Role of the Human Cytomegalovirus Major Immediate-Early Promoter's 19-Base-Pair-Repeat Cyclic AMP-Response Element in Acutely Infected Cells

Abstract

Prior studies have suggested a role of the five copies of the 19-bp-repeat cyclic AMP (cAMP)-response element (CRE) in major immediate-early (MIE) promoter activation, the rate-limiting step in human cytomegalovirus (HCMV) replication. We used two different HCMV genome modification strategies to test this hypothesis in acutely infected cells. We report the following: (i) the CREs do not govern basal levels of MIE promoter activity at a high or low multiplicity of infection (MOI) in human foreskin fibroblast (HFF)- or NTera2-derived neuronal cells; (ii) serum and virion components markedly increase MIE promoter-dependent transcription at a low multiplicity of infection (MOI), but this increase is not mediated by the CREs; (iii) forskolin stimulation of the cAMP signaling pathway induces a two- to threefold increase in MIE RNA levels in a CRE-specific manner at a low MOI in both HFF- and NTera2-derived neuronal cells; and (iv) the CREs do not regulate basal levels of HCMV DNA replication at a high or low MOI in HFF. Their presence does impart a forskolin-induced increase in viral DNA replication at a low MOI but only when basal levels of MIE promoter activity are experimentally diminished. In conclusion, the 19-bp-repeat CREs add to the robust MIE promoter activity that occurs in the acutely infected stimulated cells, although the CREs' greater role may be in other settings.

Human cytomegalovirus (HCMV) replication begins with expression of the major immediate-early (MIE) gene products, IE1 p72 and IE2 p86, which are crucial for activating subsequent viral life cycle events. The expression of these gene products is partly controlled by the MIE regulatory region, which is composed of promoter, enhancer, unique region, and modulator (reviewed in references 23 and 24). The 485-bp enhancer (base positions −65 to −550) is credited for stimulating robust transcription from MIE promoter constructs when placed in in vitro, transfection, and transgenic animal systems (23). It is also required to efficiently activate the HCMV replicative cycle (13, 21). However, enhancer functioning varies with cell type, cellular differentiation, and activity of certain signaling pathways (23, 24). Cellular transcription factors NF-κB/rel, CREB/ATF, AP-1, SP1, serum response factor, ELK-1, and liganded retinoic acid receptor are capable of binding to the enhancer and stimulating transcription from isolated MIE promoter/enhancer segments (4, 23, 24). Because NF-κB/rel, CREB/ATF, SP1, and retinoic acid receptor bind to multiple sites in the enhancer, each set of transcription factors has potential for interacting in cooperative fashion among themselves or with other types of factors. In transient-transfection studies, viral protein pp71 augments enhancer activity via CREB/ATF and AP-1 binding sites, whereas viral IE1 p72 imparts activation through the NF-κB/rel recognition site in the 18-bp-repeat element (5, 16, 18, 35).

We have shown previously that removal of the distal portion of the enhancer from base positions −300 to −580 greatly decreases both MIE gene transcription and viral replication in human foreskin fibroblasts (HFF) at a low multiplicity of infection (MOI) but not at a high MOI (21). The finding was unexpected, given that transient-transfection analyses of comparable distal enhancer deletions in reporter plasmids did not reveal a decrease in MIE promoter activity (15, 36). The 280-bp distal enhancer was determined to contain multiple cis-acting elements that operate independently to augment both MIE promoter-dependent transcription and HCMV replication at low MOIs (20). Although the identities of the elements were not discerned, at least one of the elements was found to reside in a 47-bp enhancer segment spanning base positions from −300 to −347. This short segment contains consensus recognition sites for cellular transcription factors CREB/ATF, SP1, and YY1 (Fig. 1A). The cyclic AMP (cAMP)-response element (CRE) forms a 12-bp palindrome centered in the 19-bp-repeat element, which is tightly bound by CREB/ATF (26, 27, 33, 34), whereas the SP1 and YY1 binding sites are adjoined to make up the 21-bp-repeat element (17, 23). Both the 19- and 21-bp-repeat elements are represented twice more in the distal enhancer. In contrast, the proximal enhancer contains two 19-bp-repeat elements but lacks the 21-bp-repeat. Nucleotide sequences immediately flanking the 12-bp CRE palindrome (5′-ATTGACGTCAAT-3′) vary somewhat among the 19-bp repeats, and one of five copies of the 19-bp repeat contains a variant CRE (5′-ATTTGCGTCAAT-3′) that may also be bound by AP-1 (Fig. 1B). The 19-bp-repeat CRE was determined from transient-transfection experiments to increase basal promoter activity (34) and to confer promoter activation induced by cAMP/protein kinase A (PKA) signaling (12, 26, 34). The 21-bp repeat has not been shown to function as a positive cis-acting element in transient-transfection studies. Functioning of the 19- or 21-bp repeats has not been examined in the context of the HCMV genome. Whether the presence of only one CRE or absence of 21-bp repeats in the murine CMV MIE enhancer accounts for this enhancer's poor functional performance when replacing HCMV's MIE enhancer (13) is unknown.

FIG. 1.

Schematic diagram of the HCMV MIE regulatory region and its 19-bp-repeat CREs. (A) Exploded view of enhancer components located from −300 to −347. Typical (CRE, filled diamond) and variant (AP-1/CRE, open diamond) CREs are centered in 19-bp-repeat elements. Consensus binding sites for YY1 (oval) and SP1 (rectangle) make up the 21-bp-repeat element. Base positions are depicted relative to the +1 RNA start site of the promoter for the MIE genes (arrow). The enhancer (ENH, open box) is divided into proximal (−65 to −300) and distal (−300 to −579) portions. Modulator region (MOD) is also shown. (B) Nucleotide sequences of four typical and one variant 19-bp-repeat CREs. Four 19-bp-repeat elements have identical consensus CRE sequences that form a 12-bp palindrome (underlined) but differ somewhat in flanking sequence. The variant 19-bp-repeat CRE (AP-1/CRE) contains a variant CRE (dashed underline) that also matches a consensus AP-1 binding site. R, purine; Y, pyrimidine.

Inactivated viral particles added in abundance offset the loss of a distal enhancer by restoring MIE promoter-dependent transcription in a virus grown at a low MOI (20). Virion-associated pp71 may resuscitate MIE promoter activity by acting on AP1 elements and CREs in the proximal enhancer or via other regulatory mechanisms (2, 11, 16). Alternatively, other virion components might achieve the same result by operating on remaining cis-acting sites or through different regulatory mechanisms.

Here we describe the results of studies done to test the hypothesis that 19-bp-repeat CREs enhance MIE promoter-dependent transcription and HCMV replication in acutely infected cells. Within the context of the HCMV genome, we determine the role of the 19-bp-repeat CRE in regulating basal and stimulus-induced MIE promoter activity in two different permissive cell types. The element's ability to impart promoter responsiveness to virion-associated components and to effect HCMV DNA replication is also investigated.

MATERIALS AND METHODS

Cells, viruses, and infections.

HFF were isolated and grown, as described previously (22). The hypoxanthine-guanine phosphoribosyl transferase (HGPRT)-deficient fibroblasts (GM02291) were obtained from the Coriell Institute for Medical Research (Coriell Cell Repositories, Camden, N.J.) and were grown in Eagle's minimal essential medium supplemented with 5% newborn bovine serum and 5% fetal bovine serum. HCMVs (Towne strain) were propagated, harvested, and titered, as described previously (21, 22). Titers of HCMVs with distal enhancer mutations were normalized to known titers of replication-competent viruses on the basis of cytopathic effect in HFF at 24 h postinfection (p.i.) and of the amount of viral DNA in HFF prior to viral replication (4 to 5 h p.i.) (21). Inactivation of HCMV by UV light was performed by using methods described previously (20). Forskolin (Sigma) (10 mM stock in dimethyl sulfoxide [DMSO]) was added to culture medium at the indicated concentrations. In experiments performed in the absence of serum, cells were deprived of serum for 16 h prior to infection.

Plasmids and HCMV recombination.

Plasmids p1.6 (21), pIE1 (22), pΔ-300/-579 (20), phCMV_M1-5 (37), and phCMV_WT (37) were described previously. pΔ-300/-579.C1, pΔ-300/-579.C2, and pΔ-300/-579.C3 were made by insertion of one, two, or three copies of a synthetic duplex oligonucleotide 5′-CTAGTATTGACGTCAATG-3′, respectively, into the Avr II site of pΔ-300/-579. pΔ-300/-579.19.2 and pΔ-300/-579.21.2 were constructed by insertion of two copies of oligonucleotides 5′-CTAGCGTCATTGACGTCAATGG-3′ and 5′-CTAGGGCCAGGCGGGCCATTTACCGTC-3′, respectively, into the Avr II site of pΔ-300/-579. To construct pWT and pM5, a silent single nucleotide change was first made in pIE1 at base position −35 of the MIE promoter to create a StuI site. Replacement of the StuI-to-SpeI fragment of this plasmid with equivalent fragments of phCMV_WT and phCMV_M1-5 produced pWT and pM5, respectively. All plasmids were sequenced to confirm the desired mutations. Recombinant HCMVs rΔ-300/-579.C.1, rΔ-300/-579.C.2, rΔ-300/-579.C.3, rΔ-300/-579.19.2, rΔ-300/-579.21.2, rWT, and rCRE-M5 were derived from rΔMSVgpt (22), by using homologous recombination with pΔ-300/-579.C1, pΔ-300/-579.C2, pΔ-300/-579.C3, pΔ-300/-579.19.2, pΔ-300/-579.21.2, pWT, and pM5, respectively. The method of cotransfection of plasmid (5 μg) and rΔMSVgpt DNAs (20 to 50 μg) has been detailed previously (21, 22). Recombinant viruses were selected in HGPRT-deficient fibroblasts (MOI of 0.3 to 0.5) exposed to 6-thioguanine (50 μg/ml), as described previously (21). All recombinant viruses were subjected to at least two rounds of plaque isolation. Two recombinant virus clones were obtained from independent transfection-recombination procedures to control for spurious genomic mutations.

DNA analyses.

HCMV genomic DNA was isolated and digested with the indicated restriction endonucleases, as described previously. Restriction fragment profiles were analyzed by agarose gel electrophoresis, ethidium bromide staining, and Southern blot analysis, as described previously (22). Probes were generated by either multiprime ³²P labeling of indicated double-stranded DNA fragments or ³²P end labeling of indicated oligonucleotides. Blots were stripped of probes by boiling in 0.2% sodium dodecyl sulfate prior to their reuse.

Analysis of HCMV DNA replication by Southern blotting was performed with the methods described previously (21, 22). Briefly, after virus absorption for 1.5 h, the cells were exposed for 1 min to citrate buffer (50 mM sodium citrate and 4 mM KCl) of pH 3 to inactivate extracellular virus and were then washed thrice with Hank's balanced salt solution without calcium and magnesium. Infected cell DNA was isolated, digested with HindIII, fractionated electrophoretically on a 0.7% agarose gel, and subjected to Southern blot analysis. Lambda DNA (2 μg) was added to each cell lysate sample to control for variation in processing, endonuclease digestion, and loading. Southern hybridization was carried out with a probe that is complementary to HCMV genomic termini containing TR_L or IR_L. This probe, termed T probe, was generated by multiprime ³²P labeling the 1.6-kbp BamHI-HindIII fragment of plasmid p1.6 (21). The IR_L is fused (Fused Ends [see Fig. 3]) to the short genome segment, which is contained in the 17.2- and 13-kb fragments. The TR_L is not fused (Free Ends) to the short genome segment and is contained in the 9.7-kb fragment. Hybridization signals were quantitated by image acquisition analysis (Hewlett-Packard Instant Imager). The lambda control bands were imaged by ethidium bromide staining.

FIG. 3.

Effect of the reconstituted CRE on HCMV DNA replication and MIE promoter-dependent transcription at high and low MOIs in HFF. (A) Southern blot analysis of WT, rΔ-300/-579.C3, rΔ-300/-579.C1, and rΔ-300/-579 genome amounts at an MOI of 1.0 and 0.001 at 2 and 7 days p.i., respectively. HFF were infected in parallel with equivalent input titer of viruses, as determined by two independent methods described in Materials and Methods. Cell-associated DNA was isolated, digested with HindIII, fractionated by gel electrophoresis, and subjected to Southern blot analysis with a ³²P-labeled probe corresponding to the terminal repeat-long (Free Ends) or internal repeat-long (Fused Ends) region, as detailed in Materials and Methods. The vast majority of HCMV DNA in HFF at an MOI of 1.0 and 0.001 at 2 and 7 days p.i., respectively, is newly replicated viral DNA (20, 21). Lambda DNA (Control) was added to each cell lysate prior to isolation of DNA to control for sample processing, restriction enzyme digest, and loading. (B and C) MIE RNA output at an MOI of 1.0 and 0.005 in the presence or absence of serum. Serum-nourished and -deprived HFF were infected in parallel. Input viral titers were determined by two independent methods detailed in Materials and Methods. WT, rΔ-300/-579.C3, and rΔ-300/-579 MIE RNA was quantified by multiplex real-time RT-PCR at 6 h p.i. after normalization (RNA_N) to cellular 18S RNA abundance. Results are expressed relative to the WT MIE RNA amount (REL to WT). Means and standard deviations are shown. Samples lacking RT were analyzed in parallel and did not significantly differ from the baseline.

HCMV DNA in cells at a low MOI was analyzed by multiplex real-time PCR with the ABI PRISM 7700 Sequence Detection System (Perkin-Elmer [PE] Applied Biosystems, Foster City, Calif.). Cell-associated DNA was prepared with methods described previously (22). MIE enhancer DNA was amplified with primers 5′-CATGGTGATGCGGTTTTGG-3′ and 5′-TGGAAATCCCCGTGAGTCA-3′ and was detected with reporter probe 5′-6-carboxy-fluorescein-ACCGCTATCCACGCCCATTGATGT-tetramethyl rhodamine-3′. Primers and probe used for amplification and detection of ribosomal 18S DNA, respectively, were obtained from PE Applied Biosystems. Primer and template concentrations were chosen based on results of standard curves.

RNA analyses by multiplex reverse transcriptase (RT)-PCR.

Whole-cell RNA was isolated according to the method of Chomczynski and Sacchi (6). cDNA was made by using SUPERSCRIPT II RNase H⁻ RT (Life Technologies, Gaithersburg, Md.) and random hexamers (Invitrogen, Carlsbad, Calif.), as described previously (20). The specifications of the multiplex real-time PCR procedure were reported previously (20). Primers located in MIE exons 1 and 2 produce an MIE amplicon spanning intron A. Ribosomal 18S DNA was amplified and detected with reagents supplied by PE Applied Biosystems. Primer and template concentrations were chosen based on results of relative standard curve and C_T validation experiments, where C_T is comparative threshold cycle. C_T values of samples not treated with RT but analyzed in parallel to the RT-treated samples did not appreciably differ from the baseline.

RESULTS

Reconstitution of the distal enhancer deletion with 19-bp-repeat CRE.

The distal enhancer contains three copies of a classical consensus CRE residing as a 12-bp palindrome in the 19-bp-repeat element (Fig. 1). To explore whether the CRE imparts positive cis-acting function to the distal enhancer, we initially constructed recombinant HCMVs in which the distal enhancer (−305 to −579) is replaced by one, two, or three copies of CRE (5′-CATTGACGTCAATA-3′). These viruses are designated rΔ-300/-579.C1, rΔ-300/-579.C2, and rΔ-300/-579.C3. To control for specificity, HCMV rΔ-300/-579.21.2 was made by replacing the distal enhancer with two copies of the 21-bp repeat, which contains SP1 and YY1 binding sites. The schematic layout and Southern blot analysis of rΔ-300/-579.C1, rΔ-300/-579.C2, rΔ-300/-579.C3, and rΔ-300/-579.21.2 genomes are depicted in Fig. 2 in comparison to rΔ-300/-579 and the wild type (WT). The MIE regulatory regions of recombinant viral genomes were also subjected to DNA sequencing to further confirm the desired genetic changes. Two distinct isolates of each recombinant type derived from independent recombination procedures were studied to control for spurious mutations.

FIG. 2.

Construction of HCMVs in which the distal enhancer (−300 to −579) is replaced with 19-bp-repeat CRE or 21-bp-repeat elements. (A) Schematic diagram of the WT and recombinant HCMVs rΔ-300/-579, rΔ-300/-579.C1, rΔ-300/-579.C2, rΔ-300/-579.C3, and rΔ-300/-579.21.2. The distal enhancer deletions in rΔ-300/-579.C1, rΔ-300/-579.C2, and rΔ-300/-579.C3 are reconstituted with one, two, and three copies of the conserved CRE palindrome of the 19-bp repeat. rΔ-300/-579.21.2 has two copies of the 21-bp-repeat oriented in the same way that they occur in the WT enhancer. Base positions of deletions and SphI and BsrGI sites are depicted relative to the +1 RNA start site of the MIE promoter. MIE, MIE transcription unit; ENH, enhancer; MOD, modulator; and UL128, putative UL128 gene. (B) Southern blot analysis of SphI-to-Bsr GI fragments of rΔ-300/-579, rΔ-300/-579.C1, rΔ-300/-579.C2, rΔ-300/-579.C3, WT, and rΔ-300/-579.21.2. Predicted sizes of SphI-to-Bsr GI restriction fragment length polymorphisms are depicted. Probe spans base positions from +171 to −220 with respect to the RNA start site of the WT MIE promoter.

WT, rΔ-300/-579.C1, rΔ-300/-579.C3, and rΔ-300/-579 were first examined for their abilities to replicate in HFF at MOIs of 1.0 and 0.001. Two independent methods were used to normalize input viral titers (Materials and Methods). DNA from infected cells was isolated at the indicated times p.i. and was subjected to HindIII digestion and Southern blot analysis. As shown in Fig. 3A, the four different viruses replicate DNA to similar levels at an MOI of 1.0. In contrast, rΔ-300/-579.C1, rΔ-300/-579.C3, and rΔ-300/-579 are markedly impaired in the rate of DNA replication at an MOI of 0.001 in comparison to the WT (∼30-fold difference). rΔ-300/-579.C1, rΔ-300/-579.C3, and rΔ-300/-579 are similar in ability to replicate DNA at an MOI of 0.001, as they varied by 1.2-fold or less after normalization to the internal control. The experimental outcome did not differ when the DNA replication rate was analyzed at other times after infection (data not shown). Additionally, the DNA replication kinetics of rΔ-300/-579.C2 and rΔ-300/-579.21.2, which are not shown here, did not differ from those of rΔ-300/-579.C1, rΔ-300/-579.C3, and rΔ-300/-579. The size and morphology of viral plaques in HFF did not differ among rΔ-300/-579.C1, rΔ-300/-579.C2, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 (data not shown). All viruses lacking a distal enhancer formed much smaller plaques than did the WT, consistent with prior reports.

In a parallel study, the MIE RNA produced by the WT, rΔ-300/-579.C3, and rΔ-300/-579 at 6 h p.i. in HFF was quantified for MOIs of 1.0 and 0.005 by using the real-time RT-PCR method. Concomitant measurement of cellular 18S rRNA controlled for sample-to-sample variation. Because the CRE has the potential to mediate serum-induced transcriptional activation (7), studies were performed in the presence and absence of serum. The results shown in Fig. 3B reveal that the viruses make similar amounts of MIE RNA at an MOI of 1.0, irrespective of whether serum is present or absent. However, rΔ-300/-579.C3 and rΔ-300/-579 produce 9- to 10-fold less MIE RNA at an MOI of 0.005 than does WT when serum is present (Fig. 3C). This difference is 16-fold in the absence of serum. For all of the viruses, the withholding of serum greatly decreases MIE RNA production (8- to 10-fold) at a low MOI. Comparison of rΔ-300/-579.C3 and rΔ-300/-579 indicates that three inserted copies of CRE do not significantly increase MIE RNA abundance at a low MOI regardless of the presence or absence of serum.

Thus, replacement of the distal enhancer with the 19-bp-repeat CRE does not increase basal levels of HCMV DNA replication or MIE promoter-dependent transcription in HFF.

Stimulus-induced activation of the HCMV replicative cycle by the reconstituted 19-bp-repeat CRE.

The capacity of the reconstituted 19-bp-repeat CRE to respond to cAMP/PKA-mediated signaling induced by forskolin was investigated in serum-deprived HFF at an MOI of 0.005. Relative levels of MIE RNA production were determined at 6 h p.i. for WT, rΔ-300/-579.C2, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 in the presence or absence of forskolin. As anticipated, rΔ-300/-579.C2, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 are markedly impaired in MIE RNA production compared to the WT when forskolin is omitted (Fig. 4A). However, they vary in the ability to increase MIE RNA production in response to forskolin, with the magnitude of increase commensurate with the number of CRE copies replacing the distal enhancer. As shown, forskolin induces a 32.4-, 16.5-, and 9.1-fold increase in MIE RNA production over basal levels for viruses containing three, two, and no distal enhancer CREs, respectively. After forskolin stimulation, rΔ-300/-579.C3, rΔ-300/-579.C2, and rΔ-300/-579 make MIE RNA amounts that are nearly 45, 18, and 7%, respectively, of that made by the WT. The three CREs in rΔ-300/-579.C3 increase forskolin-induced MIE RNA production by 3.6-fold compared to production in rΔ-300/-579, which lacks these elements (32.4- versus 9.1-fold change, respectively). The two 21-bp repeats of rΔ-300/-579.21.2 fail to enhance responsiveness to forskolin compared to rΔ-300/-579 (6.1- versus 9.1-fold change, respectively), implying that forskolin-induced signaling is acting specifically through the reconstituted CRE to further increase MIE promoter-dependent transcription. Similar outcomes were observed when these viruses were examined under different forskolin concentrations (data not shown). The forskolin-induced increase in MIE RNA production by rΔ-300/-579 may partly reflect the functioning of the two 19-bp-repeat CRE copies remaining in the proximal enhancer.

FIG. 4.

Forskolin-induced activation of MIE promoter-dependent transcription conferred by the reconstituted 19-bp-repeat CRE. (A) Forskolin-induced MIE RNA production in HFF at an MOI of 0.005. Forskolin (50 μM) was added to or omitted (No Add) from serum-free growth medium during the infection. MIE RNA produced by the WT, rΔ-300/-579.C2, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 was quantified at 6 h p.i. by multiplex real-time RT-PCR after normalization to the 18S RNA amount (MIE RNA_N). Results areexpressed relative to WT infection in the absence of forskolin (REL to WT_{No Add}). Means and standard deviations are shown. Samples lacking RT did not significantly differ from the baseline. (B) Forskolin-induced MIE RNA production in NTera2-derived neuronal cells at an MOI of 0.05. WT, rΔ-300/-579.C3, and rΔ-300/-579 MIE RNA was quantified at 6 h p.i. by using methods applied for panel A. Forskolin (50 μM) was added to or omitted from serum-free growth medium during the infection. (C) Comparison of the conserved CRE palindrome and entire 19-bp repeat in forskolin-induced MIE promoter activation. rΔ-300/-579.19.2 differs from rΔ-300/-579.C2 in having two tandem copies of the 19-bp repeat instead of two copies of the 14-bp palindromic CRE (data not shown). The 19-bp repeats are oriented in the same way that they occur in the WT enhancer. Serum-deprived HFF were infected with rΔ-300/-579.C2 and rΔ-300/-579.19.2 at an MOI of 0.005 in the presence or absence of forskolin (50 μM). MIE RNA was quantified at 6 h p.i. by using methods applied for panel A.

Because CREB/ATF activity varies according to cell type (25, 31), we examined 19-bp-repeat CRE functioning in NTera2-derived neuronal cells, which are also permissive for HCMV infection (9, 19). The amount of MIE RNA produced by the WT, rΔ-300/-579.C3, and rΔ-300/-579 at an MOI of 0.1 was quantified at 6 h p.i. in serum-deprived cells. In Fig. 4B, the findings reveal that the three additional copies of 19-bp-repeat CRE in rΔ-300/-579.C3 specifically increase forskolin-induced (32.11 versus 7.53) but not basal (0.27 versus 0.30) levels of MIE promoter-dependent transcription when compared to rΔ-300/-579. While these results are similar to those generated in HFF, activity of the intact MIE promoter/enhancer is subject to higher levels of induction by forskolin in NTera2-derived neuronal cells than in HFF (23.48- versus 4.44-fold induction, respectively). Nonetheless, the three reconstituted CREs in the distal enhancer are still able to elevate levels of MIE RNA production after forskolin stimulation to 37% of that of the WT (8.67 versus 23.48, respectively).

The reconstituted CRE consists of 14 out of 19 bases of the 19-bp repeat (Fig. 1B). To determine whether this difference in nucleotide composition could influence the experimental findings, the distal enhancer deletion was reconstituted with two 19-bp repeats to make virus rΔ-300/-579.19.2 (data not shown). The virus was compared to rΔ-300/-579.C.2, which has two 14-bp CRE copies instead of the two 19-bp repeats. Both viruses are equivalent to the WT in producing MIE RNA and viral DNA at an MOI of 1.0 and are as ineffective as the WT in producing MIE RNA and viral DNA at an MOI of 0.005 (data not shown). Moreover, both rΔ-300/-579.19.2 and rΔ-300/-579.C2 respond in a manner similar to that of forskolin by expressing comparable amounts of MIE RNA at an MOI of 0.005 in HFF (Fig. 4C), indicating that variability in nucleotide sequence flanking the 12-bp CRE palindrome of the 19-bp repeats is inconsequential under these experimental conditions.

We also questioned whether the reconstituted 19-bp-repeat CRE could impart an forskolin-induced increase in HCMV DNA replication. To examine forskolin's effect on initiation of the viral replicative cycle, infected cells were exposed to forskolin for 6 h in serum-free medium and were then cultured in serum-containing medium. Single-step viral DNA replication in HFF at an MOI of 0.005 was quantified by real-time PCR at 2 days p.i., before infectious viral progeny are released to surrounding cells. As shown in Fig. 5, forskolin stimulation increases rΔ-300/-579.C3 DNA abundance by 3.9-fold compared to rΔ-300/-579 (6.3- versus 1.6-fold increase, respectively) after normalization to cellular 18S gene abundance.

FIG. 5.

Forskolin-induced activation of HCMV DNA replication conferred by the reconstituted 19-bp-repeat CRE. WT, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 DNA in HFF at an MOI of 0.01 at 2 days p.i. was quantified by multiplex real-time PCR after normalization to DNA of cellular 18S genes (HCMV DNA_N). Forskolin (50 μM) was added to or omitted (No Add) from serum-free growth medium for 6 h of infection and was then replaced with growth medium containing serum. Viral DNA abundance is depicted relative to the WT in the absence of forskolin (REL to WT_{No Add}). Means and standard deviations of quadruplicate samples are shown. Uninfected cells analyzed in parallel did not significantly differ from the baseline. Relative input viral DNA amounts in cells at 4 h p.i. with WT, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 were 1, 1.9, 1.9, and 1.7, respectively. Two additional experiments performed on different lots of HFF produced similar results.

Hence, the reconstituted 19-bp-repeat CREs function in forskolin-induced activation of MIE promoter-dependent transcription and viral DNA replication at low MOIs. The reconstituted elements do not by themselves restore basal levels of MIE promoter activity in unstimulated HFF and NTera2-derived neuronal cells at low MOIs.

The 19-bp-repeat CRE confers stimulus-induced but not basal activation of MIE promoter-dependent transcription.

We next examined the collective role of all five copies of the 19-bp-repeat CRE in the enhancer. The middle CG dinucleotides of each CRE were reversed to GC, as this change abolishes CREB/ATF binding in vitro (3) and CRE-dependent HCMV MIE promoter/enhancer activation in an adenoviral transfer vector in neurons (37). Southern blot and DNA-sequencing analyses of the resultant HCMV rCRE-M5 genome (data not shown) verified the mutations schematically depicted in Fig. 6A. Remarkably, rCRE-M5 is equivalent to the WT in the rate of viral DNA replication in HFF at MOIs of 1.0 (data not shown) and 0.001 (Fig. 6B). Moreover, rCRE-M5 produces normal basal amounts of MIE RNA at MOIs of 1.0 and 0.005 in serum-deprived HFF (Fig. 6C). In contrast, forskolin (10 and 30 μM) induces two- to threefold-greater activation of MIE RNA production by the WT at 6 h p.i. than does rCRE-M5 at an MOI of 0.005 (Fig. 7A). This difference was not limited to HFF, as forskolin-induced MIE RNA production is also two- to threefold greater for WT than for rCRE-M5 in NTera2-derived neuronal cells (Fig. 7B). Similar outcomes were observed when the viruses were examined under a range of forskolin concentrations or exposure times (data not shown).

FIG. 6.

Effect of 19-bp-repeat CRE mutations on HCMV DNA replication and MIE promoter-dependent transcription in HFF. (A) Schematic diagram of base substitution mutations (boldfaced small letters) made in all five 19-bp-repeat CREs to create rCRE-M5. Mutations were confirmed by DNA sequencing of genomes of two rCRE-M5 clones made from independent recombination procedures. Nucleotide sequences of typical and variant 19-bp repeats of the WT are depicted in Fig. 1B. (B) WT and rCRE-M5 DNA production at an MOI of 0.005 in HFF at 2 and 3 days p.i. HCMV DNA in HFF was quantified and normalized to 18S DNA (HCMV DNA_N) by using methods described in the Fig. 5 legend. The analysis of HCMV DNA at 4 h p.i. controlled for equivalent input titers of the WT and rCRE-M5. The HCMV DNA amount at 2 days p.i. increased 179- to 197-fold from the viral DNA amount at 4 h p.i. (not depicted). The amount of replicated HCMV DNA is depicted relative to the WT at 2 days p.i. (REL to WT_{2 DP}). (C) WT and rCRE-M5 MIE-RNA production in HFF at an MOI of 1.0 and 0.005. MIE RNA was quantified and normalized to the amount of 18S RNA (MIE RNA_N) at 6 h p.i. by using methods described in the Fig. 3B legend. Infections were performed in parallel with experiments shown in panel B. Results are depicted relative to the WT (REL to WT).

FIG. 7.

Effect of 19-bp-repeat CRE mutations on forskolin-induced MIE promoter activation in HFF and NTera2-derived neuronal cells. (A) Forskolin-induced WT and rCRE-M5 MIE RNA production in HFF at an MOI of 0.005. Forskolin was applied or omitted from serum-free growth medium for 0.5 h and was then replaced with serum-free growth medium. A forskolin concentration of 10 or 30 μM was chosen based on prior findings of dose-response studies. MIE RNA was quantified and normalized to 18S RNA (MIE RNA_N) at 6 h p.i. by using methods described in the Fig. 3B legend. Results represent the mean and standard deviation of quadruplicate samples. They are expressed as the change (n-fold) in the amount of MIE RNA_N induced by forskolin relative to its absence. Two additional experiments performed independently on different lots of HFF produced similar results. (B) Forskolin-induced WT and rCRE-M5 MIE RNA production in NTera 2-derived neuronal cells at an MOI of 0.05. Forskolin was applied at 5 or 15 μM or was omitted from serum-free growth medium for 1.5 h p.i. and was then replaced with serum-free growth medium. MIE RNA was quantified and normalized to 18S RNA (MIE RNA_N) at 6 h p.i., as done in panel A. Mean values and standard deviations of quadruplicate samples are shown. Results are expressed as change (n-fold) in the amount of MIE RNA_N induced by forskolin relative to its absence. An independently performed experiment produced similar results.

These findings reveal that mutation of the five 19-bp-repeat CREs decreases response to forskolin-induced stimulation in two different permissive cell types but does not alter basal levels of MIE promoter activity in these acutely infected cells.

The 19-bp-repeat CRE does not contribute to activation of MIE promoter-dependent transcription by viral particles.

The mechanism(s) by which virion components such as pp71 activate the MIE promoter in acutely infected cells (2, 20) is unclear, although transient-transfection findings suggest that pp71 exploits the CRE and AP-1 cis-acting elements to attain this result (16). We therefore examined the ability of UV-inactivated isolated HCMV particles (HCMV_UV) to augment MIE promoter activity in viruses either possessing or lacking 19-bp-repeat CREs.

In the first set of experiments, HCMV_UV was applied at a high viral particle-to-cell ratio (∼3 to 5 PFU/cell prior to UV irradiation) to serum-deprived HFF during inoculation with WT, rΔ-300/-579.C3, rΔ-300/-579.21.2, or rΔ-300/-579 at an MOI of 0.005. Relative MIE RNA abundance was determined at 6 h p.i. after normalization to the cellular 18S RNA amount. As reported previously (20), the added HCMV_UV particles partially correct the MIE promoter activity defect that is imposed by a distal enhancer deletion (Fig. 8A). HCMV_UV particles activate the WT MIE promoter over basal levels to a much lesser degree than they do the rΔ-300/-579 counterpart (11.7- versus 597-fold activation, respectively), partly reflecting the higher basal activity of the WT promoter than that of rΔ-300/-579 (1.0 versus 0.007, respectively). A small amount of inactivated MIE RNA accompanies the massive inoculum of isolated HCMV_UV particles, which is consistent with previous reports (10, 20). Remarkably, the three 19-bp-repeat CREs in rΔ-300/-579.C3 do not increase activation of MIE promoter-dependent transcription by HCMV_UV compared to rΔ-300/-579.21.2 or rΔ-300/-579 (471- versus 711- or 597-fold activation, respectively), which lack these elements.

FIG. 8.

The 19-bp-repeat CREs do not enhance MIE promoter activation by UV-inactivated WT virions (HCMV_UV). (A) Activation of WT, rΔ-300/-579.C3, rΔ-300/-579.21.2, and rΔ-300/-579 MIE promoters by HCMV_UV. HFF were infected at an MOI of 0.005 with equivalent input titers of viruses in the presence or absence (No Add) of HCMV_UV at a high particle-to-cell ratio (MOI of 3 prior to UV inactivation) in serum-free medium. MIE RNA was quantified and normalized to 18S RNA (MIE RNA_N) at 6 h p.i. by using methods described in the Fig. 3B legend. Mean values and standard deviations of quadruplicate samples are depicted relative to WT infection in the absence of HCMV_UV (REL to WT_{No Add}). (B) Activation of WT and rCRE-M5 MIE promoters by HCMV_UV. These studies were performed and analyzed in the same manner described in panel A, except that HCMV_UV was at an MOI of 1 prior to UV irradiation. Mean values and standard deviations of quadruplicate samples are depicted relative to WT infection in the absence of HCMV_UV (REL to WT_{No Add}). Values for cells inoculated with HCMV_UV only and analyzed in parallel did not significantly differ from the baseline.

In the second set of experiments, the WT and rCRE-M5 were studied in parallel to determine if mutation of the enhancer's five 19-bp-repeat CREs altered promoter responsiveness to HCMV_UV. As shown in Fig. 8B, activation of MIE promoter-dependent transcription by HCMV_UV in serum-deprived HFF is not diminished by the mutations in rCRE-M5 from that in the WT (10.1- versus 9.2-fold activation, respectively). Thus, the testing of viruses produced by two different construction strategies indicates that the 19-bp-repeat CRE does not contribute appreciably to MIE promoter activation by virion-associated components.

DISCUSSION

Earlier work determined that a 47-bp distal enhancer segment containing consensus binding sites for CREB/ATF, SP1, and YY1 augments MIE promoter activity at a low MOI (20), a result not predicted by transient-transfection studies (15, 36). The possibility of multiple CREs acting in a synergistic or cooperative manner, as they do in human T-cell leukemia virus type I long-terminal-repeat promoter activation (32), was contemplated (20). Here we report that replacing the distal enhancer with 19-bp-repeat CREs does not increase basal MIE promoter activity and that mutating the five CREs of the entire enhancer does not decrease basal MIE promoter activity in acutely infected HFF and NTera2-derived neuronal cells (Fig. 3, 4, and 6). This finding suggests that 19-bp-repeat CREs do not assist in the MIE promoter activation that is imparted by incoming virion components, such as pp71 (UL82) (2, 11, 16), UL69 (38), and TRS1/IRS1 (29). The unresponsiveness of CRE at low MOIs to an excess of HCMV_UV virions further supports this inference (Fig. 8).

Several prior reports have concluded that the 19-bp-repeat CRE is readily bound in vitro by cellular CREB and ATF-1 (26, 27, 33, 34), which are expressed in most human tissues (25, 31). This finding is anticipated, given that the element contains a high-affinity binding sequence, 5′-TGACGTCA-3′, for CREB family members (25, 31). The CREB family consists of CREB, ATF-1, CREM, and their respective isoforms. Forskolin, an activator of adenylyl cyclase, induces cAMP-dependent PKA-mediated phosphorylation of CREB/ATF to allow attachment of coactivator CBP or p300, which strongly promotes transcriptional activation (25, 31). Other signaling pathways can also render CREB/ATF active by either phosphorylation-dependent or -independent mechanisms (7, 25, 31). Thus, extracellular stimuli might act through the 19-bp-repeat CREs to modify the initial regulatory events in MIE promoter activation. This notion was tested by exposing cells to forskolin at the beginning of viral adsorption, given that CREB/ATF activation occurs within 15 to 20 min of such stimulation (25). The observation that forskolin induces only a modest increase in CRE-mediated MIE promoter activity (two- to threefold) in two different cell types is somewhat of a surprise (Fig. 4 and 7). Altering the timing or duration of forskolin exposure or the concentration of forskolin did not further the increase (data not shown). Analysis of MIE RNA amounts at 2, 3, 4, and 8 h p.i. also did not reveal a greater difference (data not shown). In addition, activation of CREB/ATF through alternative signaling pathways via stimulation with epidermal growth factor, nerve growth factor, or calcium ionophores and the study of acutely infected primary human neuronal cells produced outcomes that did not surpass those represented here (data not shown). Our findings are concordant with those of transfected reporter plasmids in NTera2-derived neuronal cells (27) but not those of transfected human fibroblasts, in which the 19-bp-repeat CRE did not respond to cAMP induction (34).

Residual responsiveness to forskolin despite mutation of the five 19-bp-repeat CREs could result from any one of several possibilities: (i) cross talk between different signaling pathways is common (7, 25, 31) and can result in stimulation of diverse types of regulatory mechanisms. (ii) CREB/ATF may bind atypical CRE-like elements in a manner that is strengthened by interactions with other transcription factors (25). Such nonconsensus binding sites remain in the MIE regulatory region. (iii) Consensus CREs not located in proximity to the MIE promoter may be acting from a distance, although this would be atypical of CRE function. (iv) CREB/ATF activation quickly induces expression of c-fos (31), which may activate the MIE promoter via the AP-1 binding sites. Conceivably, CREB/ATF activation might alter expression of other cellular gene products to increase MIE RNA levels in a CRE-independent fashion. (v) The forskolin preparation may impart nonspecific effects, but this possibility seems unlikely, given that the dose-response curve largely parallels that of the WT MIE promoter over a range of forskolin concentrations (2 to 50 μM) (data not shown). Further, forskolin treatment groups were compared to groups exposed only to the DMSO vehicle (0.5%), as DMSO itself has been reported to activate transfected MIE promoter constructs (1). (vi) The mutations' failure to greatly impair CREB/ATF binding is an unlikely possibility, given that prior studies have shown that the same base changes eliminate binding in vitro (3) and abolish forskolin-induced MIE promoter/enhancer activation in neurons (37). The findings of CRE reconstitution studies (Fig. 4) are in accord with this conclusion, as they show that the CREs contribute only modestly to forskolin-induced MIE promoter activation.

Serum was omitted from some experimental groups to exclude its potential role in stimulation of CREB/ATF activity (31). We find that serum has a marked positive effect on MIE promoter activation at a low MOI (8- to 10-fold difference at an MOI of 0.005) that occurs independently of the 19-bp-repeat CRE (Fig. 3). The serum effect also is not dependent on the enhancer's serum response element and ELK-1 binding site, as their absence by removal of the distal enhancer has minimal consequences (Fig. 3). Confluent HFF arrested in the G₀ phase were used in these studies to exclude the variable of cell cycle phase effect on MIE gene expression (8). Based on these results, we propose that serum-induced augmentation of MIE promoter activity likely involves a regulatory mechanism(s) other than CRE, serum response element, and ELK-1.

The level of HCMV DNA replication in acutely infected HFF is commensurate with the level of MIE promoter-dependent transcription, a relationship that is most striking at low MOIs (20, 21). Thus, the finding that the HCMV DNA replication rate is unaltered by the presence or absence of 19-bp-repeat CREs at a low MOI in unstimulated HFF (Fig. 3 and 6) was anticipated. In contrast, the initial boosting of MIE promoter activity by cAMP stimulation might indirectly increase the level of HCMV DNA replication. Such an effect would likely be short-lived, because CREB/ATF deactivation occurs after 4 to 6 h, regardless of whether agonist is present (25). We reasoned that analysis of HCMV DNA accumulation in a single step of viral replication at a low MOI would improve chances of determining whether the initial cAMP stimulation enhances viral DNA replication. This difference might be detected as early as 2 days p.i. when the HCMV DNA amount increases by >170-fold (Fig. 6 legend) by presumed theta-form replication (21), whereas the difference might diminish subsequently with extensive rounds of rolling-circle viral DNA replication. This idea turned out to be partially correct. In a virus lacking the distal enhancer, the reconstitution of 19-bp-repeat CREs reproducibly confers a modest DNA replication advantage after forskolin stimulation compared to that in a virus without these elements or with reconstituted 21-bp-repeats instead (3.9- or 1.9-fold difference, respectively) (Fig. 5). This difference decreases at 3 days p.i. (data not shown). However, forskolin stimulation failed to augment WT DNA replication (Fig. 5), and rCRE-M5 was equivalent to the WT in this regard (data not shown). We surmise that this disparity likely is linked to the marked difference in basal MIE promoter/enhancer function, as WT MIE promoter activity is 10- to 30-fold greater than that of a MIE promoter lacking the distal enhancer when examined at an MOI of 0.005 (Fig. 3 and 4). Thus, the CRE-dependent forskolin effect on HCMV DNA replication is only appreciable for a virus in which the MIE promoter/enhancer functions greatly below capacity (20, 21). Notably, the forskolin-induced increase in viral DNA replication may not result solely from MIE promoter activation, because a variety of other key viral promoters contain CREs as well (14, 28). Whether WT DNA replication would be augmented by forskolin in a cell type in which the intact promoter/enhancer operates far below capacity with very low basal activity remains to be determined.

Because the CRE repetition contributes minimally to MIE promoter activation in acutely infected fibroblasts and neuronal cells, it may have another function. For example, its main function might be relegated to a specialized cell type, possibly equipped with unique intracellular signaling circuitry, that permits the replication of incoming viruses. Alternatively, CRE functioning might be crucial for MIE promoter reactivation in latently infected cells when the supplementary actions of virion components are lacking. In stark contrast, the CRE might not have a major role in MIE promoter activation but might instead serve another function, such as to confer MIE promoter silencing in latently infected cells by allowing binding of ICER-like repressor proteins (25, 30). The work presented here sets the stage for future studies to further unveil the role(s) of the CRE repetition.