triphosphate (IP3) and diacylglycerol (DAG) leading to an increased level of calcium and the activation of protein kinase C (PKC). Our investigation into the role of the different PKCs (α, β1, β2, γ, δ, ε, η, and θ) showed that the general PKC inhibitor GF109203X could reduce ORF50 promoter activation induced by vGPCR whereas the classic PKC inhibitor Gö6983 had no effect, suggesting that a novel PKC could be involved in ORF50 promoter regulation by vGPCR (Fig. 8). Interestingly, TPA, a frequently used KSHV lytic cycle inducer, is a well-known activator of PKC, and KSHV reactivation induced by TPA can be inhibited by the same inhibitors (GF109203X and rottlerin) blocking vGPCR’s effect on the ORF50 promoter in our studies (Deutsch et al., 2004). It has been shown that PKCδ, a member of the novel PKC sub family, translocated to the membrane after TPA treatment of BCP-1 and BCBL-1 cells. In addition, over-expression of a dominant negative PKCδ mutant was able to inhibit TPA induced reactivation (Deutsch et al., 2004). However, over-expression of PKCδ and its activation by bistratene A were unable to induce the lytic cycle suggesting that PKCδ itself was not sufficient. Whereas inhibitor treatment led us to suggest that a novel PKC could be involved in ORF50 promoter regulation by vGPCR, our study did not identify the specific member of the family. PKCδ is however a good candidate for further studies.

Downstream of PKC, we observed that a PKD dominant negative mutant could partially block ORF50 activation induced by vGPCR (Fig. 9). PKD has been involved in the regulation of replication of other herpesvirus. In the γ1-Epstein Barr virus (EBV), the switch between latency and the lytic cycle is mediated by two proteins: BZLF1 and BRFL1 (Countryman and Miller, 1985; Ragoczy, Heston, and Miller, 1998). PKD and the X-box-binding protein 1 (XBP-1) had a synergistic effect on the EBV lytic cycle (Bhende et al., 2007). The same study hypothesized that PKD disrupted EBV latency through its effect on the HDAC proteins. Indeed, it has been previously shown by many studies that inhibition of histone deacetylase activity was able to induce the EBV lytic cycle (Gruffat, Manet, and Sergeant, 2002). Similarly, the KSHV lytic cycle is rapidly induced by HDAC inhibitors like sodium butyrate and trichostatin A (Lu et al., 2006; Lu et al., 2003; Ye et al., 2007). It has been previously shown that HDAC1 was able to bind to the ORF50 promoter and inhibit its transcription (Gwack et al., 2001). In addition, it has been proposed that LANA-1 could inhibit the ORF50 promoter through an interaction between LANA-1, Sp1 and HDACs (Lu et al., 2006). The inhibition of HDACs leads to the acetylation of LANA-1 and/or Sp1 and induces a dissociation of LANA-1 from the promoter, thus blocking full activation (Lu et al., 2006). Based on our studies, we propose that vGPCR signaling reduces HDAC activity and thereby increases ORF50 promoter activity (Fig. 10).

vGPCR is a strong activator of the AP1 transcription factor (Cannon, 2007), but surprisingly, the AP1 site is not involved in vGPCR activation of the ORF50 promoter in our studies (Fig. 5). Mutation of the AP1 site even showed a moderate increase in activity. However, this result is similar to that seen with the PAN promoter (Chiou et al., 2002). On the other hand, the Sp1 family site seemed to be important for both basal activity as well as vGPCR responsiveness of ORF50. In addition, we showed that vGPCR over-expression increased Sp1 and Sp3 DNA binding activity (Fig. 5C). Further studies are essential to analyze if a decrease of vGPCR expression is responsible for a decrease of Sp1 and Sp3 DNA binding activity. Two previous studies already highlighted the role of the Sp1 family in ORF50 promoter regulation induced by butyrate (Lu et al., 2003; Ye, Shedd, and Miller, 2005). In BCBL-1 cells, Sp1 and Sp3 could bind to the ORF50 promoter (Lu et al., 2003). However, during latency the promoter could be repressed through histone deacetylase (HDAC1, HDAC5, HDAC7) association. Butyrate treatment increased histone H3 and H4 acetylation due to the recruitment of the BRG1 chromatin-remodeling complex to the ORF50 promoter (Lu et al., 2003). Our models suggest that vGPCR could decrease HDAC activity. It has already been documented in a cardiac hypertrophy model that PKC and PKD could inhibit HDAC activity through its phosphorylation (Vega et al., 2004).

The importance of the Sp1 transcription factor family has also been proposed for other herpesviral promoters. In herpes simplex virus (HSV-1), several Sp1 sites have been identified in the immediate early (IE) promoter (Jones and Tjian, 1985). In later studies, it has been shown than HSV-1 infection induced Sp1 phosphorylation and this modification could explain the decrease in thymidine kinase gene expression late in infection (Kim and DeLuca, 2002). In HCMV, Sp1 and Sp3 binding sites are crucial for the major immediate early proximal enhancer transcription and for HCMV replication in human fibroblasts (Isomura et al., 2005; Isomura, Tsurumi, and Stinski, 2004; Meier, Keller, and McCoy, 2002). EBV-RTA can induce different viral promoters through an RRE interaction. However, RTA has been found to activate some promoters without an RRE. Two Sp1 sites in the BRFL1 gene have been shown to mediate RTA activation on its own promoter (Ragoczy and Miller, 2001). In a later study, Chang et al., 2005a proposed that a complex composed of RTA, Sp1 and MCAF1 can be formed at an Sp1 site on the BRFL1 promoter allowing RTA to increase Sp1-mediated transcription. The induction of the EBV lytic cycle also requires expression of the Zta protein encoded by the BZLF1 gene. The BZLF1 promoter has been extensively studied and contains some Sp1 family binding sites (Borras, Strominger, and Speck, 1996; Liu et al., 1997).

Other transcription factors have been involved in ORF50 promoter activation. For example, HIF-1 family transcription factors are able to activate the ORF50 promoter following hypoxia (Cai et al., 2006; Davis et al., 2001; Haque et al., 2003), which is an inducer of the KSHV lytic cycle (Davis et al., 2001). Whereas Haque et al (Haque et al., 2003) showed activation of the ORF50 promoter by HIF-2 over-expression, HIF-1α over-expression was involved in ORF50 activation in a later study (Cai et al., 2006). Interestingly, vGPCR is also known to activate the transcription factor HIF-1α. It has been shown that vGPCR is able to activate the p38 and MAPK signaling pathways inducing HIF-1α phosphorylation on its regulatory domain (Sodhi et al., 2000). Whereas the ORF50 promoter contains 6 potential hypoxia responsive elements (HRE), deletion analysis showed only 3 sites were activated by HIF-1α over-expression. The more distant site (CTCAGGTT) is deleted from the p134 promoter construct and this promoter was activated by over-expression of vGPCR (Fig. 4B). The 2 other sites activated by HIF1α over-expression (GACGTGCT and TACGTGGC) are located in the first intron of ORF50. None of the constructs used in our study incorporate the first intron. Therefore, we do not eliminate the possibility that vGPCR could activate ORF50 through these two HIF1α sites in addition to the effect mediated through the Sp1 sites.

vGPCR plays a major role in KS pathogenesis (Sodhi, Montaner, and Gutkind, 2004). However, since vGPCR is not expressed in KS lesions in the absence of other viral genes and is expressed along with other lytic cycle proteins in cells that are destined to die due to progeny virus production, and is not expressed in latently infected PEL cells, its real role in tumorigenesis in the context of other viral gene expression and virus replication is debatable. Moreover, the fact that lytic cycle vGPCR proteins are also coded by the non-oncogenic β-subfamily of herpesviruses, such as HCMV, human herpesvirus-6 (HHV-6) and human herpesvirus -7 (HHV-7), strongly indicates that vGPCR’s real function must be a highly conserved function related to the viral life cycle. Our accidental observation that when vGPCR was reduced by si-vGPCR, ORF50 gene expression was also affected led to this study, and the results presented here suggest that KSHV vGPCR may play a role in the sustained expression of ORF50 by acting as part of a positive feedback loop which could lead to a continued activation of ORF50 dependent lytic cycle genes and to successful viral progeny formation. This hypothesis need to be further confirmed in future studies such as testing the effect of vGPCR knock down on the ORF50 promoter in other KSHV PEL cell lines or using a vGPCR deleted KSHV genome (BAC system), whether vGPCR deficiency affects ORF50 downstream target genes as well as new viral progeny formation, and confirmation of the importance of Sp1 binding site in ORF50 promoter in the feedback loop by mutation in BAC expressing KSHV genome.

This hypothesis raises an interesting question - why is it necessary to sustain KSHV ORF50 (RTA) expression? KSHV like other herpesviruses establish latency and a major prerequisite for this is the control of lytic replication. During the time of establishing latent infection of appropriate target cells, KSHV has obviously evolved several methods to control ORF50 (RTA) expression. Several in vitro studies have shown that LANA-1, along with other factors, can control ORF50 (RTA) expression (Cai et al., 2006; Lan, Kuppers, and Robertson, 2005; Lan et al., 2004; Li et al., 2008; Lu et al., 2006). However, these factors also need to be overcome at times of reactivation of the lytic cycle since it is paramount to propagate progeny viruses, to infect new cells and new hosts, and thus to continue the species’ existence. A successful pathogen like KSHV must have also evolved ways to regulate the reactivation of ORF50 (RTA) by overcoming the negative regulators and maintain sufficient ORF50 (RTA) expression continuously to complete the cascade of lytic gene expression. vGPCR must be one of the components fulfilling this need by activating the ORF50 promoter through the Sp1 site rather than through the AP1 site that is known to be induced by MAPK-ERK1/2 via external stimuli of latently infected cells (Chang et al., 2005a; Ford et al., 2006; Lee et al., 2008; Yu et al., 2007) or by chemical inducers such as TPA (Wang et al., 2004). The ability of vGPCR to activate the ORF50 promoter, together with the fact that vGPCR itself can be activated by ORF50 (RTA) and that the vGPCR promoter is in the opposite orientation of the major latency associated promoter of KSHV which is targeted by LANA-1 (Staudt and Dittmer, 2006), suggest that control of vGPCR is also perhaps critical to establish KSHV latency. During primary KSHV infection of endothelial cells, even though a high level of ORF50 gene expression was observed, expression of ORF50 as well the ORF50-dependent lytic genes was not sustained, and the latency genes persisted (Krishnan et al., 2004). Absence of vGPCR gene expression during the observed time points of these primary infections supports our notion that vGPCR expression is probably critical for sustaining ORF50 expression and to overcome restrictions on ORF50 expression. However, since the gene array technique used in this study had a limited sensitivity and that the time course of detection was not optimized for vGPCR expression, additional experiments including studies with vGPCR deleted virus are required to further explore the role of vGPCR during primary infection, latency establishment and reactivation, as well as virus progeny formation.

MATERIALS AND METHODS

Cells

Human embryonic kidney HEK293T cells were maintained in DMEM containing 1 mM pyruvate (Invitrogen, Carlsbald, CA, USA), 2 mM glutaMax (Invitrogen), 50 U/ml penicillin, 50 mg/ml streptomycin (Invitrogen), and 10% fetal calf serum (Fetalplex, HyClone, Logan, Utah). KSHV-carrying human B cells (BCBL-1) were cultured in RPMI 1640 (Invitrogen) medium with 10% heat-inactivated fetal bovine serum (FBS; Mediatech, Herndon, VA, USA), 2 mM L-glutaMax, 50 U/ml penicillin, and 50 mg/ml streptomycin. BC3.14 cells (HHV8 positive and EBV negative) cells were maintained in RPMI + 10% fetal calf serum, 50 U/ml penicillin, 50 mg/ml streptomycin, and 1 mg/ml G418 (Sigma, St Louis, MO, USA) (Cannon, Philpott, and Cesarman, 2003).

Reagents

Doxycycline, TPA, PTX, PD98059, U73122, Rp-cAMP were purchased from Sigma. CdTxB, Rottlerin, GF109203X and Gö6976 were purchased from Calbiochem, (La Jolla, CA). Polyclonal rabbit anti-vGPCR peptide (amino acids 4 to 16 (Y)EDFLTIFLDDDES(SC)) antibody was from Dr.Gary Hayward (Johns Hopkins School of Medicine, Baltimore) (Chiou et al., 2002). ORF50 antibody was a gift from Dr Liang (Emory University, Atlanta) (Liang and Ganem, 2003). GFP (B2) antibody was from Santa Cruz (Santa Cruz, CA). Tubulin antibody was from Sigma.

Plasmids

pSG5-vGPCR expression plasmids (WT, m8, m9 and m15) were from Dr. John Nicholas (Liu et al., 2004). The ORF50 promoter constructs (p2500Luc, p950Luc, p134Luc, p115Luc, p95Luc, p75Luc, p69Luc, and p60Luc) and the ORF50 promoter luciferase construct mutated at the Sp1 site (p134Sp1M, p134ContM) were from Dr. George Miller (Yale University School of Medicine, New Haven, CT) (Ye, Shedd, and Miller, 2005). The p2500Sp1M plasmid was constructed by directed mutagenesis on the p2500 construct using the QuikChange XL Site-Directed Mutagenesis Kit, Stratagene (GCCCCGCCCA-Sp1 site was replaced by AGGAATTCTC). The ORF50 promoter luciferase constructs mutated at the AP1 site were from Dr. Gary Hayward (Johns Hopkins School of Medicine, Baltimore) (Wang et al., 2004). The PAN promoter luciferase constructs were from Dr. Ren Sun (University of California at Los Angeles, CA) (Song et al., 2001). LanaPi and K14 promoters were from Dr.Dirk Dittmer (University of North Carolina, Chapel Hill, NC) (Jeong, Papin, and Dittmer, 2001). LanaPc promoter was from Dr.Yuan Chang (University of Pittsburgh, PA) (Sarid et al., 1999). RhoCA, Rho DN, RacCA, and RacDN expression constructs were from the Guthrie cDNA resource center (http://www.cdna.org). PKD expressing plasmids were cloned by Dr.Toker and were obtained from Addgene (Cambridge, MA) (Addgene plasmids 10809, 10810, 10814). CMV-Sp1 was cloned by Dr. Tjian and was obtained from Addgene (Addgene plasmid 12097). Sp3 WT construct was from Dr. Gilles Pages (University of Nice-Sophia Antipolis, France) (Milanini-Mongiat, Pouyssegur, and Pages, 2002). Gα12QL, Gα13QL, GαsQL, and GαiQL expression constructs were from Dr. Dhanasekaran (Philadelphia, PA). GαqQL expression constructs was from Missouri S&T cDNA resource center (plasmid #GNA0Q000C0).

Generation of siRNA expression cassettes

The following five KSHV vGPCR siRNA were designed from the vGPCR sequence (NCBI accession AF367767). These are #1: GGAATGAAACTCTAAATAT (bases 346–354); #2: GAACGTTGGAATACTCTCT (bases 443–461); #3: GGTACTGACATCCGCTGCA (bases 782–800); #4: GCATGTCAGAACCGTGTCA (bases 923–941); and #5: GTACTAAATCTACTGGACA (bases 1101–1119).

Constructs expressing shRNAs were generated as previously described (Tiscornia, Singer, and Verma, 2006a). These shRNA transcription products are processed by the cell to produce functioning siRNA sequences (Brummelkamp, Bernards, and Agami, 2002). For validation of shRNA constructs, HEK293T cells were co-transfected with a vGPCR expression plasmid (target) and the shRNA lentivector construct and immunoblot analysis was used to determine the levels of knockdown (described below).

Construction and production of lentiviral gene transfer vectors

Vector plasmids were constructed for the production of lentiviral vectors that express the desired shRNA. All vectors were designed to be self-inactivating (SIN) (Miyoshi et al., 1998) and utilize the expression enhancing woodchuck hepatitis virus post-transcriptional regulatory element (WPRE) downstream from the transgene (vector pCSC-SP-PW-GFP (Addgene plasmid 12337) modified to contain an Nhe restriction site in the LTR) (Zufferey et al., 1999). The HIV-1 central poly-purine track was also located upstream of the CMV promoter (Follenzi et al., 2000), and all shRNA cassettes were located in the 3’ remnant U3 sequence utilizing the U6 polymerase III promoter. Lentiviral vectors were produced using a four-plasmid transfection system, as previously described (Tiscornia, Singer, and Verma, 2006b). Briefly, HEK293T cells were transfected with vector and packaging plasmids, the supernatants collected, and lentiviral vectors concentrated by centrifugation. The lentiviral vector titers were estimated by flow cytometry for GFP positive cells.

Immunobloting

The effectiveness of the shRNA expression vectors was assessed by Western blot. HEK293T cells were transduced with the lentivector coding for a control siRNA (si-C) or for the siRNA against vGPCR (si1). One week later, the cells were transfected with an expression plasmid for vGPCR. 48h later, cells were harvested in RIPA lysis buffer (125 mM NaCl, 0.01 M sodium phosphate pH7.2, 0.1% SDS, 1% NP-40, 1% Sodium deoxycholate, 1 mM EDTA, and 50 mM sodium fluoride) with protease inhibitor cocktail (Sigma). Protein extracts from BCBL-1 and BC3.14 treated cells were washed in PBS and lysed in the same RIPA lysis buffer. Cellular debris was removed by centrifugation at 13,000 × g for 5 min at 4°C and equal amounts of protein samples were resolved by SDS-10% PAGE, and subjected to Western blotting with a rabbit polyclonal antibody against KSHV vGPCR (Chiou et al., 2002). To standardize the level of lentivector transduction, blots were also probed with antibodies against GFP. To confirm equal protein loading, blots were also probed with antibodies against human tubulin. Secondary antibodies conjugated to horseradish peroxidase (Santa Cruz) were used for detection. Immunoreactive bands were developed by enhanced chemiluminescence (Lumi-light PLUS Western blotting substrate, Roche).

Calcium phosphate transfection

HEK293T cells were transfected by the calcium phosphate method as previously described (Graham and van der Eb, 1973). Briefly, for transfection per well of a 6 multi-well plate, 3 µg of DNA was mixed in 100 µl of 0.25 M CaCl₂ and then added to an equal volume of BES buffered solution (BBS: 50 mM BES, 280 mM NaCl, 1.5 mM Na₂HPO₄, pH 6.95). The DNA mixture was added to the HEK293T cells in 2 ml of complete medium and incubated overnight at 3% CO₂. The medium was then changed and cells were allowed to recover 24h at 10% CO₂.

Luciferase assays

HEK293T cells were transiently transfected with the expression plasmid, a luciferase reporter construct plus a vector expressing the β-Gal reporter gene controlled by an RSV promoter (used to normalize transfection efficiency), by the calcium phosphate method. The cells were harvested and lysed in lysis reporter buffer (Promega) 36h post-transfection. Soluble extracts were assayed for luciferase and beta-Gal activities using Steady-Glo and Beta-Glo (Promega) following the manufacturer’s instructions.

RNA extraction, reverse transcription and real time PCR

Total RNA was extracted by using TRIzol reagent (Invitrogen) and quantified by densitometric analysis at 260 nm. Two to four micrograms of total RNA was treated with DNase for 1h (DNA free, Ambion) and then reverse-transcribed into cDNA by using a mix of oligo (dT)12–18 and random primer (hexamers) with the Superscript first-strand synthesis system (Invitrogen). For RT-PCR, cDNA was used as a template with gene-specific primers that were designed by Primer Express® 1.5a software. The following primer sequences were used: vGPCR sense: 5`-CCTGGTGTGTGGTGAGGAGG-3`; antisense: 5’-AGCAACAATCACCCCCCTTAC-3’; ORF 50 sense: 5`-CGCAATGCGTTACGTTGTTG-3`, antisense: 5`-GCCCGGACTGTTGAATCG-3`; ORF73 sense: 5`-CGCGAATACCGCTATGTACTCA-3`, antisense 5`-GGAACGCGCCTCATACGA-3`; PAN sense: 5`-GCCGCTTCTGGTTTTCATTG-3`; antisense 5`-TTGCCAAAAGCGACGCA-3`; ORF 26 sense: 5`-AGCCGAAAGGATTCCACCA-3`, antisense 5`-TCCGTGTTGTCTACGTCCAG-3`; ORF 57 sense: 5'-CATCCTAGAGGACTCTGT-3', antisense: 5'-TTGCTCGTCTTCCAGTGT-3'; human HPRT sense: 5`-GGACAGGACTGAACGTCTTGC-3`, antisense: 5`-CTTGAGCACACAGAGGGCTACA-3`, human CycloA sense: 5`-GTCGACGGCGAGCCC-3` antisense 5`-TCTTTGGGACCTTGTCTGCAA-3`. PCR was performed using an ABI Prism 7500 real-time PCR system utilizing SYBR Green PCR master mix (Applied Biosystems).

Nuclear extraction

BC3.14 cells were treated with 0.5 µg/ml doxycyclin for 48h, and nuclear extracts were prepared using a nuclear extract kit (Active Motif Corp, Carlsbad, CA) per the manufacturer's instructions. After measuring protein concentrations by Bradford reagent (BioRad), extracts were stored at −70°C. Purity of the nuclear extracts was assessed by immunoblotting using anti-lamin B antibodies, and cytoskeletal contamination was checked by using anti-β tubulin 1 antibodies (Sigma).

Sp1/3 DNA binding assay

1 µg of untreated or doxycycline-treated BC3.14 nuclear extracts were assayed for activated Sp1 and Sp3 by an enzyme-linked immunosorbent assay (ELISA)-based assay kit (Active Motif). This assay has been reported to be more sensitive than a gel shift assay, and uses 96-well plates coated with oligonucleotides containing the Sp1 family consensus sequence (5`-GGGGCGGGG-3`). Plates were washed 3 times in wash buffer (PBS 0.1% Tween 20), incubated with the primary antibody recognizing an accessible epitope on Sp1 or Sp3 protein upon DNA binding, followed by an incubation with a horseradish peroxidase-conjugated anti-rabbit antibody for 1h, washed 3 times and incubated with 100 µl of developing solution for 2 to 5 minutes, followed by the addition of 100 µl of developing solution according to the manufacturer’s instructions. Plates were read with an ELISA plate reader at 450 nm with a reference wavelength of 655 nm.

Histone deacetylase (HDAC) activity assay

Untreated or doxycycline-treated BC3.14 nuclear extracts were assayed for HDAC activity using a fluorescent HDAC assay kit (Active Motif Corp. cat #56200) as per the manufacturer's instructions. Briefly, 1 µg of each of the nuclear extracts were incubated for 2h with a short peptide substrate containing an acetylated lysine residue that can be deacetylated by most HDAC enzymes. Class I, II and IV HDAC could be measured directly whereas class III HDAC enzymes were measured in the presence of NAD+ (Sigma) cofactor at 250 µM concentrations. Once the substrate is deacetylated, the lysine residue reacts with the developing solution and releases the fluorophore that can be detected using fluorescence plate reader (Synergy™ HT Multi-Mode Microplate Reader) with an excitation wavelength of 360 nm and emission wavelength of 460 nm. Specificity of the HDAC activity was checked using the HDAC inhibitor trichostatin A (1 µM) provided in the kit. Specificity of the class III HDAC activity was analyzed in the presence of the metabolite, nicotinamide (5 µM; Sigma).

Viral production

BC3 and BC3.14 cells were treated with doxycycline 0.5 µg/ml for 48 hours, medium collected and the cells were removed by centrifugation at 2000 rpm for 10 minutes followed by filtration on a 0.45µm filter. The clarified medium was then concentrated 10 fold by speedvack. The DNA was extracted using DNAeasy kit, Qiagen and viral DNA copy numbers were quantified by real time DNA PCR using primers amplifying the KSHV ORF73 gene (Krishnan et al., 2004).