Human Cytomegalovirus DNA Replication Requires Transcriptional Activation via an IE2- and UL84-Responsive Bidirectional Promoter Element within oriLyt

Cells and virus.

Cos7, Vero, and human foreskin fibroblast (HFF) cells were maintained in Dulbecco's modification of Eagle's medium (catalog no. 50-003-PB; Cellgro) supplemented with 10% fetal bovine serum. HCMV strain AD169 cells (American Type Culture Collection) were maintained as frozen stocks. T-HFs were generated by infection with a retrovirus expressing hTERT (Geron Inc.) (29) followed by selection for G418-resistant colonies. Selected colonies were expanded and tested for telomerase activity using the trap assay (29, 58).

Plasmids.

The replication reporter pGEM-oriLyt, described previously (64), contains the PvuII-KpnI HCMV oriLyt fragment (nt 89796 to 94860) of AD169 (6) ligated into pGEM-7Zf(-) (catalog no. p2371; Promega). pZP13 containing full-length UL84 with its native promoter was constructed as previously described (44). pSI84 was generated by cleaving pZP13 with DraI and NotI, and the resultant 2-kb fragment containing the UL84 ORF was ligated into the mammalian expression vector pSI (catalog no. E1721; Promega). The IE2 expression construct pON2206, which contains the IE2 cDNA under the control of the SRα promoter, was a gift from E. Mocarski (Stanford University) (25).

The IE2-FLAG plasmid was constructed by PCR amplification of the IE2 ORF from pON2206 using PCR primer 5′-CGACTTAACAGATCTCGAGCTCAAGCTTCGAATTCATGGATTACAAGGATGACGACGATAAGGAGTCCTCTGCCAAGAGAAA-3′ and reverse primer 5′-CCCGGGCCCGCGGTACCGTCGACTGCAGAATTCTTACTGAGACTTGTTCCTCAGGTCCTGGATGG-3′. The reverse primer was constructed such that the IE2 was fused in frame with the FLAG epitope. The PCR product was then recombined into the phCMV Xi-cloning vector (catalog no. XC003120; Gene Therapy Systems, Inc.) as recommended by the manufacturer. pRL-CMV and pRL-SV40 were obtained from Promega (catalog nos. E2261 and E2231). The set of plasmid subclones encoding the HHV8 and HCMV replication proteins were made as described previously (3, 64).

Plasmids used in oriLyt promoter assays.

Sequences within oriLyt were PCR amplified using pGEM-oriLyt as template and the following primers containing KpnI and XhoI restriction endonuclease sites at the 3′ and 5′ ends, respectively: P1, 5′-CCGCTCGAGCATTGGAAATCATGGTTGCCCAAATTTGGTAA-3′; P2, 5′-GGGGTACCCATTGGAAATCATGGTTGCCCAAATTTGGTAA-3′; P3, 5′-GGGGTACCCAGTCCGTTTTACGTATACCGGATGCTAGGCG-3′; P4, 5′-CCGCTCGAGGCGGTAGAATACAGCGATCCCTAGTGAAGCCAC-3′; P5, 5′-CCGTCTCGAGGCTTATGACGCGTATCCGGGAGTAGCGTCTACG-3′; P6, 5′-GGGGTACCGCGGTAGAATACAGCGATCCCTAGTGAAGCCAC-3′; P7, 5′-CCGCTCGAGCAGTCCGTTTTACGTATACCGGATGCTAGGCG-3′; P8, 5′-GGGGTACCGCTTATGACGCGTATCCGGGAGTAGCGTCTACG-3′. The oriLyt subclone pGL-ORI-A was amplified by using primers P1 and P6. pGL-ORI-B was generated using primers P1 and P3. pGL-ORI-BR was generated using primers P2 and P7. pGL-ORI-C was generated using primers P5 and P6. pGL-ORI-CR was generated using primers P4 and P8. pGL-ORI-AR was generated using primers P2 and P4. All the amplified fragments were subcloned into the luciferase reporter vector pGL2-Basic (catalog no. E1641; Promega).

As a control for IE2 transactivation, a luciferase reporter, pGL-112pr, was also constructed. The UL112/113 promoter was PCR amplified by using AD169 DNA as template and the following primer set: forward, 5′-CCGGTACCAGATCTCCGCGTCACCTTTCATCGAGT-3′; reverse, 5′-CCAAGCTTGGCCGTGGAGCGAGTGCCGCCGCAGCC-3′. The 750-bp PCR product was then cleaved with KpnI and HindIII and ligated into the luciferase reporter vector pGL2-Basic. The final construct, pGL-112pr, was used in the transfection and luciferase assays.

The pGEM-oriLytSVR and pGEM-oriLytSVL plasmids were generated by cleavage of pGEM-oriLyt with NotI and SgfI and removal of the intervening sequence, followed by ligation of the PCR-amplified SV40 early promoter. The SV40 early promoter region was amplified from the vector pSI (Promega) using PCR primers 5′-CCGCGGCCGCCTCGACAGATCTGCGCAGCACC-3′ and 5′-CCGCGATCGCGACTGTTGTGTCAGAAGAATCAAGCT-3′ (to generate pGEM-oriLyt SVR) or 5′-CCGCGATCGCCTCGACAGATCTGCGCAGCACC-3′ and 5′ CCGCGGCCGCGACTGTTGTGTCAGAAGAATCAAGCT-3′ (to generate pGEM-oriLytSVL). Primer pairs were designed such that a NotI site and a SgfI site were at the ends of the primers. Two sets of primers were used in order to ligate the promoter into pGEM-oriLyt in both directions. pGEM-oriLytSVR had the SV40 early promoter ligated into HCMV oriLyt such that the SV40 early promoter directed transcription in the right-to-left direction. pGEM-oriLytSVL had the SV40 early promoter ligated into HCMV oriLyt such that the SV40 early promoter directed transcription in the left-to-right direction.

Site-directed mutagenesis.

All mutations were generated using the QuikChange II site-directed mutagenesis kit (catalog no. 200523; Stratagene). Mutagenesis was performed according to the manufacturer's protocol.

pGL-ORI-Amut was generated using pGL-ORI-A plasmid as template and the primer set 5′-CCTTATCTACCAATCACGAGAAACGGATATA-3′ (forward) and 5′-GATTGTATATCCGTTTCTCGTGATTGGTAGATAAGG-3′ (reverse), where a single G nucleotide was inserted (underlined) into the wild-type sequence. pGEM-oriLyt+G was generated using the same primers as for pGL-ORI-Amut, except that the DNA template used for PCR was pGEM-oriLyt. All mutations were confirmed by DNA sequencing.

Chromatin immunoprecipitation (ChIP) assay.

A 225-cm flask of confluent T-HF cells (6 × 10⁶ cells) was infected with AD169 at a multiplicity of infection of 1. At 48 h postinfection (hpi), cells were fixed with a 1% formaldehyde solution in 1× phosphate-buffered saline (PBS; pH 7.4) at room temperature for 10 min. The cells were washed twice with ice-cold PBS (pH 7.4) and harvested with 10 ml of PBS. The cells were then collected by centrifugation at 1,000 × g for 5 min. The pellet was resuspended in 750 μl of cell lysis buffer (50 mM Tris HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 0.5% NP-40, 2 mM MgCl, 0.2 mM EDTA, and 10 μl of protease inhibitor cocktail σ/ml) and sonicated five times at 10-s intervals. The lysis solution was diluted with PBS to a 5-ml volume and precleared with 2 μg of normal mouse immunoglobulin G-agarose (Santa Cruz Biotechnology) for 30 min at 4°C. The solution was again centrifuged to remove the normal mouse immunoglobulin G-agarose at 500 × g for 5 min. Two micrograms of IE2-specific antibody G13-12E2 (Vancouver Biotech, Ltd.) was added to 1 ml of the cell lysate and incubated for 1 h with rotation at 4°C. One milliliter of the lysate was also used for control immunoprecipitations that did not contain primary antibody. The solution was then incubated with 50 μl of protein A/G plus agarose beads (Santa Cruz Biotechnology) with rotation overnight at 4°C. The beads were pelleted and washed four times with ice-cold PBS. The beads were then resuspended in 100 μl of Tris-EDTA (TE) and treated with RNase A (50 μg/ml) for 30 min at 37°C. The complexes were treated with 5 μl of 10% sodium dodecyl sulfate and 500 μg of proteinase K/ml for 4 h at 37°C with occasional mixing. In order to disassociate the cross-linked protein-DNA complexes, the reaction mixtures were incubated at 65°C overnight. The supernatant was collected and extracted with phenol followed by chloroform and ethanol precipitation to isolate the immunoprecipitated DNA. The resulting pellet was resuspended in 20 μl of TE.

PCR was used to detect the immunoprecipitated DNA fragments. Two microliters of the TE solution was used as template for each PCR. To detect IE2 association with the cis repression signal (CRS) sequence found in the IE2 promoter, the two primers 5′-GCAGTACATCTACGTATTAGTCATCG-3′ and 5′-AGCTTAAGTTCGAGACTGTTGTGTC-3′ were used in the PCR.

To detect IE2 association with the OPAE within the HCMV oriLyt, primer pair A corresponding to nt 91928 to 91627 was used: 5′-CGTACACCGAGGACGGTGGAACCCTAACGGG-3′ and 5′-TCAACCCCGTGGCCGGCGAGGCGGTGAGCA-3′. In addition, as a negative control primer pair B was used, corresponding to nt 93128 to 92858 within oriLyt: 5′-ACGGCGCACATCTAGTGGAATTTTACCGACG-3′ and 5′-GATGGTGCTCCAGGGCGGTGGGACGGGCCCG-3′.

All reactions were tested for DNA contamination by running PCR without template DNA for the negative control. All reactions were also tested using AD169 as the template as a positive control.

Synthesis and purification of IE2-FLAG fusion protein.

Cos7 cells were plated into ten 100-mm tissue culture dishes 24 h prior to transfection. Transfection was performed using 10 μg of IE2-FLAG per plate with TransIt LT1 (catalog no. MIR2300; Mirus) according to the manufacturer's instructions. Total cell protein from all 10 dishes was harvested in 10 ml of lysis buffer (50 mM Tris HCl-150 mM NaCl-1 mM EDTA-1% Triton X-100; pH 7.4) at 48 h posttransfection (hpt). Cell lysates were sheared through an 18-gauge needle five times on ice, and cell debris was pellet at 16,000 × g for 10 min. FLAG fusion protein was purified by loading the supernatant onto a 1-ml EZview Red anti-FLAG M2 affinity gel column (product no. F2426; Sigma). The column was then washed with 5 ml of Tris-buffered saline (TBS; 50 mM Tris HCl-150 mM NaCl; pH 7.4) three times, and FLAG fusion protein was eluted with 100 μl of a 5-μg/μl solution of 3× FLAG peptide (product no. F4799; Sigma) in 5 ml of TBS. The fractions from the purification process were collected, resolved on a sodium dodecyl sulfate-10% polyacrylamide gel, and stained with Coomassie blue to monitor the protein purity and purification efficiency. Protein concentration of the purified IE2-FLAG was determined by Bradford assay.

Electrophoretic mobility shift DNA-binding assay (EMSA).

The DNA-binding probes representing the HCMV MIEP CRS were prepared by annealing the following complementary 30-mer CRS oligonucleotide pairs: 5′-GATCCCGGGAGCTCGTTTAGTGAACCGTCA-3′ and 5′-GATCTGACGGTTCACTAAACGAGCTCCCGG-3′. The OPAE site within oriLyt was represented by probes annealed using the oligonucleotide pair 5′-GATCCCTTATCTACCAATCACAGAAACGGATATA-3′ and 5′-GATCTATATCCGTTTCTGTGATTGGTAGATAAGG-3′. Labeling was accomplished by filling in the recessed ends with Escherichia coli DNA polymerase Klenow fragment and deoxynucleoside triphosphates in the presence of [α-³²P]dCTP.

For the binding reaction, IE2-FLAG-purified protein (500 ng) was preincubated in 20 μl of binding buffer (10 mM Tris-Cl, 0.5 mM dithiothreitol, 1 mM EDTA, 10% glycerol, 50 mM KCl, 8 mM MgCl₂; pH 7.5) containing 40 μg of poly(dA-dT) · poly(dA-dT)/ml for 10 min at 20°C. ³²P-labeled probe DNA (20,000 cpm) was then added for 20 min at 20°C. The mixtures were loaded onto 4% nondenaturing acrylamide gels and run at 100 V in 0.5× Tris-borate-EDTA buffer. After electrophoresis at 20°C, the gels were dried and autoradiographed. For cold competition assays, 20- or 50-fold molar excesses of nonradioactively labeled DNA fragments were added to the preincubation reaction mixtures. For antibody supershift assays, 1:100-diluted anti-IE1/2 antibody MAB810 (Chemicon) was added to the preincubation reaction mixtures.

Transient-cotransfection replication assay.

The HHV8 core replication proteins encoding helicase, primase, primase-associated factor, polymerase, polymerase accessory protein, and single-stranded DNA-binding protein (3) were cotransfected with and without UL84 and/or IE2 and/or UL36-38 expression constructs (44, 45) and pGEM-oriLyt, using the calcium phosphate coprecipitation assay (64). Total cellular DNA was harvested 5 days posttransfection, cleaved with EcoRI and DpnI, and resolved on a 0.8% agarose gel. The DNA was transferred to a Zeta-Probe nylon membrane and hybridized with ³²P-labeled pGEM7zf(-). Blots were exposed to X-ray film for 5 to 15 h.

For replication assays involving pGEM-oriLytSVR(L), cells were cotransfected with pGEM-oriLyt along with pGEM-oriLytSVR(L) and infected with HCMV 24 hpt. Total cellular DNA was harvested and treated as described above. For the cotransfection-replication assay, cells were cotransfected with pGEM-oriLytSVR and pGEM-oriLyt along with the core replication proteins for HHV8 or HCMV, with and without UL84 and/or IE2 expression constructs. Total cellular DNA was harvested and treated as described above.

In the case of the transfection of pGEM-oriLyt+G, this plasmid was transfected into HFFs and cells were subsequently infected with HCMV. Total DNA was harvested 5 days postinfection and processed as described above.

Luciferase assay.

Vero or HFF cells were seeded into six-well tissue culture plates 24 h prior to transfection. Cotransfections were performed using TransIt LT1 (catalog no. MIR2300; Mirus) according to the manufacturer's instructions. Plasmids pON2206 (25) and pSI84 were cotransfected along with the various indicated luciferase reporter constructs. The total amount of DNA and the amount of each plasmid used in the cotransfection mixture was kept constant by using pGEM-7Zf(-) as control filler plasmid DNA. For cotransfections involving the core replication proteins, plasmids encoding HCMV core replication proteins were added at 2 μg each in addition to the reporter luciferase plasmid, with and without UL84 and IE2 expression plasmids. Total cell lysates were harvested 48 hpt or at various times postinfection as indicated in the case of HFFs, and luciferase assays were carried out with a luciferase assay system (catalog no. E4530; Promega) as recommended by the manufacturer.

The gene products of UL84 and IE2 are the only additional noncore proteins required for origin-dependent DNA replication in telomerase-immortalized HFF cells.

The original replication assay performed by Pari et al. demonstrated that the core replication proteins along with the gene products of UL112-113, UL36-38, IRS1, IE1/IE2, and UL84 were all required for efficient amplification of cloned oriLyt (44, 45). In a later study, when EBV core replication genes were used in the cotransfection assay, it was shown that, in Vero cells, UL84 was the only noncore protein required for origin-dependent DNA replication (50). In HFF cells, however, the gene products of IE2, UL36-38, and UL84 were all necessary for efficient amplification of oriLyt (50). Because recent evidence showed that the UL36-38 locus encodes proteins that have antiapoptotic properties (17, 56), we investigated the requirements for origin-dependent DNA replication in T-HFs. These cells have an unlimited life span in culture but retain an otherwise normal phenotype (data not shown). In order to address the issue of differences in protein requirements between telomerase-immortalized and wild-type HFF cells, we cotransfected both T-HFs and wild-type HFFs with plasmids expressing the HHV8 core replication proteins (3), UL84, IE2, and the UL36-38 proteins, and a plasmid containing the HCMV oriLyt, pGEM-oriLyt. HHV8 core replication proteins were used to enable us to focus on those HCMV proteins required for DNA replication and that perform noncore functions. HHV8 core proteins were fully capable of providing the essential enzymatic replication functions in the presence of UL84-, IE2- and, in the case of HFFs, UL36-38-encoded proteins (Fig. 1). Omission of either IE2 or UL84 expression constructs from transfection mixtures led to failure to amplify cloned oriLyt in T-HFs (Fig. 1). In wild-type HFFs, UL36-38 was absolutely required for oriLyt amplification (Fig. 1). However, UL36-38 was not necessary in T-HFs (Fig. 1). This indicated that in T-HFs, the only factors required for efficient oriLyt-dependent DNA replication were UL84 and IE2, implying that UL36-38 gene products do not directly participate in DNA replication and may function to increase the survival of primary cells when performing the cotransfection-replication assay.

HCMV oriLyt contains a strong promoter with bidirectional activity that is constitutively active but repressed by IE2 in Vero cells.

Since we determined that oriLyt-dependent DNA replication required only two additional gene products, namely IE2 and UL84, we next wanted to elucidate the function of IE2 and UL84 in lytic DNA replication. We chose to do this by examining the HCMV oriLyt region for any clues that may lead to a replication mechanism that involves a possible interaction of IE2 with oriLyt. Inspection of the oriLyt region revealed two potential IE2 binding sites. The first putative IE2 binding sequence, termed sequence-short, is located between nt 92430 and 92444 (5′-CGTTCTACGAAAACG-3′). This site is similar to the IE2 consensus-binding site CRS element found in the promoter region of IE2 (7, 36, 47). The second sequence, termed sequence-long, is located between nt 91840 and 91902. This sequence (5′-GCCACCTTATCTACCAATCACAGAAACGGATATACAATGACCCCTCCCTAGACTCCACCCCTT-3′) was identified by Huang et al. by using a modified approach coupling the methods of cyclic amplification and selection of targets to selectively amplify IE2-binding sites (21). Figure 2A is a schematic showing the positions of these two putative IE2-binding sites in relation to the Y-block and SRT within oriLyt. Since IE2 is the major transactivator protein in HCMV, we wanted to determine if the region of oriLyt containing the two putative IE2 binding elements could act as an IE2-responsive promoter. PCR fragments were generated such that these potential IE2-binding sites, along with some flanking sequence, were subcloned into the promoterless luciferase reporter vector pGL2-Basic (Promega) (Fig. 2B).

Vero cells were transfected with pGL-ORI-A or pGL-ORI-AR, with or without the IE2 expression plasmid pON2206. These two luciferase reporter constructs contained both putative IE2-binding sites mapping to nt 91754 to 92555 but in different orientations (Fig. 2B). Cell lysates were harvested, and luciferase assays were carried out at 48 hpt. Both of these constructs displayed a high basal level of promoter activity in Vero cells (Fig. 2C, pGL-ORI-A or pGL-ORI-AR plus pGEM). The addition of IE2 to the transfection mixture resulted in a marked repression of the high basal luciferase activity for both constructs (Fig. 2C, compare pGL-ORI-A and pGL-ORI-AR plus IE2 to pGL-ORI-A and pGL-ORI-AR plus pGEM). Transfections containing control vector having the SRα promoter and pGL-ORI-A or pGL-ORI-AR had no effect on luciferase activity (data not shown). To further map this promoter region, constructs pGL-ORI-B (nt 92324 to 92555), which contained the downstream putative IE2-binding element (sequence-short), and pGL-ORI-C (nt 91754 to 92030), which contained the upstream putative IE2-binding element (sequence-long), were generated and used in the luciferase assay (Fig. 2B). As shown in Fig. 2C, only pGL-ORI-C displayed a high basal level of promoter activity in Vero cells (Fig. 2C, compare pGL-ORI-C plus pGEM to pGL-ORI-B plus pGEM), and this activity was again repressed by addition of IE2 to the transfection mixture (Fig. 2C, compare pGL-ORI-C plus IE2 to pGL-ORI-C plus pGEM). A strong luciferase signal was still observed when the DNA segment from nt 91754 to 92030 was reversed (Fig. 2, pGL-ORI-CR), whereas the reverse of the DNA segment from nt 92324 to 92555 still showed no activation (Fig. 2, pGL-ORI-BR). This result indicated that a strong constitutive promoter (oriLyt_PM) showing bidirectional activity was located within oriLyt between nt 91754 and 92030. This promoter, which contains a potential IE2-binding site, was repressed by the addition of IE2 to the transfection mixture when assayed in Vero cells.

As a control for IE2 activation of an early promoter, we cotransfected a luciferase construct pGL-112pr, which contained the UL112/113 promoter in the luciferase reporter vector pGL2-Basic, along with IE2 expression plasmid pON2206 into Vero cells. As expected, IE2 transactivated the UL112/113 promoter, which resulted in a marked increase in luciferase activity compared to that in the samples that lacked the IE2 expression construct pON2206 (Fig. 2C, pGL-112pr).

Since IE2 was shown to down-regulate its own expression by directly binding to the CRS element within the MIEP (7, 35, 36, 47, 48), we wanted to compare the observed IE2-mediated repression of the oriLyt promoter to the IE2-mediated repression of the HCMV MIEP in Vero cells. We again performed cotransfections using increasing amounts of pON2206 along with pGL-ORI-C or pRL-CMV. Basal luciferase activity from the vector pRL-CMV, which contains the MIEP along with the CRS element, was repressed in a dose-dependent manner in the presence of increasing amounts of IE2 expression plasmid (Fig. 2D). pGL-ORI-C also responded to increasing amounts of pON2206 similar to pRL-CMV, except that the degree of repression by IE2 was much greater as concentrations of pON2206 increased (Fig. 2D).

UL84 can release the IE2-mediated repression of the HCMV oriLyt_PM in Vero cells.

Since it was previously shown that UL84 interacted with IE2 in infected and transfected cells and this interaction was postulated to repress IE2 activation of the early promoter for UL112/113 (16, 57), we wanted to investigate the impact of UL84 on the IE2-mediated repression of the oriLyt_PM. Vero cells were transfected with various oriLyt-luciferase constructs, except that we now added a UL84 expression construct (pSI84) to the transfection mixture. IE2-mediated repression was inhibited when UL84 was added to the transfection mixture (Fig. 3, compare pGL-ORI-A, pGL-ORI-C, or pGL-ORI-AR plus pON2206 to pGL-ORI-A, pGL-ORI-C, or pGL-ORI-AR plus pON2206+pSI84). Cotransfection of the control parent vector, pSI, had no effect on IE2-mediated repression (data not shown). No change in promoter activity was observed when UL84 alone was transfected along with pGL-ORI-A, pGL-ORI-C, or pGL-ORI-AR (Fig. 3, compare pGL-ORI-A, pGL-ORI-C, or pGL-ORI-AR plus pGEM to pGL-ORI-A, pGL-ORI-C, or pGL-ORI-AR plus pSI84). This result indicated that the IE2-mediated repression of oriLyt_PM could be inhibited by the expression of UL84 in Vero cells. In contrast to the effect on the oriLyt_PM, IE2 activated the UL112/113 promoter (Fig. 3, pGL-112pr plus pON2206), and UL84 disrupted the IE2-mediated activation of this promoter (Fig. 3, pGL-112pr plus pON2206/pSI84), which was consistent with the results from previous studies (16).

Mutation of the novel DNA element within oriLyt_PM inhibits promoter activity in Vero cells.

Investigation of the oriLyt_PM sequence using the on-line database Transfac (63) revealed that the oriLyt_PM contained various putative transcription factor-binding sites (Fig. 4A). These sites include a CCAAT factor-binding site, SP1, CREB, and an NF-κB binding site. Also, a transcriptional repressor, YY1, binding site was found within this region. Of course, the most salient feature with the oriLyt_PM was the presence of the novel putative IE2-binding element. The sequence of the putative oriLyt IE2-binding sites (nt 91840 to 91902) was compared to the previously described consensus IE2-binding CRS element(s) by using the Align on-line program (46). As shown in Fig. 4B, a CRS-like element having a 57% similarity to the consensus CRS sequence was present within the oriLyt_PM region. Inspection of other known IE2-binding sites (A, B, C, and D) (Fig. 4B) found within the UL112/113 promoter revealed 43, 57, 57, and 43% similarity to CRS, respectively (53).

Since sequence analysis reveled the presence of a novel putative IE2-binding element, we chose to mutate this element using site-directed mutagenesis. Based on a previous study showing that a 1-bp insertion within the CRS eliminated IE2 binding (8), we generated a construct that had a similar 1-bp insertion by a designing site-directed mutagenesis primer such that a single guanosine (G), between nt 91860 and 91861, was inserted into the novel IE2-binding element in pGL-ORI-A (Fig. 4C, pGL-ORI-Amut). This mutant construct was transfected into Vero cells, and luciferase activity was measured. Interestingly, the single nucleotide insertion completely abolished the constitutive activity of the oriLyt_PM in Vero cells (Fig. 4C, sample pGL-ORI-Amut). Promoter activity from pGL-ORI-Amut was unchanged upon transfection of IE2 and/or UL84 (data not shown). These results indicated that this putative IE2-binding motif was essential for the promoter activity in an IE2-independent manner and that IE2-mediated repression in Vero cells may be due to interaction with this element, such that activation is inhibited in some undefined mechanism. Consequently, we will refer to this element (CAATCACAGAAACG) as the OPAE.

The oriLyt promoter is activated in human fibroblasts by HCMV infection.

Since we determined that the oriLyt_PM was constitutively active in Vero cells, the next step was to examine the activity of this promoter in HCMV-permissive human fibroblast cells. HFFs were transfected with pGL-ORI-A in the presence or absence of the viral DNA synthesis inhibitor phosphonoacetic acid (PAA). Cells were subsequently infected with HCMV, and luciferase activity was assayed at various times postinfection. Transfection of pGL-ORI-A, in the absence of infecting virus, resulted in a low baseline level of luciferase activity (Fig. 5, inset). However, when transfected cells were infected with HCMV, promoter activity increased markedly as a function of time postinfection (Fig. 5). A significant increase in promoter activity was observed between 48 and 72 hpi, with the highest activity occurring at the last time point, 120 hpi (Fig. 5). Promoter activity from pGL-ORI-A was decreased over 100-fold at 120 hpi in the presence of PAA but was still 100-fold over baseline levels in the absence of virus (Fig. 5 and inset). This indicated that the oriLyt_PM responded to HCMV infection and showed the most activity at times associated with DNA replication. However, there was a 10-fold increase in promoter activity between 4 and 24 hpi in the presence of PAA (Fig. 5). This response increased to about 100-fold over baseline in the presence of PAA at 96 hpi. This suggested that the oriLyt promoter was responsive to viral factors present before and after the onset of DNA synthesis. This may indicate that an initial promoter response is due to immediate early and early proteins and not totally dependent upon DNA replication.

We also evaluated the activity of the mutated construct, pGL-ORI-Amut, in HFF cells. Cells were infected 24 hpt, and luciferase activity was measured at various times postinfection. The mutated construct, pGEM-ORI-Amut, responded poorly to infection, producing luciferase activity only about 10-fold higher than the basal activity observed in uninfected cells and over 100-fold less than the activity observed in infected cell samples (Fig. 5 and inset). This result was consistent with what was observed in Vero cells with respect to the depressed constitutive expression, indicating that the OPAE is essential for promoter function independent of cell type.

As a control for HCMV activation of nonspecific promoters we transfected pRL-SV40, which uses the SV40 early promoter to drive luciferase expression. This promoter did respond to HCMV infection over the 120-h infection period (Fig. 5 and inset). The response was almost 100-fold lower than that observed for the pGL-ORI-A plasmid at 120 hpi.

The presence of the core replication proteins is necessary for full activation of oriLyt_PM in human fibroblasts.

Since the oriLyt_PM was activated by HCMV infection in human fibroblasts, we performed cotransfections using plasmids encoding all of the necessary core replication factors for oriLyt-dependent DNA replication to determine which of these factors influenced oriLyt_PM activation in human fibroblasts. In addition, we wanted to determine if IE2 alone had any effect on the oriLyt_PM in human fibroblasts. Transfection of IE2 or UL84 expression plasmids alone had a minimal effect on oriLyt_PM activity (Fig. 6). However, when both UL84 and IE2 were added together to the cotransfection mixture, a 150-fold increase in promoter activity was observed (Fig. 6). This indicated that the combination of UL84 and IE2 had a stimulatory effect on promoter activity. Further, when plasmids encoding all the core replication proteins along with IE2 and UL84 were added to the mixture, activation of the oriLyt_PM increased considerably (Fig. 6). The omission of UL84 from the mixture decreased promoter activity to basal levels (Fig. 6, All rep + IE2). Likewise, when IE2 was omitted from the cotransfection mixture, a similar decrease in promoter activity was observed (Fig. 6, All rep + 84). These data revealed that in human fibroblasts the promoter region within oriLyt was responsive to UL84 and IE2 but required the entire replication complex for optimal activity, a result consistent with the infection data. This result was the opposite of what was observed in Vero cells, where the promoter was constitutively active and repressed by the addition of IE2 alone to the transfection mixture. However, the combination of UL84 and IE2 in Vero cells did produce a small increase in promoter activity over the high basal level.

IE2 does not directly interact with OPAE.

Once we determined that there was an IE2-responsive element within HCMV oriLyt, we next wanted to investigate if IE2 directly interacted with this element in vitro. IE2 purified from mammalian cells was incubated with either the CRS element or the OPAE, and an EMSA was performed. IE2 was able to interact with the CRS element in vitro in a specific manner, as demonstrated by cold competition experiments and the observed supershifted species with IE2-specific antibody (Fig. 7A, lanes 1 to 4). However, no specific IE2 interaction was detected using the OPAE oligonucleotides as substrate under the same conditions as with the CRS oligonucleotides (Fig. 7A, lane 5). This indicated that purified IE2 was unable to bind to the OPAE in vitro and suggested that additional factors may be required for interaction.

The negative result from the EMSA prompted us to determine if IE2 interacted with the OPAE in vivo. This would mean that IE2 would require additional cellular or viral factors to bind the OPAE. In order to determine if IE2 interacted with the oriLyt region containing the OPAE, we performed a ChIP assay using HCMV-infected HFF cell extract. PCR primers specific to regions within oriLyt that corresponded to the location of the OPAE or control primers specific for another region of the origin were used to assay for an IE2-specific interaction. The ChIP assay revealed that IE2 did interact with a region of oriLyt that contained the OPAE when the A primers were used (Fig. 7B and C, lane 1). No band was amplified from samples that omitted the IE2-specific antibody from the incubation mixture (Fig. 7C, lane 3). Also, when the B primers were used (Fig. 7B), which corresponded to a downstream region of oriLyt, no amplification product was detected (Fig. 7C, lane 8). This indicated that this region of oriLyt did not interact with IE2 and also served as a negative control for the ChIP assay. As a positive control, we used primers specific for the CRS element located within the upstream region of the MIEP. Using this primer set, an amplification product was detected, confirming that IE2 interacts with the CRS element in vivo (Fig. 7C, lane 4). These results suggested that IE2 interacts with the OPAE either as a complex with another viral or cellular factor or requires other factors to be present such that binding can occur.

An intact OPAE is essential for oriLyt amplification.

Based on the findings that a strong promoter region is present within HCMV oriLyt, we compared the HCMV origin with other herpesvirus lytic origins. Figure 8A is a comparison of the HCMV, HHV8, and EBV oriLyt regions. This diagram indicates that a significant part of HHV8 and EBV lytic origins contains promoter elements, and previous studies showed that promoter activity in these lytic origins was required for origin-dependent DNA replication (3, 18). Also shown is the specific viral protein-binding sites present within HHV8 and EBV oriLyt regions; these include ORF50RE for HHV8 and ZREs and RRE for EBV (Fig. 8A).

To determine if the oriLyt_PM was required for HCMV oriLyt-dependent DNA replication, we introduced the same lethal 1-bp mutation as in the oriLyt_PM luciferase construct (pGL-ORI-Amut) into the replication-competent oriLyt plasmid. This construct, pGEM-oriLyt+G, was transfected into HFF cells, which were subsequently infected with HCMV. Five days postinfection, total cellular DNA was harvested and replication products were examined as described in Materials and Methods. When the pGEM-oriLyt+G plasmid was used in the replication assay, no replication products were detected (Fig. 8B, lane 2), whereas the control transfection which contained the wild-type oriLyt (pGEM-oriLyt) yielded the expected replication product (Fig. 8B, lane 1). This indicated that the activity of this promoter was essential for efficient amplification of oriLyt.

The SV40 early promoter can functionally substitute for the HCMV oriLyt_PM.

In the case of EBV, it was demonstrated that the promoter-enhancer region within oriLyt could be replaced by the HCMV immediate-early enhancer promoter (18). This indicated that, in at least EBV oriLyt, the presence of a nonnative strong constitutive promoter element was sufficient to replace the inducible native element. In HCMV, our study showed that oriLyt_PM mapped to a DNA sequence previously identified as the core and overlapped with one of the two essential regions (essential region I) (Fig. 9A) (66). Once we had determined that the promoter activity of the oriLyt_PM was essential for replication, we wanted to investigate the possibility that this region could be replaced with another strong constitutive promoter, such as in the case of EBV oriLyt.

Since our initial oriLyt promoter studies involved a larger region of oriLyt in transient luciferase reporter assays, we replaced the region of oriLyt between the restriction sites SgiI (91495) and NotI (92888) with the SV40 early promoter (Fig. 9A). The SV40 early promoter was ligated into oriLyt such that two constructs were generated: one having the SV40 early promoter driving expression in the right-to-left direction (pGEM-oriLytSVR) and the other driving expression in the left-to-right direction (pGEM-oriLytSVL). These new constructs were transfected into HFF cells that were subsequently infected with HCMV. Five days postinfection cells were harvested, DNA was extracted, and a Southern blot assay was performed to examine replication products by cleaving with EcoRI and DpnI. Both the original pGEM-oriLyt and each of the new origin constructs were included in the transfection mixture, so that the replication efficiency of each plasmid could be compared directly to wild-type oriLyt. Since the new oriLyt constructs removed one EcoRI site, the amplified oriLyt-SV40 early promoter-containing plasmids migrated in the gel as a larger fragment than the wild-type pGEM-oriLyt plasmid. Plasmid construct pGEM-oriLytSVR replicated to approximately the same efficiency as the pGEM-oriLyt plasmid (Fig. 9B, lane 1), whereas the pGEM-oriLytSVL construct replicated at a decreased efficiency compared to the wild-type construct pGEM-oriLyt (Fig. 9B, lane 2). Amplification of the oriLyt constructs containing the SV40 early promoter in place of the promoter region within oriLyt indicated that the mechanism of DNA synthesis is, in part, due to the activation of transcription or assembly of the transcription machinery at a promoter element within oriLyt.

pGEM-oriLytSVR amplification in a cotransfection replication assay no longer requires IE2 in human fibroblasts.

Once we had determined that oriLyt constructs containing the SV40 early promoter were able to replicate, we next wanted to investigate if IE2 was still required for lytic replication in the cotransfection replication assay in human fibroblasts. T-HF cells were transfected with both pGEM-oriLyt and pGEM-oriLytSVR along with all required core replication proteins and combinations of UL84 and IE2 expression plasmids. Both pGEM-oriLyt as well as pGEM-oriLytSVR replicated in the presence of either HHV8 core replication proteins or HCMV core replication proteins UL84 and IE2 (Fig. 9B, lanes 3 and 4, respectively). However, when IE2 was omitted from the transfection mixture, a replication product was still detected for the amplification of pGEM-oriLytSVR (Fig. 9B, lane 6, top band). As expected, no replicated product was observed for wild-type oriLyt (Fig. 9B, lane 6). The inclusion of UL84 in the transfection mixture was still necessary for amplification of both pGEM-oriLyt as well as pGEM-oriLytSVR, since a cotransfection omitting the UL84 expression plasmid failed to produce any detectable replication products (Fig. 9B, lane 5). No replication products were detected when one of the core replication proteins, helicase, was omitted from the transfection mixture (Fig. 9B, lane 7). This result indicated that the requirement of IE2 in oriLyt-dependent DNA replication was to activate the promoter within oriLyt. Because lane 5 contained a smaller amount of DNA sample, we wanted to confirm that UL84 was still required for oriLyt amplification when the SV40 promoter was substituted for the oriLyt promoter. Consequently, we again performed the cotransfection assay omitting UL84. Control samples containing all the required replication proteins plus UL84 and IE2 replicated both pGEM-oriLyt as well as pGEM-oriLytSVR (Fig. 9, lane 8). However, when UL84 was omitted, neither replication origin was amplified (Fig. 9, lane 9).