Targeted Disruption of Kaposi's Sarcoma-Associated Herpesvirus ORF57 in the Viral Genome Is Detrimental for the Expression of ORF59, K8{alpha}, and K8.1 and the Production of Infectious Virus

Kaposi's sarcoma-associated herpesvirus (KSHV) ORF57 regulatesviral gene expression at the posttranscriptional level duringviral lytic infection. To study its function in the contextof the viral genome, we disrupted KSHV ORF57 in the KSHV genomeby transposon-based mutagenesis. The insertion of the transposoninto the ORF57 exon 2 region also interrupted the 3' untranslatedregion of KSHV ORF56, which overlaps with the ORF57 coding region.The disrupted viral genome, Bac36- ${Delta}$ 57, did not express ORF57,ORF59, K8 ${alpha}$ , K8.1, or a higher level of polyadenylated nuclearRNA after butyrate induction and could not be induced to produceinfectious viruses in the presence of valproic acid, a histonedeacetylase inhibitor and a novel KSHV lytic cycle inducer.The ectopic expression of ORF57 partially complemented the replicationdeficiency of the disrupted KSHV genome and the expression ofthe lytic gene ORF59. The induced production of infectious virusparticles from the disrupted KSHV genome was also substantiallyrestored by the simultaneous expression of both ORF57 and ORF56;complementation by ORF57 alone only partially restored the productionof virus, and expression of ORF56 alone showed no effect. Altogether,our data indicate that in the context of the viral genome, KSHVORF57 is essential for ORF59, K8 ${alpha}$ , and K8.1 expression and infectiousvirus production.

INTRODUCTION

Top

INTRODUCTION

Kaposi's sarcoma-associated herpesvirus (KSHV), also known ashuman herpesvirus 8, is a gamma-2 herpesvirus. KSHV has a genomeof $~$ 165 kb that encodes up to 90 viral genes, and its genomeshares significant sequence homology with that of the Epstein-Barrvirus and herpesvirus saimiri (5, 33, 37). As with other gammaherpesviruses,KSHV infection can be either latent or lytic depending on cellularconditions (39). What makes the virus become latent during theinfection remains largely unknown, but the latent KSHV genomepersists in a nonintegrated circular episomal form as it doesin the lytic stage (1, 2, 53). The latent KSHV can be activatedin infected cells by chemical inducers, such as sodium n-butyrate(31), 12-O-tetradecanoylphorbol-13-acetate (TPA) (36), and valproicacid (VA) (15, 40), and the resulting lytic gene expressionleads to production of infectious virus. Expression of a viraltransactivator, ORF50 (RTA), as a result of chemical inductionis essential for the lytic switch from KSHV latency (26, 42).

Nuclear transcription of viral lytic genes can be robustly activatedin trans by ORF50 (25, 42), but the newly synthesized viraltranscripts must be processed correctly at the posttranscriptionallevel, including RNA 5' capping, splicing, 3' polyadenylation,and export from the nucleus to the cytoplasm, to ensure theirefficient protein translation. The members of the herpesvirusfamily have evolved a specific mechanism to facilitate thisprocess, the encoding of a posttranscriptional regulatory proteinthat promotes the expression of specific viral transcripts (43,54). KSHV ORF57, which is transactivated by ORF50 (24, 48),encodes a viral early nuclear protein of 455 amino acid residues(9, 14). ORF57 promotes viral gene expression by mediating thenuclear export of viral RNAs in a CRM1-independent manner, presumablythrough its interaction with the cellular export factor Aly/REF(30), similar to its homologs in other herpesviruses (12, 16,49). However, this function of ORF57 has never been studiedin the context of the viral genome during lytic infection. Moreover,our recent data showed that interactions between the three nuclearlocalization signals of ORF57 and Aly/REF are not essentialfor ORF57-mediated accumulation of KSHV ORF59 transcripts inliving cells (28).

To study KSHV ORF57 function in the context of the KSHV genomeduring lytic viral infection, we took the genetic approach ofdisrupting ORF57 in a KSHV genome constructed in a bacterialartificial chromosome, Bac36 (56). Since its construction, Bac36has been widely used to characterize KSHV gene function by thegenetic manipulation of individual viral genes within the viralgenome (27, 51, 53). In this study, we demonstrate that whenthe ORF57 gene is disrupted, the KSHV genome cannot expressa subset of viral lytic genes and produce infectious virionsin response to chemical induction.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Cells. Human 293 cells were obtained from the American Type CultureCollection (Manassas, VA) and were cultivated in Dulbecco'smodified Eagle medium (Invitrogen, Carlsbad, CA) supplementedwith 10% fetal bovine serum (HyClone, Logan, UT), 2 mM L-glutamine,100 U/ml penicillin, and 100 g/ml streptomycin (Invitrogen).

Plasmid construction. Construction of the KSHV ORF57 expression vector pVM7 (pORF57)was described previously (28). To create the ORF56 expressionvector pVM9 (pORF56), the KSHV ORF56 open reading frame (nucleotides[nt] 79433 to 81946 of the KSHV genome) was amplified by PCRand cloned into a pFLAG-CMV-5.1 vector (Sigma, St. Louis, MO).The insert was confirmed by enzyme digestion and sequencing.The constructed vectors express ORF57 and ORF56 with a C-terminalFLAG tag. An RTA expression vector, pORF50, was a gift fromYan Yuan of the University of Pennsylvania.

Construction of a KSHV ORF57-null mutant. We chose Bac36 to create an ORF57-null mutant by random transposon-basedmutagenesis with an EZ::TN <KAN-2> kit (Epicenter, Madison,WI). Bac36 contains a full-length KSHV genome and a hygromycinB resistance gene and expresses green fluorescent protein asa marker (56). Briefly, purified KSHV Bac36 DNA (Bac36-wt) wasincubated for 2 h with Tn5 transposase, and a 1.2-kb transposoncassette that contains a kanamycin resistance gene (Kan^r) andwas then transformed into Escherichia coli strain DH10B by electroporation.The recombinant Bac clones with transposon insertions were selectedon Luria-Bertani medium agar plates in the presence of kanamycin(50 µg/ml). Individual clones were expanded and sequencedusing the KAN-2 FP-1 (5'-ACCTACAACAAAGCTCTCATCAACC-3') and RP-1(5'-GCAATGTAACATCAGAGATTTTGAG-3') primers to determine the transposoninsertion site. A KSHV Bac clone with a transposon insert inthe ORF57 coding region was designated Bac36- ${Delta}$ 57. All Bac36 DNAswere isolated from E. coli using a QIAGEN Large-Construct kit(QIAGEN, Valencia, CA).

Genetic analysis of the KSHV ORF57-null mutant. Several methods were employed to confirm that the transposonwas inserted into the ORF57 coding region without disruptingthe integrity of the rest of the viral genome. First, we designeda pair of primers flanking the insert, oVM12 (Pr82699; 5'-CTCAGACTCCCTGCGAGCAT-3')and oVM13 (Pr83151; 5'-TCTGGTAACAAACGCATTGC-3'), and analyzedthe PCR products from the Bac36-wt and Bac36- ${Delta}$ 57 genomes by agarosegel electrophoresis. The integrity of the viral genome aftertransposon insertion was determined by restriction enzyme digestionof Bac36 DNAs. One microgram of purified Bac36 DNAs (wild typeand mutant) was digested with KpnI, and the resulting DNA fragmentswere separated in a 0.7% agarose gel together with 1-kb DNAladders (Invitrogen) at 40 V for 14 h. After separation, theDNA fragments were transferred onto a nylon membrane and analyzedby Southern blot analysis. Probe labeling and hybridizationwere carried out using an AlkPhos direct labeling kit (Amersham,Piscataway, NJ) according to the manufacturer's instructions.Briefly, 100 ng of purified ORF57-specific PCR products fromnt 82699 to 83151 of the KSHV genome or a EZ::TN XhoI-BamHIfragment was directly labeled with alkaline phosphatase usinga chemical cross-linker, and the chemically labeled probes werethen hybridized to the membrane at 55°C overnight. Afterwashes, the bound probe was detected by chemiluminescence usingthe CPD-Star detection reagent (Amersham) and exposed to X-rayfilm. After detection, the membrane was stripped in 0.5% sodiumdodecyl sulfate (SDS) at 60°C for 1 h, reconstituted in100 mM Tris-HCl (pH 8.0) for 5 min at room temperature, andreprobed with another probe.

Transient transfection and establishment of stable cell lines. Purified Bac36 DNAs were transfected into 293 cells using Lipofectamine2000 (Invitrogen). Hygromycin B (Sigma, 150 ng/ml) was thenadded for selection 24 h after transfection. Three weeks afterselection, a homogenous population of green fluorescent protein(GFP)-positive cells harboring KSHV Bac36 DNA was obtained.Butyrate (Sigma) at a final concentration of 3 mM, TPA (Sigma)at 20 ng/ml, or VA (Sigma) at 1 mM was used for lytic induction.Transient transfection of the stable cell lines with an ORF50expression vector, pORF50, was also used for lytic induction.

Western blot analysis. Protein samples were prepared by direct lysis of cells in 2xSDS protein sample buffer containing 5% 2-mercaptoethanol. Thelysed cells were boiled and separated in SDS-polyacrylamidegel electrophoresis (PAGE) gels. The following antibodies wereused in Western blot analyses: a rabbit polyclonal anti-ORF57antibody against a synthetic peptide (amino acids 119 to 132of ORF57 [unpublished data]; used at a dilution of 1:3,000),a monoclonal immunoglobulin M-type anti-ß-tubulinantibody (1:1,000; BD Pharmingen, San Diego, CA), a polyclonalanti-GFP antibody (1:3,000; Clontech, Mountain View, CA), anda monoclonal anti-FLAG M2 antibody (1:2,500; Sigma), togetherwith corresponding horseradish peroxidase-conjugated secondaryantibodies (1:10,000; Sigma). The signal on the Western blotwas detected with a West Pico chemiluminescence substrate (Pierce,Rockford, IL).

RPA. Total cell RNA was prepared using the TRIzol reagent (Invitrogen)according to the manufacturer's instructions. An antisense ORF59-specificprobe covering nt 95870 to 96171 of the KSHV genome and an antisensepolyadenylated nuclear (PAN) RNA probe covering nt 29400 to29608 were transcribed in vitro. An antisense human cyclophilinprobe (Ambion, Austin, TX) was used as an internal control forsample loading. All probes were prepared using the Riboprobesystem (Promega, Madison, WI) in the presence of [ ${alpha}$ -³²P]GTP.An RNase protection assay (RPA) was carried out using an RPAIII kit (Ambion) according to the manufacturer's instructions.Briefly, 30 to 40 µg of total RNA was hybridized with4 ng of each probe overnight at 42°C and then digested witha mixture of RNase A and RNase T₁ for 30 min at 37°C. Protectedfragments were separated in an 8% PAGE denaturing gel. The signalswere captured using a Molecular Dynamics Storm 860 PhosphorImagerand analyzed with ImageQuant software (Amersham).

Immunofluorescence staining. Before staining, stable Bac36 cells were grown on coverslipsand the viral lytic cycle was induced with butyrate. Twenty-fourhours after induction, the monolayers were washed twice withphosphate-buffered saline (PBS), fixed with 2% paraformaldehydein PBS for 20 min at room temperature, and quenched twice with100 mM glycin in PBS (10 min each). For intracellular staining,the cells were permeabilized with 0.1% Triton X-100 in PBS for10 min at room temperature. Blocking was performed using blockingsolution (3% bovine serum albumin in PBS with 0.05% Tween 20)for 1 h at 37°C. All primary and secondary antibodies werediluted 1:50 in blocking solution and incubated for 1 h at 37°C,followed by three washes (10 min each) with PBS containing 0.05%Tween 20. The following primary antibodies were used: rabbitpolyclonal anti-ORF57 antibody (as described above), mouse monoclonalanti-K8 (Promab, Albany, CA), mouse monoclonal anti-ORF59 (AdvancedBiotechnologies, Columbia, MD), rat monoclonal anti-ORF73 (anti-LANA;Advanced Biotechnologies), and rabbit polyclonal anti-K8.1 antibody(19), together with corresponding tetramethylrhodamine isothiocyanate(TRITC)-conjugated secondary antibodies (1:50; Sigma). Fluorescenceimages were collected and saved in TIFF format, and Adobe Photoshop6.0 software (Adobe Systems, San Jose, CA) was used to processthe images into composite figures.

Virus production and complementation assay. To activate the lytic phase and virion production, the stableBac36 cells were incubated with 1 mM VA for 72 h. After induction,the free-virus-containing supernatant (30 ml from a T-162 flask)was clarified by low-speed centrifugation (5,000 x g, 30 min)to remove cells and cell debris. The virus in the clarifiedculture supernatant was then concentrated by high-speed centrifugation(25,000 x g, 3 h). The virus pellets were resuspended in anequal volume (700 µl) of serum-free Dulbecco's modifiedEagle medium, and 300 µl of the resuspension was titratedin duplicate on freshly plated 293 cells with the addition of5 µg/ml polybrene (Sigma). The inoculated cells were examineddaily for the appearance of GFP-positive cells as an indicationof viral infection. Within 72 to 96 h after viral infection,the cells were trypsinized, and the cell suspension was analyzedby flow cytometry to determine the number of GFP-positive cells.The complementation assay was carried out by transient transfectionof Bac36 cells with ORF57 or ORF56 expression vectors beforeVA induction. The empty vector, pFLAG-CMV-5.1, was used as acontrol. Twenty-four hours after transfection, the lytic cyclewas induced by VA as described above.

RESULTS

Top

RESULTS

Construction of a KSHV Bac36- ${Delta}$ 57 mutant. To create a KSHV virus lacking ORF57 expression, we utilizedBac36, which contains the wt KSHV genome (Bac36-wt) (56), asa template for random transposon-based mutagenesis. The mutantKSHV genome was screened by sequencing and showed a transposoninsertion at the nt 83001 position in the N-terminal ORF57 codingregion (Fig. 1A), along with duplication of a 9-bp KSHV genomesequence on each end of the insert. The insertion took place998 nt downstream of the ORF57 transcription start site anddisrupted exon 2 of ORF57. Accordingly, the selected mutantwas named Bac36- ${Delta}$ 57. To further verify the insertion position,we carried out a PCR assay using a paired primer set, oVM12and oVM13, flanking the insertion site of Bac36- ${Delta}$ 57. The PCRamplification generated a 453-bp product from Bac36-wt withno insertion and a 1.6-kb product from Bac36- ${Delta}$ 57 containing the1.2-kb transposon cassette (Fig. 1B), confirming the resultsof the sequencing.

View larger version (50K):
[in this window]
[in a new window]

FIG. 1. Genetic analysis of KHSV Bac36-wt and the Bac36- ${Delta}$ 57 mutant. (A) Schematic diagram of the gene structure of KSHV ORF57 from nt 82003 (major transcription start site) to nt 83628 (major polyadenylation site) in the viral genome (29). ORF57 consists of two exons and a small intron (14). The exon 2 region with the insertion of an EZ::TN <Kan-2> transposon in the KSHV Bac36- ${Delta}$ 57 mutant was identified by sequencing using primers KAN RP-1 and FP-1. (B) The transposon insertion in the mutant genome was confirmed by PCR using primers oVM12 and oVM13 from the ORF57 coding region (panel A), and the PCR products were separated on an agarose gel. A plasmid, pORF57, containing a full-length ORF57 cDNA was used as a positive control. (C) Restriction analysis of purified KSHV Bac36-wt and Bac36- ${Delta}$ 57 DNA by KpnI digestion. An asterisk (*) represents a 6.0-kb DNA fragment containing the ORF57 coding region in the Bac36-wt genome or a 7.2-kb ORF57 DNA fragment after transposon insertion in the Bac36- ${Delta}$ 57 genome. (D) Southern blot analysis of Bac36-wt and Bac36- ${Delta}$ 57 DNA. After KpnI digestion, the Bac36 DNAs were separated in agarose gels, transferred onto nylon membranes, hybridized with an ORF57 probe (left), and reprobed with a transposon-specific probe (right).

We further analyzed Bac36- ${Delta}$ 57 by restriction enzyme digestionand Southern blotting. Purified Bac36-wt and Bac36- ${Delta}$ 57 DNA digestedwith KpnI were first separated in an agarose gel for ethidiumbromide staining and then transferred onto a nylon membranefor Southern blot analysis. As shown in Fig. 1C and D, the restrictionprofile of Bac36-wt was identical to the published data (56),with the ORF57 gene being cut into two digested fragments, 399bp from nt 81948 to nt 82347 (Fig. 1A; also data not shown)and 6 kb from nt 82347 to 88381. In contrast, Bac36- ${Delta}$ 57, containinga 1.2-kb transposon insertion at nt 83001, created a largerrestriction product from this region, with a size of 7.2 kb.Otherwise, the restriction profiles of the two genomes lookedthe same, indicating the integrity of the rest of the Bac36- ${Delta}$ 57genome. The specificity of the identified restriction productsgenerated from the ORF57 gene locus with or without the insertionwas also confirmed by Southern blot analysis using an ORF57-specificprobe covering nt 82699 to 83151 (left panel in Fig. 1D) ora transposon-specific probe, EZ::TN (right panel in Fig. 1D).Altogether, these data indicate that the Bac36- ${Delta}$ 57 genome hasa single transposon insertion in the ORF57 coding region.

The Bac36- ${Delta}$ 57 genome does not express the ORF57 protein during lytic induction. Previous studies (36, 56) showed that 293 cells are susceptibleto KSHV infection. When transfected with Bac36-wt, 293 cellscan support KSHV latent infection and can then be induced toproduce infectious viruses (56). We therefore transfected Bac36- ${Delta}$ 57DNA into 293 cells to determine if the transposon insertionin ORF57 exon 2 would disrupt expression of ORF57. Bac36-wtwas used as a positive control. Stable cell lines were establishedusing the presence of hygromycin B as a selection marker. Afterseveral passages of selection, all cells contained Bac36 DNAsand became GFP positive. To examine the expression of ORF57from both Bac36-wt and Bac36- ${Delta}$ 57, we developed a rabbit anti-ORF57polyclonal antibody against a synthetic peptide correspondingto amino acids 119 to 132 of the ORF57 protein, upstream ofthe insertion site. In a Western blot analysis, the antibodyrecognized a full-length ORF57 protein with the predicted sizeof 50 kDa in butyrate- or TPA-induced Bac36-wt stable cellsbut not in uninduced Bac36-wt cells (Fig. 2A, upper panel),indicating its specificity for ORF57. In contrast, no proteinof any size was detectable in butyrate- or TPA-induced Bac36- ${Delta}$ 57stable cells (Fig. 2A, upper panel), despite detection of similaramounts of tubulin protein in all samples (Fig. 2A, lower panel).To further confirm the Western blot results, we performed immunofluorescencestaining of stable Bac36-wt (Fig. 2B, upper panel) and Bac36- ${Delta}$ 57(lower panel) cells 24 h after butyrate induction. Again, onlythe stable cells with a Bac36-wt genome were found to expressORF57, although all Bac36-wt and Bac36- ${Delta}$ 57 stable cells wereGFP positive, an indication that they harbor the KSHV genome.From these data, we conclude that the insertion of the transposonin the Bac36- ${Delta}$ 57 genome completely disrupted the production ofthe ORF57 protein.

View larger version (69K):
[in this window]
[in a new window]

FIG. 2. Expression of the ORF57 protein in 293 cells transfected with Ba36-wt and Bac36- ${Delta}$ 57 mutant DNA. 293 cells were transfected with Bac36 DNAs and selected for stable cells in the presence of hygromycin B. (A) Western blot analysis of protein extracts from Bac36 DNA-harboring stable cell lines with or without 24 h of induction by 3 mM sodium n-butyrate (NB) or 20 ng/ml phorbol ester (TPA). A rabbit polyclonal anti-ORF57 antibody was used to detect the ORF57 protein in cell extracts, and the sample loading was controlled by tubulin. (B) Immunofluorescent staining of butyrate-induced ORF57 protein expression in Bac36-wt (upper panel) and Bac36- ${Delta}$ 57 (lower panel) stable cell lines. All stable cells expressed GFP (middle panels) independently of induction. Cells grown on coverslips were stained with an anti-ORF57 antibody in combination with a TRITC-conjugated secondary antibody after 24 h of butyrate induction.

ORF57 is essential for the expression of a subset of viral lytic genes. To determine whether the lack of ORF57 expression would affectthe expression of other viral lytic genes, we analyzed fiveviral genes that are expressed at different stages of the viralcycle: ORF59, K8 ${alpha}$ , K8.1, LANA, and PAN RNA. ORF59, K8 ${alpha}$ , and PANare viral early lytic genes. ORF59 encodes a viral DNA polymeraseprocessivity factor for viral DNA replication (3, 6, 21) andis a downstream target of ORF57 in transient assays (14, 28).K8 ${alpha}$ encodes a viral K-bZIP protein involved in viral gene transcriptionand DNA replication (20, 22, 57). PAN expresses an abundantviral noncoding RNA with unknown function (41, 55) that wasalso shown to be a downstream target of ORF57 in transient transfectionexperiments (14). K8.1 is a viral late gene encoding an envelopeglycoprotein (4, 19, 35, 45). LANA is a latent gene encodinga latency-associated nuclear antigen responsible for KSHV genomesegregation (2, 13, 18). Twenty-four hours after induction ofthe lytic viral cycle by butyrate, we compared the expressionof ORF57, ORF59, K8 ${alpha}$ , K8.1, and LANA in stable Bac36-wt and Bac36- ${Delta}$ 57cells using indirect immunofluorescence staining with a specificantibody against each protein. A substantial number of Bac36-wtstable cells were positive for ORF57, ORF59, and K8 ${alpha}$ , along withmany fewer cells that were positive for K8.1, indicating thatthe lytic cycle was initiated by the induction, but the majorityof the stable cells remained at the early stage of viral replication.In contrast, no Bac36- ${Delta}$ 57 cells expressed ORF59, K8 ${alpha}$ , and K8.1despite expressing LANA at levels similar to those for Bac36-wtcells (Fig. 3A). Consistent with our findings, a recent studyusing Vero cells stably transfected with an ORF57-null KSHVgenome also showed no ORF59 expression after induction withORF50 (10).

View larger version (19K):
[in this window]
[in a new window]

FIG. 3. Expression of ORF57, ORF59, K8 ${alpha}$ , K8.1, and LANA in Bac36-wt and Bac36- ${Delta}$ 57 stable cell lines after lytic induction. (A) Immunofluorescent staining of ORF57, ORF59, K8 ${alpha}$ , K8.1, and LANA proteins in Bac36 stable cells after 24 h of induction with butyrate. The stable cells were stained with specific antibodies (see Materials and Methods) in combination with TRITC-conjugated secondary antibodies. (B) The expression of ORF59 transcripts from Bac36 stable cells. RNase protection assays using an antisense ORF59-specific probe were carried out on 30 µg of total cell RNA isolated from 293 cells with or without stable Bac36 transfection after treatment with butyrate for 24 h. Cellular cyclophilin RNA was used as a loading control. Yeast total RNA was used as a negative control.

We next compared the two stable cell lines for their expressionof ORF59 and PAN RNAs, which had been identified as ORF57 targetsin another study (14) but never confirmed in the context ofviral replication. We performed an RPA on total RNA isolatedfrom the stable cell lines 24 h after butyrate induction. Aspredicted, we found that both the ORF59 transcript (Fig. 3B)and PAN RNA (Fig. 4A) were present in the Bac36-wt stable cells,but very little was present in the Bac36- ${Delta}$ 57 stable cells (comparelane 4 to lane 5 in Fig. 3B and lane 5 to lane 6 in Fig. 4A)in spite of the fact that a few stable cells showed some residualPAN RNA expression in the absence of lytic induction (Fig. 4A,lanes 3 and 4). When RTA (ORF50) was used to transactivate PANRNA expression from the two cell lines, we also found that theamount of PAN RNA expressed by the Bac36- ${Delta}$ 57 cells was much less(Fig. 4A), but RTA transactivation was less potent than thatby butyrate in induction under our conditions (compare lanes7 and 8 with lanes 5 and 6 in Fig. 4A). Similar results werealso obtained by transient transfection of 293 cells with thetwo Bac36 DNAs and butyrate induction (Fig. 4B and C). Dataindicate that PAN RNA is indeed an ORF57 target in the contextof the KSHV genome. Based on all of those observations, we concludethat the Bac36- ${Delta}$ 57 genome cannot express a subset of viral lyticgenes in the absence of the ORF57 protein.

View larger version (53K):
[in this window]
[in a new window]

FIG. 4. Expression of PAN RNA in Bac36-wt and Bac36- ${Delta}$ 57 stable cell lines or transiently transfected cells. (A) Substantial PAN RNA expression depends on ORF57. Bac36-wt and Bac36- ${Delta}$ 57 stable cells (5 x 10⁵/well) were plated into a 6-well plate, 3 wells for each. The cells at 24 h after seeding were treated with 3 mM of sodium butyrate (NB) or transfected with 3 µg of plasmid pORF50 (p50) and were harvested 24 h after NB treatment or 48 h after transfection. The cells without treatment were used as uninduced controls. Total cell RNA was prepared, and each RNA sample at 40 µg was examined by RPA for PAN RNA and cyclophilin expression with a PAN- and cyclophilin-specific probe. (B to E) PAN RNA expression in transiently transfected 293 cells with Bac36 DNAs. 293 cells at 5 x 10⁵/well were plated into a 6-well plate and transfected with 5 µg of purified Bac36 DNAs. The cells in each well were split into two wells 48 h after transfection and were induced by NB (3 mM) 24 h after the split for an additional 24 h (B and C) or cultivated for an additional 48 h without induction (E). Transfection efficiencies of 293 cells with Bac36-wt or Bac36- ${Delta}$ 57 were compared by appearance of GFP-positive cells (D). The cell lysates and total cell RNA were prepared and examined by Western blotting for ORF57 or GFP (B) or by RPA for PAN RNA or cyclophilin with a PAN- or cyclophilin-specific probe (C and E). Quantitation of PAN RNA levels in each sample (A, C, and E) was normalized to cyclophilin expression.

However, we found that both the Bac36- ${Delta}$ 57 genome and the Bac36-wtgenome, in the absence of any induction, could spontaneouslyexpress PAN RNA to a comparable level in transiently transfected293 cells independently of ORF57 (Fig. 4D and E).

Disruption of ORF57 in the KSHV genome reduces production of infectious virus. Because ORF57 disruption affects expression of viral early genesimportant for viral DNA replication, we wanted to determinewhether ORF57 is essential for the production of cell-free infectiousvirus. Initially, we compared the efficiencies of various chemicalinducers at inducing infectious virus production from the Bac36-wtstable cells and found that 1 mM VA was the most effective (datanot shown). VA is a prescription drug that has been used totreat patients with seizures and bipolar disorders (23, 34).A recent study showed that, similar to butyrate, VA inhibitshistone deacetylases (8, 32), and it is very potent for inducingthe expression of lytic KSHV genes in BCBL-1 cells (15, 40)but is much less toxic than butyrate (8, 32). This minimal toxicityallowed us to use VA for long-term induction to study virusproduction. We treated both the Bac36-wt and the Bac36- ${Delta}$ 57 stablecells with VA for 72 h, harvested the virus-containing culturesupernatants, clarified the supernatants with low-speed centrifugation,and concentrated the virus with high-speed centrifugation. Theproduction of infectious virus from each stable cell line wastitrated with fresh 293 cells, and the infected cells were examineddaily for GFP expression by fluorescence microscopy. As shownin Fig. 5A, the supernatant collected from the VA-induced Bac36-wtcells was infectious, with the infected 293 cells becoming GFPpositive 48 h after infection. The titer in the concentratedvirus resuspension was about 3.3 x 10² infectious viruses/ml.In contrast, 293 cells did not express GFP 48 h after treatmentwith the supernatant obtained from the VA-induced Bac36- ${Delta}$ 57 stablecells, indicating that the Bac36- ${Delta}$ 57 cells produced few, if any,infectious viruses. To better determine the numbers of infected293 cells in each sample, the inoculated 293 cells were collected96 h after infection, and GFP-positive cells were quantifiedby flow cytometry. Approximately 60 cells per 100,000 collectedwere GFP positive in the Bac36-wt sample, but only 15 cellsper 100,000 treated with the Bac36- ${Delta}$ 57 supernatant were GFP positive(compare bar 1 to bar 2 in Fig. 5B). These data indicate thatdisruption of ORF57 expression indeed disrupts the productionof infectious viruses.

View larger version (21K):
[in this window]
[in a new window]

FIG. 5. Virus production and complementation assays. To induce virus production from Bac36 stable cells, the cells were treated with 1 mM VA for 72 h. The virus-containing supernatants were clarified by low-speed centrifugation (5,000 x g, 30 min) and concentrated by high-speed centrifugation (25,000 x g, 3 h). The concentrated virus was titrated on freshly plated 293 cells. (A) GFP expression from 293 cells infected with Bac36-wt or Bac36- ${Delta}$ 57. A phase image and a GFP image collected 72 h after infection are shown for 293 cells treated with each type of Bac36. "Bac36- ${Delta}$ 57 + pORF57 + pORF56" indicates that the concentrated virus was collected from VA-induced Bac36- ${Delta}$ 57 stable cells cotransfected with ORF57 and ORF56 expression vectors. (B) Flow cytometry analysis of GFP-positive 293 cells 72 h after inoculation with concentrated viruses. For complementation assays, the stable cells were transiently transfected with the KSHV ORF57 or ORF56 expression vector alone or in combination for 24 h before VA induction. The panel is one representative of three repeats. (C) Western blot analysis of ORF57 or ORF56 expression in Bac36- ${Delta}$ 57 stable cells with or without ORF56 and ORF57 complementation. Cell extracts were separated with PAGE, transferred onto a membrane, and blotted with an anti-FLAG antibody to detect FLAG fusion proteins of both ORF57 and ORF56. (D) Restoration of ORF59 expression after complementation with ORF57. The Bac36- ${Delta}$ 57 stable cells (lower panel) were transiently transfected for 24 h with an ORF57 expression vector and then subjected to butyrate induction for another 24 h before immunofluorescent staining. Untransfected Bac36-wt stable cells were used as positive controls for lytic induction (upper panel). The expressed ORF57 and ORF59 proteins were examined by immunofluorescent staining using anti-ORF57 or anti-ORF59 antibody, respectively.

To ensure that the expression of GFP in 293 cells was mediatedsolely through infectious virus particles, the concentratedvirus preparations were also treated with DNase I for 30 minat 37°C to remove any possible residual cell debris contaminatedwith the Bac36- ${Delta}$ 57 genome that might have transfected the fresh293 cells. After the treatment, a flow cytometry analysis foundno GFP-positive cells among the cells treated with Bac36- ${Delta}$ 57viral preparations; however, this treatment also substantiallyreduced the number of GFP-positive cells observed among thecells infected with Bac36-wt viral preparations (data not shown).

Complementation of ORF57 deficiency in the Bac36- ${Delta}$ 57 genome by ectopic expression of a functional ORF57. Our recent analysis of the ORF56 and ORF57 gene locus revealedthat ORF56 encodes a bicistronic RNA and utilizes a poly(A)signal downstream of ORF57 (29). Consequently, our targeteddisruption of the ORF57 coding region would also disrupt the3' untranslated region (UTR) of the ORF56 RNA and affect ORF56expression, because lack of a polyadenylation signal would makethe ORF56 RNA unstable even though its coding region remainedintact. KSHV ORF56 encodes a viral DNA primase important forviral DNA replication (7, 50), and its expression increasesin the presence of ORF57 (29). Therefore, disruption of ORF56would constitute an additional defect in the Bac36- ${Delta}$ 57 genome,one that has been shown to be detrimental to infectious virusproduction (7). To complement the functions of the deleted sectionsof the Bac36- ${Delta}$ 57 genome, we cotransfected Bac36- ${Delta}$ 57 cells withexpression vectors for ORF56 and ORF57 FLAG tag fusions 1 daybefore VA induction. Four days after VA induction of the cells,the infectious virions from the cotransfected cell culture supernatantswere titrated with fresh 293 cells to test for infectious virusproduction. With this strategy, we were able to substantiallyrestore the virus production from the Bac36- ${Delta}$ 57 genome (comparebar 5 to bars 1 and 2 in Fig. 5B). Complementation with ORF57alone (Fig. 5B, bar 3) only partially restored virus productionfrom the Bac36- ${Delta}$ 57 genome, and complementation with ORF56 alonehad no effect (Fig. 5B, bar 4). To correlate expression withinfectious virus production, we also used Western blot analysisto examine the Bac36- ${Delta}$ 57 stable cells with or without cotransfectedORF56 and ORF57 expression vectors. All cells transfected withan ORF57 expression vector had abundant ORF57 expression (Fig.5C). As expected, the cells cotransfected with ORF56 and ORF57expression vectors also had both ORF56 and ORF57 expression(Fig. 5C) and better restoration of virus production than cellstransfected with ORF57 alone (Fig. 5B).

Although the complementation of ORF57 didn't fully rescue virusproduction, we were interested to see whether it could restorethe expression of ORF57 targets. Bac36 stable cells growingon coverslips transfected with an ORF57 expression vector orleft untransfected were simultaneously induced for lytic viralgene expression with butyrate. As shown in Fig. 5D, Bac36-wtstable cells expressed both ORF57 (top panel, left) and ORF59(top panel, right) when induced. The Bac36- ${Delta}$ 57 stable cells,which have no ORF57 or ORF59 response to butyrate induction(see Fig. 3A, top two panels), became ORF59 positive after transfectionwith an ORF57 expression vector, pORF57, suggesting that theectopic expression of ORF57 (Fig. 5D, left lower panel) in theBac36- ${Delta}$ 57 stable cells could rescue ORF59 expression from theORF57-deficient KSHV genome (Fig. 5D, lower right panel). Consistentwith the results of the complementation assays, however, therestoration of ORF59 expression by ectopic expression of ORF57was less efficient compared to the level seen in induced Bac36-wtcells.

DISCUSSION

Top

DISCUSSION

In this study, we genetically analyzed the importance of KSHVORF57 for the expression of lytic viral genes in the contextof the KSHV genome. Insertion of a transposon-containing cassetteinto the ORF57 exon 2 region of the KSHV genome in Bac36 renderedthe KSHV genome incapable of expressing ORF57 and producinginfectious viruses, demonstrating that KSHV ORF57 is essentialfor virus production. In addition, the mutant genome lackingORF57 expression was also unable to express ORF59, K8 ${alpha}$ , K8.1,or a higher level of PAN RNA, thus establishing a subset oflytic viral genes expressed in both the early and late virallife cycles as targets of ORF57. Since transient transfectionof an ORF56 expression vector did not lead to detectable levelsof the ORF56 protein in the absence of ORF57, we hypothesizethat ORF56 expression also depends on ORF57. The lack of anavailable anti-ORF56 antibody prevents us from pursuing thisassumption further in the context of the viral genome. LANAis not an ORF57 target, because lack of ORF57 expression inthe mutant KSHV genome had no effect on the expression of LANA.Since LANA expression in latent KSHV infection is independentof lytic gene expression, it was not surprising that ORF57 hadno effect on LANA expression in lytic induction.

The finding that substantial expression of PAN RNA is dependenton ORF57 was consistent with others' reports that KSHV ORF57acts by itself or synergistically with ORF50 to increase PANRNA expression in transient transfection (10, 14). The ORF57-enhancedPAN expression was also seen by transient transfection whenan ORF57-null KSHV genome was used in a recent study (10). Unfortunately,there was no direct comparison of the ORF57-null genome withthe wt KSHV genome for PAN expression during lytic inductionin the reported study (10). By direct comparison of the wt andmutant genome, we concluded in this report that ORF57 promotesPAN RNA expression in the context of the virus genome. However,we also found that a small amount of PAN RNA can be spontaneouslyexpressed or induced from the virus genome in the absence ofORF57. Consistently, other studies also showed that the ORF57-nullKSHV genome itself could be induced by infection with Ad50 asa source of ORF50 to express PAN RNA to a certain level in theabsence of ectopic overexpression of ORF57 (10). Together, theseobservations suggest that ORF57 promotes PAN RNA expression,most likely at the posttranscriptional level.

Various analyses of gene structure and organization in the KSHVgenome have shown that it contains multiple gene cluster regions.Several genes within a cluster region will use either a commonpromoter or a common poly(A) signal for their gene expression,and the resulting bicistronic or polycistronic transcripts overlapeach other (11, 22, 25, 38, 44, 46, 52, 57). This organizationof gene structure makes the genetic manipulation of the KSHVgenome complex, and the results from such studies could be misleading.It is certainly more difficult to complement a disrupted genewithin a cluster using complementation assays. For example,disruption of the LANA exon 2 coding region by transposon insertion(53) will simultaneously interrupt an intron within the vCyclinand vFLIP transcripts and affect their RNA splicing (38, 54);disruption of a K8.1 coding region by the same approach (27)will inevitably destroy the 3' UTR of the ORF50 and K8 ${alpha}$ transcriptsand will affect the expression of all three genes, since theyall share the same poly(A) signal downstream of the K8.1 codingregion (46, 54, 57). Since the KSHV ORF57 coding region overlapswith the 3' UTR of ORF56 (29), disruption of the ORF57 codingregion also inevitably damages the structure of ORF56 transcriptsand simultaneously knocks out two genes. Accordingly, to complementthe loss of two gene products, we cotransfected Bac36- ${Delta}$ 57 stablecells with ORF56 and ORF57 expression vectors and effectivelyrescued the production of infectious viruses from the mutantgenome to a level nearly comparable to that from the Bac36-wtgenome. Although it was puzzling, the partial restoration ofinfectious virus production in our study by the ectopic expressionof ORF57, but not ORF56 alone, might imply that the transposon-mediatedinsertion did not abrogate all ORF56 protein expression fromBac36- ${Delta}$ 57. Alternatively, a fraction of ORF56 transcripts fromthe Bac36- ${Delta}$ 57 genome might utilize any one of the three AATAAAhexamers from the inserted EZ::TN <KAN-2> transposon immediatelydownstream for polyadenylation, since the virus primase ORF56is required for viral DNA synthesis. Another report also indicatesthat the disruption of ORF57 by insertion of a Kan^r cassettein the KSHV genome could be complemented by the ectopic expressionof ORF57 alone in the presence of ORF50 (10).

Reverse genetics is a powerful tool for dissecting the functionsof viral genes. Since its development, the Bac36 system hasbeen successfully used to define the functions of several KSHVgenes (10, 17, 27, 51, 53, 56). Because of the large genomesize, genomic rearrangements could occur in both bacteria andmammalian cells. Adaptation of proper screening procedures,such as restriction analysis and Southern hybridization, couldensure the selection of the correct clones. The induction oflytic replication in 293 cells is also at times not efficient.We observed that only few such induced cells (less than 1%)reach the late stage of infection (24 h after induction), whenthe viral structural protein K8.1 is produced, and an extremelylow number of infectious viruses were present in the culturesupernatant. Even after concentration by high-speed centrifugation,only $~$ 60 cells in every 100,000 cells became infected in flowcytometry assays. While the selection of individual high-producercell clones could significantly enhance the efficiency, theuse of an entire cell population is often necessary to achievetrue representation, and this is required to definitely delineatethe function of a viral gene, as is the case in this work. Alternativecell culture systems, such as Vero cells that can support replicationefficiency, should help wider adaptation of the system in theinvestigation of the functions of KSHV genes in the contextof the viral genome (10, 47). Nevertheless, determining themechanism(s) that prevents the Bac36 genome from activatinglytic infection would provide further insights into the processof KSHV infection.

ACKNOWLEDGMENTS

This research was supported by the Intramural Research Programof the NIH, National Cancer Institute, Center for Cancer Researchand Division of Intramural Research, National Heart, Lung, andBlood Institute.

We thank Jae Jung and Bala Chandran for providing anti-K8.1antibodies and Yan Yuan for the expression vector pORF50.

FOOTNOTES

* Corresponding author. Mailing address: HIV and AIDS Malignancy Branch, Center for Cancer Research, NCI/NIH, 10 Center Dr., Rm. 10 S255, MSC-1868, Bethesda, MD 20892-1868. Phone: (301) 594-1382. Fax: (301) 480-8250. E-mail: zhengt{at}exchange.nih.gov.

Published ahead of print on 15 November 2006.

REFERENCES

Top