Phenotypic Suppression of a Herpes Simplex Virus 1 ICP27 Mutation by Enhanced Transcription of the Mutant Gene

Abstract

ICP27 is an essential herpes simplex virus 1 (HSV-1) regulatory protein that enhances viral gene expression. Although it is predominantly nuclear, it shuttles to the cytoplasm during infection using an N-terminal nuclear export signal (NES). We previously engineered an NES-negative ICP27 mutant, dLeu, that replicates poorly in cultured cells. In this study, we isolated dLeuR, a growth-competent revertant of dLeu. We show that dLeuR possesses one or more extragenic mutations that enhance ICP27 transcription, leading to overexpression of the mutant protein and restoration of viral growth. This work provides evidence of a novel pathway regulating transcription of the ICP27 gene.

TEXT

Herpes simplex virus 1 (HSV-1) encodes five immediate-early (IE) proteins. These polypeptides are synthesized at the onset of productive viral infection and regulate expression of viral delayed-early (DE) and late (L) genes. IE expression is induced within 1 to 2 h of infection by VP16, a virion protein that interacts with TAATGARAT motifs in IE promoters (25). The IE protein ICP27 is essential, being required for efficient expression of several DE/L genes (13, 18, 19, 23). It transactivates at least some of these genes by serving as an mRNA export factor for intronless DE/L transcripts (11, 20, 22). Consistent with this, ICP27 binds RNA (15, 20) and shuttles between the nucleus and cytoplasm (14, 17, 20, 22) using an N-terminal, leucine-rich nuclear export signal (NES) (20). We previously engineered an HSV-1 ICP27 mutant, dLeu, that encodes an ICP27 lacking the NES (deletion of residues 6 to 19) (12). This mutant exhibits an ∼100-fold replication defect compared to wild-type (WT) HSV-1, demonstrating the importance of the NES for ICP27 function.

In this study, we set out to isolate growth-competent revertants of dLeu. We first picked eight dLeu plaques on ICP27-complementing V27 cells (18). These viruses were amplified once in V27 cells and then serially passaged in Vero cells. After the second passage, only one culture showed a significant cytopathic effect (CPE), suggesting that it might harbor a revertant. A virus was isolated from this sample, plaque purified, and designated dLeuR. To test whether dLeuR is a true revertant, we performed yield assays at high and low multiplicities of infection (MOIs). dLeuR grew ∼10-fold better than dLeu at a high MOI (10 PFU/cell), although its growth was not restored to WT levels (Fig. 1A). At the lower MOI (0.01 PFU/cell), dLeuR's enhanced growth was more pronounced, showing a nearly 100-fold increase over dLeu. Moreover, in plaque assays, dLeuR plaques were significantly larger than dLeu plaques, although they were not as large as WT plaques (Fig. 1B). We conclude that dLeuR is a growth revertant of dLeu.

Fig. 1.

Phenotypic analysis of dLeuR. (A) Growth analysis. Vero cells were infected in duplicate at an MOI of 10 or 0.01 PFU/cell with WT HSV-1 (strain KOS1.1), dLeu, or dLeuR. Infections were terminated at 22 hpi, and titers were determined by plaque assay of the cell lysates on V27 cells. (B) Plaque morphology. Serial dilutions of the virus stocks shown were plated on Vero cells, and the cultures were incubated in medium containing 1% (vol/vol) human serum (MP Biomedicals) to neutralize cell-free virus. Plaques were allowed to develop for 3 days, at which time the monolayers were fixed and stained with Giemsa reagent (Sigma). Each image contains approximately equal numbers of plaques. (C) Viral protein expression. Vero cells were infected at an MOI of 10, and total cell proteins were harvested at 10 hpi. Equivalent fractions of the samples were analyzed by SDS-PAGE and immunoblotting. ICP27, ICP0, gD, gC, and ICP8 were detected using monoclonal antibodies (Mabs) H1113, H1112, H1103, H1104, and H1115, respectively (Rumbaugh-Goodwin Institute for Cancer Research). VP16 was detected using rabbit polyclonal antibody SW8 (24). VP5 (ICP5) was detected using MAb 3B6 (Abcam). VP22 was detected using rabbit polyclonal antibody AGV031 (5), and US11 was detected using a rabbit polyclonal antibody (4). (D) ICP27 localization. Vero cells were mock infected or infected at an MOI of 10. At 6 hpi, the cells were fixed and processed for ICP27 immunofluorescence using MAb H1113.

Previous work has shown that dLeu is deficient in DE/L gene expression (10, 12). To examine gene expression in dLeuR-infected cells, Vero cells were infected with the ICP27 null mutant d27-1 (18), dLeu, dLeuR, or WT. Total proteins were harvested at 10 hours postinfection (hpi) and analyzed for IE (ICP27 and ICP0), DE (ICP8), and L (VP5, VP16, VP22, gC, and US11) protein expression by immunoblotting (Fig. 1C). Several results were noteworthy. First, both dLeu and dLeuR expressed an ICP27 that migrates faster on SDS-polyacrylamide gels than WT ICP27, consistent with the expected NES deletion. Second, dLeuR ICP27 was expressed at higher levels than either WT or dLeu, a phenomenon that is analyzed extensively below. Third, although some viral proteins accumulated to similar levels in the dLeu- and dLeuR-infected cells, several L proteins (VP5, gC, and US11) accumulated more efficiently in dLeuR-infected cells. This indicates that dLeuR expresses certain L proteins more efficiently than dLeu does. We also analyzed the intracellular localization of dLeuR ICP27. As previously reported (3, 12), WT ICP27 localizes to both the nucleus and cytoplasm at an intermediate stage of infection, while dLeu ICP27 is restricted to the nucleus (Fig. 1D). Interestingly, cytoplasmic ICP27 is restored in the dLeuR-infected cells to a level that is higher than in the cells infected with WT virus.

The reversion mutation (or mutations) in dLeuR could be intragenic, i.e., within the ICP27 gene itself, or extragenic. To investigate this, we cloned and sequenced the dLeuR ICP27 gene, including its known promoter (272 bp upstream of the transcriptional initiation site). As expected, dLeuR retains the NES deletion. The only change found relative to dLeu was at codon 194, which is CGT in dLeuR but CGC in dLeu and WT. However, this mutation is silent, as both codons specify arginine. Moreover, plasmid-based viral complementation experiments (12) showed that the dLeu and dLeuR ICP27 genes are identical in their ability to complement the growth of d27-1 (not shown). Together, these results suggest that one or more extragenic mutations are responsible for the dLeuR reversion.

To more closely examine ICP27 overexpression by dLeuR, we performed a time course immunoblotting analysis (Fig. 2A). At 3 hpi, both dLeu and dLeuR ICP27 were reduced relative to WT. However, at 6 and 9 hpi, ICP27 levels were higher in the dLeuR-infected cells than in the cells infected by other viruses, while ICP4 and ICP0 levels were similar. Thus, ICP27 is expressed at enhanced levels in cells infected with dLeuR, and this phenomenon does not extend to other IE proteins. We used a cycloheximide (CH) reversal protocol (9) to ask whether dLeuR ICP27 overexpression requires DE/L protein expression. Cells were infected for several hours in the presence of CH, which results in a block to viral protein synthesis and the accumulation of IE transcripts. The CH-containing medium was then replaced with medium containing actinomycin D (ActD). This allows translation of accumulated IE mRNAs but prevents expression of DE/L genes (6). By using these conditions, we observed an increase in the amount of ICP27 expressed by dLeuR relative to dLeu and, to a lesser extent, WT (Fig. 2B). ICP0 was not similarly overexpressed. Thus, ICP27 overexpression by dLeuR does not require DE/L gene expression.

Fig. 2.

dLeuR exhibits enhanced expression of ICP27. (A) Protein expression time course. Vero cells were mock infected or infected at an MOI of 10 PFU/cell as shown. Total proteins were harvested at the times indicated and immunoblotted as described in the legend to Fig. 1 for viral proteins ICP27 and ICP0. ICP4 was detected with MAb H1114 (Rumbaugh-Goodwin Institute for Cancer Research), while cellular protein EEA1 (early endosome antigen 1) was detected with a commercial MAb (BD Transduction Labs). (B) Protein expression under cycloheximide (CH) reversal conditions. Vero cells were infected as described above for panel A, except that infection was carried out in the presence of 50 μg/ml CH (Sigma). At 6 hpi, the medium containing CH was removed and replaced with medium containing 10 μg/ml actinomycin D (ActD) (Sigma). Proteins were harvested 1 h later and analyzed by immunoblotting.

We next investigated ICP27 mRNA expression. At 6 hpi, dLeuR expressed approximately 2-fold-more ICP27 mRNA than dLeu did, although both viruses expressed more mRNA than the WT did (Fig. 3A, left). In contrast, ICP4 mRNA levels were similar. The increase was also seen in the presence of phosphonoacetate (PAA), an HSV-1 DNA replication inhibitor (Fig. 3A), indicating that overexpression is independent of viral DNA replication. To examine whether mRNA overexpression occurs even in the absence of DE/L expression, we again used a CH reversal. Under post-CH reversal conditions (Fig. 3B), dLeuR expressed 2.8-fold-more ICP27 mRNA than either dLeu or WT did, but ICP0 and ICP4 transcript levels were similar. Interestingly, dLeuR ICP27 mRNA was elevated even when virus infections were carried out in the continuous presence of CH (Fig. 3B, prereversal). This indicates that dLeuR exhibits enhanced expression of ICP27 mRNA even when all viral gene expression is blocked.

Fig. 3.

dLeuR exhibits elevated ICP27 gene transcription and mRNA levels. (A) ICP27 mRNA analysis. Duplicate cultures of Vero cells were mock infected or infected at an MOI of 10 PFU/cell as shown. To one set of infected cells, 400 μg/ml phosphonoacetate (PAA) was added at 1 hpi to inhibit viral DNA synthesis. RNA was harvested from all infected cells at 6 hpi and analyzed by Northern blotting using probes specific for the ICP27 or ICP4 transcripts. The numbers shown are the phosphorimager quantitations relative to the WT signal for that condition. (B) ICP27 mRNA expression under CH reversal conditions. Duplicate cultures of Vero cells were mock infected or infected at an MOI of 10 with the indicated virus in the presence of 50 μg/ml CH. At 6 hpi, RNA was harvested from one set of samples (“pre-reversal”), while in the other set of samples, CH was removed and replaced with medium containing 10 μg/ml ActD. One hour later, RNA was harvested from these samples. RNAs were analyzed by Northern analysis for ICP27, ICP0, and ICP4 mRNA expression. (C) dLeuR fails to enhance expression of an ICP27 gene in trans. Vero cells were mock infected or infected with the indicated viruses at an MOI of 10 for single infections or an MOI of 20 for coinfections (at an MOI of 10 for each virus). All infections were carried out in the presence of 50 μg/ml CH. RNA was harvested from the infected cells at 6 hpi and analyzed by Northern blotting for the expression of ICP27 or ICP27/lacZ mRNAs. Values indicate the phosphorimager quantitation of the data. ICP27 signals were normalized to that of WT for the single infections or to that of WT plus d27lacZ for the coinfections. The LacZ signals were normalized to the signal seen in the d27lacZ single infection. (D and E) Enhanced transcription of the dLeuR ICP27 gene. (D) Nuclear run-on transcription hybridization data. Vero cells were mock infected or infected as shown in the presence of 50 μg/ml CH. At 6 hpi, nuclei were harvested from infected cells. The nuclei were later processed for nuclear run-on transcription using [³²P]UTP. The run-on RNA was hybridized to filters containing complementary (c) or anticomplementary (ac) single-stranded DNA probes specific for the ICP0, ICP4, and ICP27 mRNAs. The ICP4 and ICP27 probes have been described previously (8). The ICP0 probe was constructed by cloning a 1,653-bp SmaI fragment from the ICP0 gene (corresponding to nucleotide [nt] 123696 to nt 125385 of the HSV-1 genome) into the SmaI site of M13mp19. Autoradiographs of the filters are shown. (E) Compilation of nuclear run-on transcription data from three independent experiments. The experiment shown in panel D was repeated twice more. For all three experiments, the level of gene hybridization, as determined by phosphorimaging analysis, was normalized to the cells infected with the WT virus, which was set at 1. The mean fold transcription change, relative to the WT, is shown as well as the standard deviation. R, dLeuR; D, dLeu.

To see whether dLeuR can enhance ICP27 gene expression in trans, Vero cells were coinfected with dLeuR and d27lacZ, a mutant that has the Escherichia coli lacZ gene inserted into the ICP27 coding region (18). For controls, we performed dLeu/d27lacZ and WT/d27lacZ coinfections, as well as single infections. CH was included in all infections to block gene expression. Analysis of 6 hpi RNA (Fig. 3C) indicated that, as expected, ICP27 mRNA levels were elevated in dLeuR-infected cells (∼1.6-fold increase over WT). Similarly, in cells coinfected with dLeuR/d27lacZ, the level of ICP27 mRNA was elevated compared to the level seen for cells coinfected with WT/d27lacZ (4.1-fold increase). Hybridization with the LacZ probe, however, indicated that dLeuR did not stimulate accumulation of the ICP27-LacZ transcript, suggesting that dLeuR does not act in trans to stimulate expression of another ICP27 allele.

To test whether the increase in dLeuR ICP27 mRNA is associated with increased transcription, we performed nuclear run-on transcription analysis (7, 16). Vero cells were mock infected or infected in the presence of CH. Nuclei were isolated at 5 hpi and subjected to nuclear run-on conditions, wherein preinitiated RNA polymerase II molecules elongate nascent mRNA in the presence of [³²P]UTP. The radiolabeled RNAs were hybridized to single-stranded DNA probes specific for ICP0, ICP4, or ICP27 genes (Fig. 3D and E). Notably, transcription of the ICP27 gene was highly elevated in the dLeuR-infected cells compared to the WT virus- and dLeu-infected cells, exhibiting a 12.5-fold increase over cells infected with WT virus. A slight (∼2-fold) increase was seen for transcription of ICP0 and ICP4. Thus, transcription of the ICP7 gene is specifically elevated in dLeuR-infected cells. This likely accounts, at least in part, for the observed increase in ICP27 mRNA and polypeptide.

Work in bacteria and Saccharomyces cerevisiae has shown that overexpression of a partially functional gene product can suppress a mutant phenotype (2). To test whether a similar mechanism could account for the growth competence of dLeuR, we engineered Vero cell lines that are stably transfected with the dLeu ICP27 gene, reasoning that such lines might complement dLeu by providing an additional source of dLeu ICP27. Thus, plasmids containing the dLeu ICP27 gene were cotransfected into Vero cells with a hygromycin resistance marker. After selection, three clonal lines (VdL160, VdL312, and VdL321) were positive for dLeu ICP27 expression after d27-1 infection (Fig. 4A). These lines expressed differing amounts of ICP27, with VdL321 and VdL312 expressing high levels and VdL160 expressing low levels. Interestingly, both the VdL321- and VdL312 d27-1-infected cells displayed significant amounts of cytoplasmic ICP27 (Fig. 4B), similar to what is observed in dLeuR-infected Vero cells (Fig. 1D). We next asked whether dLeu replicates more efficiently in cells that can express dLeu ICP27. The various cell lines were infected at a low MOI with dLeu for 3 days. Yields were significantly enhanced in VdL312 and VdL321 cells (Fig. 4C), demonstrating that dLeu replication is enhanced by increasing dLeu ICP27 expression. This effect was also seen in Vero cell plaque assays, where dLeu plaques were significantly larger in VdL312 and VdL321 cells (Fig. 4D).

Fig. 4.

Complementation of dLeu growth in cell lines stably transfected with the dLeu ICP27 gene. To create lines, Vero cells were stably transfected with a cloned dLeu ICP27 plasmid and a linear hygromycin resistance marker (BD Biosciences). A control hygromycin-resistant, ICP27-negative line (HygR) was also isolated. (A) Immunoblot analysis of dLeu ICP27 expression in stably transfected Vero cell clones. Infections with WT HSV-1 or d27-1 (10 PFU/cell) were carried out as shown. Total cell proteins were harvested at 6 hpi and analyzed by immunoblotting for ICP27 expression. EEA1 was analyzed as a loading control. (B) Localization of dLeu ICP27 in d27-1-infected cell lines. The cell lines shown were infected with d27-1 at an MOI of 10, and at 6 hpi, the cells were fixed, permeabilized, and stained for ICP27. (C) Enhancement of dLeu growth by increased expression of dLeu ICP27. Approximately 2 × 10⁶ control cells or dLeu ICP27-expressing cell lines were infected with 1,000 PFU of dLeu. The infected monolayers were incubated for 3 days. Viral yields were determined by plaque assay of the cell lysates on V27 cells. (D) Enhancement of plaque size by high expression of dLeu ICP27. The indicated cell lines were infected with serial dilutions of dLeu, and a viral plaque assay was performed for 3 days as described in the legend to Fig. 1. Giemsa-stained monolayers were photographed, and the plaque area was determined by ImageJ software (1) analysis of the digital image. Plaque sizes are shown relative to that of dLeu on control HygR cells.

In summary, our results indicate that the growth competence of dLeuR stems from one or more as yet unidentified extragenic mutations that increase the transcription of the ICP27 gene. This leads to overexpression of the NES-minus ICP27, allowing the virus to replicate significantly better than its parent. As mentioned above, it is well-known that suppression of a genetic defect by overexpression of a mutant but partially functional gene product occurs in yeast and bacterial systems (2), but to our knowledge, this is the first example of this genetic suppression mechanism in an animal virus.

Prior studies indicate that the N-terminal NES is critical for ICP27 function (10, 12), presumably for its mRNA export activity. It is thus surprising that this sequence is dispensable when ICP27 is overexpressed. One possible explanation is that ICP27 has a secondary route of nuclear export and that overexpression allows the NES-negative protein to better access this pathway. Alternately, it is feasible that overexpression enhances a nuclear activity of ICP27, such as the enhancement of viral mRNA accumulation (21).

Perhaps the most intriguing aspect of our study relates to HSV-1 transcription. Our results indicate that transcription of the ICP27 gene can be stimulated independently of other IE genes at a very early stage of infection, prior to the production of IE proteins. How might the reversion mutation(s) do this? As the exact sequence and arrangement of TAATGARAT motifs vary among IE promoters, it is conceivable that an alteration in VP16 could allow for the preferential induction of ICP27. However, this and other models involving altered virion components are not readily consistent with our finding that the reversion mutation does not appear to act in trans. For this reason, we favor a model in which dLeuR has a mutation in a cis-acting DNA sequence that modulates ICP27 transcription. As our sequencing indicates that the reversion maps outside the known promoter, our data suggest the existence of a DNA element that can regulate ICP27 gene transcription from a distance.