Interwoven Roles of Cyclin D3 and cdk4 Recruited by ICP0 and ICP4 in the Expression of Herpes Simplex Virus Genes

Abstract

Elsewhere this laboratory reported that (i) ICP0 interacts with cyclin D3 but not D1 or D2. The 3 cyclins independently partially rescue ΔICP0 mutants. (ii) Interaction with cyclin D3 is required for the switch from nuclear to cytoplasmic accumulation of ICP0. (iii) In infected cells cdk4 is activated whereas cdk2 is not. Inhibition of cdk4 results in nuclear retention of ICP0. Overexpression of cyclin D3 reverses the effect of the inhibitor. Here we report the following. (i) cdk4 interacts with ICP0, ICP4, and possibly with ICP8. This interaction is required to recruit cdk4 initially to ND10 and later to the viral replication compartments. (ii) cdk4 inhibitor I reduced or delayed the transcription and ultimately translation of mRNAs of ICP4, ICP27, or ICP8 and to a lesser extent that of the ICP0 gene in wild-type virus-infected cells. (iii) Overexpression of cyclin D3 resulted in a more rapid transcription of these genes. In the presence of inhibitor, the rates of accumulation of the products of these genes resemble those of wild-type virus in the absence of inhibitor. (iv) Overexpression of cyclin D3 also results in mobilization of cdk6 in nuclei of infected cells. We conclude that ICP0 encodes a function that enhances the recruitment of cyclin D3 to ND10 structures to activate cdk4 and that ICP0 along with other viral proteins recruits cdk4 to ND10 structures and ultimately to replication compartments for enhanced expression of viral genes and viral DNA synthesis.

The infected cell protein 0 (ICP0), an α protein encoded by herpes simplex virus 1 (HSV-1), interacts with many diverse cellular proteins and performs multiple functions (19). The major interactions and functions reported to date concern preemptive suppression of innate response mediated by interferons—a cellular function mediated in part by components of the ND10 nuclear bodies—and suppression of silencing of viral DNA by cellular proteins (1, 6, 11, 16, 17, 22, 24). Specifically ICP0 acts as a ubiquitin ligase that interacts with the UbcH5α conjugating enzyme and degrades the ND10 components PML and SP100, leading ultimately to the dispersal of ND10 structures (1, 12, 16). In a related event, ICP0 interacts with CoREST, a component of the histone deacetylase 1 (HDAC1) or 2/CoREST/REST/LSD1 complex (17, 18). As a consequence, the HDACs become dissociated from the suppressor complex, and ultimately all components of the complex are translocated to the cytoplasm (17, 18). In this report we marshal the evidence that ICP0 performs a third major function, that is, to mobilize cellular proteins that perform key functions in transcription, expression of viral genes, and viral DNA synthesis (22-24). Relevant to this report are the following notes.

In the course of Saccharomyces cerevisiae two-hybrid studies in which components of ICP0 were used as bait, it was noted that ICP0 interacts both in yeast and in vitro with cyclin D3 but not with cyclin D1 or D2 (25). The binding site for cyclin D3 was localized to residue D199 (35). A mutant, D199A, no longer reacted with cyclin D3 (35). A characteristic of this mutant is that in infected cells ICP0 was retained in the nucleus, although this mutant is not replication defective (35). In contrast, in wild-type virus-infected cells ICP0 is degraded in the nucleus while additional ICP0 accumulates in the cytoplasm (15). A mutant virus (R7801) that encodes and overexpresses cyclin D3 enables the accumulation of ICP0 in the cytoplasm much faster than does the wild-type virus (36). Subsequent studies have shown that cyclin D3 as well as cyclins D1 and D2 partially complements a ΔICP0 mutant (23). Moreover, all 3 colocalize with ICP0 in ND10 nuclear bodies and ultimately in the replication compartments in which viral DNA synthesis takes place (23, 36). The fundamental conclusion of these studies is that HSV-1 must mobilize a D cyclin for its replication, and it encodes in ICP0 a function designed to recruit cyclin D3. It is noteworthy that at least two herpesviruses encode cyclin D2 orthologs (4, 5, 20, 27, 30).

One function of D cyclins is to activate cdk4 to initiate a process that would lead eventually to viral DNA synthesis (23). In infected cells cdk4 is activated whereas cdk2 is not (9, 23, 36). Studies with a cdk4 inhibitor have shown that ICP0 in wild-type virus-infected cells treated with the drug is retained in the nucleus (23). In contrast, in cells infected with a mutant (R7801) that overproduced cyclin D3, ICP0 was not retained in the nucleus in the presence of the drug (23). In essence accumulation of ICP0 in the cytoplasm required the recruitment and function of cdk4 at the sites of processing (ND10) and synthesis of viral DNA (replication compartments) (23).

The objective of the studies reported here was to investigate the recruitment and function of cyclin D3 in the initial stages of HSV-1 infection. Studies published elsewhere have shown that roscovitine, a broad cyclin-dependent kinase (cdk) inhibitor, blocks transcription of viral genes (8). We report three fundamental findings. (i) cdk4 interacted with ICP0, ICP4, and ICP8 most likely by binding to a complex consisting of at least two of the three proteins. (ii) cdk4 inhibitor I reduced the accumulation of mRNAs. The effect was greater on ICP4, ICP27, or ICP8 than on ICP0. In cells infected with the R7801 mutant, which overproduces cyclin D3, transcription of all 3 genes was accelerated. In inhibitor-treated cells the pattern of transcription was similar to that of untreated wild-type virus-infected cells. (iii) Lastly, one hypothesis that could explain the effect of overexpression of cyclin D3 is activation of cdk6. We report accumulation of increased amounts of cdk6 in nuclei of cells infected with R7801 mutant virus.

MATERIALS AND METHODS

Cells and viruses.

The sources, properties, and propagation of HEp-2 and Vero cells were reported elsewhere (24). HSV-1(F), a limited-passage isolate, is the prototype strain used in this laboratory (10). The R7801 mutant carrying cyclin D3 inserted into the thymidine kinase gene and the R7910 ICP0 null recombinant have been described before (25, 36). The HSV-1(KOS)d120 mutant, a kind gift of Neal A. DeLuca (Pittsburgh Medical School, Pittsburgh, PA), lacks both copies of the α4 gene and was grown in a Vero-derived cell line (E5) expressing the α4 gene (7). The d301 mutant, a kind gift of David M. Knipe (Harvard Medical School, Boston, MA), has an internal in-frame deletion in the ICP8 open reading frame (ORF) and was grown in V5-29, a Vero-derived cell line expressing the ICP8 gene (13).

Purification of the GST-cdk4 fusion protein and GST-cdk4 pulldown assays.

The entire human cdk4 open reading frame was PCR amplified from a full-length mammalian gene collection (Invitrogen) using the primers 5′ GGAAGCTTGATATCGCTACCTCTCGATATGAGCCA 3′ (forward) and 5′ CCAAGCTTGATATCTCACTCCGGATTACCTTCATC 3′ (reverse) digested with EcoRV and inserted into the SmaI site of pGEX-4T2 in frame with glutathione S-transferase (GST). Escherichia coli XL1-Blue cells transformed with the pGEX-4T2- or with the pGEX-4T2-cdk4-expressing plasmids were induced with 0.3 mM IPTG (isopropyl-β-d-thiogalactopyranoside; Denville Scientific) for 3 h after the optical density at 600 nm reached a value of 0.5. The harvested cells were lysed by sonication in phosphate-buffered saline (PBS), and then Triton X-100 was added to a final concentration of 1%. The cell debris was removed by centrifugation, and the GST or GST-cdk4 proteins were adsorbed to glutathione-agarose beads (Sigma). After several rinses with PBS-1% Triton X-100, the proteins were eluted with 10 mM reduced glutathione (Sigma) in 50 mM Tris-HCl (pH 8), dialyzed against the same buffer, and quantified with a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA).

For the GST pulldown assays, 2 × 10⁷ HEp-2 cells, infected either with the wild-type virus or with the deletion mutants at 10 PFU/cell, were lysed in HEPES-1% Triton X-100 buffer consisting of 50 mM HEPES, pH 7.4, 250 mM NaCl, 10 mM MgCl₂, 1 mM PMSF (phenylmethylsulfonyl fluoride), and protease cocktail inhibitors (Sigma) supplemented with 1% Triton X-100. The lysates were reacted overnight with equal amount of GST or GST-cdk4 proteins immobilized on glutathione beads and rinsed three times with the HEPES-1% Triton X-100 buffer, and the bound protein complexes were subjected to electrophoretic analysis on 9% polyacrylamide gels, transferred to a nitrocellulose sheet, and reacted with the ICP0 (1:500), ICP4 (1:1,000), or ICP8 (1:1,000) mouse monoclonal antibodies (Goodwin Institute for Cancer Research, Plantation, FL).

Total RNA extraction and RT-PCR analysis.

The total RNA extraction and real-time PCR (qRT-PCR) analysis procedures were described elsewhere (24). Briefly, total RNA was extracted with the aid of TRIzol reagent (Invitrogen) according to the manufacturer's instructions. DNase treatment (Promega), phenol-chloroform-isoamyl alcohol (25:24:1) extraction (Ambion), and ethanol precipitation (Fisher Scientific) were done to remove possible DNA contamination. First-strand cDNA synthesis using 2 μg of total RNA and oligo(dT) was done with the SuperScriptIII first-strand synthesis system for RT-PCR (Invitrogen), according to the suppliers. Equal volumes from the cDNA synthesis mixtures were used for quantification of the viral gene ICP0, ICP4, ICP27, and ICP8 transcripts with the SYBR GreenER qPCR SuperMix Universal (Invitrogen), according to the manufacturer's instructions. Samples without RT-PCR served as controls. The primers for ICP0 were 5′AAGCTTGGATCCGAGCCCCGCCC 3′ (forward) and 5′AAGCGGTGCATGCACGGGAAGGT 3′ (reverse), those for ICP4 were 5′ GACGTGCGCGTGGTGGTGCTGTACTCG 3′ (forward) and 5′ GCGCACGGTGTTGACCACGATGAGCC 3′ (reverse), those for ICP27 were 5′ GTGCAAGATGTGCATCCACCACAACCTGCC 3′ (forward) and 5′ GCCAGAATGACAAACACGAAGGATGCAATG 3′ (reverse), and those for ICP8 were 5′ GACATTACGTTCACGGCCTTCGAAGCCAG 3′ (forward) and 5′ GGCCGAGTTGGTGCTAAATACCATGGC 3′ (reverse). All the transcripts were normalized to the 18S rRNA levels. The 18S rRNA primers (Universal Primers from Ambion) have been modified according to the suppliers to be suitable as an internal control for mRNA species of any abundance. The assays were performed on a StepOnePlus system (Applied Biosystems) and analyzed with software provided by the supplier.

Immunofluorescence.

The immunofluorescence procedures were described elsewhere (21, 22). Briefly, the cells were fixed in 4% paraformaldehyde at times indicated in the Results section; permeabilized; blocked with PBS-TBH solution consisting of 0.1% Triton X-100 in PBS, 10% horse serum, and 1% bovine serum albumin (BSA); and reacted with primary antibodies diluted in PBS-TBH. The anti-ICP0 exon 2 rabbit polyclonal antibody was used at a 1:2,000 dilution (25). The ICP4, ICP8, and ICP27 mouse monoclonal antibodies (Goodwin Institute for Cancer Research, Plantation, FL) were used at a 1:1,000 dilution. The cdk4 rabbit polyclonal antibody and the cdk6 mouse monoclonal antibody (Santa Cruz Biotechnology) were used at a 1:500 dilution. The four-well cultures were then rinsed several times with PBS-TBH and reacted with Alexa-Fluor-594-conjugated goat anti-rabbit or Alexa-Fluor-488-conjugated goat anti-mouse, diluted 1:1,000 in PBS-TBH. After several rinses, first with PBS-TBH and then with PBS, the samples were mounted and examined with a Zeiss confocal microscope equipped with software provided by Zeiss. For the quantification studies ≈350 cells were counted. The cdk4 inhibitor I (Calbiochem) was added at the moment of infection at a final concentration of 2.5 μM.

RESULTS

cdk4 relocalizes to nuclear compartments that recruit ICP4 and ICP8.

In recent studies we demonstrated that early in infection the distribution of cdk4 in the nucleus is similar to that of ICP0 whereas later in infection its distribution is similar in appearance to that of ICP8 in that it fills most, but not all, of the nucleus (23). The viral function that requires cdk4 redistribution was largely unknown. The objective of these series of experiments was to identify a viral partner that triggers the recruitment of cdk4 to the characteristic nuclear structures. For this purpose we examined Vero cells 7 h after exposure to HSV-1(F) or to mutants lacking the genes encoding ICP0 (ΔICP0) or ICP4 (ΔICP4) or the mutant d301 kindly furnished by D. Knipe (Harvard Medical School) and known to be deficient in the function of ICP8. The results shown in Fig. 1 may be summarized as follows.

1. In mock-infected cells, cdk4 was confined to the cytoplasm (Fig. 1p).

2. Consistent with our earlier report, in wild-type virus-infected cells cdk4 colocalizes with ICP8 in globular structures that fill most, but not all, of the nucleus (Fig. 1m to o). These structures are characteristic of early replication compartments (33, 34). Ultimately they coalesce and partially fill the nucleus (33).

3. ICP8 was present in a small fraction of cells in cultures infected with ΔICP0 or ΔICP4 mutant viruses. In these cells cdk4 was present in both nucleus and cytoplasm (Fig. 1a to c and d to f). The nuclear cdk4 was present largely in small granules or coarse dust dispersed in the nucleus.

4. In d301 mutant virus-infected cells, cdk4 was present mostly in the nucleus (Fig. 1g to l). In the nucleus cdk4 was present in a diffuse form but also in small granular structures that coincided with similar structures formed by ICP4 (Fig. 1g to i). The lack of a functional ICP8 did not interfere with the localization of cdk4 to the nucleus, but the translocation was not as efficient as that seen in wild-type virus-infected cells.

FIG. 1.

cdk4 localizes in nuclear compartments that recruit ICP4 and ICP8. Vero cells were either mock infected (p) or exposed to HSV-1(F) (10 PFU/cell) (m to o), ΔICP0 (d to f), d301 (g to l), or ΔICP4 mutants (a to c). The cells were fixed at 7 h after infection and doubly reacted with the rabbit polyclonal antibody to cdk4 and either the mouse monoclonal antibody to ICP8 (a to f and j to o) or the mouse monoclonal antibody to ICP4 (g to i). The images were captured with the same settings of a Zeiss confocal microscope.

These results indicate that efficient relocation of cdk4 appears to require ICP0, ICP4, and a functional ICP8. Whereas in the absence of ICP0 cdk4 localized in structures identifiable on the basis of the presence of ICP8 as replication compartments, in the absence of ICP8 the nuclear cdk4 was diffuse throughout the nucleus and also present in small dense structures containing ICP4. In contrast, in the absence of ICP4, cdk4 formed granular structures dispersed in both the nucleus and the cytoplasm and did not coincide specifically with the highly dispersed nuclear ICP8 (Fig. 1a to c).

GST-cdk4 pulls down independently ICP0 and ICP4.

In light of the evidence that efficient relocalization of cdk4 into the nucleus requires both ICP4 and ICP0, we examined the interaction of cdk4 with these proteins.

In these series of experiments GST-cdk4 chimeric protein was prepared as described in Materials and Methods and loaded on glutathione beads. Preliminary experiments suggested that the GST-cdk4-loaded beads pulled down both ICP4 and ICP0 (data not shown). To identify the interacting proteins more accurately, HEp-2 cells were either mock infected or exposed to 10 PFU of ΔICP0, ΔICP4, or d301 mutant virus per cell. The cells were harvested 9 h after infection, lysed, and reacted with beads loaded with GST or GST-cdk4. After overnight reaction, the beads were collected and rinsed. The bound proteins were eluted, subjected to electrophoresis on denaturing gels, and reacted with antibody to ICP4 or ICP0. The results shown in Fig. 2 were that GST-cdk4 pulls down independently both ICP0 (Fig. 2C, lane 9) and ICP4 (Fig. 2C, lane 8). The amount of protein pulled down reflected the amount of protein available in the lysates (compare, for example, the relative amounts of ICP4 or ICP0 in Fig. 2B with the amounts pulled down in Fig. 2C).

FIG. 2.

GST-cdk4 pulls down independently ICP0 and ICP4. Purified GST or GST-cdk4 proteins immobilized on beads were incubated with equal amounts of lysates from HEp-2 cells that were mock infected or infected with the wild-type virus or the deletion mutant ΔICP0, ΔICP4, or ΔICP8. (A) Ponceau S staining of the bound proteins. (B) Viral proteins present in the lysates of HEp-2 cells. (C) Immunoblots of the GST- or GST-cdk4-bound proteins eluted from the glutathione beads. The ICP4 and ICP0 mouse monoclonal antibodies were used as indicated in Materials and Methods.

The finding that cdk4 interacts with both ICP0 and ICP4 independently may explain the observation that cdk4 was relocated to the nucleus, albeit inefficiently in cells infected with either ΔCIP0 or ΔICP4 mutant viruses. We should also note that earlier this laboratory reported on the interaction of ICP0 with cyclin D3 (25).

GST-cdk4 pulls down ICP8.

Since the subcellular distribution of cdk4 in wild-type virus-infected cells appeared to coincide with that of ICP8, in that initially it formed small globular structures that later filled most, but not all, of the nucleus, our goal was to examine whether the chimeric GST-cdk4 protein could also pull down ICP8. The experimental design was similar to that described above. Thus, lysates from mock-infected Vero cells or cells that were harvested 9 h or 24 h after exposure to the wild-type virus were reacted with beads loaded with GST or GST-cdk4 proteins. After overnight reaction, the beads were rinsed several times and the bound proteins were eluted, subjected to electrophoresis on denaturing gels, and reacted with antibody to ICP8. The results shown in Fig. 3 indicate that GST-cdk4 but not GST alone pulled down ICP8 from lysates of cells harvested 9 or 24 h after infection with wild-type virus. The results indicate that cdk4 in the nucleus of infected cells is a component of a multiprotein complex that also contains ICP8.

FIG. 3.

GST-cdk4 pulls down ICP8. Purified GST or GST-cdk4 proteins as described above were incubated with equal amounts of lysates derived from Vero cells that were mock infected or infected with the wild-type virus and harvested at 9 (lanes 5 to 8) or 24 (lanes 1 to 4) h after infection. (Top) The electrophoretically separated protein complexes bound to beads were probed with monoclonal antibody to ICP8. (Bottom) Ponceau S staining of the protein-bound complexes eluted from the beads.

Inhibition of cdk4 enriches the cultures with cells that express only the ICP0 protein.

In an earlier report we showed that in cell cultures infected with wild-type virus and exposed to cdk4 inhibitor I ICP0 was retained in the nucleus (23). The effect is reversed in cells infected with a mutant virus that incorporates and overexpresses cyclin D3 (23). Here we report that in addition to the retention of ICP0 in the nucleus some ICP0-positive cells lacked demonstrable amounts of other α proteins. In these series of experiments HEp-2 cells treated with the cdk4 inhibitor I and exposed to either the wild-type virus or the virus that overexpresses cyclin D3 were analyzed by immunofluorescence at 10 h after infection. The results shown in Fig. 4 may be summarized as follows.

1. In cells infected with R7801 ICP4 was present in the cytoplasm both in mock-treated and in drug-treated cells (Fig. 4g and j).

2. In the absence of the cdk4 inhibitor I, all the cells infected with either of the viruses express concurrently ICP0 and ICP4 or ICP0 and ICP27 and approximately 80% of the ICP0-expressing cells have ICP0 in the cytoplasm [for HSV-1(F), Fig. 4A, subpanels a to c and m to o, and Fig. 4B, lanes 1, 5, 9, and 13; for R7801, Fig. 4A, subpanels g to i and s to u, and Fig. 4B, lanes 3, 7, 11, and 15).

3. Inhibition of cdk4 in wild-type virus-infected cells resulted in an enrichment of the cultures with cells that express only ICP0. Thus, quantification analysis for the concurrent expression of ICP0 and ICP4 shows that only 56% of the ICP0-expressing cells have detectable amounts of ICP4, while the analysis for concurrent expression of ICP0 and ICP27 shows that approximately 20% of the ICP0-positive cells have detectable amounts of ICP27. At the same time only 8% of the HSV-1(F)-infected cells contain ICP0 in the cytoplasm (Fig. 4A, subpanels d to f and p to r, and Fig. 4B, lanes 2, 6, and 10).

4. Inhibition of cdk4 in cells infected with the R7801 mutant virus that overexpresses cyclin D3 largely reversed both phenomena observed with the wild-type virus. Thus, approximately 95% of the ICP0-expressing cells have detectable amounts of ICP4 and approximately 90% contain detectable ICP27. Moreover, of the cells that express ICP0, approximately 90% contain this protein in the cytoplasm (Fig. 4A, subpanels j to l and v to x, and Fig. 4B, lanes 4, 8, and 12). In both untreated and treated infected cells, ICP27-expressing cells contained demonstrable ICP0 (Fig. 4B, lanes 13 to 16).

FIG. 4.

Concurrent presence of ICP0 and ICP4 or ICP0 and ICP27 proteins in untreated and cdk4 inhibitor I-treated infected cells. (A) HEp-2 cells were mock treated or exposed concurrently to cdk4 inhibitor I (2.5 μΜ) and either HSV-1(F) or R7801 (10 PFU/cell). The cells were fixed at 10 h after infection and reacted with the rabbit polyclonal antibody to ICP0 and either the mouse monoclonal antibody to ICP4 or the mouse monoclonal antibody to ICP27. The images were captured with the same settings of a Zeiss confocal microscope. (B) Approximately 350 cells from sequential fields were counted to determine the percentage of ICP0-positive cells with detectable ICP4 (lanes 1 to 4) or ICP27 (lanes 5 to 8) and the percentage of ICP0-positive cells with cytoplasmic ICP0 (lanes 9 to 12). The percentage of ICP27-positive cells with detectable ICP0 is shown in lanes 13 to 16.

The results presented here show that inhibition of cdk4 delays the efficient concurrent expression of the α proteins in the wild-type virus-infected cells. Overexpression of cyclin D3 largely reverses the effect of the drug.

Overexpression of cyclin D3 alters the pattern of transcription of viral genes.

As noted in the introduction, it has been reported that roscovitine, a broad cdk inhibitor, blocks the accumulation of viral mRNAs in cells infected with wild-type virus (8). In the section above, we reported that the accumulation of some viral proteins is at the very least delayed relative to the accumulation of ICP0 in the presence of the cdk4 inhibitor I and that this effect is at least partially reversed by overexpression of cyclin D3. The objective of the studies described in this section was to define the effect of overexpression of cyclin D3 on accumulation of mRNAs encoding ICP0, ICP4, ICP27, and ICP8 in wild-type and R7801 mutant virus-infected cells mock treated or exposed to CDK4 inhibitor I. In these series of experiments the cells were exposed (10 PFU per cell) to HSV-1(F) or R7801—a mutant that encodes and overexpresses cyclin D3. The cells were exposed concurrently to both drug and virus (10 PFU/cell). Replicate cultures were harvested at 30 min or 1, 2, 3, 4, 5, or 7 h after a 40-min adsorption time. The RNA was purified and processed for real-time PCR measurements as described in Materials and Methods. The results were normalized with respect to 18S RNA and shown as threshold cycle (ΔC_T). It should be noted that a change of C_T equal to 1 represents a 2-fold change in the amount of mRNA. The results shown in Fig. 5 were as follows.

1. In untreated infected cells, the overall patterns of accumulation of ICP4 (Fig. 5A), ICP0 (Fig. 5B), or ICP27 (Fig. 5C) mRNAs were similar. The accumulation of the mRNA transcripts continued past 4 h after infection at rates reduced relative to those prior to 3 to 4 h after infection. The rate of accumulation of ICP8 mRNA (Fig. 5D) did not show a marked increase until after 2 h of infection.

2. In cdk4 inhibitor I-treated cells, there was a marked decrease in the accumulation of all mRNAs. During the first 3 h after infection, ICP27 and ICP4 mRNAs continued to accumulate at a reduced rate (Fig. 5A and C), whereas the accumulation of ICP0 mRNA (Fig. 5B) was virtually arrested. Past 3 to 4 h after infection, ICP0 mRNA accumulated at an accelerated rate that was not matched by either ICP27 or ICP4 mRNAs. The accumulation of ICP8 mRNA was essentially flat until at least 5 h after infection.

3. The accumulation of all mRNAs was greatly accelerated in all cells infected with the R7801 mutant virus. While the largest amounts of mRNAs attained during the 7 h of infection were similar to those attained at 7 h after infection in wild-type virus-infected cells, they were attained at a much earlier time point. Thus, the highest levels of ICP4 mRNA were attained between 2 and 4 h after infection. Both ICP0 and ICP27 mRNAs reached their highest levels at about 4 h after infection, whereas ICP8 mRNA reached the highest levels at 5 h after infection. In all instances the levels of mRNAs declined past 4 to 5 h after infection, suggesting that at these times postinfection the rate of degradation of the mRNAs exceeded the rate of their synthesis.

4. In cells infected with R7801 and treated with cdk4 inhibitor, the rate of accumulation of mRNAs decreased. Interestingly, the pattern of accumulation and the amounts of accumulated mRNAs resembled those of untreated cells infected with wild-type virus.

FIG. 5.

The accumulation of α and β gene mRNAs in mock-treated or inhibitor I-treated infected cells. HEp-2 cells were exposed to HSV-1(F) or R7801 at 10 PFU/cell in the presence or absence of the cdk4 inhibitor I. The cells were harvested at 0.5, 1, 2, 3, 4, 5, and 7 h after 40 min of adsorption of virus to cells. Total RNA was extracted, reverse transcribed, and used as a template for the quantification of the ICP4, ICP0, ICP27, or ICP8 transcripts accumulating in the cells in the absence or the presence of the drug. All the data were normalized to the 18S rRNA. In each plot the C_T values of the viral mRNAs were subtracted from the respective C_T values of the 18S rRNA and the ΔC_T values are plotted over the time of the experiment.

The accumulation of viral proteins in mock-treated or infected cells exposed to cdk4 inhibitor I reflects the corresponding accumulation of viral gene transcripts.

The sections above showed the presence of cells that contain ICP0 but lack detectable ICP4 or ICP27 in cultures exposed to wild-type virus and cdk4 inhibitor I. Moreover, the cdk4 inhibitor I selectively reduced the accumulation of transcripts of viral genes and this effect was at least partially reversed by overexpression of cyclin D3. In this section we sought to determine whether the accumulation of selected α, β, and γ proteins in drug-treated cells reflects the accumulation of viral mRNAs. In these series of experiments HEp-2 cells were mock infected or exposed to 10 PFU of wild-type or R7801 virus per cell. In treated cultures the cdk4 inhibitor I was added to the medium at the time of exposure to the virus. At 4, 9, or 18 h after exposure to virus, replicate cultures were harvested and the cells were lysed, subjected to electrophoresis on denaturing gels, transferred to nitrocellulose sheets, and reacted with antibodies to β-actin, ICP0, ICP4, ICP8, U_S3, or U_S11. The results (Fig. 6) were as follows.

1. At 4 h after infection in the absence of the drug, the cells infected with R7801 mutant virus accumulated larger amounts of ICP0 than did HSV-1(F)-infected cells (on the ICP0 blot, compare lanes 3 and 5). At the same time the R7801-infected cells accumulated detectable levels of ICP4 whereas none could be detected in wild-type virus-infected cells (on the ICP4 blot, compare lanes 3 and 5). Consistent with the quantitative PCR (qPCR) data during the course of the infection, the R7801 mutant virus-infected cells accumulated larger amounts of all proteins tested and appeared to switch at earlier times to β or γ gene expression than did the wild-type virus-infected cells (in all blots, compare lanes 7 and 9).

2. The cdk4 inhibitor I had a drastic effect on the accumulation of viral proteins in cells infected with wild-type virus. Thus, at 9 h after infection wild-type virus-infected cells accumulated barely detectable ICP4 and virtually no detectable ICP8 or other β or γ proteins tested (in all blots, compare lane 8 to lane 7), even though they contained elevated levels of ICP0.

3. In contrast the virus that overexpresses cyclin D3 largely bypassed the inhibitory effect of the drug and the α proteins accumulated in amounts sufficient to drive β and γ gene expression to levels comparable to those in untreated wild-type virus-infected cells (in all blots, compare lane 10 to lane 7 and lane 9).

FIG. 6.

Accumulation of viral proteins in mock-treated and cdk4 inhibitor I-treated cells. Replicate cultures of HEp-2 cells were mock infected or exposed either to HSV-1(F) or to R7801 at 10 PFU/cell. The cdk4 inhibitor I was added to the cultures as described above. The cells were harvested at 4, 9, or 18 h after infection and lysed, and approximately 50 μg of total proteins was electrophoretically separated in a 10% polyacrylamide gel, transferred to nitrocellulose sheets, and reacted with the β-actin, ICP4, ICP0, ICP8, or U_S11 mouse monoclonal antibodies or the U_S3 rabbit polyclonal antibody (29).

The results obtained in these series of experiments are consistent with the qPCR data and suggest that the cdk4 inhibitor I selectively reduces the accumulation of α mRNAs. The effect is more pronounced on the accumulation of ICP4 and ICP27 mRNAs than on that of ICP0 mRNA, but the net effect is a drastic reduction in the accumulation of β and γ proteins. Overexpression of cyclin D3 overcomes this inhibition.

cdk6 is recruited in the nucleus of the cells infected with the wild-type virus that overexpresses cyclin D3, both in the presence and in the absence of the cdk4 inhibitor I.

In this section we examined the effect of overexpression of cyclin D3 on the accumulation of cdk6 in infected cells. The rationale for carrying out this study was based on the hypothesis that overexpression of cyclin D3 may activate cdk6, which in turn would replace the function of cdk4 in infected cells treated with the cdk4 inhibitor I. If the hypothesis was tenable, we would expect that cdk6 would colocalize with the replication compartments in a manner analogous to that of cdk4. To this aim we examined the subcellular localization of cdk6. The results, shown in Fig. 7, were as follows.

1. In mock-infected cells in the absence of the drug, cdk6 shows predominantly nuclear localization. In the presence of the drug, cdk6 became predominantly cytoplasmic and only a small amount was detected in the nucleus.

2. In the wild-type virus-infected untreated cells, a weak signal for cdk6 was observed in both the nucleus and the cytoplasm. In the presence of the cdk4 drug the cdk6 from HSV-1(F)-infected cells resembled in localization the cdk6 of drug-treated mock-infected cells in that it was retained mainly in the cytoplasm.

3. In the R7801 virus-infected cells, a significant amount of cdk6 accumulated in the nucleus and the staining from the cytoplasm was weak. In the presence of the cdk4 drug, R7801 mutant virus-infected cells contained densely packed, punctate cdk6 in the nucleus. The appearance of cdk6 in these cells is strikingly different from that of cdk6 in wild-type virus-infected cells.

FIG. 7.

cdk6 is recruited to nuclei in R7801-infected cells exposed to the cdk4 inhibitor I. HEp-2 cells exposed to HSV-1 or R7801 virus in the presence or the absence of the cdk4 inhibitor I were fixed at 6 h after infection and reacted with the mouse monoclonal antibody to cdk6 and the rabbit polyclonal antibody to ICP0. The images were captured with the same settings of a Zeiss confocal microscope.

The apparent recruitment of cdk6 to the nucleus in cells that overexpress cyclin D3 is consistent with the hypothesis that in these cells cdk6 can replace cdk4 in the presence of cdk4 inhibitor I.

DISCUSSION

A fundamental hypothesis that emerges from the studies reported here and those reported earlier (23, 25, 36) is that ICP0 encodes a function that enhances the recruitment of cyclin D3 to ND10 structures to activate cdk4 and that ICP0 along with other viral proteins recruits cdk4 to ND10 structures and ultimately to replication compartments for transcription of viral genes. Since ICP0 is retained in the nucleus in cells exposed to the inhibitor of cdk4, a corollary of this hypothesis is that accumulation of ICP0 in the cytoplasm is related to an event in which cdk4 is a key participant. The salient points that support these hypotheses and their implications are as follows.

1. ICP0 binds cyclin D3 (25). The role of this interaction may be deduced from the observation that insertion of the genes encoding cyclin D3, D2, or D1 in place of ICP0 partially complements the ΔICP0 mutant (23). Balancing this interaction is the observation that while ICP0 recruits cyclin D3, overexpression of cyclin D2 or D1 satisfies the viral need for a D cyclin (23).

2. The key partners of D cyclins are cyclin-dependent kinases (31, 32). In infected cells cdk4 is activated whereas cdk2 is not (9, 36). The hypothesis that HSV-1 recruits cyclin D3 to activate cdk4 rests on several lines of evidence. Thus, cdk4 is recruited to ND10 structures and replication compartments by viral proteins, and activation and function of cdk4 are essential for both transcription and replication of viral DNA. We cannot exclude the possibility that cyclin D3 performs other functions not shared by cyclin D2 or D1.

3. Experiments presented here show that cdk4 inhibitor I decreases the accumulation of transcription of 3 α genes (ICP0, ICP4, and ICP27) and that of ICP8. Concurrently there is a delay in the accumulation of viral proteins and decrease in the amounts of accumulated viral DNA. ICP0 accumulates at a rate higher than that of ICP4 or ICP27. We should note that cdk4 inhibitor I, which is specific for cdk4, appears to act in a manner similar to that of roscovitine (8), a general cdk inhibitor. At this time we do not know whether cdk4 inhibitor I acts selectively on transcriptional factors unique to all α gene promoters or whether it interferes with the anti-DNA-silencing functions of ICP0. The latter hypothesis is particularly attractive since it would suggest that ICP0 is a master regulator whose synthesis precedes that of other α genes, as has been reported elsewhere (2, 3).

4. The issue directly relevant to this report is that overexpression of cyclin D3 results in accelerated transcription of viral genes. In cells infected with the mutant virus that overproduces cyclin D3, the cdk4 inhibitor I reduced the rate of transcription to a level similar to that taking place in untreated cells infected with wild-type virus. As would be expected, overproduction of cyclin D3 negates the effect of the inhibitor. How does overexpression of cyclin D3 accelerate transcription and reverse inhibition of viral DNA synthesis? The simplest explanation is that overexpression of cyclin D3 activates another cdk that performs a similar function. An obvious candidate is cdk6 (14). In this report we show that the distribution of cdk6 in wild-type virus-infected cells is not different from that of mock-infected cells. In R7801 mutant virus-infected cells, cdk6 appears to aggregate in punctate nuclear structures that could roughly outline the replication compartments. Further studies will resolve the role of cdk6, which appears to be mobilized to the nucleus of cells infected with a cyclin D3-overexpressing virus.

   It is worthy of note that PML, and by extension ND10 bodies, has been reported to exert a suppressive effect on cyclin D1 under slow-growth conditions (26). It is unclear whether the effect is on all D cyclins or only on cyclin D1. Since HSV-1 appears to benefit from cyclin D1 (23) even though it specifically recruits cyclin D3 (25), this observation adds to the evolutionary pressures to degrade PML and disperse ND10 bodies—hallmarks of HSV-1 infection.

5. Lastly, a key observation that has led us to these studies is that ICP0 is retained in nuclei of cells infected with a mutant that does not bind cyclin D3 as a consequence of a single amino acid substitution (35). In contrast, in cells infected with a mutant that overexpresses cyclin D3, the switch from nuclear to cytoplasmic accumulation takes place earlier (36) consistent with the accelerated transcription and expression of viral genes reported here. We have also reported that ICP0 is retained in the nuclei of cells infected with wild-type virus and exposed to cdk4 inhibitor I but not in cells exposed to inhibitor and infected with a virus that overexpressed cyclin D3 (23). One clue that may ultimately shed light on the connection between cdk4 and ICP0 is that the switch from nuclear to cytoplasmic accumulation is partly blocked by phosphonoacetic acid at concentrations sufficient to block viral DNA synthesis (28). The sum total of the evidence then suggests that the shift from nuclear to cytoplasmic accumulation takes place in highly discrete steps. Prior to onset of viral DNA synthesis, ICP0 is translocated to the nucleus. As recently reported, the half-life of nuclear ICP0 is relatively short (15). After the onset and accumulation of late proteins, the translocation of ICP0 to the nucleus is blocked. The evidence that cdk4 inhibitor blocks cytoplasmic accumulation of ICP0 and that this is overcome in drug-treated R7801 mutant virus-infected cells (23) is consistent with this hypothesis. The challenge is to identify the mechanism by which the translocation of ICP0 is blocked.