|Home | About | Journals | Submit | Contact Us | Français|
IFI16 (interferon gamma-inducible protein 16) recognizes nuclear episomal herpesvirus (Kaposi's sarcoma-associated herpesvirus [KSHV], Epstein-Barr virus [EBV], and herpes simplex virus 1 [HSV-1]) genomes and induces the inflammasome and interferon beta responses. It also acts as a lytic replication restriction factor and inhibits viral DNA replication (human cytomegalovirus [HCMV] and human papillomavirus [HPV]) and transcription (HSV-1, HCMV, and HPV) through epigenetic modifications of the viral genomes. To date, the role of IFI16 in the biology of latent viruses is not known. Here, we demonstrate that knockdown of IFI16 in the latently KSHV-infected B-lymphoma BCBL-1 and BC-3 cell lines results in lytic reactivation and increases in levels of KSHV lytic transcripts, proteins, and viral genome replication. Similar results were also observed during KSHV lytic cycle induction in TREX-BCBL-1 cells with the doxycycline-inducible lytic cycle switch replication and transcription activator (RTA) gene. Overexpression of IFI16 reduced lytic gene induction by the chemical agent 12-O-tetradecoylphorbol-13-acetate (TPA). IFI16 protein levels were significantly reduced or absent in TPA- or doxycycline-induced cells expressing lytic KSHV proteins. IFI16 is polyubiquitinated and degraded via the proteasomal pathway. The degradation of IFI16 was absent in phosphonoacetic acid-treated cells, which blocks KSHV DNA replication and, consequently, late lytic gene expression. Chromatin immunoprecipitation assays of BCBL-1 and BC-3 cells demonstrated that IFI16 binds to KSHV gene promoters. Uninfected epithelial SLK and osteosarcoma U2OS cells transfected with KSHV luciferase promoter constructs confirmed that IFI16 functions as a transcriptional repressor. These results reveal that KSHV utilizes the innate immune nuclear DNA sensor IFI16 to maintain its latency and repression of lytic transcripts, and a late lytic KSHV gene product(s) targets IFI16 for degradation during lytic reactivation.
IMPORTANCE Like all herpesviruses, latency is an integral part of the life cycle of Kaposi's sarcoma-associated herpesvirus (KSHV), an etiological agent for many human cancers. Herpesviruses utilize viral and host factors to successfully evade the host immune system to maintain latency. Reactivation is a complex event where the latent episomal viral genome springs back to active transcription of lytic cycle genes. Our studies reveal that KSHV has evolved to utilize the innate immune sensor IFI16 to keep lytic cycle transcription in dormancy. We demonstrate that IFI16 binds to the lytic gene promoter, acts as a transcriptional repressor, and thereby helps to maintain latency. We also discovered that during the late stage of lytic replication, KSHV selectively degrades IFI16, thus relieving transcriptional repression. This is the first report to demonstrate the role of IFI16 in latency maintenance of a herpesvirus, and further understanding will lead to the development of strategies to eliminate latent infection.
The human gammaherpesvirus Kaposi's sarcoma (KS)-associated herpesvirus (KSHV), also referred to as human herpesvirus 8 (HHV-8), is an oncogenic virus etiologically associated with KS, primary effusion B-cell lymphoma (PEL) or body cavity B-cell lymphoma (BCBL), and plasmablastic multicentric Castleman's disease (pMCD) (1,–3). In vivo, KSHV DNA and transcripts have been identified in human B cells, endothelial cells, epithelial cells, macrophages, and keratinocytes (3, 4). Similar to other herpesviruses, KSHV undergoes two distinguishable phases in its life cycle: latent and lytic infection (3, 4). During primary infection, KSHV initially establishes latent infection in specific target cells, during which multiple copies of the viral genome are stably maintained as extrachromosomal episomes (3, 5). Only a few viral genes, primarily localized in the major latency locus of the genome, are expressed during this phase. These gene products bestow numerous essential functionalities to the dormant viral genome, such as evading host immune surveillance (6), promoting cellular proliferation (7,–9), maintaining the viral episome (10), and tightly suppressing viral lytic gene expression (11). However, both spontaneous reactivation and induced reactivation of the latent genome occur, resulting in the systematic expression of the full repertoire of viral genes, which leads to viral genome replication and virion production (4). Previous studies suggested that lytic reactivation of the latent genome is one of the major contributors to KS, PEL, and MCD pathogenesis with a strong relationship to the progression and prognosis of these diseases (12,–15). Thus, a better understanding of all the factors, both host and viral, regulating this latent-to-lytic transition is necessary to develop strategies to control KSHV infection and associated diseases.
Established PEL-derived B-cell lines such as BCBL-1 and BC-3 are well-validated models for the study of KSHV biology. The nucleus of these cells have >80 copies of the episomal KSHV genome during latency, and the lytic replication cycle can be induced by treatment with chemical inducers such as 12-O-tetradecoylphorbol-13-acetate (TPA), sodium butyrate, or the aminoglycoside antibiotic neomycin (16,–19). Another well-established method to induce latent B cells is overexpression of the KSHV open reading frame 50 (ORF50) gene product RTA, which functions as a crucial latent-to-lytic switch protein (20). The control of the latent and lytic phases of herpesviruses is a complex interplay and orchestration between viral and host factors. Host factors such as STAT3 (21), KAP1/TRIM28 (22), retinoblastoma (Rb) (23), and the cellular peptidyl-prolyl cis/trans isomerase Pin1 (24) were recently found to play very important roles in this latent-to-lytic life cycle switch of herpesviruses.
Interferon gamma (IFN-γ)-inducible protein 16 (IFI16) is a member of the HIN-200 (hematopoietic interferon-inducible nuclear antigens with 200-amino-acid repeats) family of proteins, which in humans includes AIM2 (absent in melanoma 2), MNDA (myeloid cell nuclear differentiation antigen), and IFIX. IFI16 is a multifunctional DNA binding protein and has been implicated in various cellular functions such as transcriptional regulation, apoptosis, autoimmunity, and cell cycle regulation (25,–27). Studies by us and others have reported the role of IFI16 as a DNA sensor that detects nuclear replicating herpesviral genomes such as KSHV, herpes simplex virus 1 (HSV-1), Epstein-Barr virus (EBV), and bovine herpesvirus 1 (BoHV-1), leading to IFI16–apoptosis-associated speck-like protein containing a CARD (ASC)–procaspase-1 inflammasome formation that results in the production of the inflammatory cytokine interleukin 1β (IL-1β) (28,–33). We have also shown that IFI16-mediated inflammasomes are activated during prolonged KSHV latency in endothelial and B cells, leading to a constitutive state of IL-1β activation (34). Recently, IFI16 was also shown to be involved in the induction of IFN-β during KSHV and HSV-1 de novo infection of target cells via the IFI16–stimulator of interferon genes protein (STING)–TANK-binding kinase 1 (TBK)–interferon regulatory factor 3 (IRF3) axis (31, 32, 35, 36).
Apart from its role in immune surveillance, IFI16 has also been shown to function as a viral restriction factor against DNA viruses. Viral restriction factors are constitutively expressed intrinsic host defense mechanisms that provide frontline protection from invading viral pathogens. Gariano et al. demonstrated that IFI16 restricts human cytomegalovirus (HCMV) replication by displacing the Sp1 transcription factor on viral gene promoters (37). Orzalli et al. found that IFI16 restricts HSV-1 immediate early (IE) protein ICP0-null virus replication and IE gene expression by heterochromatinization of the viral genome (36). Our recent studies have shown that IFI16 mediates the restriction of HSV-1 replication in part by binding to the HSV-1 transcription start sites of IE, early (E), and late (L) genes, thereby preventing the recruitment of essential transcription factors such as RNA polymerase II (Pol II), TATA binding protein (TBP), and Oct1 (38). We also found that knockdown (KD) of IFI16 increased the HSV-1 yield by 6-fold, whereas overexpression of IFI16 reduced the yield by over 5-fold. Using a Cas9-mediated IFI16 knockout, we have demonstrated that the absence of IFI16 results in increased euchromatinization of the wild-type HSV-1 genome promoters (39). Subsequently, Lo Cigno et al. demonstrated that IFI16 also restricts human papillomavirus 18 (HPV18) through epigenetic modification of the viral promoters (29). Silencing of endogenous IFI16 leads to increased HPV loads, whereas overexpression of IFI16 severely impaired HPV18 replication and transcription. Moreover, the HSV-1 IE ICP0 protein, which also functions as an ubiquitin ligase, targeted host IFI16 for degradation by as early as 6 to 8 h postinfection (p.i.) (39,–41). This removal of IFI16 is envisioned as an essential criterion for a productive HSV-1 replication cycle.
Although IFI16 has been established to function as a restriction factor of HCMV and HSV-1 lytic replication, neither of these viruses establishes successful latency in the in vitro cells used in these studies, and there are no permanent cell line models that carry these viruses in a latent state. From these observations, we asked the question, “What is the potential role of IFI16 in the life cycle of KSHV that establishes latent infection during de novo infection and maintains its latent infection in the B-lymphoma cells of PEL?” We previously observed that IFI16 is associated with chromatinized latent KSHV and EBV genomes (31, 34). However, latent gene expression continues in the presence of IFI16, and viral latency is successfully maintained. We thus hypothesized that KSHV may have evolved with the host-intrinsic restriction factor IFI16 to facilitate the establishment and maintenance of its latency.
Here, we demonstrate for the first time that IFI16 plays an important role in the maintenance of KSHV latency. Our results show that silencing of IFI16 in the latently KSHV-infected human B-cell lymphoma BCBL-1 and BC-3 cell lines results in lytic reactivation of the latent genome, resulting in increases in the levels of all classes of KSHV lytic transcripts and proteins, followed by increased viral genome replication. Consistent with this, we found that the overexpression of IFI16 reduced susceptibility to lytic reactivation by agents such as TPA. Subsequently, when we probed IFI16 levels after lytic reactivation by a variety of stimuli, we discovered that the IFI16 protein is specifically degraded during lytic reactivation of latently infected PEL cells. IFI16 is polyubiquitinated and degraded via the proteasomal degradation pathway during the late-lytic phase of reactivation. Chromatin immunoprecipitation (ChIP) assays and luciferase promoter activity assays clearly demonstrate that IFI16 functions as a transcriptional repressor and represses KSHV lytic transcripts during prolonged latency. Together, our observations demonstrate for the first time that KSHV has evolved mechanisms to utilize the multifunctional innate DNA sensor IFI16 for maintaining its latency and associated gene regulations.
The KSHV-positive PEL cell lines BCBL-1 and BC-3 and the KSHV- and EBV-negative B-lymphoma cell lines BJAB and Akata were cultured in RPMI 1640 GlutaMAX (Gibco Life Technologies, Grand Island, NY) supplemented with 10% (vol/vol) fetal bovine serum (FBS; HyClone, Logan, UT) and penicillin-streptomycin (Gibco). For TREX-BCBL-1-RTA cells (a gift from J. Jung, University of Southern California) (20), in addition to the above-described medium, the eukaryotic selection factor hygromycin B (200 μg/ml) was added. The human Kaposi's sarcoma lesion-derived SLK cell line and the human osteosarcoma-derived cell lines U2OS and U2OS clone 67 (39) were cultured in Dulbecco's modified Eagle medium supplemented with 10% (vol/vol) FBS, penicillin-streptomycin, and 2 mM l-glutamine (Gibco). SLK cells, originally from Jay A. Levy and Sophie Leventon-Kriss, were obtained through the AIDS Reagent Program, Division of AIDS, NIAID, NIH (42, 43). It has been suggested that these cells were contaminated with the clear cell renal cell carcinoma cell line Caki-1 during the establishment of the cell line (44). Human dermal microvascular endothelial cells (HMVEC-d-CC-2543; Lonza, Walkersville, MD) were cultured in endothelial growth basal medium 2 (EBM-2) supplemented with an EGM-2MV BulletKit (Lonza). All cells were regularly tested and confirmed to be mycoplasma negative by using the Mycoalert kit (Lonza).
Induction of KSHV lytic cycle was achieved with either TPA (20 ng/ml) or neomycin (200 μM) (17,–19), as indicated. TREX-BCBL-1-RTA cells were induced with doxycycline (DOX) at a final concentration of 1 μg/ml. KSHV virions were produced and purified from BCBL-1 cells according to methods described previously by us (45,–47). The KSHV genome copy number was quantitated by real-time DNA PCR using primers specific for the KSHV ORF73 gene as described previously (45,–47). For de novo KSHV infection, HMVEC-d cells were washed twice with phosphate-buffered saline (PBS), infected with the indicated numbers of DNA copies of KSHV in serum-free basal medium for 2 h, washed twice with PBS, and incubated for the indicated times in complete medium.
Five TRC (the RNAi Consortium) lentiviral human IFI16 short hairpin RNA (shRNA) constructs (Dharmacon; GE Healthcare, Lafayette, CO) (clones TRCN0000019079, TRCN0000019080, TRCN0000019081, TRCN0000019082, and TRCN0000019083) were tested for their ability to efficiently knock down IFI16. Lentiviral particles were produced by using a four-vector system in HEK293T cells as described previously (48). HEK293T cells were transfected by using the CalPhos mammalian transfection kit (TaKaRa Clontech, Mountain View, CA) according to the manufacturer's instructions. The culture medium was changed 16 h after transfection, and the lentivirus-containing supernatant was collected 24 h later. This supernatant was filtered through a 45-μm filter and used to transduce B cells. Polybrene at a concentration of 5 μg/ml was used to facilitate B-cell infection. Of the five anti-IFI16 shRNAs tested, TRCN0000019082 knocked down endogenous IFI16 in BCBL-1 cells most efficiently and was thus selected for further use.
For lentivirus-mediated overexpression of IFI16, IFI16-expressing lentivirus was produced as described previously (39).
Mouse and rabbit antibodies (Abs) against human IFI16 were obtained from Santa Cruz Biotechnology, Inc. (Sana Cruz, CA), and Sigma (St. Louis, MO), respectively. Rabbit anti-AIM2 antibody was obtained from Abcam, Inc. (Cambridge, MA). Anti-KSHV RTA was obtained from Abbiotec, LLC (San Diego, CA). Antiubiquitin antibody (P4D1) was obtained from Santa Cruz Biotechnology, Inc. Mouse monoclonal anti-KSHV gpK8.1A (4A4) (49) and anti-KSHV ORF59 (50) antibodies were generated in our laboratory. Mouse monoclonal antibodies against β-actin and β-tubulin were obtained from Sigma. The fluorescent secondary antibody Alexa 594 or 488 was obtained from Molecular Probes, Invitrogen (Carlsbad, CA). Horseradish peroxidase-conjugated secondary antibodies were obtained from KPL, Inc. (Gaithersburg, MD). TPA, MG132, PAA (phosphonoacetic acid), doxycycline hydrochloride, neomycin solution, and CHX (cycloheximide) were obtained from Millipore-Sigma (Billerica, MA).
To prepare whole-cell lysates, cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris [pH 8.0], 150 mM sodium chloride, 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with a protease inhibitor cocktail (Sigma) for 30 min on ice. The lysates were sonicated on ice and clarified by centrifugation at 13,000 × g for 10 min at 4°C. Levels of extracted proteins were estimated by using the Pierce bicinchoninic acid (BCA) protein assay kit (Thermo Fisher, Waltham, MA), and equal amounts of protein were resolved on SDS-PAGE gels, blotted onto a nitrocellulose membrane, and probed with primary antibodies as indicated. For detection, horseradish peroxidase-conjugated secondary antibodies were used, and immunoreactive bands were developed by using the SuperSignal West Pico chemiluminescent substrate (Thermo Fisher). Blots were photographed and quantitated by using FluorChemFC2 software and an AlphaImager system (Alpha Innotech Corporation, San Leonardo, CA).
Suspension cells were spotted onto 10-chamber glass slides and air dried. Cells were fixed and permeabilized with ice-cold acetone, blocked with the Image-iT FX signal enhancer (Thermo Fisher), and incubated with the indicated primary antibodies followed by the corresponding secondary antibodies conjugated with Alexa Fluor (Thermo Fisher). Slides were mounted with SlowFade Gold Antifade Mountant with 4′,6-diamidino-2-phenylindole (DAPI) (Thermo Fisher) and imaged with a Nikon Eclipse 80i microscope. Image analysis and deconvolution were performed with MetaMorph digital imaging software (Molecular Devices, Silicon Valley, CA). Imaging was performed at a ×40 magnification unless mentioned otherwise. Figures presented are representative of results from two or more independent experiments where at least five different fields were imaged.
The IFI16-overexpressing plasmid IFI16-FL (pcDNA3-FLAG) was a gift from Cheryl Arrowsmith (Addgene plasmid 35064) (51). Cells were transfected by using Lipofectamine LTX with Plus reagent (Thermo Fisher) according to the manufacturer's instructions. For the luciferase promoter assays, cells were transfected with the indicated reporter constructs with full-length KSHV promoters preceding the firefly luciferase gene. The following KSHV promoter constructs were used: ORF73 and K8/K-bZip promoters (pGL3.6 and K8 IE Luc, gifts from Yuan Chang, University of Pittsburgh) (52), an ORF50 promoter (pcDNA2500, a gift from George Miller, Yale University School of Medicine) (53), a K14/viral G protein-coupled receptor (vGCR) promoter (pDD163, a gift from Dirk Dittmer, University of North Carolina at Chapel Hill) (54), an ORF57 promoter (p57P-luc, a gift from Charles Wood, University of Nebraska) (55), and a K2/viral interleukin-6 (vIL-6) promoter (pK2p 3.2k, a gift from Ren Sun, University of California at Los Angeles) (56). The following human promoter element constructs were used: pAP1 Luc (catalog no. 631906; Clontech), pGL2-MDM2-Luc, an NFAT luciferase reporter (a gift from Toren Finkel) (Addgene plasmid 10959), and 5× NF-κB Luc (Stratagene, La Jolla, CA). To account for transfection variability and potential cytotoxicity, another construct expressing the Renilla luciferase gene under the control of a simian virus 40 (SV40) promoter was transfected into the cells. At 24 h posttransfection, luciferase assays were conducted with a dual-luciferase reporter assay system (Promega, Madison, WI) according to the manufacturer's instructions. Results were expressed as firefly/Renilla luciferase ratios and normalized to results under their respective control conditions.
Total cellular RNA was isolated by using an RNeasy minikit (Qiagen, Valencia, CA) according to the manufacturer's protocol. On-column DNase digestion was performed by using an RNase-free DNase set (Qiagen). The amount of extracted RNA was estimated by using a NanoDrop spectrophotometer (Thermo Scientific), and 1 μg RNA was reverse transcribed by using the High-Capacity cDNA reverse transcription kit (Thermo Fisher) with random primers, according to the manufacturer's instructions. For real-time quantitative reverse transcription-PCR (qRT-PCR), the synthesized cDNA was used as a template with Power SYBR green PCR master mix (Applied Biosystems) on an ABI Prism 7500 detection system (Applied Biosystems). All RNA levels were normalized to β-actin mRNA levels and calculated as the delta-delta threshold cycle (ΔΔCT). Primers used are listed in Table 1. All real-time experiments were performed in triplicate.
BCBL-1, BC-3, and BJAB cells, mock treated or treated with TPA for the indicated time intervals, were lysed in 2% SDS buffer (10 mM Tris-HCl [pH 7.5], 2 mM EDTA, 2% SDS, 150 mM NaCl, 10% glycerol, 1× protease inhibitor and phosphatase inhibitor cocktail) and boiled for 5 min, followed by sonication. Lysates were diluted 1:10 in dilution buffer (10 mM Tris-HCl [pH 7.5], 2 mM EDTA, 150 mM NaCl, 1% Triton X-100) and incubated under rotation at 4°C for 1 h. Subsequently, the lysates were clarified by centrifugation at 20,000 × g for 30 min. Equal amount of proteins were subjected to immunoprecipitation with anti-IFI16 Ab and washed with wash buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA, 1 M NaCl, 1% NP-40). Samples were boiled in SDS-PAGE sample buffer and blotted with an antiubiquitin antibody (P4D1; Santa Cruz Biotechnology).
mRNA sequencing (mRNA-Seq) was performed by using the QuantSeq 3′ mRNA-Seq library preparation kit (Lexogen GmbH, Austria) according to the manufacturer's instructions. Libraries were prepared at the DNA Services Facility at the University of Illinois at Chicago (UIC), and single-end, 100-base sequencing was performed on an Illumina HiSeq2500 instrument, employing TruSeq SBS version 4 chemistry, at the W. M. Keck Center for Comparative and Functional Genomics at the University of Illinois at Urbana-Champaign (UIUC).
Raw fastq data were quality trimmed to a minimum Phred quality score of 20 by using Trimmomatic (57). Reads were first filtered against the hg19 human reference genome by using bowtie2 (58) in the local mode to allow soft clipping of poly(A) tails. Unmapped reads were then mapped against the KSHV reference genome (GenBank accession no. NC_009333), using bowtie2 in the local mode. Gene expression counts for both human and KSHV annotations were obtained by using HTSeq (59), and differential and normalized expression levels were calculated by using edgeR. P values were corrected for multiple testing by using the false discovery rate (FDR) correction. Subsequently, differentially expressed KSHV genes were visualized on a heat map, and samples were clustered by using complete-linkage hierarchical clustering.
B-cell chromatin shearing for ChIP was performed by using the truChIP Chromatin Shearing kit (Covaris, Woburn, MA) according to the manufacturer's instructions. Chromatin was sheared in shearing buffer (truChIP Chromatin Shearing kit) by using a Covaris E220 focused ultrasonicator. After chromatin shearing, Triton X-100 and NaCl in the sheared lysate were adjusted to final concentrations of 1% and 150 mM, respectively. Shearing efficiencies were evaluated by using a 2100 Bioanalyzer instrument (Agilent Technologies, Santa Clara, CA) according to the manufacturer's instructions. The fragment size was ensured to be between 200 bp and 500 bp. For immunoprecipitation, 10 μg sheared chromatin was immunoprecipitated with 2 μg anti-IFI16 (Abcam, Cambridge, MA) or ChIP-grade control IgG (Abcam) overnight at 4°C. The chromatin-antibody complex was pulled down with ChIP-grade protein G magnetic beads (Active Motif) for 2 h at 4°C. The immunoprecipitated complex was then washed three times with low-salt and once with high-salt wash buffers (Cell Signaling Technology, Danvers, MA). To elute the chromatin, the beads were incubated in ChIP elution buffer (Cell Signaling) at 65°C for 30 min on a ThermoMixer (1,200 rpm). Following this step, the eluted chromatin was incubated with NaCl and proteinase K for 2 h at 65°C to remove all proteins and reverse the cross-linking. DNA was purified by using spin columns (Cell Signaling), and the ChIP-enriched DNAs were quantitated by real-time quantitative PCR (qPCR) using Power SYBR green PCR master mix (Applied Biosystems) and primers listed in Table 1. Enrichment of proteins on specific genomic regions was calculated as fold enrichment over control IgG ChIP.
Data are expressed as means ± standard deviations (SD) of results from at least three independent experiments (n ≥ 3), and statistical significance was calculated by using the two-tailed Student t test. A P value of <0.05 was considered significant.
To determine the potential role of IFI16 in KSHV biology, we knocked down IFI16 in the PEL cell line BCBL-1 carrying >80 copies of the episomal latent KSHV genome by transduction with a lentivirus containing shRNA for IFI16 (ShIFI16) or control shRNA (ShC). Immunoblotting showed that compared to ShC, ShIFI16 caused an ~60% reduction in the level of endogenous IFI16 by 48 h and an ~72% reduction by 96 h (Fig. 1A, top, lanes 1 to 4). In contrast, the IFI16-related innate sensor protein AIM2, which senses DNA in the cytoplasm, was not affected by ShIFI16, demonstrating the specificity and absence of off-target effects of IFI16 knockdown (Fig. 1A, middle, lanes 1 to 4). Real-time RT-PCR analysis also demonstrated an ~60% reduction in IFI16 transcript levels at the tested time points (Fig. 1B).
To determine the effect of IFI16 KD on KSHV gene expression, we performed a comprehensive high-throughput deep-sequence transcriptome analysis of mRNA collected 96 h after ShIFI16 transduction. As a control, we also induced BCBL-1 cells with TPA, a chemical inducer of the lytic cycle, for 96 h and performed similar mRNA transcriptome sequencing. Figure 1 shows a heat map representing data from three independent experiments for IFI16 KD and TPA induction. As expected, all of the KSHV genes were transcriptionally upregulated by TPA. Interestingly, IFI16 KD also resulted in a similar effect, and almost all of the lytic genes were upregulated (Fig. 1C). Data from three independent IFI16 KD experiments were compared to results for their respective control KDs (ShC), and the results are presented as relative KSHV gene expression ratios (ShIFI16/ShC) (Fig. 1D). Genes were classified as latent (La), immediate early (IE), early (E), and late (L) according to convention in the field (60), and fold changes above 1.5-fold were considered significant (Fig. 1D). The most prominent observation was that almost all of the KSHV lytic genes were transcriptionally upregulated upon IFI16 KD (Fig. 1Db to tod).d). In contrast, the KSHV latency-associated ORF73/LANA, ORF72/v-cyclin, ORF71/v-FLIP, and K12/kaposin families of protein-encoding genes (61) were not significantly affected, with only marginal upregulation (Fig. 1Da). While the ORF71, -72, and -73 genes were not affected by IFI16 KD, K12 was upregulated by about 3.5-fold compared to ShC. K12 transcripts were previously shown to be upregulated to a similar extent by TPA induction in BCBL-1 cells (62). Together, these data suggested that IFI16 KD in BCBL-1 cells induces lytic reactivation of the latent genome.
To confirm the sequencing results and to determine whether IFI16 KD also affects KSHV latency in other cells besides BCBL-1 cells, real-time RT-PCR was performed with RNA from BCBL-1 and BC-3 cells after 96 h of IFI16 KD. Genes from all four KSHV gene classes were studied, and the results correlated well with the results of our mRNA sequencing studies. Latent genes were not significantly affected, but all other lytic genes were induced by various degrees in both BCBL-1 and BC-3 cells (Fig. 2A and andC).C). KD of IFI16 was confirmed by real-time RT-PCR (Fig. 2B and andDD).
Next, we investigated whether the increased lytic transcript levels with IFI16 KD result in full-fledged lytic reactivation of the latent genome, a hallmark of which is increased replication of viral DNA. We thus determined the KSHV genomic DNA copy numbers in BCBL-1 cells after IFI16 KD or treatment with TPA. Induction by TPA resulted in an ~10-fold increase in intracellular viral genome copy numbers, and IFI16 KD resulted in an ~5-fold increase (Fig. 2E). Furthermore, to confirm that the newly synthesized DNA is packaged in the capsids to produce the virion particles, DNase treatment was performed prior to genomic DNA copy number determination. Only mature virion particles are detected in this assay, as naked, unpackaged DNA is sensitive to DNase treatment. Compared to ShC in transduced cells, we observed a significant 2.6-fold increase in virion particle-associated KSHV DNA copy numbers in IFI16 KD cells (Fig. 2F), confirming the production of KSHV virions in response to IFI16 KD.
Together, these observations suggested that IFI16 plays an important role in the maintenance of KSHV latency in PEL cell lines, and perturbation of IFI16 homeostasis disturbs latency, resulting in lytic cycle reactivation and increased lytic transcription, leading to replication of the viral genome and virion production.
To assess the effect of IFI16 KD on TPA-mediated KSHV lytic reactivation, BCBL-1 cells were transduced with a lentivirus containing ShIFI16 or a control lentivirus. Forty-eight hours later, cells were induced with TPA (20 ng/ml), and total RNA was isolated after 0, 3, 6, 9, 12, 24, 36, 48, 72, and 96 h, as indicated (Fig. 3A). KSHV transcripts representing the four gene classes were assessed by real-time RT-PCR. As expected, TPA treatment resulted in the induction of the IE ORF50 gene by as early as 6 h posttreatment. K5, representing E genes, was induced next at ~12 h posttreatment, while the L gene ORF25 was induced only after 36 h. As reported previously, the latent ORF73 gene was induced only marginally after TPA treatment (Fig. 3A, solid lines) (47). Interestingly, IFI16 depletion prior to TPA induction led to more robust and statistically significant increases in the levels of almost all lytic transcripts tested compared to those of ShC-treated cells (Fig. 3A, dashed lines). However, the ORF73 level increased only marginally, and most of the differences were not significant. Subsequently, we also assessed the increases in KSHV genomic DNA copy numbers representing replication of the viral genome at the above-mentioned time points. We observed that KSHV genome copy numbers started to increase after 36 h of TPA treatment (Fig. 3B), which was concurrent with the expression of the early and late genes. Similarly to the viral lytic transcripts, IFI16 KD resulted in significant increases in KSHV genome copy numbers over those of ShC-transduced cells (Fig. 3B). Figure 3C shows the KD efficiency of IFI16 transcripts at 0 h (first time point after 48 h of ShIFI16 transduction) and at 96 h (last time point after 48 h of ShIFI16 transduction) and IFI16 immunoblots. As shown, IFI16 expression was substantially knocked down after 48 h of shRNA treatment (Fig. 3C, compare lanes 1 and 2), which further decreased after 96 h of TPA induction (Fig. 3C, compare lanes 3 and 4).
These observations suggested that IFI16 continues to restrict lytic reactivation of latent KSHV in BCBL-1 cells during the early stages of reactivation after external lytic cycle-inducing stimuli, and KD of IFI16 relieves this restriction, resulting in higher levels of lytic transcripts and genome replication.
To further confirm the role of IFI16 in KSHV latency, we probed lytic protein expression by immunofluorescence analysis (IFA) and WB analysis after 4 days of IFI16 KD. TPA induction was used as a positive control for lytic reactivation, and two lytic proteins, the early lytic ORF59 and late lytic gpK8.1A proteins, were examined by IFA (Fig. 4A). We observed that compared to ShC-transduced cells, ShIFI16-transduced cells exhibited substantially elevated levels of both lytic proteins (Fig. 4A), which was similar to the results observed with TPA induction of the KSHV lytic cycle (Fig. 4A). Immunoblotting for the envelope-associated late gpK8.1A glycoprotein further confirmed the production of late lytic proteins after IFI16 KD (Fig. 4B, compare lane 4 to lane 3). TPA induction was used as a positive control for lytic reactivation (Fig. 4B, lanes 1 and 2).
Next, to determine the specificity of IFI16's role in KSHV latency, we overexpressed IFI16 through lentivirus-mediated transduction in BCBL-1 cells, and 24 h later, lytic reactivation was induced with TPA. Real-time RT-PCR demonstrated significantly higher levels, about 2-fold, of IFI16 transcripts in IFI16-overexpressing cells than in control lentivirus-transduced cells (Fig. 4C). Immunoblotting of IFI16 confirmed 1.9-fold overexpression of IFI16 (Fig. 4D). Real-time RT-PCR analyses of the lytic transcripts ORF50 (IE) and ORF38 (L) showed that compared to the control, the overexpression of IFI16 resulted in significant reductions in the levels of these lytic transcripts (Fig. 4C). These observations confirmed that IFI16 plays an important role in KSHV latency maintenance, and IFI16 is inhibitory to lytic transcript expression.
Our observations suggested that IFI16 is inhibitory toward lytic reactivation, and overexpression of IFI16 restricts TPA-mediated lytic reactivation and lytic gene expression. However, how TPA effectively induced lytic cycle reactivation in the presence of endogenous IFI16 was intriguing. Moreover, the additional decrease in the IFI16 protein level post-TPA treatment shown in Fig. 3C (compare lanes 2 to 4) intrigued us and prompted us to investigate the possibility of IFI16 degradation post-TPA induction. To investigate this, we probed the levels of IFI16 in BCBL-1 and BC-3 cells after TPA induction. Two different time intervals, 48 and 96 h, were selected, and as control cells, the KSHV- and EBV-negative B-cell lymphoma cell lines BJAB and Akata were included. Interestingly, we observed that as lytic reactivation progressed, the level of IFI16 was significantly reduced in BCBL-1 and BC-3 cells (Fig. 5A). By 96 h post-TPA treatment, we observed an ~40% reduction of IFI16 protein levels in both cell lines. As an indicator of lytic cycle progression, we probed for the lytic ORF50 (RTA) and gpK8.1A proteins, and the results confirmed lytic reactivation (Fig. 5A). In contrast, IFI16 levels remained unaltered by TPA treatment in control BJAB and Akata cells (Fig. 5B).
To determine whether this reduction in IFI16 protein levels is due to a loss of the IFI16 protein or a reduction and/or degradation of IFI16 mRNA, we performed real-time RT-PCR for IFI16 transcripts. The results showed that IFI16 mRNA levels did not decrease in response to TPA treatment (Fig. 5C). In fact, elevated levels of IFI16 mRNA were observed after TPA induction of BC-3 cells and, to a lesser extent, BCBL-1 cells. The expression of the lytic transcript ORF57 was monitored to confirm lytic reactivation (Fig. 5C).
Together with the results shown in Fig. 4B, these results suggested that IFI16 is probably degraded to facilitate lytic reactivation in latently infected PEL cell lines.
Chemical induction of latently infected PEL cells results in lytic induction in only about 20% of cells above the typical background levels of 0.5 to 3% spontaneous reactivation (63,–65). Therefore, we reasoned that in a population of TPA-treated latent cells, there will be some cells undergoing lytic reactivation, while others will remain uninduced (UI), and IFI16 levels in cells not lytically reactivated should remain unchanged. To gain a visual understanding of this possibility, an IFA for IFI16 was conducted with BJAB and BCBL-1 cells after 48 and 96 h of TPA induction. We used our monoclonal antibodies against the KSHV late lytic envelope glycoprotein gpK8.1A to identify cells undergoing lytic reactivation. Consistent with data from our previous report (34), IFI16 (Fig. 6A, green, yellow arrows) was observed predominately in the nucleus of BJAB cells. In contrast, due to its function in constitutive IFI16–ASC–procaspase-1 inflammasome activation in lately infected cells (34), IFI16 was distributed in both the nucleus and cytoplasm of BCBL-1 cells (Fig. 6B, yellow and red arrows, respectively). The expression of gpK8.1A (Fig. 6B, red) indicated cells undergoing lytic reactivation. As expected, only a fraction of the total BCBL-1 cells expressed gpK8.1A, demonstrating that not all cells respond to TPA induction. IFI16 fluorescence remained unchanged upon TPA treatment of BJAB cells. However, in BCBL-1 cells, we observed that after 48 h of TPA treatment, gpK8.1A-positive cells (Fig. 6B, white arrows) had significantly lower levels of IFI16 staining than did unreactivated cells. By 96 h, IFI16 was almost completely absent in reactivated cells and unchanged in unreactivated cells (Fig. 6B, enlarged image, white arrow). This suggested that IFI16 is selectively degraded only in cells that undergo KSHV lytic reactivation.
TPA is known to activate transcription by promiscuously targeting the AP-1 family of transcription factors and the protein kinase C pathway (66, 67). To demonstrate that the reduction of IFI16 levels observed in TPA-induced latently KSHV-infected cells is not an off-target effect of TPA, we used the engineered cell line TREX-BCBL-1-RTA, in which the KSHV lytic cycle can be reactivated by treatment with the tetracycline analog doxycycline (DOX). In all these cells, an epitope-tagged KSHV lytic cycle switch RTA protein cassette is expressed under the control of a tetracycline-inducible promoter (20). Upon DOX treatment, RTA is expressed in a uniform manner in all the cells, leading to the reactivation of KSHV in a higher percentage of cells than in the parental BCBL-1 cell line.
We performed IFA for IFI16 and gpK8.1A after DOX treatment of TREX-BCBL-1-RTA cells. In untreated cells, IFI16 was detected in both the cytoplasm and the nucleus (Fig. 7A, green and white arrows, respectively). The red arrows in Fig. 6 indicate gpK8.1A corresponding to cells undergoing lytic reactivation. We observed that almost all the cells underwent lytic reactivation by 48 h post-DOX treatment, and IFI16 staining in these cells was markedly reduced compared to that in untreated cells. We also observed that both cytosolic and nuclear IFI16 staining were decreased in response to lytic cycle induction. At 72 h post-DOX treatment, IFI16 staining was decreased further, indicating IFI16 degradation. To further validate these results, WB analysis for IFI16 was performed. A robust reduction in IFI16 levels was observed for both time intervals (Fig. 7B, lanes 1 to 4). The high degree of reduction of IFI16 levels in these cells compared to BCBL-1 cells can be attributed to the latter's inefficient lytic reactivation in response to TPA.
To further confirm that other modes of lytic KSHV reactivation also resulted in IFI16 degradation, we used neomycin to induce BCBL-1 cells. We have shown previously that neomycin is efficient in inducing KSHV lytic reactivation in PEL cells by inhibiting phospholipase C γ (PLC-γ) (17,–19). Therefore, we treated BJAB and BCBL-1 cells with neomycin for 48 and 96 h and determined the levels of IFI16 by WB analysis. IFI16 levels were not reduced in BJAB cells, and in contrast, we observed a gradual decrease in IFI16 levels after neomycin induction (Fig. 7C). As a positive control for the induction of the lytic cycle, we performed immunoblotting for the late KSHV protein gpK8.1A (Fig. 7C).
Together, these results confirmed that a reduction of the IFI16 protein level is a general consequence of KSHV lytic reactivation.
Next, we determined whether lytic reactivation of primary cells with latent KSHV infection could also exhibit a similar reduction of IFI16 levels. For this, we infected primary human dermal microvascular endothelial cells (HMVEC-d) with KSHV (60 DNA copies/cell). At 72 h postinfection (p.i.), when latency is established (47), cells were treated with TPA or a vehicle control for 2, 3, and 4 days (48, 72, and 96 h), and IFI16 levels were determined by WB analysis. As a control, mock-infected cells were also treated with TPA to rule out any off-target effects of TPA on IFI16 levels (Fig. 7D, compare lanes 1 and 2). We observed that after 6 days of infection (3 days for latency establishment plus 3 days of mock induction) (Fig. 7D, lane 4), IFI16 levels increased by >3.5-fold. Similar elevated levels of IFI16 were maintained at 7 days p.i. (Fig. 7D, lane 5). In contrast, after TPA induction of infected cells (Fig. 7D, lanes 6 to 8), we observed a gradual decrease in IFI16 levels at 3 and 4 days post-TPA treatment (Fig. 7D, compare lanes 7 and 8 with lanes 4 and 5). This result corresponds to lytic reactivation after the establishment of latency, and the reduction in IFI16 levels as seen in PEL cells confirmed that the reduction in IFI16 protein levels is induced during KSHV lytic reactivation.
To gain insight into the mechanism by which IFI16 is degraded after lytic reactivation, we first determined whether IFI16 is polyubiquitinated after TPA induction. BCBL-1, BC-3, and BJAB cells were induced with TPA for 48 and 96 h or mock induced, and IFI16 was immunoprecipitated. Subsequently, a WB analysis was performed by using the P4D1 antibody recognizing both mono- and polyubiquitin (Fig. 8A, top) (68). Compared to UI cells, a smeary pattern that is a hallmark of polyubiquitination was observed after TPA induction of both BCBL-1 and BC-3 cells (Fig. 8A). However, in control BJAB cells, no such IFI16 polyubiquitination was observed. WB analysis of IFI16 (Fig. 8A, bottom) confirmed the degradation of IFI16 under the indicated conditions.
Next, we employed the proteasome inhibitor MG132 to evaluate the possibility that the observed polyubiquitinated IFI16 is degraded via the ubiquitin-proteasome pathway. BCBL-1 and BJAB cells were induced with TPA for 48 h, 1 μM MG132 was added for another 18 h, and IFI16 levels were detected by WB analysis. Treatment with MG132 resulted in the stabilization of IFI16 in TPA-induced BCBL-1 cells, suggesting degradation by the proteasome during lytic reactivation (Fig. 8B). Western blotting for AIM2 also confirmed that this effect is IFI16 specific (Fig. 8B). As expected, no change was observed for BJAB cells under similar conditions. These observations confirmed that IFI16 is proteasomally degraded after lytic reactivation.
One of the unique features of KSHV de novo infection of HMVEC-d and human foreskin fibroblast (HFF) cells is that a select number of KSHV lytic genes, including ORF50 (RTA), are expressed initially along with latent genes. By 24 h p.i., the levels of these lytic transcripts are reduced, and a steady increase in latent gene expression ensues (47). We have also shown that IFI16 is not affected during de novo infection, and it interacts with the incoming KSHV genome to induce the innate inflammatory and interferon beta responses (31, 32). In addition to this, our previously reported observations showed that IFI16 colocalizes with latent KSHV genomes in BCBL-1 cells, and latent gene expression continues. Hence, we reasoned that KSHV latent gene products, and the lytic gene products observed during de novo infection, may not be involved in the degradation of IFI16 and that other lytic proteins could be involved in targeting endogenous IFI16 after reactivation.
To evaluate this possibility, we used phosphonoacetic acid (PAA), an inhibitor of herpesvirus-encoded DNA polymerase (69, 70), which consequently also specifically inhibits the generation of late lytic cycle transcripts (71). We induced BCBL-1 cells with TPA, added 100 μg/ml of PAA simultaneously, and incubated the cells for 96 h. We postulated that if IFI16 is still degraded in the presence of PAA, it would mean that the protein(s) belonging to the latent, immediate early, or early lytic cycle is instrumental in IFI16 degradation. On the contrary, if IFI16 degradation can be inhibited by PAA, this would indicate the role of one or more late lytic proteins in this process. The results clearly showed that PAA efficiently inhibited IFI16 degradation in TPA-induced cells, whereas PAA alone had no effect on IFI16 levels (Fig. 8C, compare lanes 2 to 4). To confirm that PAA inhibited KSHV DNA replication, the late lytic protein gpK8.1A was probed and was found to be completely inhibited by PAA (Fig. 8C, compare lanes 2 to 4). Real-time RT-PCR was also performed to estimate viral genome copy numbers. TPA induction resulted in an increase in the KSHV DNA copy number, but in the presence of PAA, almost no increase in the DNA copy number was observed, suggesting inhibition of the viral polymerase (Fig. 8D). PAA alone had no effect on DNA replication.
Next, we used cycloheximide (CHX) to block the synthesis of the late lytic genes and then evaluated the levels of IFI16 at different time points. For this, we chose a CHX concentration of 20 μM and a maximum time point of 8 h post-CHX treatment, as concentrations above this were toxic to the cells. We induced BCBL-1 cells with TPA and allowed the lytic cycle to continue for 36 h. According to our previously reported observations (47) and as shown in Fig. 3, IE and E genes are efficiently induced by 24 h, but late lytic genes are induced from about 36 h onward. Therefore, we blocked the expression of these late lytic proteins with CHX and then monitored IFI16 levels at different time points. To monitor the effect of CHX on viral protein synthesis, the expression of the late lytic gpK8.1A protein was examined by immunoblotting. In control dimethyl sulfoxide (DMSO)-treated cells, we observed a gradual decrease in the level of the IFI16 protein (Fig. 8Ea, lanes 1 to 4), indicating IFI16 degradation at these time points (36 h of TPA treatment plus 8 h). In contrast, the levels of the gpK8.1A protein increased gradually with time, indicating an increased synthesis of late lytic proteins at the tested time points (Fig. 8Ea). Interestingly, upon the addition of CHX, IFI16 degradation was completely inhibited (Fig. 8Eb, lanes 1 to 4), which suggested that the KSHV protein that is responsible for the degradation of IFI16 is potentially a late lytic protein. In the presence of CHX, the gpK8.1A level did not increase and rather was reduced, probably due to the absence of translation of gpK8.1A (Fig. 8Eb), thus indicating that late lytic genes were inhibited efficiently. When we tested the effect of CHX in the absence of TPA induction on BCBL-1 cells as a control, we observed that the IFI16 level was not affected at the tested time points, with very little gpK8.1A production in these cells (Fig. 8Ec, lanes 1 to 4).
Together, these experiments suggested that a KSHV late lytic protein(s) is involved in the ubiquitination and subsequent proteasomal degradation of IFI16.
Our results strongly indicated that IFI16 is targeted after KSHV lytic reactivation, while latent KSHV gene expression continues in the presence of IFI16 in PEL cell lines. IFI16 has been suggested to participate in gene regulation, but to date, its precise role in transcription and/or replication control is not well defined. It has been shown that IFI16 binds to the promoter regions of p53 and cMyc genes in the melanoma cell line M14 to regulate cell growth and apoptosis (72). Recently, we reported for the first time that endogenous IFI16 also binds to HSV-1 promoters (39). Our work demonstrated that IFI16 binds to all classes of HSV-1 promoters after infection, leading to a significant reduction in RNA polymerase II recruitment (39). In BCBL-1 cells, using fluorescent in situ hybridization (FISH), we have also shown that IFI16 colocalizes with the KSHV latent genome, leading to a state of constitutive inflammasome activation in these cells (34). Together with our observations here, we asked whether IFI16 binds to the latent and lytic KSHV gene promoters and regulates them transcriptionally.
To determine this, we performed chromatin immunoprecipitation (ChIP). IFI16 was immunoprecipitated from the nuclei of BCBL-1 and BC-3 cells, and the associated DNA was analyzed by real-time RT-PCR with primers corresponding to promoter regions of different representative genes of all KSHV gene classes. As a negative control, isotype control IgG was included in the ChIP assay. We observed that the IFI16 antibody coprecipitated all KSHV promoters tested with various efficiencies in BCBL-1 (Fig. 9A) and BC-3 (Fig. 9B) cell lines. To emphasize the specificity of the IFI16 antibody used, the results are presented as fold increases of promoter occupancy over the IgG control. We also performed conventional PCR with ChIP DNAs from BCBL-1 cells and resolved them on an agarose gel to show visually that the IFI16 antibody specifically pulled down KSHV promoter DNA over the IgG control antibody (Fig. 9C). These results demonstrated that IFI16 interacts with KSHV episomal genome DNA and occupies its latent and lytic gene promoters during prolonged latency.
Having confirmed that IFI16 binds to KSHV gene promoters, we next evaluated its possible transcription regulatory roles in KSHV gene expression. Previously reported findings suggested IFI16's role as both a transcriptional activator (73, 74) and repressor (39, 73, 75). We utilized a dual-luciferase reporter assay to measure KSHV promoter activities before and after KD of IFI16 by shRNA. We selected six different KSHV promoter luciferase constructs and transfected them into Kaposi's sarcoma-derived KSHV-negative endothelial SLK cells that were previously transfected with ShC or ShIFI16 plasmids for 24 h. The dual-luciferase assay was performed 24 h after this, and the results showed that promoter activities from most genes, except for the latent ORF73 gene, increased after IFI16 KD (Fig. 10A). These results strongly demonstrated that IFI16 suppresses transcription from most KSHV lytic gene promoters. A similar experiment was performed by using the wild-type human osteosarcoma cell line U2OS and a clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9-mediated IFI16-negative clone, U2OS clone 67 (39). In these cells also, the absence of IFI16 resulted in significantly increased promoter activity compared to that in wild-type cells (Fig. 10B). Next, to address the question of whether IFI16 can discriminate between host gene promoters and viral gene promoters, we performed a similar dual-luciferase assay with 4 unrelated human reporter luciferase constructs, namely, MDM2, Ap1, NFAT, and NF-κB, after IFI16 KD in SLK cells. We observed a small but significant increase in the transcriptional activity of all the promoters tested after IFI16 KD (Fig. 10C). Compared to some of the KSHV gene promoters tested, this increase in host promoter activity is only marginal and may signify that additional factors are involved in providing discriminatory potential toward intruder viral genes.
To further ascertain whether IFI16 acts as a transcriptional repressor on KSHV promoters, we performed an IFI16 add-back experiment using IFI16-negative U2OS clone 67 cells. These cells were transfected with either a control plasmid or an IFI16-overexpressing plasmid (IFI16-FL), and 24 h later, they were transfected with KSHV promoter luciferase constructs. After another 24 h, a dual-luciferase assay was performed, and the results showed that the add-back of IFI16 significantly reduced the promoter activities of all KSHV lytic promoters compared to the control (Fig. 10C). Collectively, these data confirm that IFI16 binds to several KSHV promoters and exerts transcriptional repressor activity.
IFI16 has emerged as an antiviral restriction factor against a number of different DNA viruses (28, 29, 37,–40, 76, 77). The antiviral function of IFI16 can be subdivided into two distinct mechanisms: (i) an immune surveillance role, where IFI16 activates the inflammasome and IFN-β pathways in response to foreign viral DNA in the nucleus (28, 34, 38, 40, 73, 77, 78), and (ii) a gene regulatory role, where it transcriptionally represses viral transcription and replication (29, 37, 39, 79). So far, this gene regulatory role of IFI16 has been demonstrated for HSV-1, HCMV, and HPV (29, 36, 37, 39). In these viruses, IFI16 has been shown to either epigenetically modify the viral chromatin so as to favor heterochromatin marks or alter the binding of transcription factors on the viral promoters. A hallmark of the herpesviral life cycle is the establishment of long-term latency in infected host cells. Periodic reactivation of this latent viral genome due to stress, trauma, or other more complex molecular events contributes to the ability of these viruses to cause diseases (80). However, to date, the role of IFI16 in the herpesviral life cycle has been studied only in relation to the lytic phase of the viral cycle, which is not the primary natural course of viral infection. This is primarily because HSV-1 and HCMV exhibit a lytic-only life cycle in cell culture, and no latency model is available. Therefore, the role of IFI16 during herpesvirus latency and its subsequent reactivation remains unexplored.
In this study, we show that IFI16 plays an important role in the maintenance of latency in KSHV-infected human PEL cell lines, and knockdown of endogenous IFI16 results in spontaneous lytic reactivation of the latent viral genome. Our observations strongly indicate that IFI16 functions as a transcriptional repressor on the viral lytic promoters, thereby aiding in maintaining a state of attenuated viral gene expression. However, during lytic reactivation, this transcriptional regulation becomes a hindrance to viral lytic gene expression and thus needs to be overcome. Our data show for the first time that KSHV has evolved a mechanism to proteasomally degrade IFI16 after the onset of the lytic cycle, thereby relieving transcriptional silencing.
For a long time, IFI16 has been speculated to participate in gene regulation, but to date, its precise role in transcription and/or replication control is not well defined (75, 81, 82). It has been shown that IFI16 binds to the promoter regions of the p53 and cMyc genes in the melanoma cell line M14 to regulate cell growth and apoptosis (72). Using an in vitro system, Johnstone et al. showed that IFI16 fused to the GAL4 DNA binding domain can transcriptionally repress a GAL4-thymidine kinase-chloramphenicol acetyltransferase (tk-CAT) cassette in HeLa cells (75). They also suggested that this function is dependent on the two HIN200 domains of IFI16. However, no mechanism of transcriptional repression was provided. Recently, we reported for the first time that endogenous IFI16 also binds to viral promoters (39). Our work demonstrated that IFI16 binds to all classes of HSV-1 promoters after infection, leading to a significant reduction in RNA polymerase II recruitment (39).
Next-generation transcriptome analysis carried out to attain a comprehensive analysis of KSHV transcription after IFI16 KD (Fig. 1 and and2)2) demonstrated a robust increase in KSHV lytic gene transcription by IFI16 KD, which was comparable to that with TPA induction. Most of the latent genes were not induced significantly, and only the K12 gene was found to be upregulated. This signifies that IFI16-mediated transcriptional regulation is viral gene class specific, and the repression mechanism can distinguish between latent and lytic genes. Using a dual-luciferase system fused to different KSHV promoters, we found that IFI16 efficiently downregulates transcription from most lytic promoters tested (Fig. 10A, ,B,B, and andD).D). The latent ORF73 promoter was not affected in SLK cells (Fig. 10A), signifying the specificity of this function. However, in IFI16 knockout U2OS cells (U2OS clone 67), in addition to the increase in lytic promoter activity, a significant increase in latent promoter activity was also observed (Fig. 10B). This difference may be attributed to the different origins and molecular signatures of the two cell lines. U2OS is a cancer cell line originating from a moderately differentiated sarcoma of the tibia, whereas SLK cells are endothelial/epithelial cells derived from a Kaposi's sarcoma tumor. This indicates that other cellular or viral factors may play a role in conferring specificity toward specific KSHV promoter classes. Recently, we have shown a similar instance where viral and cellular transcription factors cooperate to mediate viral transcription regulation (83, 84).
It has been reported that IFI16 binds DNA nonspecifically through electrostatic attraction between the positively charged HIN-200 (hematopoietic interferon-inducible nuclear proteins with a 200-amino-acid repeat) domain and the negatively charged sugar-phosphate backbone of double-stranded DNA (dsDNA) (85). Our ChIP analysis (Fig. 9) also showed that IFI16 binds to all classes of KSHV promoters, including latent promoters. This finding is in agreement with data from previous reports from our laboratory where we found that IFI16 binds to promoters of all temporal classes of HSV-1 genes (39). Thus, it is conceivable that IFI16 mediates its transcriptional regulation roles in conjunction with other transcription factors, which presumably confers specificity toward different viral gene classes. Moreover, our observation that KD of IFI16 results in a much smaller increase in the reporter activity of human promoter elements (Fig. 10C) than some KSHV promoter elements (Fig. 10A) points toward the possibility that additional transcription modulators may be involved, which helps in distinguishing between self and invading viral genes. Gariano et al. reported that IFI16 modulates the recruitment of the transcription factor Sp1 on the HCMV genome, thereby inhibiting transcription (37). However, recently, those authors found that even promoters devoid of any Sp1 binding site were inhibited by IFI16 overexpression, indicating that Sp1 is not the sole transcription factor responsible (29). Recently, we showed that IFI16 modulates interactions of RNA polymerase II, TATA binding protein (TBP), and Oct1 with the wild-type HSV-1 genome after infection (39). We and others have also shown that IFI16 interacts with several host gene promoters (36, 39, 41, 72,–75). The identities of IFI16-interacting transcription factors and other proteins involved in IFI16-mediated gene regulation in KSHV and other herpesviruses require extensive studies, and such evaluations are ongoing.
Recent developments also point toward an epigenetic route through which IFI16 regulates viral transcription. However, unlike the transcription factor-modulating function, this epigenetic regulation does not seem solely specific to foreign viral DNA. We have recently shown that IFI16 influences histone modification marks at the transcription start sites of both HSV-1 and cellular genes (39). We also found that the absence of IFI16 promoted euchromatic marks and decreased heterochromatic marks. In agreement with this, others have also shown that IFI16 promotes heterochromatinization of invading viral DNA (29). Further extensive studies are needed to evaluate the possibility of IFI16-mediated epigenetic modulation of the KSHV genome during latency and lytic reactivation.
Our present data confirm that IFI16 plays a crucial role in maintaining KSHV latency in PEL cell lines, and a disruption of IFI16 homeostasis results in successful lytic reactivation with concomitant viral genome replication (Fig. 2E). We also found that KD of IFI16 causes heightened induction of KSHV genes post-TPA treatment (Fig. 3). This signifies that IFI16 continues to exert its transcriptional silencing function even after TPA induction. KD of IFI16 also results in a greater degree of viral genome replication at a faster pace. This finding, along with our observation that overexpression of IFI16 prior to TPA induction results in decreased gene transcription and genome replication (Fig. 4B), signifies that IFI16 has a crucial function in suppressing aberrant induction of the lytic cycle. Aberrant reactivation is a deleterious option for herpesviruses. Reactivation from latency triggers various host defense mechanisms and signaling pathways, which ultimately leads to the death of infected cells and blocks viral progeny formation. Therefore, tight control of probable inducing insults is very important, and IFI16 is most likely one of the factors that confer this regulation. However, this also means that IFI16 will become a hurdle during legitimate lytic induction signals such as TPA, neomycin, or RTA overexpression in vitro and cellular stress and trauma in vivo, which are important for the ultimate propagation of the virus.
Our data show for the first time that cellular IFI16 is specifically degraded following lytic reactivation of KSHV in BCBL-1 and BC-3 PEL cells (Fig. 5 and and6).6). Degradation of cellular IFI16 by a viral factor has been reported only for HSV-1, where the immediate early gene product ICP0 targets IFI16 for degradation during de novo infection, leading to a full-fledged lytic cycle (39, 40). However, as we discuss above, HSV-1 does not establish latency in cell culture, and therefore, our report is the first to show a latent virus that targets IFI16 during lytic reactivation. HSV-1 ICP0 functions as an E3 ubiquitin ligase (86) and has been reported to induce the degradation of other cellular restriction factors, including promyelocytic leukemia (PML) nuclear body components (87). Cuchet-Lourenco et al. showed that IFI16 was not degraded if ICP0 was expressed in the absence of HSV-1 infection, indicating that other cellular or viral factors are also involved in this process (41). A similar observation was also made by Diner et al., who found that ICP0 is necessary but not sufficient for the degradation of IFI16 (78). Thus, the precise molecular events leading to IFI16 degradation after HSV-1 infection are not fully understood.
Here, we have confirmed that this degradation of IFI16 is specific to KSHV lytic reactivation and is not a TPA-specific effect (Fig. 7). We have also been able to show this effect on human primary dermal endothelial cells, which are a more natural infection model for KSHV (Fig. 7D). After de novo infection of HMVEC-d cells, we allowed latency to establish for 72 h before induction of the lytic cycle by TPA. At 6 days postinfection, we observed an ~3.5-fold increase in IFI16 levels, and the reason for this increase after KSHV infection is not clear at this time. One possibility is that KSHV infection induces reactive oxygen species (ROS) in infected cells (88), and it has been reported that ROS generation and oxidative stress can cause an elevation of endogenous IFI16 levels (89). At 2 days post-TPA induction (Fig. 7D, lane 6), a further increase in the level of IFI16 compared to the uninduced control was observed (Fig. 7D, lane 3). The cause of the increase in IFI16 protein levels 2 days after TPA treatment (Fig. 7D, lane 6) compared to the TPA-uninduced control (Fig. 7D, lane 3) is not clear at this moment. However, 3 and 4 days after TPA induction of these latently infected cells, a robust reduction of IFI16 levels was observed, indicating lytic reactivation-specific degradation of IFI16 (Fig. 7).
We also confirmed the role of ubiquitin-mediated proteasomal degradation in this restriction of IFI16. MG132, a broad-spectrum proteasome inhibitor, efficiently inhibited IFI16 degradation (Fig. 8A and andB).B). A number of E3 ubiquitin ligases have been reported for KSHV, including K5, K7, and ORF50 (RTA) (90,–92). All three of these proteins are early lytic proteins induced within 12 h of lytic induction. However, our observation that IFI16 degradation can be inhibited by PAA, a specific inhibitor of herpesviral late gene expression, points toward the possibility that the most probable candidate is a late lytic protein(s). Moreover, our data show that considerable degradation of IFI16 can be detected only after 48 h of lytic induction (Fig. 5A), indicating that early lytic proteins are not responsible for IFI16 degradation. Another observation that strengthens this notion is that during de novo infection of HMVEC-d cells, no degradation of IFI16 was observed, and in fact, we found a slight induction of IFI16 levels (Fig. 7D). We reported previously that during early phases of de novo infection (2 h p.i.), 11 immediate early/early, 8 early, and 5 late lytic genes are also induced in addition to the 4 usual latent genes. These lytic genes included ORF50 (RTA), K2, K4, K5, K6, K7, and vIRF2 and were observed up to 24 h p.i. (47). Therefore, it is reasonable to consider that these genes may not be involved in IFI16 degradation. Using CHX, we further confirmed that the inhibition of late lytic protein synthesis completely blocked IFI16 degradation (Fig. 8E). However, a more in-depth analysis needs to be undertaken to identify the KSHV protein(s) mediating the proteasomal degradation of IFI16 after lytic reactivation, which is beyond the scope of the present study.
We thus propose that KSHV has evolved to utilize the transcriptional repressor function of IFI16 to maintain its latency, during which only a small subset of viral genes are transcriptionally active. Upon lytic reactivating signals, when it no longer requires transcriptional repression, one or more late lytic gene products of KSHV degrade IFI16 to embark upon a full-fledged lytic life cycle. Further understanding of the role of IFI16 in the latent and lytic cycles of KSHV will lead to the development of effective strategies to control KSHV infection and associated diseases.
We thank Keith Philibert for critically reading the manuscript. We thank the University of Illinois Center for Research Informatics (UIC CRI) for their assistance in the bioinformatics analysis for this project.