Kaposi's sarcoma-associated herpesvirus latency-associated nuclear antigen induction by hypoxia and hypoxia-inducible factors

Abstract

Hypoxia and hypoxia-inducible factors (HIFs) play an important role in the Kaposi's sarcoma-associated herpesvirus (KSHV) life cycle. In particular, hypoxia can activate lytic replication of KSHV and specific lytic genes, including the replication and transcription activator (RTA), while KSHV infection in turn can increase the levels and activity of HIFs. In the present study, we show that hypoxia increases the levels of mRNAs encoding KSHV latency-associated nuclear antigen (LANA) in primary effusion lymphoma (PEL) cell lines and also increases the levels of LANA protein. Luciferase reporter assays in Hep3B cells revealed a moderate activation of the LANA promoter region by hypoxia as well as by cotransfection with degradation-resistant HIF-1α or HIF-2α expression plasmids. Computer analysis of a 1.2-kb sequence upstream of the LANA translational start site identified six potential hypoxia-responsive elements (HRE). Sequential deletion studies revealed that much of this activity was mediated by one of these HREs (HRE 4R) oriented in the 3' to 5' direction and located between the constitutive (LTc) and RTA-inducible (LTi) mRNA start sites. Site-directed mutation of this HRE substantially reduced the response to both HIF-1α and HIF-2α in a luciferase reporter assay. Electrophoretic mobility shift assays (EMSA) and chromatin immunoprecipitation (ChIP) assays demonstrated binding of both HIF-1α and HIF-2α to this region. Also, HIF-1α was found to associate with RTA, and HIFs enhanced the activation of LTi by RTA. These results provide evidence that hypoxia and HIFs upregulate both latent and lytic KSHV replication and play a central role in the life cycle of this virus.

Fig 1

Hypoxia increases ORF73 (LANA) promoter activity. (A) Schematic diagram of the genomic organization of the region spanning ORFK12 (Kaposin) through K14 (v-OX2) in the KSHV genome; this region includes ORF71 through ORF73 (LANA) as well as the KSHV miRNA cluster. The numbers above the closed arrows correspond to positions of initiation/termination codons of the ORFs. Two arrows between K14 (v-OX2) and LANA denote the LANA constitutive (LTc) and RTA-inducible (LTi) mRNA start sites. The location of the probes used in Northern blots for ORF73 and K12 are shown as lines above the respective genes. Shown below is an expanded diagram of LANA and the LANA promoter region surrounding LTc and LTi, with the nucleotide positions for the mRNA start sites indicated. The LTc mRNA start site has been alternatively mapped to nucleotide positions 127900 and 127948 (10, 38). The LANA promoter region contains six potential hypoxia response elements (HREs), shown as rectangular boxes labeled 1 through 6. R denotes HREs in the reverse orientation to the transcription of LANA. (B) Schematic diagram of luciferase reporter constructs of the LANA promoter region. pGL3-LANA(c&i)p-luc [LANA(c&i)p] contains a 1,201-bp DNA segment that includes LTc and LTi and the six potential HREs. pGL3-LANA(c)p-luc [LANA(c)p] contains a 795-bp DNA segment that includes the LTc start site and four potential HREs (1, 2, 3R, and 4R). pGL3-LANA(i)p-luc [LANA(i)p] contains a 570-bp DNA segment that includes the LTi promoter start site but not the LTc start site and only three of the potential HREs (4R, 5R, and 6). (C) Hypoxia and CoCl₂ treatments induce ORF73 (LANA) promoter activity in Hep3B cells. A fixed amount (700 ng) of reporter plasmid pGL3-LANA(c&i)p-luc [LANA(c&i)p] or a pGL3-basic plasmid (control) was transfected into Hep3B cells cultured in triplicate wells of 12-well culture plates. At 24 h of posttransfection, cells were exposed to hypoxia or treated with CoCl₂ for 16 h at 37°C. Cells were harvested, and the cell lysate was analyzed for luciferase activity. Mean light units were normalized to the total cellular protein estimated by the Bradford method. The fold activation of promoter activity was determined based on the ratio between the normalized mean light units and that of the pGL3-basic control construct cultured under normoxia. All values represent the means of results from three experiments, each done in triplicate. Error bars represent the standard deviations. Similar results were obtained in initial experiments when luciferase activity was normalized using a Renilla control (results not shown).

Fig 2

Hypoxia increases the production of LANA RNA and LANA protein levels. (A) Diagram of the mRNA transcripts produced from LTc and LTi as previously reported (10, 31, 33, 38, 44, 45). An unspliced transcript of 5.8 kb encoding ORFs 71 to 73 is produced from the LTc start site. Alternative splicing (as shown by dotted lines) generates mRNA transcripts of 5.4 kb and 1.7 kb for ORFs 71 to 73 and ORFs 71 to 72, respectively. A 5.5-kb transcript produced from the LTi start site also encodes ORFs 71 to 73 (31). (B and C) Northern blot analysis showing increased ORF73 (LANA) mRNA in BC-3 and JSC-1 cells exposed to hypoxia (1% O₂). Cells were cultured in normoxia (N) or hypoxia (H) or in the presence of 20 ng/ml TPA (T). After 22 h, RNA was extracted and 5 μg of RNA was probed using probes specific for ORF73 (B) or K12 (C). LANA mRNAs are upregulated in hypoxia compared to normoxia but not by TPA treatment. K12 is upregulated substantially by exposure to TPA but only minimally or not at all by hypoxia. (D) Western blot showing protein levels of LANA in BC-3 and JSC-1 cells cultured in 1% O₂ (H) for up to 72 h in hypoxia compared to those cultured in normoxia (N). After culture in hypoxia or normoxia, cells were lysed and 20 μg of protein was run on two gels. One gel (upper) was probed with antibody specific for LANA, HIF-1α, and actin, and the other (lower) was probed with antibody to HIF-2α and actin. The gels were stripped between successive probes. HIF-1α and HIF-2α levels were stabilized up to 48 h in hypoxia and then decreased in both BC-3 and JSC-1 cells. ImageJ analysis, shown in the right panel, showed LANA levels increased approximately 2- to 3-fold in BC-3 and 2.3-fold in JSC-1 cells. (E) vIRF3 levels in BC-3 and JSC-1 cells exposed to hypoxia for 24 h and through 72 h. Its levels remained essentially unchanged in hypoxia.

Fig 3

The LANA promoter is activated by HIF-2α and HIF-1α. The upper panel shows activation of the LANA promoter containing both LTc and LTi mRNA start sites [LANA(c&i)p], only the LTc mRNA start site [LANA(c)p], or only the LTi mRNA start site [LANA(i)p] by degradation-resistant HIF-1α and HIF-2α (HIF-1αm or HIF-2αm). Hep3B cells were cotransfected with pGL3-LANA(c&i)p-luc, pGL3-LANA (c)p-luc, or pGL3-LANA(i)p-luc (300 ng) and plasmids encoding degradation-resistant mutants of HIF-1α or HIF-2α or a pcDNA control (400 ng). At 48 h of posttransfection, cells were lysed and promoter activity was determined as described in the legend to Fig. 1. Fold activation for all three promoters was calculated by normalizing to the pcDNA (400 ng) control for LANA(i)p, and results are shown as normalized to the respective pcDNA control. In the lower left panel, VEGF promoter activity was assessed in a similar manner. All values represent means of triplicate determinations in one representative experiment out of two. Error bars represent the standard deviations. In the lower right panel are shown Western blots of the cells transfected with pGL3-LANA(c&i)p-luc and plasmids encoding degradation-resistant mutants of HIF-1α or HIF-2α or a pcDNA control (400 ng). The top half of each blot was probed with either anti-HIF-1α or anti-HIF2α, while the lower half was probed with anti-β actin as described in Materials and Methods.

Fig 4

HRE4R can mediate much of the activation of the LTi LANA promoter by HIFs. (A) Schematic diagram of the LTi promoter luciferase reporter constructs with progressive deletions of HRE4R [pGL3-LANA(i)p-luc(D1)] and 5R [pGL3-LANA(i)p-luc(D2)]. The wild-type LTi promoter luciferase reporter construct is shown in Fig. 1B. (B) Deletion of HRE4R but not HRE5R substantially decreases LTi LANA promoter induced by HIF-1α and HIF-2α. Hep3B cells were cotransfected with 300 ng each of wild-type pGL3-LANA(i)p-luc, pGL3-LANA(i)p-luc(D1), and pGL3-LANA(i)p-luc(D2) constructs, respectively, and degradation-resistant mutants of HIF-1α or HIF-2α or pcDNA control plasmids (400 ng). At 48 h of posttransfection, cells were lysed and promoter activity was assessed as described in the legend to Fig. 1. Fold activation was calculated by normalizing to the pcDNA (400 ng) control results of LANA(i)p. All values represent means of triplicate determinations of a representative experiment out of two. Error bars represent the standard deviations. (C) Schematic diagram of the pGL3-LANA(i)p-luc(HRE4Rm) construct. This construct was created by mutating the HRE4R sequence in the 569-bp pGL3-LANA(i)p-luc construct by site-directed mutagenesis. The sequence shown is for the DNA strand encoding LANA; the HRE is in the reverse orientation with sequence GACGTG. Mutated nucleotides are represented in bold. (D) Mutation of HRE4R attenuates LTi LANA promoter activity induced by HIF-2α and HIF-1α. Hep3B cells were cotransfected with 300 ng each of wild-type pGL3-LANA(i)p-luc or the pGL3-LANA(i)p-luc(HRE4Rm) constructs and degradation-resistant mutants of HIF-1α or HIF-2α or pcDNA control plasmids (400 ng). At 48 h of posttransfection, cells were lysed and promoter activity was assessed as described in the legend to Fig. 1. Fold activation was calculated by normalizing to LANA(i)p with pcDNA (400 ng) control. All values represent means of triplicate determinations of a representative experiment out of two. Error bars represent the standard deviations.

Fig 5

HIFs bind to the LANA promoter. (A) HIFs bind to the HRE4R in vitro as detected by electrophoretic mobility shift assay. DIG-labeled synthetic oligonucleotide containing wild-type (WT) HRE4R was incubated with nuclear extracts (NE) from Hep3B cells cultured in 10-cm dishes in normoxia (N) or in hypoxia (H) for 16 h or transfected with 5 μg degradation-resistant mutants of HIF-1α or HIF-2α for 48 h. The mixture was then analyzed on a 6% DNA retardation gel. Where indicated, unlabeled wild-type (WT) or mutant (Mut) oligonucleotides (at concentrations 100× and 150× that of the labeled probe) were added to the reaction to assess binding competition (Comp.). Protein-DNA complexes were separated, blotted onto a nylon membrane, and probed with antidigoxigenin antibody conjugated to alkaline phosphatase. The sequence of the labeled WT probe and the WT and Mut competing oligonucleotides are shown at the bottom. The WT HRE4R sequence is underlined, and nucleotide changes in the Mut sequences are shown in bold. The positions of the DIG-labeled HIF complexes and free probe are indicated with arrows. (B) Diagrammatic representation of the LANA promoter region showing the location of forward and reverse primers used in the chromatin immunoprecipitation (ChIP) assay in the region surrounding LTi. Amplification of DNA flanking the primers produces an expected band of 450 bp. Sequences of forward (F) and reverse (R) primers are shown in the lower panel. (C) HIFs bind to HRE in the region of the LANA promoter surrounding LTi in vivo as detected by ChIP assay. Duplicate samples of 10⁷ BC-3 cells were treated with CoCl₂ (100 μM) for 16 h at 37°C. Cells were cross-linked with 37% formaldehyde (1% final concentration), and unreacted formaldehyde was quenched using 2 M glycine. The nuclear extracts were sonicated (Misonix 3000; 20 s pulses for 6 min at 4°C) to obtain DNA with an average length of 200 to 1,000 bp. Sonicated lysate was centrifuged at 13,000 × g for 10 min at 4°C. Supernatant containing DNA-protein complexes was collected, and protein estimation was done by the BCA method. DNA-protein complexes in a 55-μl volume (300 μg) were diluted to 10× using ChIP dilution buffer (Upstate, California), and 20 μl (approximately 10 μg) was set aside for use as an input. The remaining DNA-protein complexes were immunoprecipitated with 8 μg of monoclonal mouse antibody to HIF-1α (lanes 3 and 4) or polyclonal rabbit antibody to HIF-2α (lanes 5 and 6) and respective control rabbit (Rb) antibody (lane 7) or mouse (Ms) (lane 8) IgG. After de-cross-linking, the DNA was purified and amplified using primers flanking the LTi LANA promoter region as diagrammed in Fig. 5B. The more-intense 450-bp bands of HIF-1α and HIF-2α immunoprecipitated samples compared to that of their IgG controls suggest the binding of HIFs to the LTi LANA promoter region. (D) HIF-1α binding to the HRE4R as detected by in vitro binding and competition assay. Nuclear extracts were prepared from Hep3B cells cultured in normoxia or treated with CoCl₂ (100 μM) for 18 h at 37°C. A total of 5 μg of nuclear extract was incubated with 20 pmol (1×), 100 pmol (5×), or 200 pmol (10×) of a wild-type (WT) 26-bp probe from the EPO gene promoter encompassing an HRE or a mutant (Mut) probe with a mutated HRE sequence. At the same time, parallel extracts were incubated with equivalent amounts of a 30-bp LANA(i)p HRE 4R synthetic oligonucleotide probe from the LANA promoter region encompassing wild-type (WT) HRE4R or a similar probe with a mutated HRE4 (Mut). The binding of HIF-1α in the Hep3B nuclear extracts to EPO HRE bound to a well in the microtiter plate was then assessed as described in Materials and Methods; results shown are the absorbance at 450 nm. All values represent means of duplicate determinations in a representative experiment out of two.

Fig 6

Hypoxia or CoCl₂ treatment enhances the RTA-mediated inducible LANA promoter activity (A) Schematic representation of the RTA-inducible LANA promoter region showing the presence of RTA response elements (RRE) in relation to the putative HRE sequences. RRE are shown as rectangular boxes filled with black color. (B) Hypoxia or CoCl₂ enhance RTA-inducible LANA promoter activity. Hep3B cells were transfected with 400 ng empty vector or an RTA expression plasmid along with a fixed amount (300 ng) of pGL3-LANA(i)p-luc. At 24 h posttransfection, cells were exposed to hypoxia or treated with CoCl₂ (100 μM) for 16 h. At the end of incubation period, cells were lysed and promoter activity was determined as described in the legend to Fig. 1. All results are normalized to the empty vector control in normoxia. CoCl₂ or hypoxia alone induced the inducible LANA promoter by approximately 1.6-fold and 2-fold, respectively. All values represent means of triplicate determinations of a representative experiment out of two. Error bars represent the standard deviation. (C) HIFs cooperate with RTA to increase the LTi LANA promoter activity. Hep3B cells were cotransfected with 300 ng of pGL3-LANA(i)p-luc along with 200 ng of degradation-resistant HIF-1α or HIF-2α and/or 200 ng RTA expression plasmid as shown. The total amount of DNA was normalized where appropriate using the pcDNA control (200 or 400 ng). At 48 h posttransfection in normoxia, cells were lysed and promoter assay was assessed as described in the legend to Fig. 1. Fold activation was calculated by normalization to the pcDNA control. All values represent means of triplicate determinations. Error bars represent the standard deviations.

Fig 7

RTA interacts with HIF-1α as detected by immunoprecipitation assay. BCBL-1 cells were exposed to hypoxia for 48 h at 37°C. At the end of the incubation period, cells were lysed and nuclear extracts were prepared. Fifty micrograms of nuclear extracts were immunoprecipitated using anti-HIF-1α mouse monoclonal antibody and respective control mouse monoclonal antibody (Ms). The immunoprecipitated material was then subjected to SDS-PAGE and Western blot analysis performed using anti-RTA rabbit serum. Total lysate (40%) was also probed with anti-RTA rabbit serum as a control. A parallel blot was also probed using anti-HIF-1α mouse monoclonal antibody. Results suggest that RTA associates with HIF-1α in vivo in BCBL-1 cells in hypoxia.

Fig 8

Schematic model showing some of the key interrelationships between HIF and LANA. A plus sign indicates upregulation, with greater activation indicated by a larger size. LANA mRNA is constitutively produced and also induced by RTA (19, 25, 31, 44). HIF is normally degraded under conditions of normoxia, but when cells are exposed to hypoxia or certain other conditions, this process is inhibited and levels of HIF increase. As shown in the present study, HIF can activate production of LANA and also enhances LANA production induced by RTA. Also, as previously shown, KSHV latent infection and LANA act to increase the levels of HIF (3, 5, 14, 31). This provides a positive feedback loop that can lead to sustained levels of HIF and LANA in KSHV-infected cells.