INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Human herpesvirus 6 (HHV-6) was first isolated from patients with AIDS and lymphoproliferative disorders (28). It was subsequently shown to be the causative agent of a common childhood disease, exanthem subitum (roseola) (1, 9, 10). HHV-6 isolates segregated into two closely related subgroups, A and B, based on their antigenic properties, T-cell line infectivities, restriction endonuclease cleavage, and genomic DNA sequences (1, 4, 7, 15, 31, 40). HHV-6A includes the prototype strains GS and U1102, and HHV-6B includes the prototype strain Z29 and isolates from roseola patients (1, 40). The genome of HHV-6 is approximately 160 kb long and consists of a unique long region of 140 kb flanked by direct-repeat regions of about 10 kb (15, 24, 25, 32). Sequence analyses of the HHV-6 genome show that it is closely related to human cytomegalovirus (HCMV), exhibiting a colinear arrangement of genes (6, 12, 24, 25).

Several studies have described potential interactions between HHV-6 and human immunodeficiency virus type 1 (HIV-1) (3, 11, 13, 14, 16, 17, 18, 21, 42). HHV-6A(GS) and HHV-6B(Z29) transactivate the expression of the chloramphenicol acetyltransferase (CAT) reporter gene under the regulatory control of the HIV long terminal repeat (HIV LTR) in human peripheral blood mononuclear cells and in T-cell lines (14, 16, 17, 42). Our studies and others have identified at least six different cloned HHV-6 genomic fragments that can transactivate the HIV LTR (39). In our studies, the highest level of transactivation was seen with a 22-kb BamHI genomic fragment (pZVB70) of HHV-6A(GS) (14, 16, 17) and with 3.8- and 1.8-kb subfragments of pZVB70 (14). Within the 1.8-kb fragment, an open reading frame (ORF) encoding a 258-amino-acid (aa) protein mediated the transactivation (14). The first methionine encoded by this ORF was in the position corresponding to 115 aa downstream from the start of the ORF, and the region of the ORF corresponding to 143-aa carboxyl terminus of the product encoded a predicted protein of about 19-kDa. This ORF was designated B701, and its product was shown to activate CAT expression from an HIV-CAT plasmid in a cotransfection assay (14). The product of ORF B701 did not show any significant homology to the other viral proteins with transactivating functions, such as ICP0 and ICP4 of herpes simplex virus type 1, but did share some weak homology with the US22 family of immediate-early (IE) and early genes from HCMV. ORF B701 is positionally homologous to HCMV UL36 ORF exon 2 (15, 25). Comparison with the HHV-6A(U1102) published sequence shows that the HHV-6A(GS) ORF B701 corresponds to ORF U16. There are two potential TATA boxes upstream of the 143-aa coding region of ORF B701 (Fig. 1). Whether the mRNA for the putative B701 protein is transcribed from these promoter regions or is part of a larger protein generated by splicing is not known. In HCMV, the UL36 coding sequence has been shown to be generated by an in-frame splicing of the UL36 exons 1 and 2 (6, 12, 36, 37), and sequence analysis of the HHV-6 genome predicts a similar splicing site between the HHV-6A ORFs U17 and U16 (25). Using reverse transcriptase RT PCR, we show evidence of splicing in the transcripts encoded by the HHV-6A(GS) ORFs U16 and U17. Our data demonstrate that a family of transcripts is generated from the IE-B region of HHV-6, similar to the HCMV UL36-UL38 family of genes. Two cDNAs of 1.9 and 1.8 kb including ORF B701 were identified from an HHV-6A(GS) cDNA library. The 5' rapid amplification of cDNA ends (5' RACE) technique was used to identify the 5' ends of transcripts generated from U17 and U16. Two different TATA boxes located upstream of ORF U17 were identified as potential transcriptional start sites. An additional cDNA generated with a unique splicing event was also identified by 5' RACE. RT-PCR results suggest that some of the transcripts from this region are IE gene products and others are late gene products. Thus, while there are some similarities to the HCMV UL36-UL38 family, the U17/U16 family of transcripts has members that are unique to HHV-6.

View larger version (29K): [in this window] [in a new window]   FIG. 1.   Comparison of HHV-6A(GS) and HHV-6A(U1102) genomes encoding ORFs U15 to U17 and the relationships of the identified cDNAs. (A) Schematic representation of HHV-6 unique long region (UL). IE-A and IE-B, locations of the IE regions (16, 25). The locations of the 22.2-kb BamHI genomic fragment pZVB70 and the 3.8-kb SalI subfragment that transactivated the HIV-1 LTR (14, 17) are indicated, as are the seven TATA sequences and the two potential polyadenylation signal sequences P. (B and C) Relationships of the ORFs identified in HHV-6A(U1102) (B) and HHV-6A(GS) (C) to comparable regions, as well as their designations, are shown. The deduced ORFs A to C in the 3.8-kb DNA sequence are in the leftward reading frame (arrow). The first methionine codon in the HHV-6A(GS) ORF B, located at nucleotide 1377, and the location of the transactivating B701 ORF, used to screen the cDNA library, are shown in the hatched box. (D) Schematic representation of the virus-specific portion (E1E2) of the 1.9-kb cDNA and the included ORF. (E) Schematic representation of the 1.8-kb cDNA and the deduced ORFs. , intron locations. The polyadenylation signal used in each transcript is indicated by AAAAA. The arrow indicates the 5'-to-3' orientation of the cDNA ORFs; the nucleotide positions are numbered from right to left. The small hatched box indicates a 96-aa ORF.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells, virus, and infection procedures. The human T-cell line J-Jhan was maintained as a suspension culture in RPMI 1640 (JRH Biosciences, Lenexa, Kans.) containing 10% heat-inactivated Fetal Clone (JRH Biosciences) and antibiotics. HHV-6A(GS) was propagated in J-Jhan cells by standard procedures described previously (2, 4, 5). The percentage of infected cells was determined by immunofluorescence staining using HHV-6 monoclonal antibodies specific for early and late proteins (2, 4, 5). Cell-free virus was prepared by concentrating the culture supernatant collected from HHV-6A(GS)-infected J-Jhan cells (2, 4, 5). For infection experiments, 5 × 10⁷ J-Jhan cells were infected with a multiplicity of infection of 5 50% tissue culture infective doses per cell. Cycloheximide (50 mg/ml) (Sigma Chemicals, St. Louis, Mo.) or phosphonoacetic acid (PAA) (200 mg/ml) (Sigma) was added to some cultures. After being incubated for 2 h at 37°C, the cells were pelleted and the supernatant was replaced with 10 ml of fresh medium. The cycloheximide and PAA were added back to the appropriate cultures. The cultures were maintained for 8 h with and without cycloheximide and for 24 h with and without PAA. Mock-infected J-Jhan cells were incubated for 8 h with and without cycloheximide or for 24 h with and without PAA. Cells were harvested and washed in diethyl pyrocarbonate (Sigma)-treated phosphate-buffered saline. The pellets were frozen at 70°C until the RNA was prepared.

Plasmids. The construction of the pZVB70 plasmid containing the HHV-6A(GS) genomic BamHI B fragment (22.2 kb) in Bluescript vector (Stratagene, La Jolla, Calif.) has been described previously (14, 17, 18). The 3.8-kb subfragment (pSal I) containing the HIV LTR transactivating fragment was generated by the digestion of pZVB70 with SalI and cloning of the resulting fragments into Bluescript (14). The HHV-6 ORF B701 (U16) was amplified from the pZVB70 fragment (Fig. 1) by PCR using the 5' primer GGGTCGACATTATGAAGTCTTGCT with a SalI restriction site and the 3' primer TCAAAGCTTGACGTATCTATTT with a HindIII restriction site. The PCR product of B701 was ligated in frame with the prokaryotic expression vectors pATH-2, pET-3b, and pGEMEX (Promega, Madison, Wis.) (14, 25, 30). The correct orientation of the insert was confirmed by restriction enzyme digestion and sequencing. The induction of fusion proteins from these vectors was done as described previously (27). HHV-6A(GS)-specific sequences including ORFs U17 and U16 in the pBKS-9.1.7.1 cDNA were amplified by PCR using the 5' primer GGTAGAGAATTCCATAACGTAG with an EcoRI restriction site and the 3' primer GGCAAGCTTACATCTCCATACATC with a HindIII restriction site. The amplified 1.1-kb cDNA fragment was cloned into the pGEM-T vector (Promega). This construct was digested with EcoRI and HindIII and ligated in frame into the prokaryotic expression vector pGEMEX.

Antibodies. The pATH-B701 plasmid in Escherichia coli strain RR-1 was induced with indoleacrylic acid to express the TrpE-HHV-6 B701 fusion protein (27). The B701 fusion protein band was cut out from the preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel, electroeluted in an Elutrap apparatus (Schleicher and Schuell, Keene, N.H.), and quantitated by bicinchoninic acid assay (27). BALB/c mice were immunized subcutaneously with 5 µg of B701 fusion protein in Freund's complete adjuvant and then given a second dose in Freund's incomplete adjuvant. Monoclonal antibodies were generated using standard procedures described previously (2). The spleen cells were fused with Sp2/O myeloma cells, and the hybridoma supernatants were screened by enzyme-linked immunosorbent assay with B701 protein-coated 96-well plates. Positive cultures were cloned and screened, and high-titer ascitic fluids were raised by intraperitoneal injection of pristane-primed BALB/c mice (2). Polyclonal antisera were raised by immunizing male New Zealand White rabbits with 100 µg of B701 fusion protein in Freund's complete adjuvant and then boosting them with the same amount of antigen in Freund's incomplete adjuvant. The rabbits were boosted 2 months after the last injection and bled 10 days later. Rabbit anti-B701 immunoglobulin G was purified by affinity chromatography on a protein A-Sepharose column.

Southern blot analysis. Southern blotting was performed by standard methods using nylon membranes (31). DNA probes were radiolabeled with [³²P]dCTP (NEN-DuPont) using a nick translation system (GIBCO/BRL, Gaithersburg, Md.) or with digoxigenin-dUTP using the Genius system (Boehringer Mannheim) according to the manufacturers' instructions. For radiolabeled probes, specific hybridizing bands on the blots were visualized by autoradiography with Kodak (Rochester, N.Y.) XAR-5 film. For digoxigenin-labeled probes, specific hybridizing bands were detected by a colorimetric procedure in which an alkaline phosphatase-tagged antibody specific for digoxigenin was incubated with the membrane, followed by the addition of nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate as substrates for the enzyme.

Screening of the cDNA library. The construction of cDNA from HHV-6A(GS)-infected T cells has been described previously (5). For library screening, PCR products of the ORF B701 (HHV-6 U16) and the HHV-6 ORF U17 were labeled with digoxigenin and used as individual probes. The ORF U17 probe was amplified from the pZVB70 genomic DNA by PCR using primers ATATGGCAGACGAACGAA (5') and TCTTGCAACTCTGCGGCAGAC (3'). Positive phages were picked and purified by four sequential steps of screening. Phage DNA containing the cDNA insert was prepared from plate lysates. The recombinant phage DNA identified by the ORF B701 probe was digested with EcoRI, separated by agarose gel electrophoresis, electroeluted, and ligated into EcoRI-digested pBluescript KS() (pBKS) (Stratagene). The resulting construct, pBKS-c1.8, was transformed into the XL-Blue strain of E. coli. The cDNA insert identified by the ORF UL17 probe was amplified by PCR using gt11 insert screening amplimers (Clontech, Palo Alto, Calif.). The cDNA insert amplified by PCR was digested with EcoRI, cloned into EcoRI-digested pBKS (pBKS-9.1.7.1), and transformed into the XL-1 Blue strain of E. coli.

Sequencing of DNA. The HHV-6A(GS) 3.8-kb SalI genomic fragment was cloned into the M13mp18 vector (Pharmacia). Overlapping deletion mutants of both orientations were generated by the Erase-A-Base system (Promega), and sequencing was performed using the dideoxynucleotide chain termination method (30). A series of deletion mutants in both orientations was also generated for the cDNA inserts and sequenced by double-stranded sequencing using the Sequenase kit (U.S. Biochemicals, Cleveland, Ohio). For the GC-rich regions that were difficult to sequence with the Sequenase kit, the Bst DNA-sequencing kit (Bio-Rad, Hercules, Calif.) was used according to the manufacturer's instructions. Two primers, ATAGGAACATTAGCACCGTC and GCATACTCTCGCAGACAT, designed from the known sequences, were used to sequence across gaps not covered by the deletion mutants. Sequencing of cDNA clones amplified by 5' RACE was done by Sequenase version 2 dideoxy sequencing and cycle sequencing using the Licor L instrument in the Biotechnology Center at the University of Kansas Medical Center. Sequences were verified by restriction digestion analysis of the predicted restriction sites and restriction fragment lengths. The sequence data were analyzed with the IBI-Pustell GENEric sequence analysis programs, and amino acid homology analysis was conducted with the FASTA program. The sequence of the HHV-6A(GS) 3.8-kb SalI fragment was used as a reference, and all the numbered coordinates for the cDNA sequences refer to the 3.8-kb genomic sequence.

Isolation of RNA. RNA was isolated by a guanidinium isothiocyanate procedure (41) or by using the Trizol reagent (GIBCO/BRL) according to the manufacturer's instructions. Contaminating DNA was removed from the RNA with RQ1 RNase-free DNase (Promega). PCR amplification of HHV-6 DNA was carried out to ensure the complete removal of the DNA from the RNA preparations. The isolated total RNA was enriched for poly(A)⁺ mRNA using the Oligotex kit (Qiagen, Chatsworth, Calif.) according to the manufacturer's instructions.

5' RACE. PCR primers and oligonucleotide probes from HHV-6 sequences were synthesized at Life Technologies, GIBCO/BRL. Two different kits from Clontech were used according to the manufacturer's instructions to determine the 5' ends of the cDNAs. For the 5'-amplifinder RACE kit, the RNA used was from HHV-6A(GS)-infected J-Jhan cells, with approximately 50% of the cells expressing viral antigens. Single-stranded cDNAs were synthesized by avian myoblastosis virus RT using an HHV-6 specific antisense primer, P1 (U16-P1; GAATGTCCTTTTCTCGATATACC). The cDNAs were purified, and single-stranded anchor oligonucleotides in the kit were ligated to the 5' ends. PCR was then performed using a sense primer specific for the anchor and an antisense gene-specific primer (P2; CAAATCGTTCCGTCGTTTCCGCC) internal to P1 (see Fig. 6). For the Marathon cDNA amplification kit (Clontech), the source of RNA was HHV-6A(GS)-infected J-Jhan cells that were approximately 30 or 80% infected. The P1 primer mentioned above was used for the PCR. The double-stranded cDNA PCR products from both protocols were ligated into the pGEM-T vector with 3' T overhangs. Clones were screened for virus-specific inserts by Southern blotting with a digoxigenin-labeled U17 gene as a probe, and positive clones were sequenced. The junction between the virus-specific sequence termination and initiation of the adapter sequence ligated on the cDNA end was identified as the 5' end of the messages. The nucleotide numbers from the 3.8-kb genomic sequence were used for numbering the cDNA sequences. Transcription start sites were considered authentic when more than one mature transcript terminated near the site. To confirm the splicing patterns found in the RACE cDNA (see Fig. 6C), PCR was carried out with the cDNA made from the Marathon kit using two virus-specific primers designed to amplify across the splice junctions. The 5' sense primer (ATCTCTGGATTGTTCCTTCCGTTG) starts at position 180, and the 3' antisense primer (GACGGTGCTAATGTTCCTAT) starts at position 1715 on the 3.8-kb map (see Fig. 5). The PCR products were cloned into pGEM-T and sequenced.

Riboprobe synthesis and Northern blot analysis. To synthesize the riboprobes, the cDNAs of ORFs U17/U16 and ORF U16 alone were cloned into the pGEM-T vector. Control riboprobes were made from the HHV-6 early gene ORF 27 cDNA (5) and the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) cDNA. Synthesis of the radiolabeled riboprobes was done using the In Vitro transcription system (Promega) according to the manufacturer's instructions. The riboprobes were labeled by incorporating [-³²P]CTP (NEN-DuPont) in the riboprobe reaction mixture. More than 10⁷ cpm of each riboprobe was added to hybridization solutions for each blot. Northern blotting was performed using standard procedures (30). HHV-6A(GS)-infected cells were collected at 96 h postinfection, when about 50% of the cells were expressing viral antigens. Poly(A)⁺-selected RNA was run on 1 or 1.5% agarose gels denatured with formaldehyde. The blots were hybridized with the HHV-6-specific probe first, stripped by boiling them in 0.1× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), and reprobed with the GAPDH-specific probe.

RT-PCR. RNA from infected and uninfected cells was treated with DNase, and the complete removal of DNA was tested by PCR using primers designed to amplify an HHV-6A(GS) IE gene and the cellular actin gene. RT-PCR was then carried out using the GeneAmp RNA PCR core kit (Perkin-Elmer, Foster City, Calif.) according to the manufacturer's instructions. Random hexamers were used to prime the RT reaction. The initial PCR was done for 40 cycles, and an aliquot was used in a nested PCR for another 40 cycles of amplification. The primers used for amplifying the various mRNAs are given in Table 1.

In vitro transcription and translation. The cDNA inserts in the vectors containing SP6 and T3 promoters were used for in vitro transcription experiments. Synthesis of sense and antisense RNA transcripts using SP6 and T3 RNA polymerases and capping of RNAs at the 5' ends were carried out using procedures described in the Riboprobe system instruction manual (Promega). RNA transcripts were translated in vitro using [³⁵S]methionine (ICN, Irvine, Calif.) and rabbit reticulocyte lysate preparations with and without canine microsomal membranes (Promega) according to the manufacturer's recommendations. Samples of in vitro-translated products were boiled with sample buffer, and equal trichloroacetic acid-precipitable radioactive counts were analyzed by SDS-PAGE. In vitro-translated products mixed with equal volumes of lysing buffer (0.05 M Tris hydrochloride, 0.15 M NaCl, 1% sodium deoxycholate, 1% Triton X-100, 100 U of aprotinin per ml, 0.1 mM phenylmethylsulfonyl fluoride) were used for immunoprecipitation with monoclonal and rabbit polyclonal antibodies. Lysates containing equal amounts of trichloroacetic acid-precipitable counts were mixed with 10 µl of antibodies and 100 µl of protein A-Sepharose beads. Samples were mixed continuously at 4°C for 2 h. The immunoprecipitates were collected, washed, dissociated by boiling them in sample buffer, and analyzed by SDS-PAGE (2, 4).

Radiolabeling procedures, radioimmunoprecipitation, and SDS-PAGE. Infection, radiolabeling, radioimmunoprecipitation, and SDS-PAGE were carried out as described previously (2, 4). Briefly, 10⁷ cells were mixed with 5 to 10 50% tissue culture infective doses of HHV-6A(GS) and incubated at 37°C for 2 h. Unabsorbed virus was removed by centrifugation, and the infected cells were incubated at 37°C. On day 3 postinfection, 10⁷ uninfected and infected cells were labeled for 20 h with 250 µCi of [³⁵S]methionine or cysteine (Tran³⁵S label; specific activity, 1,177 Ci/mmol; ICN). The lysates were used for immunoprecipitation reactions as described above. Labeled samples and molecular mass markers (Sigma) were electrophoresed in parallel lanes. The gels were stained, destained, infused with 1 M salicyclic acid, dried on filter paper, and placed in contact with XAR-5 film at 70°C for fluorography.

Transfection procedures and CAT assays. The HHV-6 ORFs U17/U16, U16+, and B701 (14) and HIV-Tat were cloned into the eukaryotic expression vector pRC/RSV (Invitrogen, San Diego, Calif.). Transfection of CV-1 monkey kidney cells or CEM human T cells with the plasmids and the [¹⁴C]chloramphenicol CAT assays were done with procedures described previously (14, 16, 17, 42). The acetylated chloramphenicol was separated from the unacetylated form by thin-layer chromatography (TLC). The percentage of CAT conversion was calculated by dividing the radioactive counts present in the acetylated chloramphenicol spots by the total radioactivity present in both the acetylated and unacetylated chloramphenicol spots on TLC (16, 42). Fold activation was calculated as the increase in CAT activity over the background levels for HIV-CAT with the pRC/RSV vector.

Nucleotide sequence accession number. The nucleotide sequence reported here has been deposited in the GenBank database and assigned accession no. AF294810.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Comparison of genomic sequences of the HHV-6A(GS) 3.8-kb SalI transactivating fragment and the homologous region in HHV-6A(U1102). The transactivating 3.8-kb SalI subfragment of the 22-kb BamHI genomic fragment (Fig. 1A) of HHV-6A(GS) was sequenced, and the results were compared with the sequence of the same region of HHV-6A strain U1102 (15, 24) (Fig. 1B and C). Although the sequences of the two strains are homologous in this region, there are significant differences, and for that reason, the base pair coordinates given refer to the 3.8-kb sequence derived in this study. Three leftward ORFs designated GS-ORF A to C are included within the 3.8-kb DNA (Fig. 1C), with ORFs A and B in reading frame 3 and ORF C in reading frame 2. The HHV-6(GS) ORF A initiates at nucleotide position 702 and ends at nucleotide 978, with the first methionine codon at nucleotide 720. The predicted protein is 86 aa in length. HHV-6A (U1102) ORF U17, formerly EFLF1 (15, 25), corresponds to GS-ORF A, but U1102-U17 encodes a larger protein of 133 aa (Fig. 1B). Except for one amino acid, the 88-aa product of GS-ORF A is identical to the first 88-aa stretch of the product of U1102-ORF U17. GS-ORF B starts at nucleotide 1032, ends at nucleotide 1805, and potentially encodes a protein of 258 aa. The first methionine codon in GS-ORF B is at nucleotide 1377 (Fig. 1C). The HHV-6A (U1102) ORF U16, formerly EFLF2 (15, 25) corresponds to GS-ORF B and encodes a protein of 258 aa. The predicted protein of GS-ORF B differs from the U1102-U16 protein by two amino acids. GS-ORF C initiates at nucleotide position 2000 and ends at nucleotide 2398, with a first methionine codon at nucleotide 2069. The predicted protein is 110 aa long and does not show any significant homology with other known herpesvirus sequences. The HHV-6A(U1102) ORF U15, formerly EFLF3 (15, 25), corresponds to GS-ORF C and encodes a protein of 110 aa. It is 100% identical to GS-ORF C. Within the 3.8-kb SalI genomic fragment, seven potential TATA sequences have been identified at nucleotide positions 95, 335, 640, 790, 1249, 1355, and 2005 (Fig. 1A). Two potential polyadenylation signal sequences are at nucleotide positions 1869 and 2892.

Transcripts encoding HHV-6A ORFs U16 and -17 are spliced. In the HHV-6A(GS) ORF A (U17) and ORF B (U16) sequence, a potential splice donor site [T/(GT)AAGT] at position 948 and an acceptor site [A/(AG)G] at position 1032 were proposed, based on the sequence homology to HCMV. To determine whether the HHV-6A(GS) ORFs A and B are joined by splicing of the transcript, HHV-6A(GS)-infected total cell RNA was reverse transcribed with random hexamers, and the resulting cDNAs were subjected to PCR using specific primers (F and R) designed to direct amplification across the predicted splice junction (Fig. 2A, primer set 1, and Table 1). Amplifications of 570- and 486-bp products, representing unspliced and spliced RNA, respectively, were predicted. The PCR products were Southern blotted and probed with a digoxigenin-labeled oligonucleotide probe (O) that is specific for ORF B (U16) and internal to the primers used for the cDNA amplification (Fig. 2A, and Table 1). RT-PCR was also performed using primers designed to amplify a 458-bp fragment encoding the B701 ORF (Fig. 2A, primer set 2, and Table 1). This was done to detect potential splicing within this region. RT-PCR, using primers specific for the cellular actin transcripts (35), was performed as a control reaction (Fig. 2A and Table 1).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   RT-PCR showing evidence for splicing in the transcripts encoding HHV-6A ORFs U16 and U17. (A) Schematic locations of primer sets and oligonucleotides used in RT-PCR and expected sizes of DNA products amplified from spliced and unspliced transcripts. The ORFs U17 and U16 correspond to Fig. 1 ORFs A and B, respectively. The arrowhead labeled F designates the location of the forward primer, and the arrowhead labeled R designates the location of the reverse primer. The line labeled O indicates the location of the probe used to detect the specific PCR products. (B) Southern blot of RT-PCR products. Lane 1, RT-PCR with total RNA from uninfected J-Jhan cells. Lane 2, RT-PCR with total RNA from HHV-6A(GS)-infected J-Jhan cells. The sizes of the DNAs amplified from spliced and unspliced transcripts by the primers shown in panel A are indicated.

Identification of a 1.9-kb cDNA encoding the HHV-6 ORFs U16 and U17. To clearly define the splicing event detected above and describe the resulting ORFs, a cDNA library from HHV-6A(GS)-infected cells, constructed in gt11 (2), was screened by plaque hybridization using digoxigenin-labeled B701 DNA (458 bp) as a probe. After the fourth screening, two recombinant phages with 1.8- and 1.9-kb inserts were isolated and characterized further. The 1.9-kb insert identified by the B701 (U16) probe also hybridized with the digoxigenin-labeled U17 probe, indicating the presence of ORF U17 in this transcript. In contrast, the 1.8-kb cDNA hybridized only with the B701 (U16) probe and not with the U17 probe (data not shown). The 1.9-kb cDNA insert was cloned into Bluescript, sequenced, and compared with the genomic sequences. DNA sequence analysis revealed that the cDNA insert consisted of approximately 1,100 bp of virus-specific cDNA (Fig. 1D). This portion of the cDNA was referred to as E1E2. As predicted, the 1.1-kb virus-specific portion of the 1.9-kb cDNA contained both U17 and U16 ORFs (Fig. 1D) spliced together into one large ORF by the removal of the 84-bp genomic intron sequence. The resulting ORF was predicted to encode a protein 335 aa long with a molecular mass of 38 kDa. There were four base pair differences in the virus-specific cDNA compared with the 3.8-kb genomic sequence, but the reading frame was not affected. The polyadenylation signal immediately 3' to the U16 ORF, nucleotide position 1869 (Fig. 1), was used to terminate this transcript. Approximately 800 bp at the 5' end of the 1.9-kb cDNA insert sequence did not share any homology with the HHV-6A(GS) 3.8-kb sequence or with the complete genomic sequences of HHV-6A(U1102) in the data bank (data not shown). However, this sequence was 97% homologous with the human mitochondrial gene coxIII. This indicated that the 1.1-kb portion including U17 and U16 was the only HHV-6A(GS)-specific portion of the 1.9-kb cDNA. The coxIII gene sequences identified in the 1.9-kb cDNA included an ORF, but it was in the orientation opposite to that of the virus-specific ORF. The coxIII portion was probably the result of a cloning artifact that occurred during the generation of the cDNA library.

Identification of a 1.8-kb cDNA including the HHV-6 U16 and U15 ORFs. DNA sequence analysis of the 1.8-kb cDNA revealed that the insert consisted of 1,734 bp and included three ORFs (Fig. 1E). Comparison with the genomic 3.8-kb SalI sequence showed only four base pair differences, which did not affect the reading frames, and demonstrated that this cDNA lacked ORF U17. The ORF at the 5' end of the cDNA started at nucleotide position 1020 and ended at nucleotide position 1805, and the first methionine codon was located at nucleotide position 1377. This ORF was in reading frame 1 and corresponded to the HHV-6A(GS) ORF B, or HHV-6A(U1102) U16 (Fig. 1C), which includes the B701 ORF used to screen the cDNA library. The next ORF in the cDNA was in reading frame 3, starting at position 2001 and ending at 2398. This ORF corresponded to the HHV-6A(GS) ORF C, or HHV-6A(U1102) U15 (formerly EFLF3), in the genomic sequence (16, 25) (Fig. 1B, C, and E). The first methionine codon for U15 was at position 2069 and was in a good context for a translational start site. Toward the 3' end of the cDNA (at bp 1511), 160 bp of the genomic sequence was spliced out, and the splice donor and acceptor consensus sequences were identified at the boundaries (from positions 2531 to 2690 of the 3.8-kb SalI sequence). The third ORF generated by this splicing is predicted to encode a protein of 96 aa, but the first methionine codon is not in the appropriate context for a translational start site. The polyadenylation signal identified at the 3' end of this cDNA corresponds to the polyadenylation signal found at position 2892 of the 3.8-kb SalI genomic DNA fragment.

In vitro expression of cDNA inserts containing HHV-6 ORFs U16/U15 and U17/U16. To determine the specificity of the proteins encoded by the 1.8- and 1.9-kb cDNAs, we raised rabbit polyclonal antisera and generated murine monoclonal antibodies specific for the 148-aa carboxyl terminus of the U16 (B701) protein that is common to both cDNA products. These antibodies specifically reacted with the B701 protein expressed in bacteria, as demonstrated by the Western blot shown in Fig. 3A, lane 1. Antibodies recognized a protein with a molecular weight of approximately 20,000 (20K), as well as higher-molecular-weight proteins, probably representing dimeric and multimeric forms of the B701 protein. These antibodies were subsequently used in radioimmunoprecipitation and Western blot reactions with proteins expressed from the cDNA inserts. In the 1.8-kb cDNA, the first ORF, U16, encodes a protein 261 aa long, with the first methionine codon at position 119. Translation of this ORF should give rise to a protein of about 19K (B701). In vitro translation of RNA generated from the 1.8-kb cDNA insert including the U16 and U15 ORFs resulted in three polypeptides of about 24, 20, and 14K (Fig. 3B, lane 2). No specific polypeptides were synthesized from RNA transcribed from the T7 promoter, representing the antisense RNA (Fig. 3B, lane 3), or in control reactions without RNA (Fig. 3B, lane 1). In immunoprecipitation reactions (Fig. 3B, lanes 4 to 9), B701-specific polyclonal rabbit antiserum (Fig. 3B, lanes 4 to 6) and murine monoclonal antibody (Fig. 3B, lanes 7 to 9) immunoprecipitated the 24, 20, and 14K polypeptides generated from the Sp6 transcript (Fig. 3B, lanes 6 and 9). No specific reaction was seen with control preimmune rabbit serum and normal mouse serum (data not shown). This indicated that the in vitro-translated polypeptides from the 1.8-kb cDNA were encoded by the U16 ORF. The 24K protein probably represented the primary translation product of U16 and the observed molecular weight larger than the predicted 19K may have resulted from migration characteristics of this protein in SDS-PAGE. The smaller polypeptides probably represent products from incomplete translation, incomplete transcription, or internal initiation of translation.

View larger version (50K): [in this window] [in a new window]   FIG. 3.   (A) Reactivity of monoclonal antibody against the B701 (U16) protein. Lane 1, reactivities of antibodies in the Western blot reaction with the B701 protein expressed in the pET-3b vector. The 20K protein and the higher-molecular-weight dimeric and/or multimeric forms of B701 protein recognized are indicated by the arrowheads. (B) SDS-PAGE analysis of in vitro-expressed polypeptides from the 1.8-kb cDNA insert. In vitro-synthesized mRNA, transcribed from the cDNA insert in the pGEM-T plasmid, was translated in vitro using rabbit reticulocyte lysate. Lanes 1, 4, and 7, translation reaction components without the addition of RNA. Lanes 2, 6, and 9, translation from RNA transcribed with SP6 RNA polymerase. Lanes 3, 5, and 8, translation from RNA transcribed with T7 RNA polymerase. Lanes 1 to 3, total translated products. Lanes 4 to 6, immunoprecipitations of translated products using a monoclonal antibody against the B701 (U16) protein. Lanes 7 to 9, immunoprecipitations of translation products using polyclonal rabbit antisera raised against the B701 protein. Samples were analyzed on 12% acrylamide cross-linked with bisacrylamide, and standard molecular mass markers were included in parallel lanes. The numbers indicate approximate molecular masses (in kilodaltons) of the prominent polypeptides identified.

View larger version (40K): [in this window] [in a new window]   FIG. 4.   (A) SDS-PAGE analysis of in vitro-expressed polypeptides from the 1.1-kb E1E2 (U17/U16) cDNA insert. In vitro-synthesized mRNA, transcribed from the cDNA insert in the pGEMEX plasmid, was translated in vitro using rabbit reticulocyte lysate. Lanes 1, 3, 5, and 7, translation from RNA transcribed with SP6 RNA polymerase. Lanes 2, 4, 6, and 8, translation from RNA transcribed with T3 RNA polymerase. Lanes 7 and 8, total translated products. Lanes 1, 2, 5, and 6, immunoprecipitations of translated products with two different polyclonal rabbit antisera raised against the B701 protein. Lanes 3 and 4, immunoprecipitations of translation products with a monoclonal antibody against the B701 (U16) protein. Samples were analyzed on a 12% acrylamide gel cross-linked with bisacrylamide, and standard molecular mass markers were included in parallel lanes. The numbers indicate the approximate molecular masses (in kilodaltons) of the prominent polypeptides identified. (B) Expression of U17/U16 from the 1.1-kb cDNA in pGEMEX-E1E2. Bacteria transformed with either pGEMEX or pGEMEX-E1E2 were induced with IPTG for various lengths of time, and the lysate pellets and soluble supernatants were collected. The proteins were analyzed on SDS-12% PAGE, transferred to nitrocellulose membranes, and reacted with monoclonal antibody against B701 (U16) protein. Lanes 1 and 2, pellet and supernatant from bacteria with control pGEMEX plasmid induced with IPTG for 6 h. Lanes 3 and 4, pellet and supernatant from bacteria with pGEMEX-E1E2 plasmid induced with IPTG for 2 h. Lanes 5 and 6, pellet and supernatant from bacteria with pGEMEX-E1E2 plasmid induced with IPTG for 4 h. The fusion protein recognized is indicated by the arrowhead on the right.

Identification of HHV-6 U17/U16 gene products in infected cells. To identify the HHV-6 U16/U17 gene products in the infected cells, uninfected and HHV-6-infected cells collected after 3 days postinfection were labeled for 20 h with [³⁵S]methionine and used in immunoprecipitation assays. Rabbit polyclonal antibodies against part of the U16 (B701) protein immunoprecipitated HHV-6(GS)-specific polypeptides with approximate molecular weights of 51, 45, 38, and 34K (Fig. 5A, lane 4). These polypeptides were not recognized from uninfected cells (Fig. 5A, lane 3) and were not recognized by preimmune serum (Fig. 5A, lanes 1 and 2). Similar-size polypeptides from HHV-6-infected cells were also recognized by the monoclonal antibody against U16 (B701) (Fig. 5B, lane 2).

View larger version (57K): [in this window] [in a new window]   FIG. 5.   Identification of HHV-6 U17/U16 gene products in infected cells. (A) Lanes 1 and 3, uninfected J-Jhan cells. Lanes 2 and 4, HHV-6A(GS)-infected J-Jhan cells. The cells were labeled for 20 h with [35S]methionine and used for immunoprecipitation. PB (lanes 1 and 2), preimmune serum. RaB701 (lanes 3 and 4), rabbit antibodies raised against the B701 protein. (B) Immunoprecipitation with monoclonal antibody against B701 protein. Lane 1, uninfected J-Jhan cells. Lane 2, HHV-6A(GS)-infected J-Jhan cells. The numbers on the right indicate approximate molecular masses (in kilodaltons) of HHV-6-specific polypeptides.

Transcripts originating from HHV-6 U17/U16 genes are differentially spliced and are initiated from two potential transcriptional start sites. To determine whether the 1.8-kb cDNA encoding ORFs U16 and U15 represents a complete or incomplete transcript and to determine the transcriptional start site for the U17/U16-encoded transcripts, 5' RACE was done using mRNA from HHV-6A(GS)-infected cells. The sequencing analysis of the 3.8-kb SalI genomic fragment from HHV-6A(GS) identified numerous TATA sequences, located upstream or within the coding regions of ORFs 17 and 16 (Fig. 6), that could potentially serve as transcription initiation sites. Figure 6 shows a schematic representation of ORFs U16 and U17, the 5' untranslated region, the potential TATA boxes, and the locations of the P1 and P2 primers used for 5' RACE. Sequencing of several cDNAs generated by 5' RACE showed that the cDNAs represented both spliced and unspliced messages (Fig. 6A to C). In the spliced transcripts represented in Fig. 6A, an 84-bp intron sequence was removed in frame between ORFs U17 and U16, generating the combined U17/U16 ORF previously identified in the 1.1-kb cDNA, with the first methionine codon at the 5' end of U17. The 5' ends of two of these spliced transcripts initiated at nucleotide position 680, which is 39 bp downstream of the TATA-1 sequence at nucleotide positions 637 to 641 in the SalI 3.8-kb genomic fragment (Fig. 6A). Since this TATA-1 sequence is 83 bp upstream of the first methionine in ORF U17 (Fig. 6A), this suggested that the TATA box could function as the promoter site for these messages (Fig. 6A). A third spliced transcript and an unspliced transcript, which initiated 54 bp downstream of the TATA-1 site, were also identified, and they could represent incompletely extended cDNAs generated by the 5' RACE technique.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Identification of 5' ends of cDNAs. A schematic representation of the ORFs U16 and U17 from HHV-6A(GS) is shown at the top. The numbers correspond to the sequence numbers of the 3.8-kb SalI genomic fragment. The bent arrows indicate the positions of potential TATA sites for transcription initiation. Transcription of this region occurs in the leftward orientation, and the AAA at the left side of U16 indicates the position of the polyadenylation signal used by the U17/U16 spliced transcript. P1 and P2 indicate the positions of the nested antisense oligonucleotide primers used for the 5' RACE procedures. (A and B) Schematic representations of the cDNAs that encode the U17/U16 splice product. The vertical lines mark the transcriptional start sites, and the numbers above correspond to the 3.8-kb map. The 84-bp intron was removed from both transcripts, generating one continuous ORF represented by a hatched box underneath each transcript. The protein products encoded by these cDNAs are identical, with the first methionine codon at position 720 in U17 (indicated by M). (A) Schematic representation of the 5' RACE products that terminated near the TATA-1 promoter at position 640. The 5' end of the cDNA was located at approximately position 680. (B) Schematic representation of the 5' RACE products that terminated near the TATA-2 promoter at position 95. The 5' end of the cDNA was found at approximately position 175. (C) Schematic representation of a second 5' RACE product that terminated near the TATA-2 promoter. The 5' end of the cDNA was at approximately position 175; an 84-bp intron between U17 and U16 and a 508-bp intron, which included part of U17 and the upstream untranslated region, were removed. The ORF is indicated by a thick line, and the first methionine codon, indicated by M, is at the end of U17. The hatched box underneath the transcript represents the protein translated from the transcript. The bent line indicates intron locations.

Analysis of transcripts generated from the HHV-6(GS) U17/U16 genes. The 5' RACE results indicated heterogeneity in the transcripts and suggested potential differential regulation of the transcripts from the HHV-6 U17/U16 region. Northern blot reactions were performed to determine the extent of the heterogeneity and to evaluate the kinetics of expression of these transcripts. Radiolabeled antisense riboprobes from the HHV-6 U16, U17/U16, and U27 (early-late) genes and the cellular GAPDH genes were used to probe poly(A)⁺-selected RNA from HHV-6A(GS)-infected and uninfected J-Jhan cells. The U16 riboprobe hybridized with multiple transcripts of 3.9, 2.8, 2.5, and 1.5 to 1.6 kb from RNA purified from infected J-Jhan cells (Fig. 7A, lane 1), and no specific binding was detected with RNA from uninfected J-Jhan cells (Fig. 7A, lane 2). Transcripts of similar sizes were also detected in RNA from infected cells by the radiolabeled U17/U16 riboprobe (Fig. 7B, lanes 1 and 2). Northern blot analysis using an antisense riboprobe generated from the U17 ORF alone was not successful, due to nonspecific binding of the probe to residual rRNA, both in uninfected and infected cells (data not shown). However, there were no transcripts unique to the U17/U16 probe, suggesting that transcripts encoded by the U17 gene were not expressed independently of the U16 gene. As a control for virus-specific transcripts, we used the HHV-6 early gene U27 (5), and the transcripts recognized by the U27 riboprobe (Fig. 7C) were of sizes comparable to those reported previously (5). Though similar amounts of RNA samples and radioactive probes were used in these Northern blot reactions, the U16 and U17 transcripts were visualized only after 72 h of exposure of the autoradiographs while the U27 transcripts were detected by 4-h exposure of the autoradiograph. This suggested that the relative abundances of the U16- and U17-specific transcripts were quite low compared with the expression of the U27-specific transcripts. After the HHV-6-specific transcription pattern was determined, these blots were stripped and rehybridized with a riboprobe specific for GAPDH (Fig. 7). The GAPDH riboprobe hybridized to one species of RNA from both infected and uninfected J-Jhan cells, indicating that the various sizes of viral transcripts detected were not the result of degradation of the poly(A)⁺-selected RNA.

View larger version (33K): [in this window] [in a new window]   FIG. 7.   Northern blot analysis of the transcripts encoded by U16 and U17/U16. (A) Transcripts detected by radiolabeled antisense riboprobe generated from the U16 ORF of HHV-6A(GS). The approximate sizes (in kilobases) of the transcripts detected are indicated next to the arrows. (B) Transcripts detected by radiolabeled antisense riboprobe generated from the HHV-6A(GS) U17/U16 spliced cDNA. (C) Transcripts detected by radiolabeled antisense riboprobe generated from the HHV-6A(GS) U27 early gene. (A, B, and C) Each blot was stripped and reprobed with an antisense riboprobe generated from the GAPDH host cell gene. The GAPDH-specific Northern blot is shown below its respective viral-gene blot. Lanes 1, poly(A)+-selected RNA from HHV-6A(GS)-infected J-Jhan cells expressing viral antigens in more than 50% of the cells. Lanes 2, RNA isolated from uninfected J-Jhan cells. Lanes 3 to 7, kinetics of RNA expression from the U17/U16 region. J-Jhan cells were infected with HHV-6A(GS) as described in Materials and Methods, and samples were collected at various times after the initiation of infection. Lane 3, 8-h infections in the presence of cycloheximide. Lane 4, 8-h infections. Lane 5, 24-h infections. Lane 6, 24-h infections in the presence of PAA. Lane 7, 48-h infections.

RT-PCR identified both IE and late transcripts from the HHV-6A(GS) U17/U16 IE-B region. The percentage of infected cells in the Northern blot time course experiment was low, resulting in low expression of the HHV-6A(GS) U17/U16-specific transcripts. Since cycloheximide can decrease expression of some IE genes (36, 37), evaluating the expression of HHV-6 U17/U16 with a more sensitive method was necessary. RT-PCR was carried out with RNA from HHV-6A(GS)-infected J-Jhan cells collected at different time points. Three different primer pairs (Table 1) were used to detect U17- and/or U16-containing transcripts. Figure 8C shows a schematic diagram of the primers used to detect B701 (U16) and the sizes of the expected PCR products. Since the 3' antisense PCR primer for the B701 protein was at the 3' end of the ORF, detecting it by PCR from the cDNA primed with the random hexamers was difficult. We therefore modified the RT-PCR procedure slightly and used the B701-specific 3' antisense primer to prime the cDNA synthesis by RT. Subsequently, the same antisense primer and the 5' sense primer (Fig. 8C, primer pair I, and Table 1) were used for the PCR. The B701-specific PCR products were detected by Southern blot analysis using a digoxigenin-labeled oligonucleotide probe hybridizing to a site internal to the primers (Table 1). The transcript including ORF B701 (458-bp PCR product) was detected 24 h after infection (Fig. 8A, lane 6), and expression was not affected by the presence of PAA (Fig. 8A, lane 7). The expression of this transcript was maintained throughout the observed 72 h of infection (Fig. 8A, lanes 8 and 9). We reasoned that the sensitivity of the RT-PCR would be increased by using an antisense primer for the PCR that was internal to the primer used to reverse transcribe the cDNA. The cDNA generated using the initial antisense primer was subjected to PCR using a 3' antisense primer internal to the RT primer and the same 5' sense primer, amplifying a 344-bp PCR product (Fig. 8C, primer pair N, and Table 1). In these nested PCRs, a low level of the B701 transcript was detected at 8 h after infection (Fig. 8B, lane 5), and the expression of the transcript was not inhibited by cycloheximide (Fig. 8B, lane 4). This indicated that, similar to the HCMV UL36 products, the HHV-6 U16 ORF-encoded transcripts were expressed at low levels under IE conditions. RT-PCR was negative for virus-specific transcripts in the control RNAs from uninfected cells (Fig. 8A and B, lanes 1 to 3), demonstrating the specificity of these procedures.

View larger version (21K): [in this window] [in a new window]   FIG. 8.   Detection of B701-specific transcripts by RT-PCR in a single round of PCR amplification (A) and a second nested round of PCR amplification (B). The PCR products were subjected to Southern blot analysis to show B701-specific RT-PCR products. Lanes 1 to 3, RNA isolated from uninfected J-Jhan cells cultured for 8 h in the presence of cycloheximide (lane 1), for 24 h in the presence of PAA (lane 2), and untreated (lane 3). Lanes 4 to 9, RNA isolated from HHV-6A(GS)-infected J-Jhan cells 8 h postinfection in the presence of cycloheximide (lane 4), 8 h postinfection (lane 5), 24 h postinfection (lane 6), 24 h postinfection in the presence of PAA (lane 7), 48 h postinfection (lane 8), and 72 h postinfection (lane 9). (C) Schematic representation of the U16 and U17 ORFs, with the B701 portion of U16 indicated by hatch marks. The numbers correspond to the sequence of the 3.8-kb SalI genomic fragment. The arrows mark the location of the 3' antisense RT primer, the set of primers used for the initial PCR in panel A (I), and the set of primers used for the nested PCR in panel B (N). The line with circles marks the location of the oligonucleotide probe. The expected PCR product is indicated by a thick line, and the predicted size of the product is shown to the right. The number in parentheses is the size of the nested PCR product.

View larger version (25K): [in this window] [in a new window]   FIG. 9.   Detection of U17/U16-specific transcripts by RT-PCR. Southern blots of U17/U16-specific RT-PCR products detected by a single round of PCR amplification (A) and a second nested round of PCR amplification (B) are shown. (A and B) Details of the identification of RNA samples used and the lanes with each sample's products are given in the legend to Fig. 7. (C) Schematic representation of ORFs U16 and U17 and expected sizes of PCR products. The arrows mark the locations of the primers used for PCR in panel A (I) and the nested PCR in panel B (N). The line with circles internal to both sets of primers marks the location of the oligonucleotide used to detect the specific transcripts. The thick black lines below indicate the two transcripts predicted, spliced and unspliced, and the predicted sizes of the products are shown to the right. The number in parentheses is the size of the nested product. These numbers correspond to the sizes of the bands marked by arrows in panels A and B.

View larger version (30K): [in this window] [in a new window]   FIG. 10.   Detection of U16+-specific transcripts by RT-PCR. Southern blots of U16+-specific RT-PCR products by a single round of PCR amplification (A) and a second nested round of PCR amplification (B) are shown. (A and B) Details of the identification of RNA samples used and the lanes with each sample's products are given in the legend to Fig. 7. (C) Schematic representation of U16 and U17 and the upstream noncoding region. The arrows mark the two sets of PCR primers used, initial primers (I) used in panel A and the nested primers (N) used in panel B. The line with circles marks the location of the oligonucleotide used to detect the specific products on Southern blots. The thick lines underneath represent the four possible transcripts, and the sizes of the PCR products are shown to the left. The number in parentheses is the size of the nested-PCR product.

View larger version (51K): [in this window] [in a new window]   FIG. 11.   Detection of virus-specific and host cell-specific control transcripts by RT-PCR. Details of the identification of the RNA samples used and the products are given in the legend to Fig. 7. The lanes for each blot are arranged so that identical RNA samples are vertically aligned. (A) Detection of U89 IE gene products by a single round of PCR. The sizes of the products, spliced and unspliced, are shown at the right. RNAs from cells infected for 8 h with or without cycloheximide were subjected to a second round of nested-PCR amplification, and the Southern blot of each product is shown below the corresponding lane of the initial amplification. (B) Detection of the multiply spliced transcripts encoded by U100 by Southern blotting of RT-PCR products. (C) Detection of the host cell-specific actin gene by RT-PCR.

View larger version (14K): [in this window] [in a new window]   FIG. 12.   Summary of transcripts detected by RT-PCR. (A) Schematic representation of U17, U16, and the 700-bp untranslated region upstream of U17. The two TATA box promoters used to generate the transcripts detected by RT-PCR are indicated by bent arrows, and the numbers underneath correspond to the sequence locations on the 3.8-kb SalI genomic fragment. (B) Transcripts detected in RNA samples generated from cells infected for 8 h in the presence of cycloheximide. (C) Transcripts detected in RNA samples generated from cells infected for 24 h with no inhibitors present. These transcripts were not seen in RT-PCRs on RNA generated from cells infected for 24 h in the presence of PAA.

Transactivation of HIV LTR CAT by ORFs U17/U16 and U16+. To determine the transactivating ability of the ORFs U17/U16 and U16+, the HIV-CAT expression vector and a eukaryotic expression vector containing either U17/U16, U16+, or B701 were cotransfected into CV-1 or CEM cells and the CAT expression was measured (Fig. 13). HHV-6 A(GS) infects the human T-cell line CEM (20), and the transactivating activities of B701 in HSB-2 T cells and CV-1 fibroblasts were shown to be comparable (14). Both the U17/U16 and U16+ ORFs transactivated the HIV LTR (Fig. 13), increasing CAT expression about four- to sevenfold. This level of activation was similar to the transactivation observed previously when B701 was shown to transactivate the HIV LTR (14).

View larger version (29K): [in this window] [in a new window]   FIG. 13.   Activation of HIV LTR by U17/U16 and U16+ ORFs. CV-1 cells (A) and CEM cells (B) were transfected with 2.5 µg of HIV-CAT (A and B, lanes 1) and cotransfected with pRC/RSV vector (V) alone (A, lane 5; B, lane 4) or with different concentrations of plasmids containing the various HHV-6 genes (A, lanes 2, 3, 4, and 6 to 9; B, lanes 2, 3, and 5) and the HIV-Tat gene (A, lane 2; B, lane 6) in the pRC/RSV vector. At 36 h after transfection, the cells were harvested and CAT activities were measured by TLC. Cm, unacetylated [14C]chloramphenicol. 1-, 3-, and 5-AC, acetylated forms of [14C]chloramphenicol. The percentage of CAT conversion (activity) was calculated by dividing the radioactive counts present in the acetylated chloramphenicol spots by the total radioactivity present in both the acetylated and unacetylated chloramphenicol spots. Fold activation was calculated as the increase in CAT activity over the background levels for HIV-CAT with the pRC/RSV vector.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of the DNA sequence of the genome of the HHV-6A strain U1102 (16, 25) demonstrates the close relationship of HHV-6 with HCMV. These two betaherpesviruses have a colinear arrangement of genes in several different regions of their genomes. In HHV-6, one of these, designated the IE-B gene block, contains an ORF (U16) which has been previously shown by us to activate the HIV LTR (14, 17). It was predicted that the HHV-6A(U1102) ORF U16 would be expressed with ORF U17 as a spliced gene product (15, 25). Several observations argue in favor of this prediction. The HHV-6 U17/U16 splice product bears positional homology to the HCMV UL36 IE gene. The UL36 gene is part of the UL36-UL38 gene family from HCMV (6, 12, 15), and the products of this family are transcriptional activators (8, 33, 36). The predicted sequences of the U17/U16 and UL36 proteins both contain several US22 protein-specific motifs first described by Kouzarides et al. (19). The US22 motifs are unique to the betaherpesviruses and are found in several IE gene products of HCMV. In the HHV-6 U17/U16 and HCMV UL36 predicted proteins, these homologous motifs are colinear in their arrangement and are brought together by comparable splicing events in the two genes (6, 15, 25).

We have identified a family of transcripts generated from the IE-B region of the HHV-6A(GS) genome. The identification of HHV-6 U17/U16 (E1E2) from the 1.9-kb cDNA demonstrates that HHV-6 expresses the U17 and U16 ORFs as the predicted spliced gene product. The splicing event removes an intron of 84 bp, bringing together U17 and U16 into one ORF. The protein predicted from the nucleotide sequence was estimated to be around 38K, and in vitro transcription and translation of this cDNA confirmed the size. The protein generated in vitro and by expression in a prokaryotic expression vector was recognized by polyclonal antisera and by a monoclonal antibody generated against the B701 (U16) protein.

Our Northern blot data confirm that multiple transcripts are at least partially encoded by the U17/U16 region of the genome and are generated from the leftward strand of the HHV-6 genomic DNA. Given the number and sizes of transcripts detected on the Northern blots by U16 and U17/U16 riboprobes, it seems quite likely that additional overlapping transcripts, initiated from alternative promoters, have yet to be identified. While we are not able to conclude which TATA sequence was used to initiate transcription of the 1.8-kb transcript identified, our results show that the 1.8-kb transcript is incomplete. If the complete 1.8-kb transcript initiated from TATA-1 or TATA-2 (Fig. 5), its size would be approximately 2.1 or 2.5 kb, respectively. It is also possible that the 1.8-kb cDNA is part of a much larger transcript initiating from a promoter far upstream of the promoters identified here. The predicted sizes of transcripts expressing the U17/U16 splice product range from roughly 1.2 to 1.7 kb, depending on the precise transcriptional initiation site and the size of the poly(A)⁺ tail. The U16+ transcript would also be at least 1.2 kb fully spliced if the 3' end is at the 3' end of U16. However, the 3' end of this transcript is not known. Allowing for variation in transcript size due to the lengths of the poly(A) tails, these sizes can all be correlated with the estimated sizes of the transcripts identified on the Northern blots. The largest transcript, approximately 3.9 kb, seen on Northern blots does not fit with the cDNAs identified to date. The HCMV UL37 transcript is a large transcript of approximately 3.4 kb (8, 36, 37, 38). This transcript initiates from a promoter also used to transcribe HCMV UL37X1, but because of differential splicing, UL37 terminates much farther downstream and includes a large noncoding region. This noncoding region overlaps the HCMV UL36 cDNA, and hence, a UL36-specific probe also detects the large UL37 transcripts in Northern blot analysis. It is possible that a similar situation also occurs in HHV-6.

The level of expression of the entire family of HHV-6 U17/U16 transcripts was low when compared with those of other IE (U89) and early (U27) genes from HHV-6. Since cycloheximide can decrease expression of some IE genes (36, 37), we used a more sensitive RT-PCR technique. Our results show that transcription of the U17/U16-containing transcripts was initiated at one of two different TATA sequences upstream of U17 (Fig. 6 and 12). These transcripts were differentially spliced and differentially expressed. At least some of these transcripts were expressed as IE transcripts, and their expression continued throughout infection. One transcript, U16+, was described as a late gene, as its expression was dependent on DNA replication. This transcript demonstrated a unique and unexpected splicing pattern that would allow the expression of the complete U16 ORF independent of U17 (Fig. 11).

Using RT-PCR with RNA from HHV-6A(U1102)-infected cells, Mirandola et al. (22) classified the U16/U17 transcript as an early transcript, since it was absent in cells treated with inhibitors of protein synthesis. The virus subtype, host cells, method of infection, and RT-PCR protocol described in our studies were all different from the methods described by Mirandola et al. (22) and could account for the differences in the results. RNA samples in the study by Mirandola et al. (22) were said to be free from contaminating DNA, but the detection of an unspliced gp82/105 RT-PCR product at IE time points would indicate the presence of DNA. The presence of DNA in RT-PCRs can interfere with the function of the RT and result in PCR amplification of erroneous transcripts. In our study, all RNA samples were subjected to nested rounds of PCR without prior reverse transcription to be certain all DNA was removed. Furthermore, the quantity of RNA that was subjected to this nested PCR was equivalent to the quantity used in the complete RT-PCR. In the Mirandola et al. (22) studies, only 200 ng of RNA was subjected to PCR, yet 1 µg was used for the RT-PCR. Additionally, the cDNA generated from the reverse transcription was precipitated before PCR. This could have had the affect of concentrating any genomic DNA present.

The result presented by Mirandola et al. (22) supporting the conclusion that the U17/U16 transcript was an IE transcript in HHV-6B but an early transcript in HHV-6A is not entirely convincing. Perhaps the RT-PCR performed by Mirandola et al. (22) was not sensitive enough to detect the low-level expression. Our results indicate that HHV-6A(GS) U17/U16 transcripts are expressed in the presence of protein synthesis inhibitors but at a lower level, similar to the HHV-6B results presented (22). One might conclude that the decrease in expression of U17/U16 that we observed is due to a "leaky transcript," but an alternative explanation is that there are two pathways regulating expression of this transcript, one requiring the synthesis of another viral protein while the other does not.

The regulation of expression of IE genes in HHV-6 and other herpesviruses is complex. The major IE complex in HHV-6 is encoded by ORFs U86, U89, and U90, and two overlapping transcripts, the 3.5-kb IE-1 and the 4.7-kb IE-2 gene products, are expressed as the result of differential splicing (32). HCMV UL36 is part of the UL36-UL38 gene family that expresses at least four different but overlapping transcripts from three different transcription initiation signals (36, 37, 38), and these transcripts are also differentially regulated. HCMV UL36 is regulated by a promoter that is immediately upstream of the first exon of this gene, and two other promoters that are found approximately 2.5 and 3.0 kb upstream of UL36 exon 1 regulate expression of UL37, UL37X1, and UL38. Since UL37 and UL37X1 are expressed from the same promoter, the differential expression suggests additional mechanisms are involved in the regulation. Some of these mechanisms could include posttranscriptional means of regulation, such as splicing- and translation-coupled degradation of the transcript (8, 36, 37, 38).

Our studies demonstrate that while the HHV-6 IE-B region and the HCMV UL36-UL38 gene family have some similarities, there are also several differences. Two transcriptional start sites have been identified upstream of HHV-6 U17, which differs from the situation in the homologous HCMV gene. The U16+ cDNA isolated by 5' RACE studies is also unique to HHV-6. Unlike HCMV, there are two possible polyadenylation signals found in this region. Similar to HCMV UL36, the first one is located immediately 3' of U16 and was the termination signal used to generate the U17/U16 transcript identified. The other is located about 800 bp downstream of the 3' end of U16 and was identified as the termination signal in the 1.8-kb cDNA. Since this transcript also encodes HHV-6 U15, a gene unique to HHV-6, it is possible that U15 expression is regulated by one of the two promoters identified here. The promoter immediately upstream of the first methionine of U15 could also regulate expression of ORF U15, but it was not investigated here. The HHV-6 U18 and U19 genes are homologous to the HCMV UL37 and UL38 genes, respectively (15, 25), and are part of the HHV-6 IE-B gene block. Since U18 and U19 are not part of the 22-kb BamHI genomic fragment that was originally identified as a transactivating fragment (14, 17), these ORFs were not included in these studies.

Temporal regulation of transcript expression is a hallmark of herpesviruses. The IE, or , transcripts are detected within a few hours after initiation of infection, and de novo protein synthesis is not required for their expression. Specific viral protein synthesis is required for the expression of the early, or , transcripts, and the late, or , transcripts are expressed when viral DNA synthesis begins. Our results show that expression of the U17/U16 gene products is tightly regulated and complex. The U17/U16 spliced transcript is expressed immediately at a low level and then increases until 24 h after infection and stays high throughout the rest of the replication cycle of the virus (Fig. 8A and B). This is similar to the expression of the HCMV UL36 gene (8, 36, 37). While this gene is also expressed as an IE gene, the presence of protein synthesis inhibitors during the culture period reduces the level of UL36 expression but does not eliminate it. The expression pattern of HHV-6 U17/U16 could be the result of expressing the same transcript from two promoters. It is possible that the 5' RACE product we detected and show schematically in Fig. 6A is expressed under IE conditions while the product shown in Fig. 6B is expressed under late conditions. Initiation from the TATA-2 promoter (Fig. 6B and C) produced multiply spliced transcripts, and the RT-PCR data show that one is the approximate size of U17/U16. Thus, the TATA-1 promoter would express U17/U16 immediately after infection without the need for viral protein synthesis, and when the synthesis of some unidentified viral gene regulator begins, the TATA-2 promoter is activated. This would allow the maintenance of expression of U17/U16 throughout the HHV-6 replication cycle. However, we cannot rule out the possibility that the 5' RACE product identified in Fig. 6B is actually an intermediate or incompletely spliced transcript of U16+ (Fig. 6C). Our data also strongly suggest that generation of the spliced transcripts regulated by TATA-2 is dependent on DNA replication (Fig. 10A and B). Furthermore, the regulation of expression of the U16+ gene product may involve splicing and protein synthesis, making U16+ a late gene product. The unspliced HHV-6 RT-PCR product detected in Fig. 10 was seen at IE times in the presence of cycloheximide (Fig. 10A), suggesting that the inhibition of protein synthesis was allowing the transcript to accumulate. It is conceivable that this PCR measured a larger transcript originating from a promoter upstream of U17 but continued through the U17/U16 region of the genome, similar to HCMV UL37.

Our present transactivation studies and other studies (14, 16, 17) show that the U17/U16 region of the HHV-6 genome encodes products capable of activating transcription under the control of the HIV promoter (14, 17, 25). Some of the members of the HCMV US22 family, such as HCMV IRS1 and TRS1 (28, 34), and HCMV UL36 (8) have been shown to function as gene regulators. IRS1 and TRS1 have been shown to work in a cooperative fashion with HCMV IE1 and IE2 for regulating expression of HCMV UL44 (34) and the HCMV major IE gene promoter (28). Similarly, transactivation mediated by UL36 alone was reported to be minimal, but when UL36 was included in cotransfection experiments with other HCMV IE genes, the synergistic response was much greater than the transactivation mediated by the IE gene alone (8). Other studies have shown that HCMV UL36 and possibly UL37 are essential for viral DNA replication (32). In fact, an antisense oligonucleotide specific for the splice donor site common to both transcripts will inhibit the growth of the virus almost completely. These observations suggest that the gene regulation function of the UL36-UL38 gene family may be secondary to the DNA replication function. Given the similarities between UL36 and U17/U16, the possibility that U17/U16 might be involved in DNA replication and/or may function in a cooperative manner with other HHV-6 IE genes to regulate viral gene expression needs to be investigated.

The function of the HHV-6 U16+ ORF should also be investigated. If the transcript terminates at the poly(A)⁺ signal found at the 3' end of the U16 ORF, the protein encoded by this cDNA would be identical to the carboxyl end of the U17/U16 protein. The structural homology to other US22 family members is found within the U16 ORF, so U16+ would also be a US22 family member. The protein encoded by U16+ would be about 29 kDa and, like U17/U16, has several potential phosphorylation sites and glycosylation sites. The expression of truncated forms of regulatory proteins is seen in other gene families of HCMV. IRS1 is expressed at two different proteins, the smaller one generated from the carboxyl-terminal domain of the larger (28). The smaller protein, IRS1²⁶³, is expressed throughout infection, but predominately at early and late times after infection, and was shown to antagonize the activation functions of both TRS1 and IRS1. The IE2 gene of HCMV is also expressed as multiple products resulting from differential splicing. One of these, IE2_338aa, is the carboxyl-terminal ORF of IE2_579aa. This smaller protein has been shown to inhibit the transactivating function of IE2_579aa. However, IE2_338aa also has its own transactivating function (reviewed in reference 23). It is interesting to speculate that, like IRS1²⁶³ and IE2_338aa, U16+ may be a repressor of the U17/U16 function. Further study in this area will be required to evaluate the functions of proteins encoded by the transcripts generated from the U17/U16 region of the HHV-6 genome.