Heterologous Viral Promoters Incorporated into the Human Cytomegalovirus Genome Are Silenced during Experimental Latency

Abstract

Human cytomegalovirus (HCMV) lytic phase gene expression is repressed upon entry into myeloid lineage cells where the virus establishes latency. Lytic infection is not initiated because the tegument-delivered transactivator protein pp71 fails to enter the nucleus and inactivate the Daxx-mediated cellular intrinsic defense that silences the viral genome. When pp71 is expressed de novo in THP-1 monocytes, it localizes to the nucleus, inactivates the Daxx defense, and initiates lytic infection. We speculated that replacing the native viral promoter that drives pp71 expression with one that is highly and constitutively active in myeloid cells would permit pp71 de novo expression upon infection and that this newly expressed pp71 would accumulate in the nucleus, inactivate the intrinsic defense, and initiate the cascade of lytic gene expression. Surprisingly, we found that this promoter was still subject to normal silencing mechanisms in THP-1 monocytes and primary CD34⁺ cells, two independent myeloid lineage cells. A second constitutively active heterologous viral promoter located in a different region of the HCMV genome was also silenced in THP-1 and CD34⁺ cells. Furthermore, these two independent heterologous viral promoters inserted into three different regions of the HCMV genome in three different viral strains all required prior expression of the viral immediate early proteins for activation in fibroblasts. From this, we conclude that incorporation within the HCMV genome impacts the proclivity of heterologous viral promoters to initiate transcription. These observations have mechanistic implications for the expression of viral genes and transgenes during both lytic infection and latency.

INTRODUCTION

Upon entry into completely differentiated cells such as fibroblasts, human cytomegalovirus (HCMV) initiates a productive, lytic replication program where infectious progeny virions are released and cells are lysed and killed (1). This type of infection allows virus to spread throughout and among hosts and causes disease in patients with immature or impaired immune function. Lytic infection is triggered when the viral pp71 protein is introduced into cells from the tegument layer of the virion during entry, migrates to the nucleus, and activates viral immediate early (IE) gene expression (2). Synthesis of the IE proteins is required for the propagation and completion of the productive replication cycle. Tegument-delivered pp71 transactivates the IE genes by inactivating a cellular intrinsic immune defense (by degrading the transcriptional corepressor Daxx) that otherwise silences lytic phase gene expression from the viral genome (3).

Upon entry into incompletely differentiated cells of the myeloid lineage, HCMV establishes latency (4, 5). In latently infected cells, lytic phase gene expression is suppressed, and nascent virions are not produced. However, latent genomes retain the capacity to reactivate upon cellular differentiation to fully productive, lytic infections. Latency allows the virus to be maintained for the life of the infected host even in the face of intense immune surveillance and is in part responsible for the pandemic and indelible success of HCMV as an infectious pathogen. Viral IE genes are not expressed when latent infections are established in part because tegument-delivered pp71 fails to reach the nuclei of infected, incompletely differentiated myeloid cells, and thus, the cellular intrinsic defense encounters no resistance and effectively silences the viral genome (6–8, 49).

Thus, the subcellular localization of pp71 upon delivery by the tegument controls the fate of the infected cell: lytic if it enters the nucleus and latent if it remains cytoplasmic (9). The mechanism through which pp71 subcellular localization is controlled is unknown. However, it is clear that the localization of the tegument-delivered protein and newly (de novo) expressed pp71 are regulated differently (10). In the incompletely differentiated myeloid cells where tegument-delivered pp71 remains cytoplasmic, the de novo expressed protein efficiently enters the nucleus where it degrades Daxx. Functionally, this can result in IE gene activation, as is the case in THP-1 monocytes expressing pp71 from a transduced recombinant adenovirus that are then subsequently infected with HCMV. Under such circumstances, THP-1 cells ectopically expressing wild-type pp71 show evidence of Daxx degradation and initiate viral IE gene expression, whereas those that ectopically express a mutant pp71 protein that localizes to the nucleus but fails to bind (11) or degrade (3) Daxx do not initiate IE gene expression (6).

Controlling the subcellular localization of pp71 represents an attractive step at which the course of an HCMV infection could be manipulated. Delivering pp71 to the nuclei of myeloid cells, for example, by de novo expression, could force the virus to eschew latency, activate IE gene expression, and initiate lytic infection. Accomplishing this in CD34⁺ cells by coinfection with two independent viruses, as we have previously in THP-1 cells (6), is complicated because CD34⁺ cells are inefficiently infected and transduced. A more direct way to achieve de novo pp71 expression in CD34⁺ cells would be to generate a recombinant HCMV in which constitutive pp71 expression was unimpeded. The gene encoding pp71 is located within the unique long (UL) region of the viral genome and is expressed with early-late kinetics (12, 13). Thus, de novo pp71 expression is completely dependent on prior viral IE protein production for its synthesis, which is inhibited in CD34⁺ cells by the Daxx-mediated intrinsic defense.

We reasoned that a potential way to uncouple pp71 expression from intrinsic repression and IE dependence in the undifferentiated myeloid cells where HCMV establishes latency would be to exchange the putative native pp71 promoter with the long terminal repeat (LTR) of the spleen focus-forming virus (SFFV). This promoter has been used in retroviral (14) and lentiviral (15, 16) transduction of primary human CD34⁺ cells, and where it was examined, it was more active than the HCMV major immediate early promoter (MIEP) in primary human CD34⁺ cells (17, 18) and multiple myeloid cell lines (19). Surprisingly, we found that this constitutive myeloid-active promoter is silenced in the context of the HCMV genome upon infection of THP-1 and primary CD34⁺ cells and is still dependent upon prior viral IE protein function after infection of fibroblasts (thus, it is expressed with early kinetics). Similarly, we found that activation of the early promoter from simian vacuolating virus 40 (SV40) located at the beginning of the unique short (US) region of the viral genome is also silenced in THP-1 cells. Furthermore, the SFFV promoter in the middle of the UL segment of the HCMV genome and the SV40 promoter located at either end of the US region all depend upon viral IE protein synthesis for their activation during lytic infection of fibroblasts (they are each expressed with early kinetics). These observations negate the utility of this approach to test the role of pp71 preexpression on the establishment of latency. Nevertheless, they reveal the genome-wide and dominant nature of transcriptional silencing imposed upon otherwise constitutive heterologous viral promoters when incorporated into the HCMV genome, illuminating provocative aspects of viral gene expression and repression during both lytic replication and latency.

MATERIALS AND METHODS

Cells and viruses.

Human foreskin fibroblasts (HFs) were cultured in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen) supplemented with 10% (vol/vol) fetal bovine serum (Gemini), 100 U/ml penicillin, and 100 μg/ml streptomycin plus 0.292 mg/ml glutamine (Gibco) in a 5% CO₂ atmosphere at 37°C. THP-1 cells were cultured as described above in RPMI 1640 medium (Invitrogen) supplemented as described above. CD34⁺ cells (Lonza) were cultured as previously described (7). The wild-type HCMV strains were bacterial artificial chromosome (BAC)-derived AD169-GFP (AD169 labeled with green fluorescent protein [GFP]) (20), FIX-GFP (21), and TB40/E-GFP (22). Virus was UV inactivated on ice in a UV Stratalinker 2400 (Stratagene) by exposure to the 254-nm light source at 0.12 J/cm² for 2 min. Crude viral stocks were pelleted through a cushion (20% [wt/vol] d-sorbitol, 50 mM Tris [pH 7.2], 1 mM MgCl₂), at 20,000 rpm for 1 h at 20°C and then resuspended in DMEM. AD169 and BAD-SF82 virions were further purified in 10 to 50% sucrose gradients by centrifugation at 20,000 rpm for 1 h at 20°C. The native UL82 promoter was replaced with the SFFV LTR (from pSR-CMV-MLV-SIN [CMV stands for cytomegalovirus, and MLV stands for murine leukemia virus] [23]) using the two-step Red recombination mutagenesis protocol of Tischer et al. (24) to produce BAD-SF82. Primers utilized are listed in Table 1. Virus stocks and growth curve time points were titrated by plaque assay.

Table 1.

Oligonucleotide primers or probes used for the construction of plasmids, recombinant BACs, and probe synthesis

Primer or probea	Sequenceb	Purpose
Kanr For	CAACCCTCAGCAGTTTCTTAAGACCCATCAGATGTTTCCAGGCTCCCCCAAGGACCTGAAATGACCGCGGTAGGCGTGTACGGTGG	Kanr cassette
Kanr Rev	AACTGCTGAGGGCCAGTGTTACAACCAATTAACC
SFFV-UL82 For	TGAGCCACCCGCCGCACGCGCTTAGGACGACTCTATAAAAACCCACGTCC	BAD-SF82
	ACTCAGACACGTAACGCCATTTTGCAAGGCATGGA
SFFV-UL82 Rev	GGCGGGATGGGGGGAGGGTCAGGGGATGCGCAAAGGTGAACGGGTCTTC
	GTGGGAGGTCGGCCGAGTGAGGGGTTGTGAG
UL83-UL82 For	CCGAATTGGAAGGCGTATGG	Screening primers for BAD-SF82
UL83-UL82 Rev	GTGCAGGCGACCAAAGCTTC
UL83 probe For	GTCAGAATCTGAAGTACCAGGAAT	UL83 probe
UL83 probe Rev	taatacgactcactatagggTCAACCTCGGTGCTTTTTGGGCG
UL82 probe For	ATGTCTCAGGCATCGTCCTCGCC	UL82 probe
UL82 probe Rev	taatacgactcactatagggCTCACCTCTATGTTGGTACGCAGG
IE1 For	CGTCCTTGACACGATGGA	mRNA for IE1
IE1 Rev	TCTCCTCGAAAGGCTCATGA
GAPDH For	GAGCCAAAAGGGTCATC	mRNA for GAPDH
GAPDH Rev	GTGGTCATGAGTCCTTC

^aFor, forward; Rev, reverse.

^bThe T7 promoter sequence is indicated by lowercase type.

Assays for proteins.

Western blotting was performed as described previously (25). Primary antibodies were specific for Daxx (catalog no. D7810; Sigma), UL44 (catalog no. CA006-100; Virusys), pp71 (IE-233) (26), IE1 (1B12) (27), pp65 (10-C50K; Fitzgerald), or pp28 (CMV157) (28). Secondary horseradish peroxidase (HRP)-linked antibodies were from Millipore (AP124P and AP132P). Indirect immunofluorescence was performed as described previously (6, 29). Primary antibodies were specific for pp71 (CMV355), or IE1 (1B12). Secondary Alexa Fluor 546-linked antibodies were from Life Technologies (catalog no. A11003). Stained cells were mounted in ProLong with DAPI (4′,6-diamidino-2-phenylindole dihydrochloride) (Invitrogen). Images were captured with a Zeiss microscope and camera (Axiovert 200 M).

Assays for RNA.

Total RNA was isolated with RNeasy minikits (Qiagen) and quantitated with a Nanodrop 8000 instrument (Thermo Scientific). For Northern blot analysis, RNA was separated in 1% (wt/vol) agarose–MOPS (morpholinepropanesulfonic acid)–formaldehyde gels and then transferred onto a Nytran nylon membrane (catalog no. 09-301-149; Fisher) by capillary blotting. Single-stranded, ³²P-labeled probes were generated with a MAXIscript T7 kit (Ambion) using primers listed in Table 1. Hybridization conditions included prehybridization with Ultrahyb (catalog no. AM8670M; Ambion) for 2 h at 60°C, overnight hybridization with probes at 60°C, followed by subsequent washings (15 min, 60°C) with 2× SSC(1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.2% SDS, then 0.2× SCC–0.2% SDS. Dried membranes were visualized and quantitated with a Molecular Dynamics Typhoon 9410 variable-mode imager. For reverse transcription-PCR (RT-PCR), RNA was treated with RNase-free DNase (Promega), converted to cDNA, and amplified with the SuperScript one-step system with Platinum Taq (Life Technologies) using the primer pairs listed in Table 1. PCR products were separated in agarose gels and visualized by ethidium bromide staining. The drug treatments were actinomycin D (1 μg/ml), cycloheximide (200 μg/ml for HFs, 100 μg/ml for THP-1 cells), or valproic acid (VPA) (1 mM).

RESULTS

A recombinant HCMV in which the UL82 promoter is replaced with the SFFV LTR (BAD-SF82) is viable.

The gene encoding pp71 (UL82) is transcribed (Fig. 1A) both as the second open reading frame of a bicistronic 4.0-kb mRNA, as well as a 1.9-kb monocistronic mRNA that initiates within the short intergenic region between UL83 (pp65) and UL82 (12). Presumably, a promoter driving the expression of this 1.9-kb, pp71-encoding transcript exists within this intergenic region. To generate a recombinant HCMV in which UL82 expression is independent of its putative native promoter but UL83 expression is unaltered, we simultaneously deleted the sequences distal to the putative polyadenylation signal sequence for UL83 and proximal to the start codon of UL82 (this deleted region contains a putative TATA box for the initiation of the 1.9-kb transcript) and replaced them with the SFFV promoter using BAC mutagenesis (30) to create BAD-SF82. Restriction enzyme digestion (HindIII) demonstrated that BAD-SF82 displayed an identical banding pattern to that of AD169 except for the expected conversion of a 821-bp fragment to an 1,133-bp fragment resulting from the deletion of the 63-bp putative native promoter and insertion of the 375-bp SFFV promoter (Fig. 1B). This genetic replacement was confirmed by PCR amplification (Fig. 1C) and sequencing (Fig. 1D).

Fig 1.

BAD-SF82. (A) Schematic representation of the UL83-UL82 locus. Nucleotide positions are labeled relative to the start site (+1) of the UL83 open reading frame coding for pp65. The positions of HindIII sites are indicated. Bicistronic (encoding UL83/pp65 and UL82/pp71) and monocistronic (encoding pp71) mRNAs are produced from this locus. In the recombinant BAD-SF82, a putative TATA box-containing promoter between the UL83 poly(A) addition sequence and the UL82 start codon was replaced by the SFFV promoter. (B) BAC DNA from wild-type AD169 (AD) or BAD-SF82 (SF) was digested with HindIII and separated by agarose gel electrophoresis. The positions of markers (in base pairs) are shown to the left of the gels. An expansion of the bottom of the gel is shown to accentuate the expected different banding patterns. (C) DNA isolated from cells infected with AD169 or BAD-SF82 was subjected to PCR with primers specific to the 3′ end of UL83 and the 5′ end of UL82 to amplify the intergenic region (see the floating arrows in panel A). The expected size (in base pairs) of the amplified products is shown to the right of the gel. (D) The sequences of the PCR products generated in panel B were determined. The labeled region indicates the junctions between AD169 and SFFV sequences. The numbers are nucleotide positions as shown in panel A.

Virus (BAD-SF82) was recovered upon transfection of the recombinant BAC into fully permissive fibroblasts. No differences from wild-type virus were detected during high (Fig. 2A) or low (Fig. 2B) multiplicity of infection (MOI) growth curve analysis. Likewise, the kinetics of protein accumulation for immediate early (IE1), early (UL44), and late (pp28) gene products were not remarkably different between AD169 and BAD-SF82 at either high (Fig. 2C) or low (Fig. 2D) MOIs. The only notable difference was the substantial increase in pp71 protein steady-state levels achieved upon BAD-SF82 infection (as lysates were compared on the same gel, pp71 levels during AD169 infection appear underexposed). This correlated with increased incorporation of pp71 into BAD-SF82 particles compared to AD169 (Fig. 3A). Significant differences in virion incorporation of pp65, UL44, or pp28, other tegument proteins (31), were not observed (Fig. 3A).

Fig 2.

BAD-SF82 replicates with wild-type kinetics. (A) HFs were infected with the indicated virus at an MOI of 1. At the indicated day postinfection (d.p.i.), virus was prepared from supernatants and attached cells and titrated by plaque assays. Error bars show standard deviations. (B) A growth curve after infection at an MOI of 0.1 was performed as described above for panel A. (C) HFs were infected with the indicated virus at an MOI of 1. At the indicated hour postinfection (hpi), lysates were prepared and analyzed by Western blotting with the indicated antibodies. M, mock. (D) Protein expression was monitored after infection at an MOI of 0.1 as described above for panel C.

Fig 3.

SFFV-driven UL82 leads to increased incorporation of pp71 into virions. (A) One hundred thousand PFU of the indicated virus was lysed and analyzed by Western blotting with the indicated antibodies. AD, AD169; SF, BAD-SF82. (B) Total RNA extracted at the indicated hour postinfection (hpi) from HFs infected with the indicated viruses at an MOI of 1 was analyzed by Northern blotting with the indicated probes. Numbers under the UL82 blot indicate quantitation of the 1.9-kb band relative to the mock (M) control. Ethidium bromide-stained 28S rRNA served as a loading control.

The increased accumulation of pp71 protein within infected cells and virus particles likely results from increased UL82 transcription driven by the SFFV promoter. Northern blot analysis indicated that the 1.9-kb monocistronic transcript encoding UL82 was dramatically increased in cells infected with BAD-SF82 compared to wild-type virus, while there was qualitatively little change in the bicistronic transcript encoding both UL83 and UL82 (Fig. 3B). From these data, we conclude that the incorporated SFFV promoter within BAD-SF82 drives the expression of a UL82 transcript and results in higher-level accumulation of pp71 protein in comparison to that of the native locus but does not dramatically alter the lytic growth properties of the virus in fibroblast cells.

The SFFV promoter within BAD-SF82 is silenced upon infection of myeloid cells.

As expected (6, 7), pp71 incorporated into the tegument layer of BAD-SF82 was capable of entering the nucleus upon infection of fibroblasts (Fig. 4A) but remained in the cytoplasm upon infection of THP-1 (Fig. 4B) and CD34⁺ (Fig. 4C) cells. Thus, despite the increased concentration within the virions of BAD-SF82, tegument-delivered pp71 was still apparently subject to the normal control measures for subcellular localization upon infection of completely or incompletely differentiated cells. Likewise, viral genomes delivered by BAD-SF82 apparently enter the nucleus and are transcriptionally competent, as treatment of THP-1 cells with the histone deacetylase inhibitor VPA permitted IE1 RNA (Fig. 5A) and protein (Fig. 5B) accumulation. However, the HCMV major immediate early promoter (MIEP) that drives IE1 production and the SFFV promoter that drives pp71 were inactive in the absence of VPA (Fig. 5A). Thus, the constitutively myeloid-active SFFV promoter located within the UL region of the HCMV genome was silenced upon infection of THP-1 cells.

Fig 4.

BAD-SF82 tegument-delivered pp71 localizes properly. (A) HFs were infected with the indicated live or UV-inactivated virus at an MOI of 1 for 5 h, after which time pp71 (red) was visualized by indirect immunofluorescence microscopy. Nuclei were counterstained with DAPI (blue). (B) Infected (MOI of 3) THP-1 cells were analyzed as described above for panel A at 6 hpi. (C) Infected (MOI of 1) CD34⁺ cells were analyzed as described above for panel A at 6 h. Nuclei were counterstained with Hoechst dye (blue).

Fig 5.

THP-1 cell activation of heterologous viral promoters inserted into recombinant HCMV genomes requires VPA treatment and viral IE protein synthesis. (A) Total RNA prepared from THP-1 cells infected (+) for 6 h with the indicated viruses at an MOI of 1 was amplified by PCR with primers specific for the indicated genes. Products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) serves as a control. (B) THP-1 cells were infected with the indicated virus at an MOI of 1 for 36 h, after which time IE1 (red) was visualized by indirect immunofluorescence microscopy. Nuclei were counterstained with DAPI (blue). Where indicated, VPA was added 3 h prior to infection. (C) Expression of UL82 (pp71), IE1, and GFP RNA was analyzed as described above for panel A. Where indicated, VPA (V) was added 3 h prior to infection, and cycloheximide (C) was added at the time of infection. (D) Total RNA prepared from THP-1 cells infected for 36 h with the indicated viruses at an MOI of 3 was analyzed by Northern blotting with UL82-specific primers. Ethidium bromide-stained 28S rRNA served as a loading control. Where indicated, VPA was added 3 h prior to infection. (E) pp71 was analyzed in THP-1 cells by indirect immunofluorescence as described above for panel B. (F) pp71 was analyzed in THP-1 cells infected with wild-type AD169 at an MOI of 2 for 6 h by indirect immunofluorescence. Nuclei were counterstained with DAPI (blue). Where indicated, VPA was added 3 h prior to infection.

While the MIEP and the SFFV promoter are activated without the need for trans-acting viral transcription factors when present in heterologous expression systems (15, 32), only the MIEP showed transcriptional independence in THP-1 cells treated with VPA. The protein synthesis inhibitor cycloheximide blocked SFFV-driven pp71 mRNA accumulation in THP-1 cells treated with VPA, but not the accumulation of the MIEP-driven IE1 RNA (Fig. 5C). The decrease in IE1 transcripts in the presence of cycloheximide is expected in THP-1 cells where the MIEP is ∼10-fold weaker than it is in fibroblasts (7) and because the synthesis of the IE proteins, which would normally transactivate the MIEP, is inhibited by the drug. Northern blot analysis confirmed that the pp71-encoding transcripts detected by PCR in the presence of VPA and the absence of cycloheximide were of a length (1.9 kb) consistent with transcriptional initiation from the SFFV promoter (Fig. 5D). When UL82-encoding RNA was expressed (in the presence of VPA), it produced de novo pp71 protein that was capable of entering the nucleus (Fig. 5E), even though tegument-delivered pp71 (the only type of pp71 present in the absence of VPA) remained in the cytoplasm (Fig. 5E) at 36 h postinfection. Prior to de novo pp71 expression in the presence of VPA, the tegument-delivered protein still localizes in the cytoplasm of drug-treated THP-1 cells (Fig. 5F), indicating that HDAC inhibition does not induce cellular differentiation, which would permit the rapid trafficking of tegument-delivered pp71 to the nucleus (6).

Importantly, the SFFV promoter in BAD-SF82 is also silenced upon infection of CD34⁺ cells. In the absence of VPA, pp71 transcripts fail to accumulate (Fig. 6A) and pp71 protein remains in the cytoplasm (Fig. 6B), implying that it is tegument delivered and not de novo expressed (which would localize to the nucleus) (7). We conclude that the constitutively myeloid-active SFFV promoter, when incorporated into the UL83-UL82 intergenic locus of HCMV, is silenced upon infection of undifferentiated myeloid cells where HCMV establishes experimental latency and requires both VPA treatment and prior HCMV IE protein production to be activated.

Fig 6.

Activation of SFFV and SV40 promoters in recombinant HCMV genomes in primary CD34⁺ cells requires VPA treatment. (A) Total RNA prepared from primary CD34⁺ cells infected for 24 h with the indicated viruses at an MOI of 1 was amplified by PCR with primers specific for the indicated genes. Products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. Where indicated, VPA was added at the time of infection. GAPDH serves as a control. (B) CD34⁺ cells were infected with the indicated virus at an MOI of 1 for 36 h, after which time pp71 (red) was visualized by indirect immunofluorescence microscopy. Nuclei were counterstained with Hoechst dye (blue).

A second heterologous constitutively active viral promoter located at a different position within the HCMV genome is also suppressed during the establishment of latency.

While the SFFV promoter located in the middle of the UL region of the HCMV genome is clearly inactive when latency is established (Fig. 5 and 6), it is unclear whether another viral promoter in another location would also be subject to such silencing. Conveniently, both BAD-SF82 and the wild-type AD169 clone from which it was derived, encode a GFP transgene replacing the viral US2 and US3 genes and under the transcriptional control of the SV40 promoter. Interestingly, we found that GFP transcription in THP-1 cells required both histone deacetylase (HDAC) inhibition and prior viral IE protein expression (Fig. 5C). Likewise, GFP transcription was not activated in CD34⁺ cells in the absence of VPA (Fig. 6A). Thus, a different heterologous constitutively active viral promoter located in a separate region of the viral genome is also silenced upon HCMV infection of undifferentiated myeloid cells and the establishment of experimental latency.

Expression from heterologous viral promoters within recombinant HCMV genomes in fibroblasts also requires viral IE proteins.

Cycloheximide blocked the ability of VPA to indirectly activate the SFFV promoter in THP-1 cells, but not the ability of VPA to directly activate the HCMV MIEP (Fig. 5C). This indicates that, with the caveat of requiring VPA to inactivate the intrinsic immune defense because tegument-delivered pp71 remains cytoplasmic and fails to degrade Daxx, the MIE locus is indeed expressed with IE kinetics in THP-1 cells, but the SFFV and SV40 promoters are expressed with early kinetics (viral IE protein function [presumably IE1 or IE2] is required for their activation). Likewise, the SFFV promoter in BAD-SF82 was also dependent upon prior viral RNA and protein expression for activation during lytic infection of fibroblasts (Fig. 7A). UL82-encoding RNA was blocked from accumulating in fibroblasts infected with BAD-SF82 when treated with cycloheximide (Fig. 7A), but not when treated with the DNA replication inhibitor phosphonoacetic acid (PAA) (33) (Fig. 7B). From this, we conclude that, despite the heterologous nature of the incorporated SFFV promoter, it was expressed with early kinetics in all cell types tested when inserted prior to the UL82 gene, a locus normally expressed with early kinetics.

Fig 7.

Heterologous viral promoters within HCMV genomes are expressed with early kinetics during lytic infection of fibroblasts. (A) Total RNA isolated from HFs pretreated for 2 h with dimethyl sulfoxide (DMSO) (D), actinomycin D (A), or cycloheximide (C) and then infected with the indicated virus at an MOI of 1 for 5 h was amplified by PCR with primers specific for the indicated genes. Products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. (B) Total RNA prepared from HFs infected for 36 h with the indicated viruses at an MOI of 3 was analyzed by Northern blotting with UL82- or UL83-specific probes. Ethidium bromide-stained 28S rRNA served as a loading control. Where indicated, phosphonoacetic acid (PAA) was added 3 h prior to infection. (C) Total RNA isolated from HFs pretreated for 2 h with DMSO or cycloheximide and then infected with the indicated virus at an MOI of 0.5 for 5 h was amplified by PCR with primers specific for the indicated genes. Products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining.

In congruence with its early expression in VPA-treated THP-1 cells (Fig. 5C), the SV40 promoter driving GFP from the US2-US3 locus also depends upon viral protein production for its activation (Fig. 7C). Though we have not specifically tested its kinetic class with drug experiments, we predict that the GFP gene in this context is expressed with early kinetics. The dependence upon de novo viral proteins for activation extended to SV40 promoter-driven GFP transgenes recombined into either end of the US region in two clinical strains of HCMV, TB40/E (where it is inserted between US34 and TRS-1) and FIX (where it replaces US3 to US5). Much like the UL82 locus, most genes in these loci are expressed with early kinetics, although US3 and TRS-1 are IE genes (34). Thus, two constitutively active heterologous viral proteins become subject to HCMV-specific transcriptional control when inserted at three disparate locations within the genomes of three different viral strains.

DISCUSSION

The expression of herpesvirus genes that encode proteins that drive lytic replication is silenced during viral latency (35). Driving the production of such proteins in cells where they would normally not be expressed and in which latency would be established seems a logical way to force the initiation (if not completion) of the lytic cycle in these cells. This has recently been achieved for the gammaherpesvirus Kaposi's sarcoma-associated herpesvirus (KSHV) by replacing the endogenous ORF50/replication and transcription activator (RTA) promoter, which is activated with IE kinetics, with the cellular phosphoglycerate kinase (PGK) promoter in a recombinant virus (36). We are not aware of similar experiments with alphaherpesviruses. Our attempts to replace the HCMV MIEP with a heterologous promoter has generated only nonviable recombinants (Q. Qin and R. F. Kalejta, unpublished observations), unlike the KSHV study discussed above, which yielded viable virus.

We thus devised an alternative strategy designed to activate IE gene expression in the undifferentiated cells where HCMV establishes latency without the need to replace the MIEP. By replacing the putative natural promoter driving expression of the tegument-delivered transactivator of IE gene expression, the pp71 protein (encoded by the early/late gene UL82), with one we suspected would be constitutively expressed in undifferentiated myeloid cells, we anticipated pp71 would be expressed de novo from infecting recombinant viral genomes, localize to the nucleus, inactivate the Daxx-mediated intrinsic cellular defense, and thus activate viral IE gene expression. This strategy is essentially the inverse of one recently used successfully to block the animation of latent herpes simplex virus type 1 (HSV-1) genomes by replacing the native promoter of its tegument transactivator, VP16 (35, 37). Our experiments unexpectedly revealed silencing of the SFFV promoter driving pp71 expression in THP-1, CD34⁺, and HF cells, where it was expressed with early kinetics. The SV40 promoter, located at two different locations and in three different viral strains, was similarly regulated. As the HDAC inhibitor VPA could disrupt this normal regulation, heterologous viral promoters inserted into HCMV genomes seem to be under a similar chromatin-based control as are native HCMV promoters (38).

The regulation of exogenous promoters when incorporated into herpesviral genomes has been studied previously, but few if any universal truths seem to apply. The rabbit β-globin promoter demonstrated early kinetics within the HSV-1 genome (39, 40), whereas the PGK promoter appears to be expressed with IE kinetics from the KSHV genome (36). Furthermore, the rat neuron-specific enolase (NSE) promoter lost cell type specificity when incorporated into the HSV-1 genome (41). Thus, parameters for the expression of cellular promoters from the context of herpesviral genomes seem difficult to predict.

Here we have analyzed the expression pattern not of a cellular promoter but of heterologous viral promoters within the context of a herpesvirus genome. Such a scenario is more common than may be initially expected, as many recombinant herpesviruses express a GFP transgene from the SV40 promoter to permit the facile monitoring of viral infections with fluorescence microscopy (42). In fact, GFP fluorescence from such a transgene (located adjacent to either the UL21 or US34 loci) has been utilized for flow cytometry-based sorting of CD34⁺ cells presumably latently infected with recombinant HCMVs (43, 44). Our data indicating that SV40-driven GFP expression requires new viral protein production appears to indicate that cells selected in such manner are either abortively lytically infected or are detected as GFP positive for a reason other than de novo GFP expression. HCMV virions nonspecifically incorporate proteins and RNAs abundantly present in cells during virion assembly (45), and these virion mRNAs, when introduced into newly infected cells, have the capacity to be translated. Therefore, the fluorescence basis for sorting in these experiments, presumably generated by GFP, may originate from tegument-delivered protein or protein translated from tegument-delivered mRNAs. In this scenario, such selection might still be a viable means to enrich for experimentally latently infected cells. Were tegument-delivered mRNA to be the source of the GFP fluorescence in these experiments, it is unclear why a similar mode of expression was not observed here for SFFV-driven pp71.

Thus, questions remain regarding how heterologous promoter selection and placement within recombinant HCMV genomes affect true de novo expression of the genes they are designed to activate, especially in undifferentiated myeloid cells. Answers may lead to the improvement of already promising CMV-based vaccine vectors (46). Provocatively, herpesviral promoters seem, in general, to maintain their kinetic properties when moved to or reproduced at a different location within the extant genome (47, 48), perhaps making this type of expression system the most predictable and preferable.