The Kaposi's Sarcoma-Associated Herpesvirus ORF34 Protein Binds to HIF-1α and Causes Its Degradation via the Proteasome Pathway

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV) is the causative agent for Kaposi's sarcoma (KS) and two other lymphoproliferative disorders, primary effusion lymphoma (PEL) and multicentric Castleman's disease (MCD). Kaposi's sarcoma is a highly vascular tumor, and recently both hypoxia-inducible factor 1α (HIF-1α) and HIF-2α were detected in KS samples, indicating a role of HIFs in the KSHV life cycle. Previously, we showed that ORF34, a lytic gene of unassigned function, was activated by hypoxia and that ORF34 transcription was upregulated by both HIFs (M. Haque, D. A. Davis, V. Wang, I. Widmer, and R. Yarchoan, J Virol. 77:6761–6768, 2003). In the present study, we show that coexpression of ORF34 with HIF-1αm (degradation-resistant HIF-1α) caused substantial reduction in HIF-1α-dependent transcription, as evidenced by reporter assays. Two-way immunoprecipitation experiments revealed that ORF34 physically interacted with HIF-1αm in transient expression experiments. Deletion analysis revealed that three different ORF34 domains interacted with the amino-terminal domain of HIF-1α. Also, purified HIF-1α and ORF34 proteins interacted with each other. The observed transcriptional inhibition of HIF-1α-dependent promoters was attributed to degradation of HIF-1α after binding with ORF34, since the overall amount of wild-type HIF-1α but not the degradation-resistant one (HIF-1αm) was reduced in the presence of ORF34. Moreover, ORF34 caused degradation of HIF-1α in a dose-dependent manner. Inhibition of the ubiquitin-dependent pathway by the chemical proteasome inhibitor MG132 prevented HIF-1α degradation in the presence of ORF34. These results show that ORF34 binds to HIF-1α, leading to its degradation via the proteasome-dependent pathway.

INTRODUCTION

Kaposi's sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus 8 (HHV-8), is a gamma-2 herpesvirus that is related to Epstein-Barr virus (EBV) and herpesvirus saimiri (1, 2). KSHV is the causative agent for Kaposi's sarcoma (KS) and two other lymphoproliferative disorders, primary effusion lymphoma (PEL) and multicentric Castleman's disease (3–5).

KSHV shares significant sequence homology with EBV (6). As with other herpesviruses, KSHV has two distinct phases of life cycles, latent and lytic replication. During latency, a limited number of viral genes are expressed that play a key role in the replication and maintenance of the viral genome in KSHV-infected cells (7–9).

Lytic gene expression can be induced by treatment of latently infected cells with chemical agents, such as 12-0-tetradecanoyl 13-acetate (TPA) and sodium butyrate (10, 11). Recently, it has been reported that hypoxia, a physiologic stimulus, can also induce lytic replication of KSHV in PEL cells (12). The term “hypoxia” refers to the presence of low (inadequate) levels of oxygen in cells or tissues. This condition represents a physiological stress that allows KSHV to replicate and cause tumors in human body extremities, where oxygen levels typically are low. Recent results from different laboratories indicated that hypoxia and hypoxia-inducible factors (HIFs) play important roles in the KSHV life cycles (13–18).

HIFs are heterodimeric transcription factors composed of one of the three oxygen-sensitive HIF-α subunits (HIF-1α, HIF-2α, and HIF-3α) and a constitutively expressed HIF-1β subunit, also known as aryl hydrocarbon receptor nuclear translocator (ARNT) (19–23). Under normoxic conditions, HIF-α proteins are expressed but targeted for proteasomal degradation by interaction with von Hippel-Lindau (pVHL) tumor suppressor protein and subsequent polyubiquitination (24–27). The interaction of pVHL and HIF-α proteins is regulated by the hydroxylation of specific proline residues in the oxygen-dependent degradation domain (ODD) of HIF-α (28–30). In mammalian cells, three prolyl hydroxylase domain (PHD) enzymes, PHD1, PHD2, and PHD3, also known as HIF prolyl hydroxylases (HPHs), have been shown to hydroxylate HIF-α proteins (31–33). In the presence of oxygen, the activity of HIF-α proteins is also regulated by hydroxylation of specific asparagine residues within the C-terminal transactivation domain of HIF-α by a specific asparaginyl hydroxylase, called factor inhibiting HIF (FIH) (34–36). Under hypoxic conditions, PHD activity is limited, which increases the stability of HIF-α proteins, leading to rapid accumulation of the protein and an increased ability to recruit coactivators (37, 38). These enzymatic modifications can also be inhibited by iron chelation and cobalt ions (38). Once stabilized, HIF-α proteins translocate to the nucleus, heterodimerize with constitutively expressed HIF-β subunit, bind to hypoxia-responsive elements (HREs) within the promoter or enhancer region of hypoxia-responsive target genes, and upregulate gene expression (39). The core consensus sequence that HIF heterodimers bind to has been identified as 5′-R (A/G) CGTG-3′ (40). HIF-1α and HIF-2α have a high degree of sequence similarity and biochemical properties (19). Both proteins dimerize with ARNT and bind to the same DNA sequences but have distinct biological functions (41, 42). Less is known about HIF-3α compared to HIF-1α and HIF-2α. While HIF-1α and HIF-2α act as transcriptional activators for their target genes, HIF-3α has been shown to act as a transcriptional repressor (39).

In KSHV, there is increasing evidence that hypoxia and HIFs play an important role in KSHV latent and lytic replication. Kaposi's sarcoma is a highly vascular tumor, and recently both HIF-1α and HIF-2α were detected in KS samples, indicating a role of HIFs in the KSHV life cycle (14, 43). It was recently reported that two KSHV latent proteins, latency-associated nuclear antigen 1 (LANA1) and viral interferon regulatory factor 3 (VIRF3), also known as LANA2, stabilize HIF-1α through protein-protein interactions (13, 17). Latent KSHV infection of endothelial cells increases the level of both HIF-α mRNA and protein (14). More recently, hypoxia and HIFs have been shown to induce transcription of LANA in KSHV-infected PEL cells, and the LANA promoter can be transcriptionally activated by both HIF-α proteins acting through at least one HIF binding site within the promoter region (44). The viral G protein-coupled receptor (vGPCR) carried by ORF74 is a KSHV lytic gene shown to upregulate vascular endothelial growth factor (VEGF) production by stimulating the activity of HIF-1α (18). Moreover, both HIF-α proteins can induce transcription and expression of the KSHV RTA (15), and several HIF binding sites have been identified in the promoter region of RTA (13).

We have reported that ORF34, a lytic gene of unknown function, was activated by hypoxia. Moreover, both HIFs upregulated ORF34 transcription through an HRE located in the ORF34 promoter region (15). In addition, we demonstrated that the same ORF34P HRE could upregulate transcription of downstream genes carried by ORF35 to ORF37 (16). In the present study, we have investigated functional interactions between ORF34 and HIF-1α and identified ORF34 as the interacting partner of HIF-1α. Our data demonstrate that ORF34 promotes degradation of HIF-1α acting through the ubiquitin proteasome pathway.

MATERIALS AND METHODS

Cell culture and transfection.

Human embryonic kidney 293 (HEK293) and 293T (HEK293T) cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% antibiotics. The cells were maintained at 37°C in a 5% CO₂ and 95% air incubator. For immunoprecipitation and immunoblotting, subconfluent cells were transfected in 10-cm or 6-well plates. For luciferase assay, cells were transfected in 24-well plates. All transfections were carried out using Genjet in vitro transfection reagent from SignaGen Laboratories (Ijamsville, MD) as recommended by the manufacturer.

Expression plasmids.

Full-length ORF34 was amplified by PCR with specific primers containing BglII and EcoRI from BCBL-1 genomic DNA and cloned into green fluorescent protein (GFP)-tagged pEGFP-C1 (Clontech, CA) in the same sites. Different deletion mutants of GFP34 were constructed with specific primers containing BglII and SalI sites using GFP34 as the template. The specific sites of different deletions are shown schematically in Fig. 3A.

Fig 3.

Identification of ORF34 regions that interact with HIF-1α. (A) Schematic diagram showing the different deletion mutants of ORF34 (D1, D2, and D3) used in the IP assay. (B) HEK293 cells were transiently transfected with Flag-tagged HIF-1αm along with different deletion mutants (DM) of GFP34 as indicated. After 30 h of transfection, whole-cell extracts were prepared and immunoprecipitated with anti-GFP antibody. The immunoprecipitates were analyzed by immunoblotting with anti-HIF-1α antibody. The presence of GFP34 and HIF-1α in the whole-cell extracts was analyzed by immunoblotting using specific antibodies.

Full-length wild-type human HIF-1α (pHA HIF-1αW) and its mutant plasmid containing two alanine substitutions in place of proline at positions P402A and P564A (pHA HIF-1αm), which escape pVHL-mediated degradation, were provided by Eric Huang (University of Utah School of Medicine, Salt Lake City, UT). These plasmids were used as templates in PCR to construct full-length Flag-tagged wild-type and degradation-resistant HIF-1α using gene-specific primers. These gene constructs were subsequently inserted into HindIII-XbaI sites of p3XFLAG-CMV14 expression vector (Sigma, MO) and designated Flag-tagged HIF-1αw and Flag-tagged HIF-1αm. To construct mammalian glutathione S-transferase (GST)-fused expression vectors encoding full-length HIF-1αm and its different deletion mutants, the full length and specific segments of HIF-1α were subcloned into the KpnI-NotI sites of a pEBG expression vector (plasmid 22227; Addgene, MA) using Flag-tagged HIF-1αm as a PCR template. The specific sites of different deletions are shown schematically in Fig. 4A. Hemagglutinin (HA)-tagged ubiquitin expression plasmid was a gift from Clinton Jones (University of Nebraska, Lincoln, NE).

Fig 4.

Identification of HIF-1α domains that interact with ORF34. (A) Schematic diagram showing the different deletion mutants of GST-HIF-1α (D1, D2, D3, and D4) used in the IP assay. (B) HEK293 cells were transiently transfected with GFP34 along with different deletion mutants (DM) of GST-HIF-1α as indicated. After 30 h of transfection, whole-cell extracts were prepared and immunoprecipitated using anti-GFP antibody. Immunoprecipitates were analyzed by anti-GST antibody to detect the specific segment of GST-HIF-1α bound to ORF34 protein. Arrowheads indicate specific binding, and an asterisk denotes nonspecific binding by the antibody used. (C) The presence of GST-HIF-1α and GFP34 in the whole-cell extracts was analyzed by immunoblotting using specific antibodies. The asterisk in the upper panel denotes nonspecific binding by the antibody used.

For prokaryotic expression, full-length ORF34 and HIF-1α were cloned into the EcoRI-XhoI site of pGEX4T-1 (GE Healthcare, NJ) and the BamHI-NotI site of pET-32a (EMD Millipore, MA) expression vector, respectively.

Reporter plasmids.

The wild-type human erythropoietin (EPO) HRE-luciferase reporter (pEpoE-luc) containing a 50-bp EPO 3′ enhancer sequence, as well as its mutant (pEpoEm-luc), containing a 3-bp substitution in the HIF-α binding site (45), were obtained from Eric Huang (University of Utah School of Medicine, Salt Lake City, UT). The human VEGFp12 reporter (46) was constructed by inserting a 99-bp fragment containing the HIF-α binding site upstream of the simian virus 40 (SV40) promoter of the pGL3 promoter vector into NheI-BglII sites using primers VEGFp12 F (5′-CGA GCT AGC CCT TTG GGT TTT GCC AGA-3′) and VEGF p12 R (5′-AGC AGA TCT ACG GGA AGC TGT GTG GTT-3′; boldface letters indicate restriction endonuclease sites).

Antibodies and inhibitors.

Rabbit polyclonal and mouse monoclonal antibodies against GFP (A01388 and A00185) and monoclonal antibody to GST (A00865) were purchased from GenScript (Piscataway, NJ). Mouse monoclonal anti-FLAG M2 (F3165) and anti-β-actin (A5316) were from Sigma (St. Louis, MO). Rabbit polyclonal antibody to GST (SC-459) and mouse monoclonal antibody to HA (SC-7392) were from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse monoclonal antibody to anti-FLAG M2 (200472) was from Agilent Technologies, Inc. (Wilmington, DE). Mouse monoclonal anti-human HIF-1α (610959) was from BD Biosciences (San Jose, CA). For secondary antibodies, where appropriate, blots were incubated with either alkaline phosphatase-conjugated goat anti-mouse (S3721) and goat anti-rabbit (S3721) antibodies (Promega, WI) or horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (SC-2005) from Santa Cruz Biotechnology. The proteasome, lysosome, and protein synthesis inhibitors MG132 (C2211), chloroquine (CQ), and cycloheximide, respectively, were from Sigma.

Transfection and reporter assay.

The day before transfection, HEK293 cells were plated at 60 to 70% confluence in a 24-well plate and transfected with 250 ng each of VEGFp12, pEpoE-luc, and pEpoEm-luc promoters along with 125 ng of Flag-tagged HIF-1αw or Flag-tagged HIF-1αm plasmid in the presence or absence of 125 ng of GFP34 expression plasmid. Cells were also transfected with 2 ng/well of an internal control plasmid, Renilla luciferase (pRL-SV40), to normalize transfection efficiency, and the total amount of DNA was kept constant by using the appropriate control vector. After transfection, cells were incubated in normoxic conditions for 30 h and lysed using passive lysis buffer. Firefly and Renilla luciferase activities were measured with a Victor-2 luminometer (PerkinElmer, MA) using a dual-luciferase kit (Promega, WI) and were normalized to an internal control plasmid. Luciferase activity was calculated by dividing the activity for each reporter by that of the empty vector and compared to the results obtained for HIF-1α or the combination of HIF-1α and GFP34.

Immunoblot analysis.

Cell lysates were prepared in 0.5% IGEPAL lysis buffer (50 mM Tris-HCl, pH 8.00, 150 mM NaCl, 0.5% IGEPAL CA-630, 1 mM EDTA) supplemented with protease inhibitor and clarified by centrifugation for 10 min at 4°C. Cleared lysates were mixed with sample buffer, heated at 92°C for 5 min, and electrophoretically separated through 4 to 20% Tris-HEPES-SDS gels (Thermo Scientific, IL). Following electrophoresis, gels were transferred to nitrocellulose membranes using an iBlot dry transfer device (Invitrogen, CA) and blocked for 60 min in Tris-buffered saline containing 0.1% Tween 20 (TBST) plus 5% nonfat dry milk. Membranes were probed with the specific primary antibodies and then with either goat anti-mouse secondary antibody conjugated with alkaline phosphatase or HRP-conjugated goat anti-mouse IgG. After washing the membranes 3 to 4 times with phosphate-buffered saline-Tween 20 (PBS-T), the specific signal was detected with appropriate substrate.

Immunoprecipitations.

At the indicated times posttransfection, cells were washed once with PBS and whole-cell extracts were prepared in 0.5% IGEPAL lysis buffer supplemented with protease inhibitor. Cell extracts were incubated at 4°C for 10 min on a shaker and centrifuged at 14,000 rpm for 10 min, and supernatants were collected into new tubes. Immunoprecipitation reactions were performed by using protein G Dynabeads (Invitrogen, CA) or protein A/G-plus-agarose beads (Santa Cruz, CA) according to the manufacturer's instructions. Briefly, cell extracts were incubated with specific antibodies overnight at 4°C. The next day, 20 μl of Dynabeads or protein A/G plus agarose was added to the antigen-antibody complex and incubated at room temperature (RT) for 20 min. After being washed with PBS-T 3 or 4 times, the samples were eluted in 35 μl of sample buffer heated at 72°C for 10 min, subjected to SDS-PAGE using 4 to 20% Tris-HEPES-SDS gels, and transferred to a nitrocellulose membrane using an iBlot dry transfer device. After blocking the nonspecific binding sites, the membranes were analyzed by immunoblotting with specific antibodies as described above. The presence of specific proteins in transfected cells was confirmed by immunoblotting with specific antibodies to each protein.

Expression and purification of fusion proteins.

Full-length HIF-1α was expressed from pET-32a expression vector as 6His- and thioredoxin (trx)-tagged fusion proteins and purified using Bio-Rad's automated Profanity IMAC (immobilized metal affinity chromatography) column according to the manufacturer's instructions. Briefly, a single colony of Escherichia coli BL21 star (DE3) (Invitrogen, CA) harboring the gene was grown in 2 ml of LB medium at 37°C overnight with appropriate antibiotics on a rotating incubator. The next day, 2 ml of overnight culture was added to 98 ml of fresh LB medium and grown at 37°C until the optical density at 600 nm (OD₆₀₀) of the cultured cells reached 0.6 to 0.8. Expression of the fusion protein was induced with 0.25 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 3 h at 32°C. The induced bacterial cultures were harvested, washed once in STE buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA), and treated with 1 mg/ml of lysozyme at RT for 30 min, followed by the addition of 50 U/ml benzonase and incubation at RT for another 60 min. Bacterial lysates were then pelleted by centrifugation at 4,600 rpm for 15 min and the supernatant discarded. Pellets were lysed in lysis buffer (10 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10 mM imidazole) containing 0.25 to 1% N-lauroylsarcosine (Sarkosyl) until the lysis was completed. Cells were pelleted by centrifugation at 14,000 rpm for 20 min at 4°C, and supernatant was collected. Triton X-100 was added to the clear lysate at a final concentration of 3 to 4% and incubated at RT for 30 min, after which it was subjected to purification by a Profinia instrument.

Full-length ORF34 was expressed as a GST fusion protein and purified by binding to glutathione-Sepharose beads using column chromatography. Briefly, E. coli BL21 star (DE3) harboring individual colonies of GST and GST34 was induced, and cell lysates were prepared as described above. To each clarified and Triton X-100-treated cell lysate, 500 μl of glutathione Sepharose beads previously washed with PBS-T was added, and the mixture was incubated for at least 1 h to overnight at 4°C. The bead-bound protein was loaded into gravity flow columns and washed extensively with PBS-T. Finally, the protein bound to beads was eluted with GST elution buffer containing 10 mM reduced glutathione in 50 mM Tris-HCl, pH 8.0.

In vitro pulldown assays.

For pulldown assays, equal amounts of GST and GST34 proteins were mixed with 250 ng of pET-32a-HIF-1α in 0.5 ml of pull-down buffer (50 mM Tris-HCl, pH 8.00, 150 mM NaCl, 0.5% IGEPAL CA-630, 1 mM EDTA, 1 mM dithiothreitol [DTT], and 10% glycerol) and incubated at 4°C for 20 min. Glutathione-Sepharose 4B (15 μl) was added to each tube and incubated for an additional 30 min. The beads were washed extensively with pull-down buffer, and the bound proteins were eluted by boiling in 2× sample buffer at 92°C for 5 min. The eluted proteins were analyzed by SDS-PAGE followed by Western blotting using antibody to HIF-1α.

Protein degradation and stability assays.

HEK293 cells were cotransfected with Flag-tagged HIF-1αw or HIF-1αm with or without GFP34 in 6-well plates using 1 μg of DNA per well, and the total amount of DNA was adjusted (where appropriate) using control vector. After 30 h of transfection, cells were washed once with culture medium and supplemented with 10 μM MG132, 100 μM chloroquine, or control dimethyl sulfoxide (DMSO) in specific wells for 16 h and harvested 46 h posttransfection. For protein stability assays, after 24 h of transfection cells were treated with 50 μg/ml of cycloheximide for the indicated time periods to block new protein synthesis. Whole-cell extracts were prepared in 0.5% IGEPAL lysis buffer and immunoblotted with specific antibodies.

In vivo ubiquitination assay.

For ubiquitination assays, HEK293 cells were cotransfected with plasmids encoding GFP34 or Flag-tagged wild type HIF-1α in the presence of HA-ubiquitin or a combination of these plasmids as shown in Fig. 7. After 30 h of transfection, cells were washed once with culture medium and supplemented with 10 μM MG132 for 16 h before harvest. After 46 h of transfection, cell lysates were prepared in 0.5% IGEPAL lysis buffer, and equal amounts of cell lysates (800 μg) were immunoprecipitated with anti-Flag, anti-HA, or anti-GFP antibody at 4°C overnight. The next day, 20 μl of protein A/G plus agarose was added to the antigen-antibody complex and incubated at 4°C for 60 min. After washing 3 or 4 times with lysis buffer, the samples were eluted in 30 μl of sample buffer heated at 92°C for 5 min and subjected to SDS-PAGE. Immunoprecipitates were then analyzed by immunoblotting with specific antibody as indicated. The presence of each protein in the transfected cell lysates was analyzed by immunoblotting of equal amounts of total cell lysates with specific antibody.

Fig 7.

Ubiquitination of HIF-1α and ORF34. (A) HEK293 cells were cotransfected with HA-tagged ubiquitin (HA-Ub; 300 ng) along with GFP-tagged ORF34 (100 ng) or Flag-tagged HIF-1αw (600 ng) or with a combination of both as indicated. The total amount of DNA in each well was kept constant using a control vector. After 30 h of transfection, cells were treated with MG132 for 16 h. Whole-cell extracts were prepared and immunoprecipitated with anti-Flag, anti-HA, or anti-GFP antibody and analyzed by immunoblotting with specific antibody as indicated. (B) Expression of ubiquitin, HIF-1α, and ORF34 in total cell extracts was analyzed by immunoblotting with the indicated specific antibody. The asterisk in the middle panel denotes nonspecific binding by the antibody used.

Statistical analysis.

Statistical analyses were performed using GraphPad Prism 5 software. For comparisons of control reporters under different treatments, data were analyzed using one-way analysis of variance (ANOVA) followed by Bonferroni's multiple-comparison test, and P values of less than 0.05 were considered statistically significant.

RESULTS

ORF34 inhibits transcriptional activity of HIF-1α.

KSHV ORF34 is an early-late gene encoding 327 amino acids with unassigned function (6). ORF34 is the first viral gene, not only for KSHV but also among all viral genes, shown to be activated by hypoxia (15). Previously, we showed that both HIF-1α and HIF-2α strongly upregulated the ORF34 promoter acting through a functional HRE in the promoter region of ORF34 (15, 16). In the present study, we investigated whether transcriptional activity of HIF-1α is affected by ORF34. We selected two prototypic HIF-α-responsive promoters. VEGFp12 is an HIF-1α-specific (46) luciferase reporter which contains one copy of an HIF-α binding site inserted upstream of an SV40 promoter (pGL3 promoter; Promega). The EPO HRE luciferase reporter is an HIF-2α-specific promoter (47, 48) containing a 50-bp EPO 3′ enhancer sequence inserted upstream of the SV40 promoter (45). HEK293 cells were cotransfected with VEGFp12 or with wild-type pEpoE-luc HRE luciferase reporter plasmid along with Flag-tagged HIF-1αw or HIF-1αm alone or in combination with GFP34 and Flag-tagged HIF-1αm (Fig. 1A and B). These experiments revealed that both wild-type VEGFp12 and pEpoE-luc promoters were moderately activated by Flag-tagged HIF-1αw plasmids, and this activation was further enhanced in cells cotransfected with Flag-tagged HIF-1αm-expressing plasmids. However, coexpression of GFP34 substantially decreased the transcriptional activity of HIF-1α for both reporting promoters. In cells transfected with the mutant pEpoEm-luc reporter, the luciferase activity was not affected by either Flag-tagged HIF-1αw or Flag-tagged HIF-1αm alone or in combination with GFP34 (Fig. 1C). These results indicate that ORF34 acts as a negative regulator of HIF-1α.

Fig 1.

ORF34 suppresses the transcriptional activity of VEGF and EPO luciferase reporters mediated by HIF-1α. Human VEGFp12 (A), an HIF-1-specific promoter, pEpoE-luc (B), and its mutant, pEpoEm-luc (C), an HIF-2-specific promoter, were cotransfected into HEK293 cells with 250 ng of each of these reporters and the indicated amount of Flag-tagged HIF-1α wild-type or mutant plasmid in the absence or presence of GFP34. After 30 h of normoxic incubation, cells were harvested and assayed for luciferase activity. Results were normalized to cells transfected with empty vector. Statistical analyses were done using GraphPad Prism 5, and P values (P < 0.05) of each treatment were compared to the control vector and are indicated where significant. Each result is the mean from two independent experiments done in triplicate. EV, empty vector.

Interaction of ORF34 and HIF-1α protein.

Based on the results that ORF34 expression caused substantial reduction in HIF-1α-dependent transcriptional activation, we tested whether HIF-1α physically interacted with ORF34 protein in transient coexpression experiments. The full-length coding sequences of ORF34 and HIF-1α were cloned into GFP- and GST-tagged expression vectors (Fig. 2A). For coimmunoprecipitation experiments, 293T cells were transfected with GFP34 alone or in the presence of GST-HIF-1αm, a plasmid with specific proline mutations at the ODD domain of HIF-1α (Fig. 2A) that causes resistance to pVHL-mediated degradation while remaining functional in normoxia (25, 28). Cell lysates from transfected cells were first immunoprecipitated with anti-GFP polyclonal antibody and then immunoblotted with anti-GST monoclonal antibody to detect the presence of GST-HIF-1α bound to GFP34. The anti-GFP antibody coimmunoprecipitated GST-HIF-1α from the cell extract obtained from cells cotransfected with GST-HIF-1αm plasmid but not from the cell lysates transfected with control vector (Fig. 2B). In the reverse experiment, anti-GST polyclonal antibody coimmunoprecipitated GFP34 only in the presence of GST-HIF-1αm but not from control cell lysates (Fig. 2C). The presence of GFP34 and GST-HIF-1α in the cell extracts was detected with specific antibodies, and actin was used to monitor the equal loading control (Fig. 2B and C).

Fig 2.

In vivo and in vitro interaction of ORF34 with HIF-1α. (A) Schematic diagram showing coding potentials of ORF34 and HIF-1α protein along with different regulatory domains of HIF-1α. KSHV ORF34 is encoded by 327 amino acids with unknown function, and no functional domains have been identified. HIF-1α protein is 826 amino acids long. It has bHLH and PAS domains at the N terminus, which mediate DNA binding, dimerization, and protein-protein interactions; an ODD domain at the central region, which regulates the oxygen-dependent stability of HIF-1α protein through the hydroxylation of two specific proline residues, as indicated; and two transactivation domains at the C terminus (N-TAD and C-TAD) that are involved in the transregulatory function of HIF-1α protein. (B and C) In vivo interactions between KSHV ORF34 and human HIF-1α. Human 293T cells were transiently transfected with GFP34 or cotransfected with mammalian GST-tagged degradation-resistant HIF-1α as indicated. After 30 h of transfection, whole-cell extracts were prepared and immunoprecipitated (IP) using either anti-GFP (B) or anti-GST (C) antibody. The immunoprecipitates were analyzed by immunoblotting (IB) with anti-GST or anti-GFP antibody. The presence of GFP34 and GST-HIF-1α in the cell extracts was analyzed by immunoblotting with specific antibodies. Arrowheads indicate specific binding, and an asterisk denotes nonspecific binding by the antibody used. (D) The relative purities of GST, GST34, and pET-32a-HIF-1α proteins used in the binding experiment are shown, as evaluated by SDS-PAGE and Coomassie blue staining. (E) In vitro interaction of HIF-1α and ORF34 in a GST pulldown assay. His-tagged purified HIF-1α was incubated with equal amounts of purified GST and GST34 protein, subjected to Western blotting, and probed with anti-HIF-1α antibody.

To confirm this interaction in a cell-free system, we performed an in vitro GST pulldown experiment using purified, full-length GST34 and pET-32a-HIF-1α proteins expressed in E. coli (Fig. 2D). Specifically, purified GST and GST34 proteins were incubated directly with pET-32a-HIF-1α in a pulldown buffer (50 mM Tris-HCl, pH 8.00, 150 mM NaCl, 0.5% IGEPAL CA-630, 1 mM EDTA, 1 mM DTT, and 10% glycerol) as described in Materials and Methods. HIF-1α bound specifically to GST34 but not to control GST protein (Fig. 2E). Collectively, these results demonstrate that KSHV ORF34 interacts with HIF-1α in vivo and in vitro.

Identification of interacting domains of HIF-1α and ORF34 proteins.

In order to identify the specific interacting regions of ORF34 for HIF-1α, we constructed different deletion mutants of the ORF34 gene (Fig. 3A). GFP34 D1, D2, and D3 appeared to be expressed in the cytoplasm of transfected cells (not shown). We have tested the ability of each ORF34 domain to bind to HIF-1α after coexpression in HEK293 cells using coimmunoprecipitation experiments (Fig. 3B). The three GFP34 deletion mutants were cotransfected with full-length Flag-tagged HIF-1αm plasmid (Fig. 3B). Cell extracts were immunoprecipitated with anti-GFP polyclonal antibody, and the immunoprecipitates were probed with anti-HIF-1α antibody to detect the regions of ORF34 bound to HIF-1α. Results showed that the C terminus of ORF34 (GFP34 D3), which encompass the amino acids 217 to 327 of ORF34, bound strongly to HIF-1α, whereas GFP34 D1 and D2 bound weakly to HIF-1α (Fig. 3B).

To identify the binding domains of HIF-1α to ORF34, different deletion mutants of HIF-1α were constructed using a mammalian GST-tagged vector (Fig. 4A). The predicted size of HIF-1α is 93 kDa, and that of GST is 25 kDa. Therefore, the predicted size of HIF-1α fused with GST is 118 kDa. These GST-HIF-1α deletion mutants were cotransfected with GFP34 into HEK293 cells, and cell extracts were immunoprecipitated with anti-GFP polyclonal antibody (Fig. 4B). The immunoprecipitates were probed with anti-GST monoclonal antibody to detect the HIF-1α domains bound to GFP34. Results revealed that bHLH and PAS domains of HIF-1α (GST-bHLH D1 and GST-PAS D2) bind to ORF34 (Fig. 4B).

ORF34 downregulates HIF-1α protein levels.

Initial experiments in which GFP34 was coexpressed with Flag-tagged HIF-1αm plasmid indicated a potential degradation of HIF-1α in the presence of GFP34 (not shown). To further investigate this phenomenon, HEK293 cells were cotransfected with HIF-1αw or mutant plasmids in the presence or absence of GFP34 (Fig. 5A). After 30 h of transfection, cell extracts were prepared and the levels of HIF-1α and ORF34 proteins were evaluated by immunoblotting with specific antibodies. Expression of wild-type HIF-1α was completely inhibited in the presence of ORF34 (Fig. 5A, lanes 2 and 4). Moreover, a reduced level of HIF-1α was detected in cells cotransfected with GFP34 and Flag-tagged HIF-1αm compared to cells transfected without the GFP34 expression plasmid (Fig. 5A, lanes 3 and 5), indicating that ORF34 causes HIF-1α degradation. The observed GFP34-induced degradation of HIF-1α occurred in a dose-dependent manner. Specifically, increasing amounts of GFP34 tested against a constant amount of Flag-tagged HIF-1αw led to a dose-dependent decrease in HIF-1α protein levels (Fig. 5B).

Fig 5.

ORF34 reduces the level of HIF-1α protein in cotransfected cells. (A) HEK293 cells were cotransfected with expression plasmids encoding Flag-tagged HIF-1α wild-type or mutant plasmids (600 ng) in the presence or absence of GFP34 (300 ng) as indicated. The total amount of DNA in each well was kept constant. Whole-cell extracts were prepared at 30 h posttransfection and immunoblotted with anti-HIF-1α and anti-GFP antibodies to detect the levels of their expression. (B) Cotransfection of ORF34 downregulates HIF-1α protein in a dose-dependent manner. HEK293 cells were cotransfected with a fixed amount (600 ng) of Flag-tagged HIF-1αw expression plasmid and with increasing amounts (0 to 400 ng) of GFP34 plasmid. Whole-cell extracts were prepared at 30 h posttransfection and immunoblotted with specific antibodies as indicated to detect the levels of their expression.

ORF34 promotes the proteasomal degradation of HIF-1α.

In mammalian cells, two major pathways mediate protein degradation, the ubiquitin-proteasome pathway and the lysosomal degradation pathway. HIF-1α protein undergoes proteasomal degradation by interacting with pVHL and many other cellular proteins (25, 28–30, 49). In order to understand the degradation pathway of HIF-1α in the presence of ORF34, we used specific inhibitors of either proteasome- or lysosome-dependent degradation to treat cells cotransfected with HIF-1α and ORF34. HEK293 cells were cotransfected with GFP34 and Flag-tagged HIF-1αw and treated with the specific inhibitors of proteasome- and lysosome-dependent degradation, MG132 (MG) and chloroquine (CQ), respectively. Immunoblot analysis of whole-cell lysates revealed that chloroquine and dimethyl sulfoxide (D; used as a control) did not block the ORF34-mediated degradation of HIF-1α (Fig. 6A, compare lane 1 to lanes 4 and 5), whereas treatment of cells with MG132 restored the expression of HIF-1α in the presence of ORF34 (Fig. 6A, lanes 1 and 3). We next examined the effects of MG132 in cells cotransfected with degradation-resistant HIF-1α plasmid with ORF34. GFP34 was coexpressed with Flag-tagged HIF-1αw or HIF-1αm expression plasmids and maintained in the absence or presence of MG132. Immunoblot analysis revealed that, in the absence of GFP34, MG132 increased the accumulation of HIF-1α expression compared to that of untreated cells (Fig. 6B, lanes 1 and 2). In the presence of GFP34, expression of HIF-1α was inhibited but treatment of cells with MG132 partially restored the accumulation of HIF-1α expression (Fig. 6B, lanes 3 and 4). In contrast, a steady expression level of HIF-1α was maintained in cells cotransfected with a Flag-tagged HIF-1αm plasmid regardless of MG132 treatment (Fig. 6B, lanes 5 and 6). Collectively, these results suggest that the proteasomal degradation pathway is involved in ORF34-mediated degradation of HIF-1α protein.

Fig 6.

ORF34 protein induces degradation of HIF-1α via the proteasome pathway. (A) HEK293 cells were cotransfected with 600 ng of Flag-tagged HIF-1α wild-type plasmid along with 100 ng of GFP34 and control empty vector. After 30 h of transfection, cells were left untreated or were treated with specific inhibitors of either the proteasomal (MG132; MG) or lysosomal (CQ) degradation pathway and harvested 16 h after treatment. Whole-cell extracts were prepared and immunoblotted with specific antibodies. (B) HEK293 cells were cotransfected with 600 ng of Flag-tagged HIF-1α wild-type or mutant plasmids along with 300 ng of GFP34, and the total amount of DNA was kept constant with the control vector. After 30 h of transfection, MG132 or control DMSO was added to specific wells and harvested after 16 h of treatment. Whole-cell extracts were prepared and immunoblotted with specific antibodies as indicated. An asterisk denotes nonspecific binding by the antibody used. (C) ORF34 protein decreases the stability of wild-type HIF-1α protein. HEK293 cells were cotransfected with 600 ng of Flag-tagged wild-type HIF-1α or mutant plasmid along with 100 ng of GFP34, and the total amount of DNA was kept constant with the control vector. After 24 h of transfection, cycloheximide was added for the indicated periods, and whole-cell lysates were prepared and immunoblotted with specific antibodies as indicated.

To confirm that the HIF-1α degradation in ORF34-transfected cells is mediated by cellular proteases, we used cycloheximide to block new protein synthesis. Since our dose-response experiment (Fig. 5B) revealed that as little as 100 ng of GFP34 can reduce the expression of wild-type HIF-1α significantly, we used this lowest dose of GFP34 to cotransfect HEK293 cells with 600 ng of Flag-tagged HIF-1α wild-type or mutant plasmids in the presence or absence of GFP34. After 24 h of transfection, cell lysates were prepared at different time points after cycloheximide treatment. An enhanced degradation of wild-type HIF-1α was observed in cells transfected with GFP34 compared to cells transfected with control vector (Fig. 6C). In contrast, the HIF-1α protein level remained relatively constant in cells cotransfected with mutant HIF-1α and GFP34 (Fig. 6C). These results indicate that ORF34 targets HIF-1α for proteasome-dependent degradation.

To confirm that ORF34 protein promotes the degradation of HIF-1α through the ubiquitin pathway, we examined the ubiquitination state of HIF-1α by an in vivo ubiquitination assay. The HA-ubiquitin expression plasmid was coexpressed with Flag-tagged wild-type HIF-1α in the presence or absence of GFP34 in HEK293 cells. After 30 h of transfection, cells were treated with MG132 for 16 h before harvest. Cells were washed once in PBS and lysed in 0.5% IGEPAL lysis buffer. Equal amounts of total cell lysates (800 μg) were immunoprecipitated with anti-Flag or anti-HA antibody and then immunoblotted with anti-HA or anti-Flag antibody to detect the ubiquitination level of HIF-1α. In the absence of GFP34, a higher level of ubiquitinated HIF-1α was observed when the immunoprecipitates were probed with either anti-HA or anti-Flag antibody. However, the level of ubiquitinated HIF-1α appeared to be substantially reduced in cells cotransfected in the presence of GFP34 (Fig. 7A, left and middle). In contrast, when the cell lysates were immunoprecipitated with anti-GFP antibody, the ubiquitination level of ORF34 remained unaffected regardless of the presence of HIF-1α (Fig. 7A, right). Immunoblotting of cell extracts with specific antibody revealed that expression of GFP34 caused reduction in the overall levels of ubiquitin as well as HIF-1α, while the level of actin remained unaffected (Fig. 7B).

DISCUSSION

We have previously shown that ORF34 is an early-late lytic gene activated by hypoxia via a functional HRE sequence located within the ORF34 promoter region. Hypoxia-induced transcriptional activation of ORF34 leads to transcriptional activation of the downstream genes ORF35 to ORF37 using the same HRE sequence (15, 16). In this study, we show that ORF34 downregulates HIF-1α protein levels and therefore HIF-1α-dependent transcription. This inhibitory activity apparently is due to direct binding of ORF34 to HIF-1α, leading to its degradation by the ubiquitin-dependent proteasome pathway. In this regard, ORF34 acts as a feed-back inhibitor of HIF-1α-dependent transcription.

Initial experiments in transfected cells revealed that expression of ORF34 reduced HIF-1α-dependent transcription of two prototypic HIF-α-responsive promoters (VEGFp12 and pEpoE-Luc) (45, 46, 48). In these experiments, HIF-1αw and HIF-1αm (hydroxylation-resistant mutant) plasmids were utilized. HIF-1αm-mediated transcriptional activation of these promoters was inhibited by ORF34. This is the first report of a KSHV gene that appears to downregulate HIF-1α-dependent transcription. Coimmunoprecipitation assays revealed that ORF34 and HIF-1α physically interacted. Apparently, this interaction requires the N-terminal bHLH and PAS domains of HIF-1α. Interestingly, this HIF-1α amino-terminal domain interacted with truncated versions representing the amino terminus, central region, and carboxyl terminus of the ORF34 protein, with the strongest interaction observed with the ORF34 carboxyl terminus (GFP34 D3). The three ORF34 fragments that interact with HIF-1α do not appear to contain repeat domains that could explain the observed binding. Therefore, these three domains appear to independently bind to HIF-1α, and as such it is likely that they represent linear-discontinuous binding sites for different HIF-1α domains.

Examination of the stability of HIF-1α in the presence of ORF34 revealed that HIF-1α was degraded in an ORF34 dose-dependent manner. Apparently, HIF-1α degradation was enhanced in the presence of cycloheximide, indicating that HIF-1α degradation by ORF34 did not require de novo protein synthesis. Therefore, the observed downregulation of HIF-1α transcription was largely due to HIF-1α degradation after binding to ORF34 protein. Regulation of HIF-1α protein stability is critical in maintaining the function of HIF-1α. HIF-1α undergoes rapid ubiquitination by binding to the von Hippel-Lindau (pVHL) protein complex followed by proteasomal degradation, resulting in inactive protein production (29, 36, 49–51). Hydroxylation of specific proline residues of HIF-1α to alanine (P402A and P564A) prevents ubiquitination, leading to stable HIF-1α protein production. Stability and regulation of HIF-1α protein have been more extensively studied with other cellular proteins. Recently, heat shock protein 70 (HSP70) and carboxyl terminus of Hsc70-interacting protein (CHIP) were identified as the HIF-1α interacting proteins and were shown to promote the ubiquitination and proteasomal degradation of HIF-1α (2). Initial evidence that HIF-1α stability is functionally linked to its ubiquitination status was obtained in experiments where the ubiquitin-proteasome pathway inhibitor MG132 prevented HIF-1α degradation in the presence of ORF34. HIF-1α appeared to be less ubiquitinated in the presence of ORF34. However, this reduction in ubiquitination is probably due to lower levels of HIF-1α due to ORF34-dependent degradation.

Interestingly, ORF34 protein appeared to be ubiquitinated in transient expression assays; however, the presence of HIF-1α did not affect the overall levels of ORF34 ubiquitination (Fig. 7). The fact that ORF34 protein is not degraded in the presence of HIF-1α indicates that the ORF34-HIF-1α protein complex does not cause degradation of both proteins. Therefore, these results suggest that physical interaction of ubiquitinated ORF34 with HIF-1α leads to HIF-1α degradation. Ubiquitination and proteasomal degradation of HIF-1α is mediated by the pVHL protein. Alignment of pVHL amino acid sequence with that of ORF34 protein did not reveal any significant sequence similarities. It is not known whether pVHL interacts with HIF-1α in the presence of ORF34. Alternatively, it is possible that ORF34 uses a novel strategy to regulate HIF-1α stability and ubiquitination independently of pVHL. Further studies are needed to elucidate the mechanism by which ORF34 regulates HIF-1α stability.

Other KHSV-encoded proteins have been found to physically interact with HIF-1α. The KSHV latency-associated nuclear antigen 1 (LANA1) was shown to physically interact with HIF-1α protein (13). Apparently, LANA1 stabilizes HIF-1α, causing enhanced RTA expression in PEL cell lines expressing LANA1. The KSHV viral interferon (IFN) regulatory factor 3 (vIRF3), known as latency-associated nuclear antigen 2 (LANA2), also was shown to interact with HIF-1α protein. In this instance, this interaction increased HIF-1α stability and HIF-1α-dependent transcriptional activity (17). Furthermore, expression of vIRF3 led to increased levels of VEGF expression and facilitated endothelial tube formation (17). Finally, the ORF74-encoded lytic protein, also known as G protein-coupled receptor (vGPCR), was shown to upregulate the expression of VEGF by stimulating the activity of HIF-1α, resulting in the activation of p38 mitogen-activated protein kinase (18). Thus, both latent and lytic genes of KSHV interact with and stabilize HIF-1α, enhancing HIF-1α-dependent transcriptional activation. HIF-1α stability has been shown to be negatively regulated via a few cellular genes, including HSP70, CHIP, and spermidine/spermine N¹-acetyltransferase-1 (SSAT-1) (2, 52). In all of these cases, these proteins bind to HIF-1α, causing increased HIF-1α ubiquitination and degradation (2, 52).

The results described here show, for the first time, that HIF-1α can be destabilized by the KSHV ORF34 protein. Collectively, these results suggest that KSHV possesses multiple mechanisms for tightly controlling HIF-1α transcriptional activity. Hypoxia can induce lytic replication of KSHV in PEL cells. Apparently, the KSHV latent protein LANA interacts with HIF-1α to upregulate RTA expression and KSHV lytic replication during hypoxia (13). Therefore, it is possible that ORF34 competes with LANA for HIF-1α binding to inhibit KSHV lytic replication. This intricate HIF-1α regulation may be crucial to control the lytic versus latent status of KSHV within tumors.