ICP0 and the U_S3 protein kinase of herpes simplex virus 1 independently block histone deacetylation to enable gene expression

Abstract

SK-N-SH cells exposed to low ratios of ICP0-null (ΔICP0) mutants of herpes simplex virus per cell express the viral α proteins, but the progression to β and γ gene expression does not ensue. In these restrictive cells, post-α gene expression can be induced after exposure of the infected cells to sodium butyrate, an indication that VP16 brought into cells by the virus and the α gene products made after infection cannot block the silencing of viral post-α genes by histone deacetylases (HDACs). This observation is consistent with evidence reported earlier that ICP0 dissociates HDAC1/2 from the CoREST/REST complex. In permissive U2OS cells, replication is independent of the ratio of ΔICP0 mutant per cell. To determine whether other viral genes are involved in blocking HDACs, we used a surrogate system consisting of baculoviruses carrying viral or cellular genes driven by CMV immediate–early promoter. Expression of these genes requires blocking of histone deacetylation. We report that (i) cotransduced U_S3 or U_S3.5 protein kinase substitutes for sodium butyrate in enabling the expression of a reporter gene in restrictive cells and enhancing it in permissive cells; (ii) HDAC1 is phosphorylated concomitant with the expression of reporter genes; and (iii) the amounts and appearance of HDAC1 are altered in transduced cells expressing U_S3 protein kinase in the absence of other viral proteins. We conclude that the U_S3 protein kinase blocks histone deacetylation by a mechanism distinct from that of ICP0 and that debilitated histone deacetylation contributes to the permissiveness of U2OS cells for ΔICP0 mutants.

The central theme of this article is that, immediately upon release of herpes simplex virus (HSV) 1 DNA from capsids, a competition ensues between cellular factors whose mission is to silence viral DNA and viral proteins that aim to block the cell. In cells infected at low ratios of wild-type virus per cell, α-transinducing factor (or VP16) transactivates α genes, the first set to be expressed after infection. Of the six α proteins (ICP0, ICP4, U_S1.5, ICP22, ICP27, and ICP47), two, ICP4 and ICP0, are required for expression of β genes. ICP4 represses its own gene and other genes and most likely recruits host factors to the machinery that transcribes viral DNA (1). In cells infected at low ratios of infectious virus per cell, viral replication takes place in a few cell lines (e.g., U2OS cells) but is blocked in most cell lines (e.g., SK-N-SH and Vero cells) (2–5). Numerous reports have shown that herpes viral gene expression depends on the inhibition of histone deacetylases (HDACs) (5–11). Association of ICP0 orthologs with HDACs has been reported for other herpesviruses (12). The equivalent functions of ICP0 have been reported for immediate–early protein 1 of murine and human cytomegaloviruses, respectively (13, 14). In an earlier report we showed that ICP0 interacts with the complex CoREST/REST/HDAC1/2 and dissociates the HDACs from the complex (15).

In this report we present evidence that ICP0 enables the transition from α to β gene expression by blocking histone deacetylation inasmuch as in SK-N-SH cells exposed to low ratios of ΔICP0 plaque-forming units (pfu) per cell, sodium butyrate, a HDAC inhibitor, complements the mutant in enabling post-α gene expression. We report that HSV-1-encoded protein kinases U_S3 and U_S3.5 also act to block histone deacetylation. Finally, we show that, in SK-N-SH cells that restrict post-α gene expression, the U_S3 protein kinase mediates the phosphorylation and rearrangement of HDAC1. The results suggest that the permissivity of U2OS cells for ΔICP0 mutants may reflect the regulation of HDAC1/2 in U2OS cells.

Relevant to this report are the following: Both ICP0 and the U_S3 protein kinase are multifunctional proteins (1). ICP0 is a 775-residue protein encoded in three exons (16). An important biologic activity of ICP0 is that it transactivates genes introduced by infection or transfection but does not bind DNA (17–19). In addition to the functions described in this and the earlier report (15), ICP0 also blocks interferon (IFN) pathway signaling (20–22). One of the pathways by which ICP0 blocks antiviral responses activated by IFN is to degrade PML (promyelocytic leukemia protein) and SP100 and to disperse the ND10 nuclear structures inasmuch as the antiviral effects of exogenous IFN are signaled through ND10 nuclear structures (23). ICP0 ubiquitin ligase activity mapping in exon II causes degradation of PML and dispersal of ND10 structures, thereby blocking the IFN signaling pathway (16).

Imbedded in the transcriptional unit for the U_S3 protein kinase is a small transcriptional unit encoding a shorter, carboxyl-terminal, colinear U_S3.5 protein (24). Both U_S3 and U_S3.5 act as protein kinases targeting amino acid residues similar to those phosphorylated by protein kinase A (PKA) (25). Both U_S3 and U_S3.5 kinases mediate the phosphorylation of HDAC1 and HDAC2 (24, 26, 27). Whereas the U_S3 protein kinase blocks apoptosis induced by defective viruses or exogenous agents, the U_S3.5 enzyme does not (27–32).

Results

In SK-N-SH Cells Infected at Low Multiplicity of Infection with the ΔICP0 Mutant Virus, the Expression of ICP8, a β Gene, Requires Inhibition of Histone Deacetylation.

SK-N-SH or U2OS cells grown on four-well slides were exposed to 0, 1, 5, or 20 mM sodium butyrate 1 h before infection with 0.005 pfu of ΔICP0 virus per cell (titer was determined on U2OS cells). At 18 h after infection, the cells were fixed and reacted first with polyclonal anti-CoREST antibody and monoclonal anti-ICP4 or anti-ICP8 antibody and subsequently with FITC-conjugated anti-rabbit and Texas red-conjugated anti-mouse antibodies. Approximately 200 cells illuminated by the anti-CoREST antibody were counted with the aid of a Zeiss confocal microscope, and the percentage of cells expressing ICP4 or ICP8 was determined. The results (Fig. 1A and B) were as follows.

Fig. 1.

Sodium butyrate enables expression of ICP8 in SK-N-SH cells exposed to ΔICP0 mutant virus and enhances expression of transduced genes in SK-N-SH and U2OS cells. (A and B) SK-N-SH or U2OS cells seeded on four-well slides were exposed to concentrations of sodium butyrate shown for 1 h and then infected with 0.005 pfu of ΔICP0 virus per cell. At 18 h after infection the slides were fixed and reacted with polyclonal anti-CoREST plus monoclonal anti-ICP4 (A) or anti-ICP8 (B) antibodies. Secondary antibodies were FITC-conjugated anti-rabbit and Texas red-conjugated anti-mouse antibodies. The images collected with a Zeiss confocal microscope at ×40 were tabulated, and the percentages of cells expressing ICP4 or ICP8 were determined. (C) SK-N-SH and U2OS cells seeded on four-well slides exposed to sodium butyrate as shown were transduced with baculovirus expressing myc-tagged U_L34. After 9 h of incubation the cells were fixed and reacted with polyclonal anti-CoREST and monoclonal anti-myc antibodies. Images collected as above were tabulated, and percentages of cells expressing myc-tagged U_L34 were determined. (D) SK-N-SH and U2OS cells grown in 25-cm² flasks were transduced with 10 pfu of baculovirus carrying myc-tagged U_L34 per cell at shown concentration of sodium butyrate for 24 h. The cells were lysed, electrophoretically separated on 10% denaturing polyacrylamide gel, and reacted with monoclonal anti-myc antibody as described previously (43).

In the absence of sodium butyrate, 19% of total infected SK-N-SH cells expressed ICP4, and only 2.6% expressed ICP8. At the optimal dose of sodium butyrate (5 mM), the number of ICP4-positive cells increased to 36%, whereas the number of ICP8-positive cells increased to 16%. Thus, sodium butyrate increased the number of ICP4-positive cells ≈2-fold and that of ICP8-expressing cells 6-fold. We may conclude from these results that, in SK-N-SH cells, the expression of α genes and especially that of β genes is enhanced by inhibitors of histone deacetylation.

In contrast to the results obtained with SK-N-SH cells, the number of U2OS cells expressing ICP4 and ICP8 (58% and 27%, respectively) was significantly higher, and these numbers did not increase significantly in cells exposed to sodium butyrate. We conclude from this experiment that inhibition of histone deacetylation is critical for the expression of ICP8 in SK-N-SH cells but not in U2OS cells and that sodium butyrate compensates at least in part for the absence of ICP0 in SK-N-SH cells.

Expression of Genes Transduced with the Aid of Baculoviruses also Requires Inhibition of Histone Deacetylation.

Genes driven by CMV immediate–early promoter and carried by baculoviruses can be expressed in mammalian cells. In most cell lines tested, the expression of the inserted gene required inhibition of histone-deacetylating enzymes (33). The following experiments were designed to compare the requirement for inhibiting histone deacetylating enzymes in SK-N-SH and U2OS cells.

In the first series, SK-N-SH and U2OS cells grown on four-well slides exposed for 1 h to BC2901 expressing a myc-tagged U_L34 (40 pfu per cell) were incubated for 9 h in medium containing 0–20 mM sodium butyrate, then fixed and reacted with monoclonal anti-myc antibody for detection of the U_L34 protein and polyclonal anti-CoREST antibody to visualize the fixed cells. As shown in Fig. 1C, <2% of SK-N-SH cells transduced with BC2901 expressed U_L34 protein in the absence of sodium butyrate as compared with 20% of cells exposed to 10 mM sodium butyrate. In contrast, between 65% and 75% of U2OS cells transduced with BC2901 expressed U_L34 protein in a manner independent of the concentration of sodium butyrate.

In the next series of experiments, electrophoretically separated proteins from replicate cultures of SK-N-SH or U2OS cells transduced with BC2901 (10 pfu per cell) and maintained for 24 h in medium containing 0–20 mM sodium butyrate were reacted with anti-myc antibody. The results (Fig. 1D) were that only trace amounts of the myc-tagged U_L34 protein accumulated in untreated SK-N-SH cells in contrast to huge amounts detectable in U2OS cells. Sodium butyrate caused an increase in the accumulation of U_L34 protein in both cell lines. But, even at the most effective dose of sodium butyrate, the amount of myc-tagged protein accumulating in SK-N-SH cells was lower than the amount detected in untreated U2OS cells. We conclude the following: (i) Optimal expression of genes introduced by transduction into SK-N-SH and U2OS cells requires inhibition of histone-deacetylating enzymes. This requirement is more stringent for SK-N-SH cells than for U2OS cells, suggesting that either histone deacetylation is more active in SK-N-SH cells or defective in U2OS cells. (ii) While the number of U2OS cells capable of expressing transduced genes was higher than in SK-N-SH cells, this number remained constant independent of sodium butyrate concentration. Nevertheless, inhibition of histone deacetylation was essential for optimal expression of the transduced gene. The results suggest the possibility that the total fraction of cells capable of expressing U_L34 was limited and independent of inhibition of deacetylation by sodium butyrate. The situation appears to be different in SK-N-SH cells in that both the amount of U_L34 protein and number of cells accumulating the protein required inhibition of histone deacetylation. Furthermore, because the number of SK-N-SH cells expressing U_L34 protein was relatively low and sodium butyrate-dose dependent, the results suggest that some SK-N-SH cells were incapable of expressing U_L34 protein even under the highest concentrations of sodium butyrate.

The U_S3 Gene Can Substitute for Sodium Butyrate for Expression of Baculovirus-Mediated Transduction of Reporter Gene.

Viral protein kinase U_S3 mediates the phosphorylation of histone-deacetylating enzymes HDAC1 and HDAC2 (24, 26). In this series of experiments, the expression of four reporter genes encoded in baculoviruses was investigated in SK-N-SH and U2OS cells in the presence or absence of sodium butyrate and compared to cotransduction with the U_S3 gene. The results (Fig. 2) were as follows.

Fig. 2.

U_S3 substitutes for sodium butyrate in enhancing reporter gene expression. Replicate cultures of U2OS (lanes 1–8) or SK-N-SH cells (9–16) in 25-cm² flasks were exposed to insect cell medium (lanes 7, 8, 15, and 16), 8 pfu per cell of BC2619 expressing cdc34, 2 pfu per cell of BC2620 expressing HDAC4, 2 pfu per cell of BC2618 expressing U_S11, or 2 pfu per cell of BC2617 expressing U_S1.5 (lanes 1, 2, 9, and 10) alone, or mixture containing 8, 4, 2, or 1 pfu per cell of BC2602 expressing Flag-tagged wild-type U_S3 (lanes 3–6 and 11–14). After 1 h the inoculum was replaced with medium supplemented with 10% newborn bovine serum (lanes 1–8) or FBS (lanes 9–16) in the presence (lanes 1, 8, 9, and 16) or absence (lanes 2–7 and 10–15) of 6 mM sodium butyrate. The cells were harvested at 22 h after transduction, rinsed three times with PBS containing protease inhibitor mixture (Roche), and then solubilized in 150 μl of disruption buffer (50 mM Tris·HCl, pH 7/2% SDS/710 mM 2-mercaptoethanol/3% sucrose). Portions (50 μl) of lysates were boiled for 5 min, and the solubilized proteins were subjected to electrophoresis in 11% (A, B, and D) or 12% (C) denaturing polyacrylamide gel, transferred to nitrocellulose sheets, blocked with 5% nonfat milk, reacted with mouse monoclonal antibody against the Flag epitope (for detection of U_S3, cdc34, and HDAC4 in A and B) or U_S11 (C), or rabbit polyclonal antibody W2 against U_S1.5 (D), followed by appropriate secondary antibody conjugated to alkaline phosphatase (Bio-Rad), and visualized according to the manufacturer’s instructions.

In SK-N-SH cells, the products of the transduced reporter genes either failed to accumulate (e.g., U_S1.5) or accumulated at a low level (cdc34, HDAC4, or U_S11) in the absence of either sodium butyrate or U_S3. In each test, sodium butyrate produced maximum accumulation of the reporter gene product (Fig. 2, lane 9). U_S3 enabled the accumulation of reporter proteins to levels above those obtained in the absence of any treatment. In the case of cdc34 and HDAC4, the highest accumulation of the reporter proteins was in cells transduced with the lowest amounts of Bac-U_S3 (1 pfu per cell). The overall impression is that, at 8 pfu of Bac-U_S3 per cell, the yields of reporter proteins were lower than in cells transduced with lower ratios of Bac-U_S3 per cell. The largest amounts of HDAC4 and U_S11 obtained in cells cotransduced with Bac-U_S3 were comparable to those obtained in cells treated with sodium butyrate. In U2OS cells, the amounts of some reporter proteins produced in cells cotransduced with 8 pfu of Bac-U_S3 per cell (e.g., HDAC4 and cdc34) were lower than in cells transduced with lower ratios of Bac-U_S3 per cell, whereas others (e.g., U_S1.5) were optimal at 4 pfu per cell but were reduced at either higher or lower ratios of Bac-U_S3 per cell. As a general rule, at ratios <8 Bac-U_S3 pfu per cell, the expression of the reporter genes was U_S3-dose-dependent. At high ratios of Bac-U_S3 pfu per cell, U_S3 may be toxic to the transduced cells. The largest accumulations of reporter proteins in cells cotransduced with U_S3 were comparable to those obtained in cells exposed to sodium butyrate.

The key conclusion of the results shown in Fig. 2 is that U_S3 protein kinase can substitute for sodium butyrate to enable expression of genes introduced into cells by transduction at relatively low multiplicities of infection.

U_S3 Is More Effective If Transduced into Cells Several Hours Before the Transduction of the Reporter Gene.

This series of experiments was done in two parts. In the first (Fig. 3A) U2OS cells were exposed to insect cell medium or transduced singly or cotransduced with baculovirus encoding Flag-tagged wild-type U_S3 (0, 0.5, 1, 2, or 4 pfu per cell) or HDAC4 (2 pfu per cell) and processed as described in the legend to Fig. 3. In the second part (Fig. 3B) the experiment was repeated by using baculoviruses encoding U_S3 (1 pfu per cell) or HDAC4 (2 pfu per cell). In addition, two cultures transduced and maintained for 6 h or 3 h with U_S3 were then transduced with HDAC4. The salient features of the results (Fig. 3) were as follows. (i) The expression of the reporter gene HDAC4 in U2OS cells was U_S3-dose-dependent. (ii) Expression of HDAC4 was vastly higher in cells transduced with U_S3 3 or 6 h before transduction with HDAC4 than in cells cotransduced with U_S3 and HDAC4. (iii) HDAC1 was posttranslationally modified in all cell cultures transduced with U_S3 alone or in combination with HDAC4 independent of the dose of U_S3.

Fig. 3.

U_S3 is more effective if transduced into cells several hours before transduction of reporter gene. (A) Cotransduction. Replicate cultures of U2OS were exposed for 1 h to insect cell medium (lane 1), to 4, 2, 1, or 0.5 pfu of BC2602 expressing Flag-tagged wild-type U_S3 (lanes 2–5) per cell alone, to a mixture containing 2 pfu per cell of BC2620 expressing HDAC4 (lanes 7–10), or to 2 pfu per cell of BC2620 alone (lane 6), then maintained for 24 h in 10% serum, harvested, and processed as described in the legend to Fig. 2. The electrophoretically separated proteins were reacted with mouse monoclonal antibody against the Flag epitope (for detection of U_S3 and HDAC4) or rabbit polyclonal antibody against HDAC1, as described in the legend to Fig. 2. (B) Preactivation. Replicate cultures of U2OS were exposed to 1 pfu per cell of BC2602 expressing Flag-tagged wild-type U_S3 alone (lane 5), 2 pfu per cell of BC2620 expressing HDAC4 alone (lane 4), or mixture containing both BC2602 and BC2620 (lane 3). Two cultures preexposed for 6 h (lane 1) or 3 h (lane 2) to 1 pfu per cell of BC2602 were also transduced with 2 pfu per cell of BC2620 expressing HDAC4. Cells were maintained in 10% serum for 21 h, then harvested and processed as described above.

The Kinase Activity of U_S3 Is Essential for Expression of the Reporter Gene Transduced into U2OS Cells.

Replicate cultures of U2Os cells were exposed to insect cell medium, empty baculovirus, or baculoviruses encoding Flag-tagged U_S3, Flag-tagged U_S3.5, or K220N mutant of U_S3 (8, 4, or 2 pfu per cell). Five hours later they were transduced with U_S11 (2 pfu per cell), maintained for 22 h in the presence or absence of 6 mM sodium butyrate, and then harvested and processed as described in the legend to Fig. 4. The results (Fig. 4A) were as follows. (i) The amounts of accumulation of wild-type U_S3 protein was greater than that of the mutant U_S3 protein in cells exposed to the same number of pfu (compare lanes 10–12 with lanes 13–15). The most likely explanation is that U_S3 enhances its own gene expression. (ii) The amounts of U_S11 protein accumulating in transduced cells preactivated with the mutant U_S3 gene (lanes 13–15) were small and were not much greater than that accumulated in cells preactivated with the empty baculovirus (lanes 1–3). In contrast, appreciable amounts of U_S11 protein accumulated in transduced cells preactivated by either wild-type U_S3 (lanes 10–12) or U_S3.5 (lanes 7–9) at levels higher than that accumulated in cells treated with sodium butyrate (lane 6). (iii) HDAC1 was posttranslationally modified in all cell cultures cotransduced with the wild-type U_S3 (lanes 10–12) or U_S3.5 (lanes 6–9) but not the mutant U_S3 gene (lanes 13–15). These results suggest that the expression of U_S3 and of the reporter gene U_S11 required a functional U_S3 kinase domain.

Fig. 4.

The U_S3 kinase activity is required for enhancement of gene expression. (A) Replicate cultures of U2OS cells were exposed to insect cell medium (lanes 4–6), 8, 4, or 2 pfu per cell of BC2623 (control baculovirus; lanes 1–3), BC2608 (expresses Flag-tagged wild-type U_S3.5; lanes 7–9), BC2602 (expresses Flag-tagged wild-type U_S3; lanes 10–12), or BC2621 (expresses K220N-mutant U_S3; lanes 13–15) for 1 h, then maintained in medium supplemented with 10% FBS. Five hours later the cells were transduced with 2 pfu per cell of BC2618 (expresses U_S11; all lanes except lane 4). The cultures were maintained for 22 h in 10% serum in the presence (lane 6) or absence (lanes 1–5 and 7–15) of 6 mM sodium butyrate, then harvested and processed as described in the legend to Fig. 2. The electrophoretically separated proteins were reacted with monoclonal antibody against U_S11 or polyclonal antibody against U_S3 or HDAC1. (B) Replicate cultures of U2OS cells were exposed for 1 h to insect cell medium (lanes 1–4), to lower ratios of pfu of BC2602 (expresses Flag-tagged wild-type U_S3; lanes 5–10) per cell, or to 8, 4, or 2 pfu of BC2621 (expresses K220N-mutant U_S3; lanes 11–13) per cell, maintained in 10% serum for 5 h, and then transduced with 2 pfu per cell of BC2618 (expresses U_S11; all lanes except lanes 1,2). Cultures were maintained for 22 h in 10% serum in the presence (lanes 1 and 3) or absence (lanes 2 and 4–13) of 6 mM sodium butyrate, then harvested and processed as described above.

The results of the experiment described above did not exclude the possibility that the absence of U_S11 protein was due to the low level of expression of U_S3 mutant protein rather than the lack of kinase function. In the experiment shown in Fig. 4B, the cells were transduced with lower ratios of wild-type U_S3 (2, 1, 0.5, 0.25, 0.125, or 0.06 pfu per cell) to enable comparison of cultures expressing equal amounts of mutant protein. The results (Fig. 4B) were as follows. (i) Comparison of cultures expressing equal amounts of wild-type and mutant U_S3 (e.g., lanes 7 and 11) shows that U_S11 accumulation was higher in cells transduced with the wild-type U_S3 gene. (ii) The amounts of U_S11 accumulating in cells transduced with the smallest doses of wild-type U_S3 were higher than in cells transduced with the mutant U_S3 gene (compare lanes 9 and 10 with lanes 11–13). (iii) HDAC1 was posttranslationally modified in all cell cultures cotransduced with the wild-type gene (lanes 5–10) but not the mutant U_S3 gene (lanes 11–13). The key conclusion is that the kinase activity of U_S3 is necessary for the enhancement of expression of transduced reporter genes.

Changes in the Appearance and Distribution of HDAC1 Correlated with Accumulation of the U_S3 Protein Kinase.

SK-N-SH or U2OS cells seeded on four-well slides were mock-transduced or transduced with baculovirus BC2623 expressing myc-tagged U_S3, fixed 20 h after transduction, and then reacted with polyclonal anti-HDAC1 and monoclonal anti-myc antibodies followed by FITC-conjugated goat anti-rabbit and Texas red-conjugated goat anti-mouse antibodies. Images were collected with a Zeiss confocal microscope. The striking feature observed in this (Fig. 5) and other experiments was that SK-N-SH cells formed two populations, one with prominent amounts of HDAC1 and the other with little or no HDAC1. In contrast, the distribution of HDAC1 in U2OS cells was relatively uniform. Moreover, in SK-N-SH cells transduced with U_S3 and expressing the viral protein kinase, the amounts of HDAC1 were low or barely detectable in contrast to cells not expressing the kinase. In U2OS cells transduced with U_S3 and expressing the viral protein kinase, the intensity of HDAC1 fluorescence was decreased and had a speckled appearance but did not vary much from cell to cell. It is not clear whether U_S3 was expressed preferentially in SK-N-SH cells containing less HDAC1 or whether cells that were transduced ended up with less HDAC1 irrespective of the amounts present at the time of transduction. In a duplicate experiment the monoclonal antibody against HDAC1 was also tested to ensure that these observations did not reflect the properties of specific antibodies, and the same results were obtained.

Fig. 5.

Distribution of HDAC1 in cells expressing baculovirus transduced with U_S3. SK-N-SH and U2OS cells seeded on four-well slides were mock-transduced or transduced with myc-tagged U_S3. After 20 h of transduction, cells were fixed and treated as described in Fig. 1. The slides were incubated with polyclonal anti-HDAC1 plus monoclonal anti-myc antibodies and then reacted to FITC-conjugated goat anti-rabbit plus Texas red-conjugated goat anti-mouse antibodies.

Discussion

The evidence collected in the past three decades indicates that, upon release of viral DNA from capsids into the nucleus, α-transinducing factor (VP16) activates the transcription of α genes. The six products of the initial transcription are ICP0, U_S1.5, ICP4, ICP22, ICP27, and ICP47. Expression of β and γ genes ensues. At low multiplicity of infection, the transition from α to β and γ protein synthesis requires ICP0 and ICP4. In the absence of ICP0, the four regulatory proteins ICP4, ICP22, U_S1.5, and ICP27 enable viral gene expression in a few selected cell lines exemplified by U2OS cells but not in most cell lines exemplified by SK-N-SH cells (2, 4, 16). Central to the understanding of the mechanism by which the virus insures the expression of its genes is the role of ICP0 in enabling the expression of post-α genes and also why it is critical in cells exemplified by SK-N-SH cells and not in U2OS cells. The model we propose is as follows.

On entry of the viral DNA into the nucleus, a competition ensues between cellular proteins that attempt to silence the viral genome and viral gene products, which ensure that viral gene expression is unencumbered. As noted above, α genes are transactivated by α-transinducing factor, whereas β genes require both ICP4 and ICP0 for their expression. We propose that HSV encodes two proteins, ICP0 and U_S3, that independently block silencing of viral DNA to enable the transition from α to β gene expression.

The role of ICP0 is apparent from three series of experiments. First, this laboratory reported that the expression of viral genes by a virus carrying cDNA copies of ICP0 is delayed by several hours. Ultimately, however, the synthesis of proteins and accumulation of virions in mutant-infected cells catches up with those of wild-type virus-infected cells (34). One interpretation of the delay is that the synthesis of ICP0 directed by the cDNA copy is impaired and slower than that directed by the wild-type gene. Inhibitors of histone deacetylation accelerated the synthesis and accumulation of viral gene products in cells infected with the ICP0-cDNA mutant virus (26).

The second line of evidence is based on the earlier report that ICP0 binds CoREST/REST and dissociates this complex from HDAC1 and HDAC2. Sequentially, in wild-type virus-infected cells, the CoREST/REST complex and HDAC1/2 are independently transported to the cytoplasm. Furthermore, both HDAC1/2 and CoREST become posttranslationally modified (15, 26). It was also reported that ICP0 interacts physically with HDAC4, HDAC5, and HDAC7 (35).

The third line of evidence presented in this report centers on the gene expression of ΔICP0 mutant virus in SK-N-SH and in U2OS cells. At low multiplicity of infection, ΔICP0 mutant virus is arrested in SK-N-SH cells at α gene expression. We show that the transition to post-α gene expression can be induced by exposure of cells to sodium butyrate. In U2OS cells that are permissive for ΔICP0 mutants, the transition does not require sodium butyrate. In essence, sodium butyrate compensates at least in part for the absence of ICP0. These three series of experiments support the hypothesis that one function of ICP0 is to block the repression of the viral genome by histone deacetylation.

The hints that U_S3 and U_S3.5 protein kinases may play a similar role emerged from observations that both kinases mediated the phosphorylation of HDAC1 and HDAC2 (26, 27). In this report we show that both U_S3 and U_S3.5 protein kinases enable the expression of viral or cellular genes transduced into mammalian cells with the aid of baculoviruses. Concurrently, HDAC1 undergoes changes in intracellular distribution or amounts. The advantage of the baculovirus system is that it does not introduce into cells transcriptional factors active in mammalian cells, and, in most cell lines, expression of transduced reporter genes requires exposure to sodium butyrate. Our data show that cotransduction of U_S3 readily enables the expression of reporter genes in SK-N-SH cells and enhances the expression of transduced reporter genes in U2OS cells in the absence of sodium butyrate. It would appear that U_S3 and ICP0 share similar objectives to block the silencing of the viral genome, although the mechanisms by which these proteins accomplish their goals are different. Cotransduction of U_S3 or U_S3.5 protein kinase with a desired gene is a convenient and rapid way of expressing genes in mammalian cells in a dose-dependent manner. The accumulation of reporter gene product was observed as early as 6 h after transduction, and the amounts of protein continued to increase for at least 22 h (data not shown).

In essence, HSV encodes one site-specific transcriptional activator for α genes (VP16) and three proteins whose function is to block cellular silencing machinery (ICP0, U_S3, and U_S3.5). In addition, ICP4 acts as a sequence-specific repressor and as a sequence-independent transcriptional factor, based on the observations that in infected cells ICP4 does not transactivate post-α genes in the absence of ICP0 or sodium butyrate and that there is little homology among the DNA binding sites of ICP4 in post-α genes (3, 36). Moreover, post-β gene expression is enhanced by the modification of DNA templates by ICP8 (37). Viral gene expression is also measurably aided by the selective degradation of cellular mRNA by virion host shutoff protein encoded by U_L41 and by inhibition of splicing by ICP27 (38, 39).

Last, the properties of U2OS cells that enable them to support the replication of ΔICP0 mutants remain a puzzle. The data suggest that U2OS cells are only partially effective in silencing HSV DNA upon release from capsids into the nucleus.

Materials and Methods

Cells and Viruses.

Protocols for U2OS (human osteosarcoma cells), SK-N-SH, and insect cell line Sf9 (Spodoptera frugiperda) were reported elsewhere (27, 40). The recombinant baculoviruses BC2820, BC2602, BC2608, BC2617 (U_S1.5), and BC2618 (U_S11) were described (24, 27, 41, 42). The properties and the derivation of deletion mutant R7910 (ΔICP0) were described elsewhere (40).

Antibodies.

Rabbit polyclonal antibodies against U_S3 and the carboxyl-terminal region of ICP22 (W2) and the mouse monoclonal antibody against Us11 were described previously (24). We purchased monoclonal antibodies against ICP4 and ICP8 from the Goodwin Cancer Research Institute (Plantation, FL), polyclonal antibody against HDAC1 and monoclonal antibody against actin and Flag-epitope antibody from Sigma, polyclonal antibody for CoREST and monoclonal antibody against HDAC1 from Upstate Biotechnology, and the monoclonal antibody against c-myc from Santa Cruz Biotechnology.

The methods for generation and amplification of recombinant baculoviruses, confocal microscopy, preparation of cell lysates, electrophoretic separation of proteins, and immunoblotting were previously described (26, 34, 43).

Plasmids for Construction of Recombinant Baculoviruses.

PRB5999, pRB6000, pRB6001, pRB6003, and pRB5914 were used for construction of baculoviruses BC2619 (5′ Flag-tagged cdc34 cut from pRB5997), BC2620 [5′ Flag-tagged HDAC4 cut from plasmid pBJ5.1-HDAC4 kindly provided by Bradley Bernstein (Harvard University, Cambridge, MA)], BC2623 [5′ Myc-tagged U_S3 amplified from HSV-1(F) viral DNA], BC2901 [5′ Myc-tagged U_L34 amplified from pRB5709 (44)], and BC2621 [K220N U_S3 mutant kindly provided by J. Munger (Princeton University, Princeton, NJ)], respectively. The baculovirus transfer vector pRB5850 (42) was used for cloning cdc34, Myc-tagged U_S3, and K220N U_S3 mutant, MTS1-Myc (43) was used for U_L34, and the StuI/SacI-collapsed of pRB5850 was used for HDAC4.

Abbreviations

HDAC

histone deacetylase

HSV

herpes simplex virus

pfu

plaque-forming unit.