Alternatively Initiated Gene 50/RTA Transcripts Expressed during Murine and Human Gammaherpesvirus Reactivation from Latency

Abstract

In the process of characterizing the requirements for expression of the essential immediate-early transcriptional activator (RTA) encoded by gene 50 of murine gammaherpesvirus 68 (MHV68), a recombinant virus was generated in which the known gene 50 promoter was deleted (G50pKO). Surprisingly, the G50pKO mutant retained the ability to replicate in permissive murine fibroblasts, albeit with slower kinetics than wild-type MHV68. 5′-rapid amplification of cDNA ends analyses of RNA prepared from G50pKO-infected fibroblasts revealed a novel upstream transcription initiation site, which was also utilized during wild-type MHV68 infection of permissive cells. Furthermore, the region upstream of the distal gene 50/RTA transcription initiation site exhibited promoter activity in both permissive NIH 3T12 fibroblasts as well as in the murine macrophage cell line RAW 264.7. In addition, in RAW 264.7 cells the activity of the distal gene 50/RTA promoter was strongly upregulated (>20-fold) by treatment of the cells with lipopolysaccharide. Reverse transcriptase PCR analyses of RNA prepared from Kaposi's sarcoma-associated herpesvirus- and Epstein-Barr virus-infected B-cell lines, following induction of virus reactivation, also revealed the presence of gene 50/RTA transcripts initiating upstream of the known transcription initiation site. The latter argues that alternative initiation of gene 50/RTA transcription is a strategy conserved among murine and human gammaherpesviruses. Infection of mice with the MHV68 G50pKO demonstrated the ability of this mutant virus to establish latency in the spleen and peritoneal exudate cells (PECs). However, the G50pKO mutant was unable to reactivate from latently infected splenocytes and also exhibited a significant reactivation defect from latently infected PECs, arguing in favor of a model where the proximal gene 50/RTA promoter plays a critical role in virus reactivation from latency, particularly from B cells. Finally, analyses of viral genome methylation in the regions upstream of the proximal and distal gene 50/RTA transcription initiation sites revealed that the distal promoter is partially methylated in vivo and heavily methylated in MHV68 latently infected B-cell lines, suggesting that DNA methylation may serve to silence the activity of this promoter during virus latency.

Herpesviruses contain large, double-stranded DNA genomes that encode an extensive array of viral proteins required for efficient infection and persistence in host organisms. These viruses establish life-long latent infection in a variety of cell types, a characteristic that is often accompanied by periodic virus reactivation and resumption of lytic replication and egress. Unlike the alpha- and betaherpesviruses, which exhibit neuronal or broad cellular tropism, the gammaherpesviruses are mainly lymphotropic, infecting and establishing latency in B or T lymphocytes. Gammaherpesviruses are associated with the development of lymphomas in both their natural host and in animal models and have therefore been subject to intensive study in an effort to understand, treat, and prevent disease.

The gammaherpesvirus type 1 Epstein-Barr virus (EBV) is implicated in the development of several human malignancies, including Burkitt's lymphoma, Hodgkin's lymphoma, and nasopharyngeal carcinoma. Kaposi's sarcoma-associated herpesvirus (KSHV), a gammaherpesvirus type 2, is likewise associated with tumor development in Kaposi's sarcoma, primary effusion lymphoma, and multicentric Castleman's disease. Although cell culture systems exist to study cells latently infected with these viruses, the narrow tropisms of both EBV and KSHV have necessitated the use of animal models to intimately study the process of primary virus infection and events that contribute to the perpetuation of viral latency in vivo. Murine gammaherpesvirus 68 (MHV68) is a naturally occurring rodent pathogen that mimics several key aspects of both EBV and KSHV infection following experimental infections of inbred mice (5, 34, 41).

Substantial sequence homology is shared among the murine and primate gammaherpesviruses in the region of open reading frame 50 (ORF50), which encodes the key lytic switch protein RTA (45). Importantly, in all the characterized gammaherpesviruses the gene 50 transcript contains a short first exon that splices to a long coding exon (Fig. 1). For the type 2 gammaherpesviruses the ORF50/RTA reading frame is extended by the exon 1/exon 2 splice, with the translation initiation site located near the 3′ end of exon 1. In the case of the type 1 gammaherpesvirus EBV, exon 1 does not contribute to the RTA coding sequence. Finally, the position of the short first exon is also well-conserved among all the gammaherpesviruses, being located in the region between ORFs 48 and 49 (Fig. 1). RTA plays a pivotal role in lytic replication in both human and murine gammaherpesviral infections, and ectopic expression of RTA is sufficient to trigger reactivation into the lytic cycle in latently infected cell lines (33, 40, 49). Because gammaherpesviruses are thought to first infect permissive cells, followed by establishment of a latent infection within lymphocytes, potentially antigenic lytic gene expression must be carefully controlled. This is important both during initial infection, to allow for efficient establishment of latency, as well as during latency when the capability to reactivate remains. This level of regulation is accomplished for both latent and lytic genes by several mechanisms shared between EBV, KSHV, and MHV68, including multiple promoter usage, extensive alternative splicing, and epigenetic modifications (1, 9, 12, 35-37, 47). We demonstrate here that expression of the MHV68 RTA homolog can be driven from a newly defined promoter upstream of the promoter previously identified and that this expression is sufficient to allow for lytic replication and establishment of latency. We also identified an additional exon included in the RTA-encoding transcript generated from this promoter, and we provide evidence that the existence and utility of both the distal promoter and additional exon are conserved among the murine and human gammaherpesviruses. We provide evidence that DNA methylation may serve as a regulatory mechanism for silencing the distal gene 50 promoter during latent infection and may potentially account for the severe reactivation defect observed in mice infected with a gene 50 mutant MHV68 lacking the proximal promoter. Finally, we demonstrate that although the G50pKO virus is defective for reactivation from latently infected splenocytes, this virus is capable of reactivation from peritoneal exudate cells (PECs), indicating that activity of the newly identified distal promoter is regulated in a cell-type-specific manner during latent infection.

FIG. 1.

Conservation of G50/BRLF1/Rta coding region among gammaherpesviruses. Schematic representation of the G50/BRLF1/Rta coding regions of MHV68, EBV, KSHV, and HVS. The nucleotide positions for G50 exons 1 and 2 are provided. The box drawn around exon 1 demonstrates the conserved juxtaposition of exon 1 relative to open reading frames encoded on the opposite strand.

MATERIALS AND METHODS

Viruses and tissue culture.

Both virus mutants were generated from the wild-type MHV68 bacterial artificial chromosome (BAC) provided by Ulrich Koszinowski. Virus was passaged and titers were determined as previously described (26). Mouse embryonic fibroblasts (MEFs), Vero-Cre, and NIH 3T12 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 2 mM l-glutamine, 100 U penicillin per ml, and 100 mg streptomycin per ml. Vero-Cre cells were cultured with the addition of hygromycin (300 μg/ml). S11E, A20-HE1, A20-HE2, Akata, Clone-13, and BCBL-1 cell lines were maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM l-glutamine, 100 U penicillin per ml, and 100 mg streptomycin per ml. A20-HE1 and A20-HE2 cells were maintained under hygromycin selection as previously described (11). All tissue cultures were performed in a 5% CO₂ tissue culture incubator at 37°C.

Generation of G50pKO and G50pKO.MR viruses.

The MHV68.G50pKO virus was generated by first using wild-type BAC DNA as a template in two separate round 1 50-μl reaction mixtures. PCR was performed with Vent polymerase (New England Biolabs) with 25 cycles under the following parameters: 94°C for 30 seconds, 57°C for 30 seconds, and 72°C for 1 min. The internal primers were designed to introduce HindIII sites, which would later be used for Southern blotting confirmation of mutant and marker rescue viruses. Reaction 1 used forward primer 5′-GATCATGACTTCTAGTCATATCC-3′ and reverse primer 5′-GCTAGCTAAGCTTCTAGCCAGGGAATTTTGTTATGTGC-3′, corresponding to nucleotide (nt) 66402. Reaction 2 used forward primer corresponding to nt 66548, 5′-GCTAGAAGCTTAGCTAGCCTTCACGGGTTTCAAGGTCC-3′, and reverse primer 5′-GTTTCCTGACCTCTGTAGACG-3′. One microliter from each reaction mixture was then used as template for a round 2 reaction using the reaction 1 forward primer and the reaction 2 reverse primer. The resulting product was gel purified, cloned into pCRBlunt (Invitrogen), and sequenced to verify the deletion of the G50 promoter region. The fragment was then excised from pCRBlunt and cloned in the suicide vector pGS284 using NsiI. Positive clones were identified using BglII and NotI digestion. Allelic exchange using the MHV68 BAC was performed as previously described (26). The MHV68.G50pKO.MR virus was generated by using the round 2 primers to amplify wild-type BAC DNA. The product was cloned into pGS284 as above and used in the allelic exchange protocol with MHV68.G50pKO.MR BAC. Mutant and marker rescue BACs were confirmed by Southern blot analyses following HindIII digestion and probed with a PCR-generated fragment encompassing the original site of amplification. Mutant and marker rescue BAC DNAs were then transfected into Vero-Cre cells and propagated to generate virus stocks.

Plaque assays, in vitro growth, and determination of viral titers.

NIH 3T12 cells were plated in six-well plates at 2 × 10⁵ cells per well the day prior to infection. Virus stocks were diluted to a 200-μl volume in cMEM at the indicated multiplicities of infection (MOIs) and added to 3T12 monolayers. Plates were rocked every 15 min for 1 h at 37°C. For single-step growth analysis, the viral inoculum was removed. After infection, cMEM was added and plates were incubated under normal cell culture conditions until the indicated times postinfection, whereupon plates were stored at −80°C until titer determinations. Intracellular virus was liberated by freeze-thaw lysis, cell lysates were diluted, and viral titers were determined by plaque assay as previously described (44).

RACE analysis.

RNA was prepared from NIH 3T12 cells 16 h after infection with MHV68 G50pKO or wild-type virus (MOI, 1) by guanidine isothiocyanate-phenol extraction or from untreated A20-HE1 cells or A20-HE1 cells 24 h after treatment with phorbol-12-myristate-13-acetate (TPA; 20 ng/ml) using Trizol reagent (Invitrogen) per the manufacturer's instructions. For 3T12 analysis, poly(A) RNA was purified using the Sigma gel elute kit, and 300 ng was converted to cDNA using the BD Smart RACE cDNA amplification kit (Clontech). For A20-HE1 and A20-HE2 analysis, random amplification of cDNA ends (RACE) was performed using the GeneRacer system (Invitrogen). For 3T12 analysis, nested PCR was performed using Vent polymerase with the 5′ universal forward primer (round1), 5′ universal nested forward primer (round 2), and reverse primer 5′GGTTGAGGTAGCTGTACC-3′ at an annealing temperature of 60°C. One microliter of round 1 product was used as the template for the round 2 reaction. For A20-HE1 analysis, nested PCR was performed using Amplitaq Gold DNA polymerase (see PCR mix below in “Bisulfite PCR analyses”) under the following cycling conditions: for round 1 (using 1 μl of modified cDNA per 50-μl reaction mixture), 94°C for 8 min, 10 cycles with denaturation at 95°C for 30 s, annealing starting at 68°C and decreasing 0.4°C every two cycles, and extension at 72°C for 2 min, then 25 cycles with denaturation at 95°C for 30 s, annealing at 65°C for 30 s, and extension at 72°C for 2 min, followed by a final extension at 72°C for 7 min. Round 2 (using 1 μl of round 1 product per 50-μl reaction mixture) was performed under the conditions desrcibed above. Both reactions were performed with the reverse gene-specific primer 5′-CCTTCTCATGGTCACATCTGTGTCTCACTGAAAAC-3′. PCR products were visualized by ethidium gel electrophoresis and excised bands purified using a Qiagen gel extraction kit (Valencia, CA). Purified PCR products were ligated into a pGEMT-Easy vector (Promega) and analyzed by DNA sequencing.

Reporter plasmids.

DNA fragments upstream of E0 were amplified from wild-type MHV68 BAC using AccuPrime Pfx supermix (Invitrogen) at 64°C with the same reverse primer in E0 (5′-GATCGTCTAGAGTGCTGGGTTGTGAAG-3′) and the following forward primers: −100bp (5′ GATCGGCTAGCACTCGAAGTGTCCAGC-3′), 250bp (5′-TGGCGGCTAGCTTAATCCTATATGGAGAT-3′), 500bp (5′-GATCGCTAGCGAAAACGCGGAGAG-3′), and −1000bp (5′-GATCGCTAGCGAAAAGCGGGAGAG-3′). Products were gel purified and cloned into pCR-Blunt (Invitrogen) for sequence verification. Reporter fragments in pCR-Blunt were digested with NheI and XbaI, and the luciferase vector pGL4.10[luc2] (Invitrogen) was digested with NheI and treated with calf intestinal phosphatase (CIP) prior to ligation. Digested products were gel purified and extracted using the Qiagen gel extraction kit (Valencia, CA), resuspended in water, and ligated overnight at 16°C. Ligation mixtures were transformed and screened for the presence and orientation of inserts by digestion with either NheI and EcoRV or PvuII. Positive clones were cultured and plasmid DNA isolated using the Qiagen EndoFreeMaxi kit (Valencia, CA).

Transfections and luciferase assays.

One day prior to transfection, 2 × 10⁵ NIH 3T12 fibroblasts or RAW 264.7 macrophages were plated in six-well plates in cMEM. Immediately prior to transfection, cells were washed 2× in phosphate-buffered saline and covered with serum-free medium. NIH 3T12 transfections were performed with Lipofectamine and Lipofectamine Plus reagent (Invitrogen) according to the manufacturer's instructions. RAW 264.7 transfections were performed with TransIT transfection reagent (Mirus) according to the manufacturer's instructions. All transfections were performed with the recommended quantity of reporter plasmid DNA (2 to 2.5 μg) as well as 5 ng of pHR-Luc Renilla luciferase vector as a transfection control. Empty pGL4.10[luc] and pGL4.13[luc] were used as negative and positive controls, respectively. The green fluorescent protein (GFP) plasmid pMaxGFP (Amaxa) was used to monitor transfection efficiency. After 24 h, macrophage cultures were stimulated with 5 μg lipopolysaccharide (LPS; Sigma-Aldrich). After 48 h, cell lysates were harvested and luminometric assays performed using the dual luciferase assay system (Promega) according to the manufacturer's instructions. Each transfection was performed in triplicate, and data are presented as the firefly:Renilla luciferase ratio versus the empty vector.

Cell lines and treatments.

Induction of A20-HE1 cells at 1 × 10⁶ cells/ml was performed by treatment with TPA (20 ng/ml). Akata cells at 1 × 10⁶ cells/ml were induced by treatment with goat-anti-human immunoglobulin G (IgG; Jackson ImmunoReseach Laboratories, Inc.) at 100 μg/ml as previously described (10a). Clone-13 cells were treated at 1 × 10⁶ cells/ml with TPA (20 ng/ml) for 24 h. BCBL-1 cells at 2 × 10⁵ cells/ml were treated with TPA (20 ng/ml) for 48 h.

Quantitative RT-PCR.

RNA from A20-HE2 or NIH-3T12 cells infected with MHV68 (MOI, 10) was isolated using either Trizol reagent (Invitrogen) or the Qiagen RNeasy Maxi kit. Three or five micrograms of RNA was treated with DNase I (Invitrogen) according to the manufacturer's instructions in a total volume of 50 μl. Twenty microliters of DNase-treated RNA was subsequently used in first-strand cDNA synthesis using SuperScript II reverse transcriptase (Invitrogen). Five microliters of the cDNA reaction mixture was used in each quantitative amplification reaction mixture. Quantitative PCR was performed using iQ Supermix (Bio-Rad) with the following primers (at 900 nM): E1-E2, forward (5′-GGAATTTCTGCAGCGATGGCCTCT-3′) and reverse (5′-CCTCTTTTGTTTCAGCAGAGACTCCA-3′); E0-E1-E2, forward (5′-GCAGTCCGTAGCCGCTGGAGTGT-3′) and reverse (5′-GCAGAGACTCCAATCAACTGGCTCAA-3′). Taqman probe was used (at 250 nM; 5′-6-carboxyfluorescein-CTGGCACGGATCGAAGCAGGTCTAC-6-carboxytetramethylrhodamine-3′). PCR was performed with the following cycle parameters: 95°C for 3 min, 40 cycles at 95°C for 30 seconds, 58°C for 30 seconds, and 72°C for 30 seconds, and then 95°C for 1 min. A standard curve was generated using a spliced E0-E1-E2 PCR product amplified from TPA-treated A20-HE2 cDNA under the conditions described below for MHV68 E0-E2 reverse transcription-PCR (RT-PCR) cloned into the pGEMT-Easy vector (Promega). Quantitative RT-PCRs (qRT-PCRs) were performed in a Becton-Dickinson iCycler and analyzed using Bio-Rad iCycler software.

MHV68, EBV, and KSHV G50/RTA transcript analysis.

EBV and KSHV genomic sequence (GenBank accession nos. NC007605 and U75698) were examined for the presence of putative splice donor sites. RNA was prepared from treated or untreated cells using Trizol reagent as described above. cDNA synthesis was performed with random hexamers using SuperScript II reverse transcriptase (Invitrogen). All reactions were performed using GoTaqFlexi DNA polymerase (Promega) in a PCR mix of 1× GoFlexi buffer, 1.5 mM MgCl₂, deoxynucleotide triphosphates (0.2 mM each), and forward and reverse primers (0.2 μM each) with cycling parameters of 94°C for 5 min, 30 cycles at 95°C for 30 seconds, annealing at the indicated temperature for 30 seconds, and extension at 72°C for 1.5 min, followed by a final extension at 72°C for 7 min.

For MHV68, the E0-E1 reaction used forward primer 5′-CCAGGTCATCAAGGGTCCAAATACTC-3′ and reverse primer 5′-GGAATCCGAGTCAGAGGCCAT-3′; annealing was at 59°C. For the E0-E2 reaction, the forward primer was the same as for the E0-E1 reaction and the reverse primer was 5′-GTCTGGTGGGATGTTGATGGCGT-3′; annealing was at 58°C. For the E1-E2 reaction, the forward primer was 5′-CTGCTGGCAACCACCACCTTCA-3′ and the reverse primer was the same as for the E0-E2 reaction; annealing was at 58°C.

For EBV, the forward primer 5′-CCATGGGTGATAACGTCCTGAACG-3′ and reverse primer 5′-GCCCGTCTTCTTACCCTGTTGTTTCG-3′ were used; annealing was at 54°C.

For KSHV, the forward primer was 5′-GGCAAGCAAGCTGGTGTTCTGGAT-3′ and the reverse primer was 5′-CCTCCGATTGCAGACGAGTCG-3′; annealing was at 56°C. PCR products were visualized by ethidium-gel electrophoresis and cloned and sequenced as above.

Mice, organ harvest, and tissue preparation.

Female C57BL/6 mice (The Jackson Laboratory) and Dnmt3b conditional knockout mice (43) were housed at the Yerkes or Whitehead vivarium in accordance with university and federal guidelines. C57BL/6 mice 8 to 12 weeks of age or Dnmt3b conditional knockout mice 3 to 4 months of age were inoculated intranasally with 1,000 PFU of either wild-type MHV68, MHV68.G50pKO, or MHV68.G50pKO.MR (C57BL/6 infections) or MHV68-Cre.MR (Dnmt3b conditional knockouts) (28) virus in 20 μl cMEM following isofluorane anesthetization. Mice were sacrificed by asphyxiation or isofluorane inhalation and cervical dislocation. PECs were harvested by peritoneal lavage with 10 ml cMEM, and spleens were harvested and splenocytes prepared by homogenization and treated with Tris-ammonium chloride for red blood cell elimination as described previously (13). PECs or splenocytes were immediately used for reactivation analyses, genomic DNA isolation, or stored in cMEM-10% dimethyl sulfoxide at −80°C until prepared for quantitative or limiting dilution PCR analyses.

Ex vivo limiting dilution assays.

To determine the frequency of genome-positive cells in preparations of splenocytes or PECs, serially diluted cells were subjected to nested PCR using primers to detect G50 as previously described (13). To determine the frequency of cells reactivating from latency, single-cell suspensions of splenocytes or PECs from mice at days 16 to 18 postinfection were plated in twofold serial dilutions onto MEFs in 96-well plates as previously described (13). At day 21 postplating, each well was assessed for the presence of cytopathic effect (CPE). To determine the frequency of cells reactivating, the percentage of wells with CPE at each dilution was used in a nonlinear regression analysis to calculate the frequency of reactivation per cell by Poisson distribution. Disrupted splenocytes and PECs were also plated in parallel as previously described to determine the contribution of preformed infectious virus to reactivation (13).

Quantitative PCR to determine frequencies of viral genomes.

Quantitative PCR was performed as described by Moorman et al. (26). Briefly, DNA was extracted from splenocytes using the Qiagen DNeasy kit. DNA was quantitated in a fluorometer, and 0.5 μg of DNA was included in each PCR mixture. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)-specific PCR was also performed on each sample to control for variation in input DNA. A two-step PCR was used which consisted of the following steps: 2 min at 50°C, 10 min at 95°C, and 50 cycles of 95°C for 15 s, followed by 60°C for 30 s. The same PCR conditions were used for both the ORF50 and GAPDH reactions. The number of copies of ORF50 and GAPDH in each sample was determined by comparison to a series of standard curve reactions using a plasmid control containing appropriate sequences. The standard curve dilutions used represented a range from 10⁸ to 10¹, in serial 10-fold dilutions, and were performed in a background of 0.5 μg of splenic DNA from naive mice. All real-time PCRs were performed on a Becton Dickinson iCycler and analyzed in a minimum of two reactions. Data points represent the log square means of the replicate samples.

Genomic DNA isolation and bisulfite modification.

Splenocytes were incubated for 12 to 18 h at 50°C in 100 mM NaCl, 10 mM Tris-Cl, 25 mM EDTA, 0.5% sodium dodecyl sulfate, and 0.1 mg/ml proteinase K, followed by phenol-chloroform extraction and ethanol precipitation. Genomic DNA (10 to 20 μg) was digested for 12 to 18 h with HindIII and ethanol precipitated. Digested DNA (1 to 2 μg) in a volume of 50 μl was denatured at 37°C for 10 min in 0.2 M NaOH prior to bisulfite modification. Modification was performed by the addition of 30 μl of 10 mM hydroquinone, 520 μl of 3 M sodium bisulfite, and a mineral oil overlay and 16-hour incubation at 50°C. Modified DNA was purified using the GeneCleanII system (QBiogene), eluted in 50 μl double-distilled H₂O, and desulfonated by incubating in 0.3 M NaOH for 5 min at 37°C. DNA was ethanol precipitated and stored at −80°C until use in PCR.

Bisulfite PCR analyses.

Bisulfite-modified DNA was amplified using AmpliTaqGold DNA polymerase (Applied Biosystems). Each nested and heminested PCR mixture contained 1× Amplitaq Gold buffer, 3 mM MgCl₂, deoxynucleoside triphosphates (0.2 mM each), and forward and reverse primers (0.2 μM each). All reactions were performed in a Becton Dickinson iCycler with the following parameters: 95°C for 10 min (hot start); 30 cycles of 95°C for 30 seconds, annealing at the indicated temperature for 30 seconds, and 72°C for 1 min; followed by a final extension at 72°C for 7 min. For round 1, 2 μl of bisulfite-modified DNA was used in each 25-μl reaction mixture, and annealing was performed using a temperature gradient in four 3.2°C increments from 42 to 55°C for 30 cycles. Round 1 products were combined, and 1 μl was used in 25-μl round 2 reaction mixtures.

For the distal promoter region round 1, the forward primer 5′-ATGATGATTTATTAAAGAATTATGTTTT AGGT-3′ and reverse primer 5′-CAACCTCACCAACTTTTACAATAAATA-3′ were used. For round 2, the forward primer was the same as for round 1 and the reverse primer was 5′ CCCTTAATAACCTAATAAAAAACC CAATA-3′. Round 2 annealing was performed at 50°C with 30 cycles. For the proximal promoter region round 1, the forward primer was 5′-GTTGGTGAGGTTGGGAAGTTAT-3′and the reverse primer was 5′-CTCTCTCCTCAACCTTTAAAAAAACAT-3′, with reaction conditions as above. For round 2, the forward primer was 5′-GAATAGAAGGTAATTTTTTGAATAGAGT-3′ and the reverse primer was 5′-CTACCAAATTTCCATAAAAATAAAAAACTA-3′. Round 2 annealing was performed at 53°C with 30 cycles. PCR products were visualized by ethidium gel electrophoresis, purified using the GeneCleanII system as above, and ligated into the pGEMT-Easy vector (Promega). Plasmid DNA was sequenced by Macrogen USA (Rockville, MD) and analyzed using VectorNTI AlignX software (Invitrogen).

RESULTS

A recombinant MHV68 with a deletion spanning the known gene 50 promoter is replication competent in vitro.

During the generation of a gene 50 null virus (G50.Stop), the ability to produce mutant virus stocks was confounded by the strict dependence of MHV68 on gene 50 expression for virus replication (48). To overcome this problem, the G50.Stop mutant was propagated on a stable cell line harboring a plasmid containing an intact copy of MHV68 gene 50 under the control of the human cytomegalovirus IE promoter (30). However, the presence of a wild-type copy of gene 50 within the complementing cell line led to the generation of recombinant viruses harboring intact gene 50 (revertants), a problem that was exacerbated by the slow growth of the G50.Stop mutant on the complementing cell line. As such, any revertants that were generated had a significant growth advantage and quickly dominated the culture. In an attempt to circumvent this problem, we generated an MHV68 mutant that contained a 183-nt deletion (Δ66401-66584), removing the entire core promoter sequence previously shown to drive G50 transcription (19) (Fig. 2). It was expected that this mutant virus, termed G50pKO, would be phenotypically identical to the G50.Stop virus (i.e., replication null) due to its inability to drive G50 expression. Furthermore, growth of this mutant on the gene 50-complementing cell line would not lead to the generation of wild-type revertants, since the RTA expression construct in this cell line lacked any gene 50 promoter sequences.

FIG. 2.

Generation and confirmation of MHV68.G50pKO and MHV68.G50pKO-MR viruses. (A) Schematic depicting the organization of open reading frames in the G50 region. The upright arrow indicates the relative position of the characterized G50 promoter, and the hatched area indicates the region of the deletion in the MHV68.G50pKO virus. The indicated splicing of E1 to E2 gives rise to a full-length G50-encoding transcript. Wild-type sequence contains HindIII sites at bp 64898 and bp 57541; the MHV68.G50pKO BAC contains an additional HindIII site at 66549 (indicated by the asterisk). (B) HindIII digest of MHV68 wild-type, G50pKO, and G50pKO-MR virus BACs. The arrows indicate the fragments produced by the introduction of the additional HindIII site in panel A. (C) Southern blot hybridization using the PCR-generated probe indicated in panel A.

During the propagation of the G50pKO virus, we unexpectedly observed that not only was the mutant virus able to replicate in noncomplementing NIH 3T12 fibroblasts, it was able to do so at appreciable levels compared to wild-type virus (Fig. 3). In single-step growth analyses (MOI, 10), the MHV68G50pKO virus exhibited a mild defect in both the initiation and magnitude of lytic replication compared to wild-type virus, which resulted in a nearly 2-log defect in viral replication at late times postinfection (Fig. 3A). Multistep viral growth curves (MOI, 0.1) revealed a similar defect, with delayed initiation of replication and a ca. 2-log deficit in viral titers at late times postinfection (Fig. 3B). These analyses revealed that, although the G50pKO virus is attenuated for lytic replication compared to wild-type virus at both high and low MOIs, deletion of the previously defined core G50 promoter does not result in a replication null phenotype. This suggested the presence of a second promoter capable of driving gene 50 expression in the absence of the characterized promoter.

FIG. 3.

MHV68.G50pKO virus replicates in vitro. (A) Single-step growth analysis. NIH 3T12 cells were infected with wild-type or MHV68.G50pKO virus. Cells were infected at an MOI of 10 and harvested at the indicated times postinfection for determination of viral titers. (B) Multistep growth analysis. Cells were infected at an MOI of 0.1 and harvested at the indicated times postinfection.

Identification of alternatively initiated gene 50 transcripts.

Previous analyses demonstrated that MHV68 RTA, like RTA in other type 2 gammaherpesviruses, is encoded by a transcript generated from the splicing of a short exon, E1, to a long downstream exon, E2 (19-22, 40, 46) (Fig. 1). The observation that the G50pKO mutant was capable of replicating in permissive fibroblasts argued for the presence of an alternatively initiated gene 50 transcript. 5′-RACE analyses of RNA prepared from NIH 3T12 cells infected with the G50pKO virus, using a primer in exon 2, revealed a distinct transcription initiation site mapping upstream of the previously identified gene 50 transcription initiation site (Fig. 4A). The upstream initiated transcript gave rise to a spliced transcript containing an extended first exon (Fig. 4A). To assess whether this upstream initiation site was utilized by wild-type MHV68, 5′-RACE was performed with RNA prepared from NIH 3T12 fibroblasts infected with wild-type MHV68. These analyses revealed that this transcription initiation site is utilized by wild-type virus and gives rise to a novel spliced gene 50 transcript containing three exons: a short exon, which we have termed E0, that splices to the previously identified exon 1, which in turn is spliced to exon 2 (Fig. 4A). Notably, the splice acceptor site for the E0/E1 splice is not present in the G50pKO virus due to the deletion introduced, and as such the E0/E1 splice was not observed in cells infected with this mutant virus.

FIG. 4.

RACE and RT-PCR analyses identify an additional upstream G50 exon. (A) RACE analyses were performed using cDNA generated from G50pKO-infected 3T12 or A20-HE2 cells 24 hours after treatment with TPA. The 5′ end of E0 was mapped using the primer located in exon 2 (indicated at nt 68071) for 3T12 analyses (top), and the splice junction-spanning primer II was used for A20-HE2 analyses (bottom). Both experiments identified the 5′ terminus for E0 at nucleotide position 65909. The E0-E1-E2 spliced product is depicted. (B) RT-PCR analyses of A20-HE2 cells treated with TPA. cDNA was prepared from A20-HE2 cells either untreated or at the indicated time after TPA treatment and used in PCR analyses to detect E0-containing transcripts generated upon reactivation. The positions of the forward and reverse primers used in each of the three panels are indicated. Products a and e represent unspliced transcripts, and product c represents a partially spliced transcript containing unspliced E0-E1 sequence and spliced E1-E2; the spliced transcripts are indicated by asterisks. Products b, d, and f are artifacts of the amplification reaction and represent hybrid species (heteroduplexes) formed between RT-PCR products arising from spliced and unspliced/partially spliced transcripts (verified in separate PCRs using purified plasmids containing either the fully spliced and/or unspliced/partially spliced templates; data not shown). Note that these transcripts are derived from a wild-type G50 locus with both distal and proximal promoters intact.

We extended these analyses to assess whether the alternative gene 50 transcripts were present during MHV68 reactivation from latency. For these analyses we utilized a recently characterized MHV68 latently infected A20 B-cell line (A20-HE1) (11) treated with TPA to induce virus reactivation. 5′-RACE analyses were conducted on cDNA from A20-HE1 cells 24 h after treatment with TPA; to eliminate contamination from unspliced message, amplification of A20-HE1 cDNA was performed with a genome-specific primer encompassing the predicted E0-E1 splice junction (Fig. 4A). This analysis also identified a transcriptional start site at nt 65909, indicating that the E0-encoding transcript initiating upstream of E1 is produced both during infection of permissive fibroblasts as well as during TPA-induced reactivation from latently infected B cells. Importantly, the E0-containing spliced transcripts utilize a consensus splice donor site [(A/C)AGGT(A/G)AGT]) at the 3′ end of the E0 exon (nt 66088) and a consensus splice acceptor site ([(Y)_n(C/T)AG/G]) located just upstream of the previously identified proximal gene 50 transcription start site (E1 exon transcriptional start site at nt 66494). Notably, the E0/E1 splice product could readily be detected by RT-PCR with RNA prepared from A20-HE1 cells treated with TPA, employing an upstream primer located in the E0 exon and downstream primers located in either the E1 or E2 exons (Fig. 4B). In addition to the expected splice product (Fig. 4B), unspliced and partially spliced products were also detected (for details see the legend to Fig. 4B). Taken together these data demonstrate that in addition to the E1/E2 spliced gene 50 transcript produced by a promoter proximal to the E1 exon, there is an additional gene 50 promoter upstream of the E1 exon which gives rise to a spliced transcript containing the E0, E1, and E2 exons. This promoter is utilized during virus replication in permissive fibroblasts and also during virus reactivation from B cells.

Identification of promoter activity mapping in the region immediately upstream of the E0 exon.

To validate the existence of a bona fide promoter capable of producing E0-containing transcripts, we examined the region immediately upstream of the E0 transcriptional start site for promoter activity. Fragments containing the upstream region were cloned into the pGL4.10 luciferase reporter vector, and the resulting reporter constructs were transfected into either NIH 3T12 fibroblasts or the murine macrophage cell line RAW 264.7. While the −100bp fragment conferred little or no promoter activity, the −250bp fragment exhibited a 20-fold increase in activity over empty vector in NIH 3T12 fibroblasts, compared to a more modest 5-fold increase in unstimulated RAW 264.7 macrophages (Fig. 5). In both NIH 3T12s and RAW 264.7 cells, the inclusion of additional upstream sequences repressed the activity observed with the −250bp fragment alone. In 3T12s, the −1000bp fragment exhibited only ∼3-fold higher activity than empty vector, and while this fragment still had some activity in stimulated macrophages (∼18-fold), it was greatly reduced compared to the −250bp fragment. This suggests the existence of a repressive element(s) in the more distal region of the E0 promoter that can attenuate activity of the distal promoter in these cell types. NIH 3T12 fibroblasts are a permissive cell line for MHV68 infection, while macrophages have been identified as a reservoir for latent MHV68 infection that are capable of reactivating virus in response to various stimuli. To determine if macrophage activation would increase promoter activity, we added LPS to the RAW 264.7 cultures for 24 h prior to harvest. This stimulation led to a 30-fold increase in promoter activity with the −250bp reporter construct (ca. 150-fold over empty vector), −500bp construct (ca. 30-fold over empty vector), and −1000bp constructs (ca. 15-fold over empty vector). These data demonstrate that the region upstream of the MHV68 gene 50 E0 exon displays significant promoter activity under conditions which are known to support lytic viral replication and gene 50 expression. In addition, since these transfections were performed in the absence of viral infection, the observed activity is independent of the expression of any viral proteins, and thus suggests a possible role for this promoter in MHV68 reactivation from latency in stimulated macrophages.

FIG. 5.

Promoter activity in the region immediately 5′ to MHV68 E0. NIH 3T12 or RAW 264.7 cells were cotransfected with phRL-Luc (Renilla luciferase) and pGL4.10[luc] luciferase reporter constructs containing either 100-, 250-, 500-, or 1,000-bp fragments immediately upstream of E0. RAW 264.7 cells were stimulated 24 h after transfection with LPS (5 μg/ml). Luciferase assays were performed 48 h after transfection, and data are presented as the fold difference in the ratio of firefly:Renilla luciferase versus the empty vector control. Data are representative of three independent transfections.

E0-containing transcripts are of low abundance relative to total gene 50 transcripts during permissive infection and reactivation from a latently infected B-cell line.

The identification of a second gene 50 promoter producing E1-E2 spliced transcripts raised questions regarding the relative abundance and kinetics of gene 50 transcripts arising from the distal versus the proximal gene 50 promoter. To address this issue, we used quantitative RT-PCR to determine the contribution of distal promoter-driven E0-E1-E2 transcripts to total E1-E2 transcripts in RNA from A20-HE2 cells at various times post-TPA treatment, as well as permissive NIH 3T12 fibroblasts 24 h postinfection with wild-type MHV68. To detect total E1-E2 transcripts, we used primers in E1 and E2 to amplify cDNA from TPA-treated A20-HE2s and infected NIH 3T12 cells and quantitated the amplification product using a Taqman probe specific for a region within the E2 exon. At 0 h postinduction, very few copies of spliced E1-E2 transcripts were detected in the latently infected A20-HE2 cells, verifying the relatively low level of spontaneous reactivation previously seen in these cells (11) (Fig. 6). TPA treatment resulted in a rapid accumulation of E1-E2 spliced transcripts by 4 h postinduction, which peaked between 8 and 12 h postinduction kinetics in keeping with the rate of viral replication seen previously upon TPA stimulation of this cell line and the classification of gene 50 as an immediate-early gene. E1-E2 spliced transcripts were also detected in NIH 3T12 fibroblasts 24 h after MHV68 infection. This amplification reaction could also theoretically detect unspliced or antisense transcripts, an issue of concern given the high degree of complex, bidirectional transcription known to occur in the G50 region. These species could inflate the quantitation of E1-E2 spliced transcripts arising from the distal or proximal promoter. However, when visualized by gel electrophoresis, only a single 218-bp amplified product corresponding to the spliced E1-E2 transcript was observed, while distinct higher molecular weight species representing unspliced or antisense transcripts were not seen (data not shown). We therefore concluded that spliced E1-E2 transcripts comprise the majority of E2 exon-containing transcripts detected in this assay, and any contribution by other transcript species was considered negligible for the purposes of this assay.

FIG. 6.

Quantitative RT-PCR analysis of distal versus proximal promoter-driven transcripts. Relative copy numbers of E1-E2 versus E0-E1-E2 transcripts in cDNA from TPA-treated A20-HE2 or 3T12 cells infected with wild-type MHV68. Mean copy number was calculated from two independent experiments comprised of at least two independent cDNA synthesis reactions. Each qRT-PCR was performed in triplicate, and data are presented as means for at least four independent cDNA preparations with calculated standard errors of the means. nd, not detected.

To detect E0-E1-E2 spliced transcript, generated from the distal gene 50 promoter, we used the same Taqman probe within the E2 exon, as described above, but with primers in E0 and E2. Unlike the readily quantifiable E1-E2 spliced transcripts, spliced E0-E1-E2 transcripts were not detected at early times following TPA induction of A20-HE2s and only sporadically detected in two of five separate cDNA preparations from these cells at 24 h. E0-containing transcripts were also undetectable at 48 h, arguing against the possibility of the distal promoter exhibiting strong late promoter kinetics (data not shown). E0-E1-E2 spliced transcripts were detected at variable levels in NIH 3T12 cells 24 h after infection but at less than 1% of the level observed for E1-E2 spliced transcripts from the same cDNA preparation. The larger size of the E0-E1-E2 amplicon versus E1-E2 (218 versus 377 bp) was most likely not responsible for the discrepancy in copy number, as standard curve correlation coefficients and PCR efficiency were generated using the same plasmid containing E0-E1-E2 cDNA and were very similar for both reactions (data not shown). Notably, previous attempts to detect E0-containing transcripts by Northern blotting, using a splice junction-specific oligonucleotide probe complementary to the E0-E1 splice junction, were unsuccessful even when performed with significant quantities of purified mRNA. The scarcity of spliced E0-containing transcripts in both reactivating A20-HE2 and permissively infected NIH 3T12 cells suggests that spliced G50 transcripts generated from the distal promoter are significantly less abundant than those generated from the proximal promoter during both reactivation from latency and permissive infection. It should be noted that, although in low abundance, transcripts originating from the distal promoter were detected in NIH 3T12 cells infected with wild-type MHV68. This corroborates the distal promoter activity seen with the NIH 3T12 reporter.

Evidence for a conserved distal RTA promoter in human gammaherpesviruses.

Although sequence variation exists among both type 1 and type 2 gammaherpesviruses, the organization and function of lytic genes in the region encoding gene 50 are highly conserved (Fig. 1). The expression of E1/E2 spliced transcripts under the control of an E1 exon-proximal promoter is observed in both type 1 (EBV) and type 2 (MHV68, KSHV, and HVS) gammaherpesviruses and appears to represent a conserved strategy for expression of gene 50. In all four characterized gammaherpesviruses, open reading frames are arranged on opposite coding strands in a slightly overlapping orientation as illustrated in Fig. 1. The 5′ end of ORF 48/BRRF2 is positioned adjacent to the 5′ end of the E1 exon, and ORF 49/BRRF1 lies between the E1 and E2 exons in the opposite orientation. The high degree of organizational homology in this region led us to investigate whether the human gammaherpesviruses also express an alternatively initiated gene 50 transcript(s). To do this, we adopted the strategy outlined in Fig. 7A: the region 1,000 bp upstream of the characterized transcriptional start site of both the KSHV and EBV E1 exons was examined for potential splice donor sites [using the consensus sequence (A/C)AGGT(A/G)AGT, which would represent the 3′ terminus of an E0 exon], as well as for potential splice acceptor sites [using the consensus sequence (Y)_n(C/T)AG/G] near the 5′ end of the defined E1 exon. We then designed primers to flank these potential splice donor and acceptor sites and used them to perform RT-PCR on RNA from KSHV- (BCBL-1) or EBV-infected B-cell lines (Clone-13; Akata) treated with TPA or anti-IgG, respectively, to induce virus reactivation. RT-PCR products were gel purified, cloned, sequenced, and analyzed for the presence of spliced gene 50 transcripts.

FIG. 7.

Identification of upstream-initiated transcripts from treated EBV or KSHV latent cell lines. (A) Strategy to identify putative E0-containing transcripts from EBV and KSHV latent cell lines. To detect putative E0-containing transcripts, RNA was prepared from untreated cells or cells treated with reactivating stimuli for the indicated time. cDNA was generated and amplified with a forward primer positioned upstream of the G50/Rta/BRLF1 exon 1-proximal promoter and reverse primer in E1 or E2. The indicated products were cloned and sequenced to confirm the identity of spliced upstream-initiated transcripts. (B) cDNA generated from Akata cells treated with anti-IgG was used for PCR with the indicated primers. The indicated products are schematically represented. (C) cDNA generated from BCBL-1 cells treated with TPA for the indicated times was used for PCR with the indicated primers. The exon structures of the amplified products are schematically represented.

Treatment of the latent EBV-positive Burkitt's lymphoma cell line Akata with antiimmunoglobulin led to the identification of two distinct transcript species (products A and B in Fig. 7B) utilizing the same splice acceptor site in the BRLF1 coding exon, but two distinct splice donor sites for splicing the E0 exon. Sequence analysis revealed that both products initiated at nt 94954 (F primer binding site), contained a portion of the putative E0 sequence, spliced to the E2 acceptor site at nt 92897, and ended at nt 92615 (R primer binding site). However, product B utilized a splice donor located at nt 94862, while the product A splice utilized a splice donor located at nt 94830. A third larger amplification product (product C) was identified that contained sequence from nt 94954 to nt 94693 and from nt 93371 to nt 92615 (R primer binding site). Examination of the splice junction present in product C revealed that this corresponds to a spliced antisense transcript which contains a portion of ORF49 (BRRF1) and ORF48 (BRRF2). Finally, a 1.4-kb product (product D) was observed that corresponds to a partially spliced transcript (E0/E1-E2 product) in which the intron between exon 1 and exon 2 has not been removed. (Note that this same partially spliced product was detected following RT-PCR from TPA-treated A20-HE2s [Fig. 4B].) As shown, the presence of the observed amplified products was dependent on both antiimmunoglobulin stimulation of virus reactivation and the inclusion of reverse transcriptase in the cDNA reaction (Fig. 7B). An analysis of potential translation products from the transcript corresponding to product B, an E0/E2 exon spliced transcript, revealed a possible extension of the BRLF1 open reading frame (Fig. 8B). This would be predicted to add a unique 31 amino acids (aa) extension at the N terminus of RTA. Thus, one possible consequence of initiating from the distal promoter during EBV infection may be the generation of a functionally distinct form of RTA. Future studies will be required to assess whether this predicted product is expressed and whether it has an altered function(s).

FIG. 8.

Exon 0 extends the EBV and KSHV G50 reading frames. (A) Translation of the observed KSHV transcripts using an ATG in exon 0 (gray) provides an additional 31 amino acids to the N terminus of the EBV BRLF1 exon 2 reading frame (black). (B) Translation of the observed EBV exon 0-containing transcript A. Exon 0 and exon 1 (gray) both extend the Rta exon 2 reading frame (black).

To assess whether alternatively initiated gene 50 transcripts were also expressed during KSHV reactivation, RNA was prepared from the KSHV latently infected BCBL-1 cells 48 and 56 h after treatment with TPA. As depicted in Fig. 7C, a forward primer positioned upstream of a putative splice donor site and a reverse primer in the gene 50 E2 exon identified a spliced transcript (product A) spanning nt 70436 (F primer binding site) to nt 70728 and nt 72572 to nt 72722 (R primer binding site). The presence of consensus splice donor and acceptor sites at nt 70728 and nt 72572, respectively, identifies this transcript as an authentic spliced transcript, rather than an artifact of the amplification reaction. The splice acceptor site at nt 72572 is the same as that used to generate the E1/E2 spliced gene 50 transcript (20), confirming the validity of the junction produced between the putative E0 exon and the E2 exon. Another product (product B) was also detected that shared the same splice acceptor site as product A but utilized a noncanonical splice donor at nt 70823. The presence of both bands following RT-PCR was dependent on the inclusion of reverse transcriptase in the cDNA reaction (data not shown). The identification of upstream-initiated spliced transcripts containing the gene 50 E2 exon suggests that, like MHV68, the human rhadinovirus KSHV also encodes an additional distal gene 50 exon and promoter capable of generating full-length gene 50 transcripts. Importantly, unlike MHV68, the identified transcripts lack the E1 exon, raising the question of whether they give rise to a distinct RTA species. Translation of the lower-molecular-weight product A in Fig. 7C revealed the presence of a putative initiating methionine that would generate a product in frame with ORF50 and would thus be predicted to give rise to a functional RTA product (Fig. 8A). The E0 exon is predicted to encode 10 aa at the N terminus of RTA, while the E1 exon of E1/E2 spliced transcripts contributes 6 aa at the N terminus of RTA (Fig. 8A). The inclusion of the additional sequence between nt 70728 and nt 70823 introduces stop codons, and thus translation of the longer product B would not be predicted to encode an extended RTA product.

The proximal gene 50 promoter knockout mutant, G50pKO, establishes splenic latency in vivo but is severely impaired for virus reactivation.

Evidence of a distal gene 50 promoter capable of driving RTA expression prompted us to evaluate the relevance of the distal gene 50 promoter during virus infection in vivo. C57BL/6 mice were infected intranasally with 1,000 PFU of either wild-type MHV68, the G50pKO virus, or a genetically repaired marker rescue virus, G50pKO.MR. Levels of latent infection in splenocytes were determined by measuring the amount of viral genome present in the spleen at day 16 postinfection by quantitative PCR. The viral genome load was very similar between wild-type-, G50pKO-, and G50pKO.MR-infected mice, with slightly lower levels observed in mice infected with the G50pKO (Fig. 9A). We subsequently confirmed, by using a limiting dilution analysis, that the G50pKO robustly establishes latency but at a slightly lower level than wild-type or marker rescue virus (data not shown). Since we have previously shown that expression of RTA is required for establishment of latent infection in the spleen following intranasal infection (27), the presence of virus genome-positive splenocytes in G50pKO-infected mice therefore corroborates the in vitro observations that, in the absence of the proximal gene 50 promoter, the distal gene 50 promoter is capable of driving sufficient levels of RTA expression to allow virus replication.

FIG. 9.

G50pKO virus establishes latency in vivo but exhibits a severe reactivation defect. (A) Real-time PCR analysis of splenocytes from mice at day 16 following infection with the indicated virus. Establishment of latent infection was determined by quantification of relative G50 copy number (compared to GAPDH) in splenocytes from infected animals. (B) Ex vivo reactivation analysis to determine the frequency of cells reactivating from latency upon explant to MEF monolayers. Reactivation was scored by the presence of CPE.

In addition to the presence of genomes in latently infected cells, another parameter for assessing latent infection is to determine the frequency of cells able to reactivate from latency and induce CPE when plated in serial dilutions onto MEF monolayers. Remarkably, despite harboring nearly wild-type frequencies of viral genomes, splenocytes from G50pKO virus-infected animals displayed a complete inability to reactivate from latency (Fig. 9B). Splenocytes from both wild-type- and G50pKO.MR-infected mice reactivated at a frequency of approximately 1 in 6,200 and 1 in 10,300, respectively, while no detectable CPE was observed on MEFs cultured with splenocytes from G50pKO-infected mice (Fig. 9B). Notably, no preformed infectious virus was detected in these splenocyte samples following mechanical disruption of the cells (data not shown). The G50pKO reactivation defect was not related to route of infection, as the same phenotype was observed upon intraperitoneal infection with an equivalent viral dose (data not shown). That the G50pKO mutant can establish latency at a relatively normal frequency, yet is completely unable to reactivate ex vivo, indicates that the distal G50 promoter can compensate for the proximal promoter during the establishment of latent infection in vivo, but not during reactivation from latently infected splenocytes, and corresponds to a defect in reactivation from B cells, the predominant latency reservoir in the spleen.

The distal gene 50 promoter is progressively methylated in splenocytes during in vivo infection and in latently infected B-cell lines.

The reactivation defect exhibited by the G50pKO mutant suggests that distal gene 50 promoter activity may be differentially regulated during reactivation from latency and primary lytic infection. This temporal discrepancy in promoter function suggested a possible role for epigenetic regulation of distal G50 promoter activity. A wealth of evidence exists supporting a role for DNA methylation in the regulation of both latent and lytic gene expression in KSHV and EBV (4, 7, 17, 18, 24, 29). To determine if DNA methylation regulates the MHV68 proximal and distal G50 promoters, we used bisulfite PCR analyses to examine the methylation status of CpG dinucleotides within the proximal and distal gene 50 promoters in viral genomes recovered from MHV68 latently infected B-cell lines, as well as latently infected splenocytes recovered from mice at various times postinfection. Remarkably, the 110-nt gene 50 proximal promoter extending from nt 66442 to nt 66552 contains no CpGs but is flanked by two CpGs located at nt 66430 and 66555. In contrast, the −250 region of the putative distal promoter and the 5′ end of E0 contains five CpGs (Fig. 10). In splenocytes from wild-type MHV68-infected animals at day 16 postinfection, the two sites in the proximal promoter region were unmethylated in all clones examined, while a significant number of the CpGs in the distal promoter were methylated. At day 42 postinfection, the percentage of methylated CpGs increased in both regions, but to a greater degree at the distal promoter sites (18% in proximal versus 42% in distal). At day 90 postinfection, the percentage of methylated CpGs had increased to 80% in the distal promoter region but remained nearly equivalent to day 42 levels in the proximal promoter region (20%). This partitioning was also readily apparent in viral genomes present in MHV68 latently infected B-cell lines. In the MHV68-positive lymphoma cell line S11E, hypermethylation of the CpGs in the distal gene 50 promoter was apparent, along with consistent methylation of the CpG upstream of the core proximal gene 50 promoter. This same pattern of methylation was also observed in the two independently generated A20 B-cell lines latently infected with an MHV68-hygromycin-EGFP virus (A20-HE1 and A20-HE2). These analyses demonstrate that not only does viral CpG methylation accumulate in the gene 50 region throughout the duration of latent viral infection, it is more concentrated in the distal than in the proximal gene promoter region. Methylation of CpGs in promoter regions is most often associated with silencing of the associated gene (25). Given that methylated CpGs accumulate in the distal promoter region during latent infection, we hypothesize that methylation-induced silencing of the distal gene 50 promoter reflects a general lack of distal promoter usage in latently infected splenocytes and may contribute to the severe reactivation defect observed in latently infected splenocytes from G50pKO-infected animals. The absence of significant methyl-CpG accumulation in the proximal promoter regions suggests that this promoter is refractory to DNA methylation and may therefore retain functional capacity in latently infected splenocytes, even at late times postinfection.

FIG. 10.

Bisulfite PCR analysis of CpG methylation in regions containing the proximal and distal G50 promoters. Bisulfite-treated DNA from MHV68-infected splenocytes at the indicated times postinfection or from MHV68-positive latent cell lines was amplified, cloned, and sequenced to determine the frequency of methylated CpGs in the distal versus proximal promoter regions. A circle represents a CpG dinucleotide and the genomic position is indicated above each column. Open circles represent unmethylated cytosines, and filled circles represent methylated cytosines. Each row represents the sequence of an individual clone.

G50pKO virus reactivates from peritoneal exudate cells.

Given the pronounced reactivation defect of G50pKO virus-infected splenocytes, we questioned whether this mutant virus was capable of ex vivo reactivation from latently infected cells. Peritoneal exudates are comprised mostly of macrophages and dendritic cells, both of which have been shown to harbor latent MHV68 (10). PECs from animals infected intraperitoneally typically exhibit a slightly enhanced frequency of reactivation during early latency compared to splenocytes. We therefore examined PECs from mice infected with G50pKO virus following intraperitoneal infection to determine if the distal promoter was sufficient to drive reactivation from this cell population. Frequencies of viral genome-positive cells were reduced in splenocytes from G50pKO-infected animals relative to those infected with G50pKO.MR following intraperitoneal infection. This roughly 10-fold reduction in establishment is greater than would be expected given the slight reduction in viral genome copy number determined by qPCR following intranasal infection and may reflect a differential requirement for the proximal G50 promoter during establishment of latent infection in splenocytes following intraperitoneal infection. (Fig. 11A). Splenocytes from G50pKO.MR virus exhibited a similar frequency of reactivating cells as observed following intranasal infection, while PECs from G50pKO.MR-infected animals reactivated at a slightly higher frequency than splenocytes, as expected based on previous observations (15). The reactivation defect seen in splenocytes from G50pKO-infected animals following intranasal infection was also present following intraperitoneal infection (Fig. 11B). Importantly, the establishment defect was not evident in PECs from G50pKO-infected animals, as the frequency of viral genome-positive cells was equivalent to that in PECs from G50pKO.MR-infected animals (Fig. 11A). Remarkably, PECs from animals infected with G50pKO virus were able to reactivate from latency at a frequency nearly identical to that of G50pKO.MR splenocytes (Fig. 11B). This is in striking contrast to the complete absence of reactivation in G50pKO-infected splenocytes and demonstrates that the distal promoter is sufficient to drive both establishment of latency and virus reactivation from PECs. These data suggest that the distal gene 50 promoter is regulated in a cell-type-dependent manner and may reveal a specific role for the distal promoter in driving the establishment and reactivation of latent infection in macrophages or dendritic cell priypopulations.

FIG. 11.

G50pKO virus reactivates from peritoneal exudate cells following intraperitoneal infection. Female C57BL/6 mice 6 to 8 weeks of age were infected with 1,000 PFU of G50pKO or G50pKO.MR virus by intraperitoneal injection. Splenocytes and PECs were harvested at day 18 following infection and assessed for establishment of and reactivation from latency by limiting dilution assays. (A) Splenocytes and PECs were plated in serial dilutions and subjected to nested PCR to detect G50. The percentage of genome-positive cells in each dilution was used to calculate the frequency of genome-positive cells as in reference 13. (B) Splenocytes and PECs were plated in serial dilutions onto MEF monolayers as in Fig. 9. The percentage of wells exhibiting cytopathic effect in each dilution was used to calculate the frequency of cells reactivating from latency. Mechanically disrupted cells were plated in parallel for each virus and cell type and verified the absence of preformed infectious virus (data not shown). Data are representative of at least two independent experiments with at least four mice per group, and error bars were generated using the standard errors of the means.

DISCUSSION

Roles for alternatively initiated gene 50/BRLF1 transcription in gammaherpesvirus infection.

In this paper we demonstrate that MHV68 gene 50 transcription is more complex than previously characterized and involves transcription initiation from at least two independent promoters. This complexity appears to be shared among the murine and human gammaherpesviruses, and therefore most likely represents an evolutionarily conserved transcriptional strategy important to gene 50/BRLF1 transcription and successful gammaherpesvirus infection. However, pathogenesis experiments demonstrated that the distal gene 50 promoter was not sufficient to compensate for loss of the proximal gene 50 promoter during reactivation from latently infected splenocytes but was able to drive an attenuated frequency of reactivation in PECs relative to G50pKO.MR virus. This suggests that the two promoters have evolved to have distinct regulatory functions and that these functions are likely to be cell type specific.

The identification of an E0 exon upstream of E1 and E2 exons elaborates the characterized G50 coding sequence. In MHV68, E1 contains an ATG initiation codon which, when spliced to E2, extends the G50 open reading frame by an additional 94 amino acids. This mechanism is shared among the other type 2 gammaherpesviruses KSHV and herpesvirus saimiri (HVS), such that the utilization of the ATG in E1 as a translational start site produces an augmented RTA protein with a high level of sequence conservation at the amino terminus (40, 46). For EBV, a type 1 gammaherpesvirus, the E1/E2 spliced transcript contains the initiating ATG in the E2 exon, and as such the E1 exon is noncoding. However, the overall organization of gene 50/BRLF1 transcription still mirrors that of the type 2 gammaherpesviruses in that E1, although not providing any additional G50 coding sequence, is retained.

The splicing of KSHV RTA E1 and E2 provides an additional 6 amino acids at the N terminus of RTA and has been observed by several groups (21, 40). We were unable to detect a cDNA species containing the putative E0 exon spliced to the previously identified E1-E2 spliced transcript using this analysis. However, a transcript containing the putative E0 and E2 contains an ATG that, if used as the translational start site, encodes a protein with 10 unique N-terminal amino acids that splices in frame with the E2 reading frame previously characterized by the E1-E2 spliced product. We were also unable to detect a transcript containing the putative E0 exon spliced to the BRLF1 E1 exon in reactions from IgG-treated EBV-positive Akata cells. However, the EBV transcript using the upstream consensus splice donor site also contains an ATG within the putative E0 exon sequence presumably capable of translating an extended RTA protein that would extend the previously characterized BRLF1 reading frame. Analogous to the KSHV E0-E2 transcript, the addition of the EBV E0 exon sequence provides an additional 31 amino acids to the N terminus of RTA. This configuration may be conserved in MHV68 as well, as translating a hypothetical E0-E1-E2 spliced product using an ATG in E0 cannot produce a full-length amino acid sequence in any frame, while translating an E0-E2 product provides 14 additional amino acids from E0 that would splice in-frame with the previously identified E2 reading frame used by the E1-E2 spliced product.

The observation of an anti-IgG, reverse transcriptase-dependent product (C in Fig. 7B) in the Akata RT-PCR corroborates the complex bidirectional gene transcription in this area of the genome observed for both KSHV and EBV (8, 31) and perhaps represents a form of lytic gene self-regulation by the generation of antisense gene 50 transcripts during reactivation from latency. RT-PCR analysis was also performed in a second latent EBV-positive Burkitt's lymphoma cell line, clone-13, following treatment with TPA to induce reactivation. Only the longer E0-E2 species containing the splice site at nt 94830 was detected in RT-PCR analysis of this cell line; neither the shorter splice variant nor the spliced antisense transcript was detected despite a lack of any sequence variation between the two cell lines. This suggests that regulation of G50 splicing in the context of reactivation from latent EBV infection may be strain or cell type dependent.

The existence of multiple gene 50 promoters has been reported by Whitehouse et al. upon the observation that HVS uses two different promoters to drive transcription of two distinct RTA products (46). The promoter immediately upstream of E1 drives transcription of the spliced E1-E2 product which encodes the functional RTA protein, while a second promoter situated within E2 gives rise to an unspliced transcript encoding a truncated RTA product of unknown function. The activity of these promoters is temporally regulated, with the first having immediate-early kinetics while the unspliced transcript produced by the second promoter was sensitive to cycloheximide treatment. These observations suggest the intriguing possibility that different RTA promoters have unique kinetic profiles and produce distinct gene 50 transcripts. Given the lytic growth defect and reactivation defect of the proximal promoter knockout virus, it was not surprising to find that the distal gene 50 promoter generates significantly fewer spliced gene 50 transcripts than the proximal promoter. It may be that generation of spliced transcripts encoding functional gene 50 is only one of the roles for the distal gene 50 promoter. In addition to contributing to the pool of gene 50 protein-producing transcripts, it is possible that transcripts from the distal gene 50 promoter, either spliced or unspliced, play a role in regulating transcription from the proximal promoter following initiation of the lytic replication program. This may occur either through competition for transcriptional machinery or direct interference with transcription from the proximal gene 50 promoter, a mechanism used to explain promoter switching and exclusion of the “proximal” Wp promoter in favor of the “distal” Cp promoter during the establishment of EBV latency (32). Although no detectable levels of E0-E1-E2 spliced transcripts were detected at 48 h postinduction of A20-HE2 cells, the presence of E0-E1-E2 transcripts in 3T12 cells 24 h postinfection, along with low-level detection of E0-initiated transcript in A20-HE cells at 24 h postinduction at a time when E1-E2 transcripts appear to be decreasing, could still support the hypothesis that distal promoter-generated transcripts arise with delayed kinetics relative to proximal promoter-generated transcripts. These transcripts may in fact serve to modulate gene 50 expression in favor of dampening the lytic cascade during establishment or reactivation from latency. In the context of the G50pKO virus, it may be that distal promoter activity is increased or is sustained at low levels for such time that sufficient G50 transcripts accumulate to allow for lytic growth both in vitro and in vivo. Another possibility is that gene 50 transcripts initiating from the distal promoter serve as an alternative mechanism for modifying gene 50 expression at a distinct stage of virus infection or in distinct cell populations (e.g., macrophages). The ability of the G50pKO virus to reactivate from latently infected PECs but not splenocytes strongly supports this latter hypothesis. Finally, it is possible that altering the 5′ untranslated region present in the gene 50 transcript impacts either transcript stability or efficiency of translation initiation. None of these possibilities are mutually exclusive, and further studies on these alternative gene 50 transcripts and distinct forms of RTA will be required to determine their importance during gammaherpesvirus infection. The relationship between the previously characterized gene 50 promoter and the distal gene 50 promoter identified here may provide valuable insights into the regulation of gene 50 transcription in MHV68 as well as in the human gammaherpesviruses and requires further characterization.

CpG suppression and DNA methylation in gammaherpesviruses.

Gammaherpesviruses are unique among the herpesvirus family in that they exhibit overall CpG suppression, suggesting that their genomes have been subjected to extensive CpG methylation throughout their evolution within vertebrate hosts (14, 16). The MHV68 genome is particularly sparse in terms of CpG distribution, with only a 0.4 ratio of CpG:GpC dinucleotide frequency, compared to roughly 0.6 for EBV and 0.8 for KSHV. As was described for the region encompassing the KSHV lytic genes RTA and K8 (7), the locus containing MHV68 gene 50 is also CpG suppressed to a greater degree than the overall genome, with an overall CpG:GpC ratio of only 0.26. Even more remarkably, while the 250-nt region upstream of the MHV68 E0 exon (containing the core distal gene 50 promoter) exhibits a CpG:GpC ratio of 0.33, the ratio for the 250-nt region upstream of the E1 exon (containing the core proximal gene 50 promoter) is only 0.09. The 110 bp within this region previously shown to be required for proximal gene 50 promoter activity by reporter assays and for efficient reactivation from splenocytes in this study are completely devoid of CpGs. This may represent a characteristic viral adaptation such that the proximal promoter is refractory to this form of host-directed epigenetic silencing and thus more readily responsive to reactivation stimuli. The distal promoter region, though still suppressed relative to the entire MHV68 genome, has a higher frequency of CpGs than the ORF48/49/50 locus as a whole. This suggests that the few CpG dinucleotides that have been retained amid a presumed sea of cytosine nucleotide substitutions may represent key elements of MHV68 gene regulation and that their methylation status is likely of significant relevance to the control of MHV68 lytic gene expression.

During lytic replication in the early stages of G50pKO infection, when the majority of viral DNA is packaged in virions and therefore unmethylated, it appears that the distal gene 50 promoter drives RTA expression to sufficient levels to allow for the establishment of splenic latency at levels comparable to wild-type infection. However, as the viral genome is maintained in the relatively quiescent memory B-cell latency reservoir, it accumulates higher proportions of methylated CpGs. This progressive methylation of the viral genome has been seen at the latency promoters Wp and Cp during the generation of lymphoblastoid cell lines following EBV-induced transformation of B lymphocytes (42) and in seropositive individuals (29). With respect to RTA expression, reactivation of latent gammaherpesvirus infection has been shown to be accompanied by a loss of CpG methylation in the KSHV RTA promoter upon treatment of latent BCBL-1 cells with TPA (7). Similarly, treatment with the demethylating agent 5-azadeoxycytidine induces reactivation in both KSHV and EBV latently infected cell lines, presumably due to demethylation of key lytic gene promoters (2, 7). Biopsies from KSHV-positive tumors demonstrated that the majority of viral genomes in these samples contain unmethylated gene 50 promoters, suggesting that reactivation of latent KSHV either requires or results in loss of CpG methylation at lytic gene promoters (7).

In the context of wild-type MHV68 lytic infection, the core proximal gene 50 promoter is not subjected to methylated CpG-induced silencing and therefore may be required to drive G50 expression from the methylated latent state to induce reactivation and subsequent lytic replication. A possible scenario is that in a wild-type infection, rapid viral DNA synthesis induced upon reactivation overwhelms the capabilities of host methyltransferases and as such results in a passive “demethylation” of the distal promoter, rendering it once again active during the later stages of virus reactivation. In a G50pKO infection, however, the proximal promoter is no longer present to initiate gene 50 transcription and subsequent reactivation from the methylated viral genome. The distal gene 50 promoter, which was unmethylated and therefore able to facilitate lytic replication in the absence of the proximal promoter during establishment, is now repressed by CpG methylation and therefore no longer able to drive gene 50 expression and thus unable to promote virus reactivation from latency.

It is worth noting that the RTA promoters defined for EBV and KSHV contain a higher density of CpGs than either the proximal or distal promoter regions of MHV68 and that these CpG “islands” are heavily methylated in latent cell lines (3, 7). There does not appear to be a “methylation-free” region in these promoters analogous to that seen in the 110-bp region of the MHV68 proximal G50 promoter. However, EBV and KSHV differ from MHV68 in that they encode a bZIP protein (Zta and K8, respectively) immediately downstream of gene 50/BRLF1. Recent studies have demonstrated that the EBV Zta protein, unlike many mammalian transcription factors whose binding is inhibited by methylated cytosines within their recognition site, preferentially binds to and transactivates a methylated RTA promoter (3, 4, 17). Although Zta and K8 differ in both their ability to bind to DNA and induce lytic replication, the two proteins share significant homology with respect to host-cell protein interactions (38). It is possible that cellular factors associated with these bZIP proteins may confer a unique resistance to promoter methylation, and thus may circumvent the need for a “methylation-free” EBV-BRLF1 or KSHV-RTA promoter in order to initiate virus reactivation and subsequent viral replication. It is hypothesized that the ability to reactivate from latency is an important aspect in the maintenance of long-term gammaherpesvirus infection. Therefore, MHV68, which does not encode a known bZIP protein, may have evolved a CpG-free gene 50 promoter to retain resistance to the host-mediated epigenetic silencing that occurs during latent infection.

Mutations in Orf48 do not recapitulate the G50pKO phenotype.

The original intention in generating a gene 50 promoter knockout virus was to obtain a mutant virus that was essentially G50-null and could be used in place of the G50.Stop virus to avoid recombination during propagation on RTA-expressing stable cell lines. For this reason, the G50pKO virus was engineered to completely ablate the region containing the core 110-nt proximal gene 50 promoter and thus contains a large deletion of the first 820 nt of ORF48, a tegument-associated protein of unknown function encoded on the opposite strand (6) (Fig. 1). However, it is unlikely that the splenic reactivation phenotype observed with the G5OpKO virus is due to disruption of ORF48, since previous studies using an ORF48 deletion virus did not reveal a splenic latency defect (23). In contrast, a study examining MHV68 mutants generated from transposon insertions reported that ORF48 was essential for lytic growth in vitro, as an insertion at nt 66462 within the first 200 nt of the gene resulted in a replication-incompetent phenotype (39). However, this likely represents a complex phenotype since this transposon insertion disrupts the proximal gene 50 promoter, and the introduction of a large transposon between the distal gene 50 promoter and exon 1 undoubtedly interferes with RTA expression driven by the distal gene 50 promoter. Notably, the absence of CPE observed upon infection with the transposon mutant is a much more severe than the defect in virus replication observed with the G50pKO virus. With respect to the replication defect observed with the G50pKO mutant there are two likely explanations. First, the deletion removed the E0/E1 splice acceptor site, leading to production of a transcript from the distal gene 50 promoter containing the E0-E1 intron in addition to E0 and E1 exon sequences. Retaining the E0-E1 intron introduces a number of short open reading frames that likely interfere with efficient translation of RTA from this transcript. Second, the kinetics of gene 50 expression from the proximal promoter may not be suited to efficiently express RTA in the absence of transcription initiation from the proximal gene 50 promoter. Thus, these seemingly incongruous results likely arise from the fact that, like many other gammaherpesvirus genes, transcription of MHV68 gene 50 and the surrounding genes is extremely complex and varies according to infected cell type. Much further investigation is required to fully understand the regulation and contribution of transcription in this critical region of the viral genome to virus replication and reactivation.