Kaposi’s sarcoma associated herpesvirus (KSHV) Rta and cellular HMGB1 proteins synergistically transactivate the KSHV ORF50 promoter

Abstract

Kaposi’s sarcoma-associated herpesvirus ‘replication transcriptional activator’ (Rta) plays a critical role in the switch from latency to lytic replication. Rta upregulates several lytic KSHV genes, including its own, through multiple mechanisms. We demonstrate that cellular HMGB1 binds and synergistically upregulates the ORF50 promoter in conjunction with Rta. No direct interaction between Rta and HMGB1 was observed, however a ternary complex is formed in the presence of Oct1. Furthermore, deletion of an Oct-1 binding site within the ORF50 promoter ablates the HMGB1-mediated synergistic response. These results suggest Rta autostimulation may be mediated by a transient complex involving Oct1 and HMGB1.

Introduction

Kaposi’s sarcoma associated herpesvirus (KSHV) is the etiological agent of several of AIDS-related malignancies including Kaposi’s Sarcoma [1, 2]. In common with all herpesviruses, KSHV has two distinct phases within its life cycle; latency and lytic replication. The KSHV ORF50 gene product, known as the ‘replication and transcriptional activator’ (Rta), is conserved within all gamma-2 herpesviruses and plays a critical role in the switch from latency to lytic replication [3–5]. Expression of Rta in KSHV-latently infected cells is necessary and sufficient to induce lytic replication [6]. Rta initiates lytic replication through transcriptional activation of a variety of lytic genes [7–11], via direct binding to promoter DNA sequences, or through interactions with cellular transcriptional control proteins including RBP-Jκ [12], C/EBPα [13], AP1 [14] and HMGB1 [15].

The high mobility (HMG) protein HMGB1, is an abundant nuclear non-histone chromatin-associated protein. HMGB1 is composed of three domains consisting of two tandem box domains (A and B boxes) and an acidic carboxy-terminus. Both HMG boxes share a common three dimensional structure which can bind DNA through the minor groove resulting in DNA bending [16, 17]. This bending of DNA creates a favourable DNA confirmation allowing the binding of other proteins to DNA. HMGB1 has been shown to enhance the binding of a variety of sequence-specific DNA-binding proteins including p53, Hox domain proteins and octamer binding factors (Oct1/2) [18]. HMGB1 has also been shown to stimulate both Epstein Barr virus (EBV) lytic transactivating proteins, Rta and ZEBRA [19, 20]. Interestingly, stimulation of these two proteins is via two distinct methods. HMGB1 binds DNA in a sequence specific manner to promote the HMGB1 and ZEBRA nucleoprotein complex. Conversely, HMGB1 has no sequence specificity or direct protein-protein interaction when promoting EBV Rta binding to DNA.

Recent analysis has shown that HMGB1 can upregulate Rta-mediated transactivation by enhancing Rta binding to a range of KSHV lytic promoters [15]. Rta has also been shown to autostimulate its own promoter [21, 22]. Therefore herein, we determined whether HMGB1 has a role in the autostimulation of the ORF50 promoter.

Materials and Methods

Cell Culture, plasmids and transfection

293T cells were cultured in DMEM supplemented with 10% FCS, glutamine, and penicillin-streptomycin (Invitrogen). The ORF50 promoter and Rta expression constructs [23] and HMGB1 constructs [15] have been previously described. pET-HMGB1 was produced by PCR amplification and cloning of the HMG boxes into pET-21b. Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen), a maximum of 1μg of each DNA was used per transfection in a maximum total of 3μgs per combination transfection as the manufacturer’s instructions.

Luciferase assays

Cells were harvested 48 hr post-transfection and lysed with passive lysis buffer (PLB) (Promega), following the manufacturer’s instructions. The luciferase activity of the lysates was determined using LUCII reagent (Promega) according to the manufacturer’s protocol using a Lumat LB 9507 luminometer (EG&G Berthold, Germany).

Chromatin Immunoprecipitation (ChIP) Assays

Cell lysates were harvested and ChIP assays performed using the ChIP assay kit (Upstate Biotechnology). Chromatin extracts, cross-linking, sonication, immunoprecipitation, agarose bead elution and protein removal were carried out based on the manufacturer’s protocol. DNA recovered from immunoprecipitates with the appropriate antibodies was used as a template for PCR amplifications using primers to the ORF50 promoter (5’-CAAAGAGCTTGGGGGGGCAGA-3’ and 5’-TGCCACCCAGCTACTGGTTTC-3’).

In vitro pulls downs assays

Recombinant His-tagged HMGB1 was expressed and immobilised to Ni-NTA agarose beads as previously described [24]. Protein expression of Rta and Oct-1 was performed by in vitro transcription/translation using the TNT system (Promega) according to the manufacturer’s instructions.

Results

HMGB1 and KSHV Rta synergistically upregulate the ORF50 promoter

To determine whether HMGB1 has a role in the autostimulation of the ORF50 promoter, 293T cells were transfected with pORF50Δ6 (which contains an ORF50 promoter fragment encompassing 568 bps upstream of the transcription start site cloned upstream of the luciferase reporter gene [23]) in the absence or presence of pHMGB1 and/or pRta-GFP. Cells were harvested 48 hr post transfection and the protein concentration of each sample calculated. An equal amount of cell extract was then assayed for luciferase activity. Moreover, to confirm an equal transfection efficiency, immunofluorescence was undertaken to ensure a comparable number of cells (approximately 60%) were transfected (data not shown). Results showed that HMGB1 alone had little effect on the ORF50 promoter, whereas autostimulation was observed in the presence of Rta, as previously described [21, 22]. However, in the presence of both HMGB1 and Rta a dramatic synergistic response was observed (Fig. 1a). This suggests that HMGB1 enhances the Rta-mediated autostimulation of the ORF50 promoter, encompassing a promoter region of 568 bps. It must be noted that this was the largest deletion construct tested in this assay.

Figure 1. HMGB1 and Rta synergistically upregulate the ORF50 promoter.

293T cells were transfected with (a) pORF50Δ6 or (b) pORF50Δ7-9 in the presence of pGFP, pHMGB1 and pRta-GFP as indicated. Cells were harvested 48 hrs post-transfection and cell lysates were assayed for luciferase activity. The variations between 3 replicated assays are indicated and data is presented as fold activation versus empty vector control.

To map the HMGB1 responsive regions within the ORF50 promoter, a series of promoter deletions, pORF50Δ7-9 were utilised [23]. 293T cells were transfected with each construct in the absence or presence of pHMGB1 and/or pRta-GFP and assayed for luciferase activity as described above. A synergistic response in the presence of both Rta and HMGB1 was observed using pORF50Δ7 and pORF50Δ8. However, this stimulatory response was lost between Δ8 and Δ9 constructs. (Fig. 1b). This suggests the region required by HMGB1 and Rta to synergistically activate the ORF50 promoter is located within the 128 bps between Δ8 and Δ9, comprising 71299-71427 bps of the published sequence [1].

HMGB1 associates with the ORF50 promoter

To determine whether the synergistic response on the ORF50 promoter in the presence of HMGB1 and Rta is through direct interactions of HMGB1 with the ORF50 promoter, chromatin immunoprecipitation (ChIP) assays were performed. 293T cells were transfected with pORF50Δ6 in the absence or presence of either pHMGB1 and/or pRta-GFP. After 24 hours, the cells were harvested and chromatin extracts, cross-linking, sonication, immunoprecipitation, agarose bead elution and protein removal were performed. DNA recovered from immunoprecipitates with HMGB1- or GFP-specific antibodies were then used as templates to specifically amplify the ORF50 promoter. Results demonstrate that HMGB1 can associate with the ORF50 promoter in both the absence and presence of Rta (Fig. 2). Moroever, ChiP analysis demonstrates a positive interaction between endogenous HMGB1 and the ORF50 promoter in GFP-transfected cells. These results suggest that the synergistic response is possibly due to the HMGB1 protein binding the ORF50 promoter.

Figure 2. HMGB1 associates with the ORF50 promoter.

PCR amplification of the KSHV ORF50 promoter from HMGB1-, GFP-specific or no antibody immunoprecipitates using cell extracts transfected with the pGFP, pHMGB1 and/or pRta-GFP in the presence of pORF50Δ6.

Oct-1 binding to the ORF50 promoter is required for the HMGB1/Rta synergistic response

The deletion analysis experiments shown in Figure 1b highlight a region within the ORF50 promoter required for the HMGB1/Rta synergistic response. This region encompasses an Oct-1 binding site, essential for Rta-mediated autostimulation [25]. Interestingly, the HMG boxes of HMGB1 and 2 proteins have been shown to stimulate the binding of sequence-specific binding proteins, including Oct-1, in vitro. Moreover, Rta has been shown to directly interact with Oct-1 and this interaction is critical for transactivation of lytic promoters [26]. Therefore, we tested whether a multi-protein complex is formed comprising HMGB1, Oct-1 and Rta and also assessed the direct interactions between these proteins. Pulldown experiments were performed using the recombinant histidine-tagged HMGB1 protein immobilised to beads and incubating with radio-labelled Oct-1 or Rta proteins. Results showed that Oct-1 bound directly to the recombinant HMG boxes, in contrast no direct interaction was observed between HMGB1 and Rta (Fig. 3a). However, upon repeating the pulldown, incubating HMGB1 protein in the presence of both radiolabelled Oct-1 or Rta a ternary complex was observed (Fig. 3a). This suggests that Oct-1 may function as a bridge between HMGB1 and Rta to facilitate assembly of an enhanceosome.

Figure 3. Oct-1 binding to the ORF50 promoter is required for the HMGB1/Rta synergistic response.

(a) Ni-NTA bead bound HMGB1 or beads alone were incubated with ³⁵S-Met-labelled Oct-1, Rta or empty vector control produced by ITT. Following washes, bound proteins were separated by SDS-PAGE and the dried gel was exposed to autoradiograph film for 16 hrs. (b) PCR amplification of the ORF50 promoter from HMGB1-specific or no antibody immunoprecipitates using cell extracts transfected with the pGFP, pHMGB1 and/or pRta-GFP in the presence of pORF50ΔOct. (c) 293T cells were co-transfected with (a) pORF50Δ6 or (b) pORF50ΔOct in the presence of pGFP, pHMGB1 and pRta-GFP as indicated. Cells were harvested 48 hrs post-transfection and cell lysates were assayed for luciferase activity. The variations between 3 replicated assays are indicated and data is presented as fold activation versus empty vector control.

To further investigate the role of this multi-protein complex on the activation of the ORF50 promoter, we assessed what effect mutating the Oct-1 binding site within the ORF50 promoter would have on the synergistic activation by HMGB1/Rta. A mutant ORF50 promoter construct upstream of the luciferase reporter gene, pORF50-ΔOct, was therefore generated by site-directed mutagenesis which prevented the binding of Oct-1 as previously described [25]. ChIPs assays were then performed to assess whether HMGB1 could still bind with the ORF50 promoter-lacking the Oct-1 binding site. Results showed that HMGB1 can still associate with the ORF50 promoter in the absence of Oct-1 (Fig. 3b).

The ability of HMGB1/Rta to synergistically activate the ORF50 promoter in the absence of Oct-1 was then assessed. 293T cells were transfected with pORF50-ΔOct in the absence or presence of pHMGB1 and/or pRta-GFP. Cell extracts were then assayed for luciferase activity. A decrease in autostimulation of the pORF50-ΔOct promoter construct in the presence of Rta was observed compared to the wild type promoter, as previously observed [25]. However, no synergistic response was observed in the presence of Rta and HMGB1 using the ORF50 promoter construct lacking the Oct-1 binding site (Fig. 3c). Therefore, even the presence of endogenous or overexpressed HMGB1 has limited effect on the ORF50 promoter in the absence of Oct-1 and suggests that Oct-1 binding is required by HMGB1 and Rta to synergistically activate the ORF50 promoter.

The latent-lytic switch is essential for KSHV virion propagation and has critical implications in disease pathogenesis [2]. KSHV Rta is the key regulator of the latent-lytic switch, thus the regulation of the ORF50 promoter and Rta’s ability to autostimulate is key to lytic reactivation. Herein, we have demonstrated that HMGB1 augments Rta-mediated autostimulation of the ORF50 promoter. HMGB1 has previously been shown to enhance the ability of Rta to bind DNA to four different Rta-responsive sequences, thus leading to their transactivation. Results herein suggest that HMGB1 can also interact with the ORF50 promoter suggesting that HMGB1 has little sequence specificity and is consistent within the hypothesis that HMGB1 probably recognises DNA structure rather than sequence [18]. In an attempt to identify any sequence specificity of HMGB1 we utilised a deletion series of the ORF50 promoter and analysis suggested that the synergistic response observed was lost upon deletion of the Oct-1 binding site within the ORF50 promoter. Further analysis to assess a possible role of Oct-1 in HMGB1-mediated activation identified a possible direct interaction between HMGB1 and Oct-1. It has previously been shown that HMG proteins can functionally interact with the POU domains of Oct factors [27]. Moreover, analysis suggests that Oct-1 may also interact with Rta to form a ternary complex with HMGB1. However, we believe this ternary complex may be transient, as we have been unable to detect a HMGB1/Oct-1/DNA or HMGB1/Rta/DNA complexes with sufficient stability to assay using EMSAs (data not shown). It will be of interest to determine whether this synergistic response and complex formation exists in primary effusion lymphomas and what effect mutating the Oct-1 binding site with the ORF50 promoter has in the context of the wild type virus. Interestingly, recent analysis has shown that the Oct-1/Rta interaction is critical for regulating KSHV reactivation [26]. Furthermore, Oct-1 binding to the lytic K-bZIP promoter stimulates Rta DNA binding and also augments Rta-mediated transactivation. At present, however, the mechanism of this stimulation is unknown. It will also be interesting to determine whether HMGB1 has any role to play in enhancing the Oct-1/Rta augmentation of this and other KSHV lytic promoters.