Mechanisms Governing Maintenance of Cdk1/Cyclin B1 Kinase Activity in Cells Infected with Human Cytomegalovirus

Cells and virus.

Human foreskin fibroblasts (HFF) were passaged and maintained in minimum essential medium (MEM)-Earle's medium containing 10% fetal bovine serum and supplemented with glutamine, penicillin, streptomycin, and amphotericin as previously described (38). Cells were grown to confluence and allowed to arrest for 3 days prior to trypsinization and replating at a lower density (49). Cells were infected at a multiplicity of infection (MOI) of 5 with the Towne strain of HCMV at the time of release from confluence (G₀ infection) or were mock infected with conditioned medium. For experiments performed in the absence of serum, confluence-synchronized cells were replated at a lower density in serum-free MEM-Earle's medium supplemented with antibiotics. HFF were infected with serum-free virus (47) at an MOI of 5 at 2 days postplating, and cultures were maintained in serum-free medium throughout the duration of the experiment.

Western blotting.

Confluence-synchronized HFF were infected at a MOI of 5 as described above. At the time points indicated, cells were trypsinized, counted, and frozen. Pellets were lysed in reducing sample buffer (RSB) (50 mM Tris [pH 6.8], 2% sodium dodecyl sulfate, 10% glycerol, 5% 2-mercaptoethanol, 25 mM sodium fluoride, 1 mM sodium orthovanadate, 5 mM β-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 50 μM leupeptin, and 100 μM pepstatin A) containing bromophenol blue and then were sonicated and boiled for 5 min. Approximately 7.5 × 10⁴ cells per lane were loaded onto polyacrylamide gels as previously described (50). Proteins were transferred to nitrocellulose. The following antibodies were used to probe filters: anti-cyclin B1 (BD Biosciences), anti-IE1/2 and anti-UL44 (the Goodwin Research Institute), anti-Cdk1 and anti-phospho-Tyr15 Cdk1 (Santa Cruz Biotechnology, Inc.), anti-Rsk1 (Santa Cruz Biotechnology, Inc.), anti-phospho-Ser381 Rsk1 (Cell Signaling Technology), anti-Wee1 (Santa Cruz Biotechnology, Inc.), anti-Cdc25B (Oncogene Research Products), anti-Cdc25C (Santa Cruz Biotechnology, Inc.), and anti-β actin (Sigma, St. Louis, Mo.). The antibody against G6PD was a generous gift from Rod Nakayama (University of California, Irvine). These primary antibodies were diluted in BLOTTO (5% nonfat milk and 0.05% Tween 20 in Tris-buffered saline [pH 7.4]). Antibodies against Myt1 were a generous gift from Robert Booher (Onyx Pharmaceuticals) and were diluted in BLOTTO containing 15% nonfat milk and 0.1% Tween 20. Horseradish peroxidase-conjugated secondary antibodies were purchased from Calbiochem and diluted 1:1,000 to 1:10,000. SuperSignal chemiluminescent reagents were purchased from Pierce and used per the manufacturer's instructions.

Inhibition of the proteasome by MG132.

G₀-arrested cells were released from confluence and simultaneously infected with HCMV Towne at an MOI of 5 or mock infected with conditioned medium as described above. At the time points indicated, mock- and virus-infected cultures were treated with 2.5 or 10 μM MG132 (Calbiochem) dissolved in dimethyl sulfoxide (DMSO) or were treated with DMSO (control) for 3 h prior to harvesting of cells. Lysates were prepared in RSB as described above, and samples were processed for Western blotting.

Cycloheximide block and release experiments.

Cycloheximide block and release experiments were performed as described by McElroy et al. (38). Briefly, G₀-synchronized HFF were released from confluence and replated at a lower density. Cycloheximide (Sigma) was added to a final concentration of 100 μg/ml at 1 h postplating. After this 1-h pretreatment, cells were mock infected or infected with HCMV Towne at an MOI of 5 in the presence or absence of 100 μg of cycloheximide per ml. The cells were washed three times at 3 h p.i. to remove the cycloheximide block and were refed with medium with or without cycloheximide. At the designated time intervals, actinomycin D (Sigma) was added to a final concentration of 20 μg/ml. All samples were harvested at 18 h p.i. Cell pellets were lysed in RSB and processed for Western blotting.

Immunofluorescence.

Confluence-synchronized HFF were seeded on glass coverslips and infected in G₀ with HCMV Towne at an MOI of 5. At the time points indicated, coverslips were washed twice in phosphate-buffered saline (PBS) and fixed in ice-cold methanol for 10 min. Immunofluorescence staining was done as previously described (50). Fixed cells were blocked in 10% normal goat serum (NGS) in PBS for 20 min at room temperature (RT) prior to incubation with primary antibodies. Antibodies were diluted in PBS containing 10% NGS as follows: anti-Cdk1 antibody (Santa Cruz Biotechnology, Inc.), 1:50; and anti-cyclin B1 antibody (BD Biosciences), 1:50. After a 30-min incubation at RT, coverslips were washed three times with PBS, for 3 min per wash. Coverslips were then incubated with fluorescein isothiocyanate- or tetramethyl rhodamine isocyanate-conjugated donkey anti-mouse isotype-specific secondary antibodies in PBS containing 10% NGS and Hoechst stain for 30 min at RT. Coverslips were washed three times in PBS, for 3 min per wash, prior to mounting onto slides with SlowFade Light (Molecular Probes) mounting medium to prevent photobleaching. Images were captured by using MetaMorph Software (Universal Imaging Corporation, Downingtown, Pa.) and were processed with Adobe Photoshop.

Biochemical fractionation.

Confluence-synchronized HFF cells were infected with HCMV Towne at an MOI of 5. At 24 and 48 h p.i., cells were digested with trypsin and counted. Fractionation was conducted as previously described (55). Cell pellets were resuspended on ice in isotonic lysis buffer to a concentration of 10⁶ cells/50 μl. Digitonin was added to samples to a final concentration of 0.2 mg/ml. Samples were incubated on ice for 15 min with frequent vortexing. Nuclei were pelleted by spinning at 3,000 rpm (700 × g) for 5 min. The supernatant (fraction 1) was carefully removed from the pellet, which was washed two times with isotonic lysis buffer. Fraction 1 was further clarified by centrifugation at 5,000 rpm for 10 min. The clarified supernatant was collected and represents the soluble cytosolic fraction (C). The nuclear pellet was resuspended in 50 μl of isotonic lysis buffer and layered on 50 μl of 37% sucrose. Nuclei were spun through the cushion at 5,000 rpm (2,000 × g) for 10 min. The supernatant was collected and the sucrose was carefully removed from the pellet. The pellet was washed once with isotonic lysis buffer. The pellet was then extracted with digitonin (2 mg/ml) in isotonic lysis buffer for 5 min on ice with frequent vortexing. Insoluble material was pelleted at 3,000 rpm for 5 min. The supernatant was collected and the pellet was washed twice in isotonic lysis buffer. The pellet represents the detergent-resistant fraction containing predominantly the nuclei and other insoluble material (N). The fractions were mixed with an equal volume of 2× RSB to make 1× RSB. Approximately equal cell numbers were loaded onto sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels.

In vitro kinase assays.

In vitro kinase assays were performed as previously described (24). Cell pellets containing 3 × 10⁵ cells were solubilized on ice in lysis buffer (10 mM Tris [pH 7.4], 5 mM EDTA, 130 mM NaCl, 10 mM NaH₂PO₄, 1 mM dithiothreitol, 1% Triton X-100, 25 mM sodium fluoride, 1 mM sodium orthovanadate, 5 mM β-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 50 μM leupeptin, and 100 μM pepstatin A) for 30 min. Clarified supernatants were precleared with protein G beads (Santa Cruz Biotechnology, Inc.) for 1 h. Ten micrograms of antibody against cyclin B1 (clone GNS1; Lab Vision) coupled to protein G beads was incubated with lysates for 4 h at 4°C. Immunoprecipitates were washed two times with lysis buffer and split into two samples, one for Western blotting and the other for kinase assays. Immunoprecipitates were washed once in kinase assay buffer (20 mM HEPES [pH 7.4], 10 mM MgCl₂). In vitro kinase assays were performed at RT for 30 min in kinase assay buffer containing 1 mM dithiothreitol, 10 μM ATP, 5 μg of histone H1 (Upstate USA, Inc., Charlottesville, Va.), and 2 μCi of [³²P]ATP. Reactions were stopped by adding an equal volume of 2× RSB. Samples were resolved in SDS-12% PAGE gels. For Western blots, samples were resolved on SDS-10% PAGE gels and processed as described above. Horseradish peroxidase-linked isotype-specific secondary antibodies (Southern Biotechnology Associates, Inc., Birmingham, Ala.) were used for blots probed with anti-Cdk1 antibody to prevent detection of light chains of immunoglobulin.

Steady-state levels of cyclin B1 are maintained in HCMV-infected cells in the absence of serum.

Previous work from our laboratory described the induction of cyclin B1 and its associated kinase activity in human fibroblasts infected with HCMV (24). Those experiments were performed in cells that were previously synchronized by serum starvation and given serum stimulation at the time of infection. In order to determine if the effect on cyclin B1 accumulation was a direct effect of viral infection or a result of serum stimulation, we assessed the expression of cyclin B1 in cells infected with HCMV in the absence of serum. Cells were grown to confluence and allowed to arrest for 3 days before replating at a lower density in the absence of serum. Two days later, fibroblasts were infected at an MOI of 5 with the Towne strain of HCMV prepared in serum-free medium. As shown in Fig. 1, cyclin B1 accumulated in HCMV-infected fibroblasts in the absence of serum stimulation as early as 24 h p.i. Minimal cyclin B1 expression was detected in mock-infected cells. These data indicated that cyclin B1 accumulation was not an artifact of the experimental design but was induced by viral infection. All subsequent experiments were performed in medium containing 10% fetal bovine serum with confluence-synchronized cells.

Accumulation of cyclin B1 is controlled at the levels of both synthesis and degradation.

In earlier studies, we had obtained evidence that suggested that cyclin B1 accumulated earlier in infected cells than in uninfected cells. The kinetics of cyclin B1 expression were determined by Western blotting. Figure 2A shows the fluctuation of cyclin B1 expression over time in mock- and virus-infected cells. In mock-infected cells, a low level of cyclin B1 was detectable in G₁ through 10 h p.i. The level then dropped slightly at 14 h p.i., consistent with degradation of cyclin B1 mediated by the E3 ubiquitin ligase APC. As the cells leave G₁ phase, the levels of cyclin B1 begin to rise again (Fig. 2A, 18 h p.i.). The expression of cyclin B1 was highest at 24 h p.i. in mock-infected samples, consistent with S phase. The levels dropped slightly by 48 h p.i. and disappeared by 96 h p.i., when mock-infected cells reached confluence (Fig. 2A and data not shown). In contrast, high levels of cyclin B1 were detected as early as 10 h p.i. in virus-infected cells (Fig. 2A). In addition, the steady-state level of cyclin B1 in HCMV-infected cells did not fluctuate significantly at late times during the infection, and cyclin B1 was detected as late as 96 h p.i. As a loading control, lanes containing equivalent cell numbers were probed with an antibody against β-actin, which does not fluctuate during the cell cycle.

Previous work from our laboratory demonstrated that expression of cyclin B1 mRNA was slightly increased in HCMV-infected cells at 12 h p.i.; however, this slight increase was not sufficient to account for the early appearance and accumulation of cyclin B1 protein (49). We inferred that another mechanism might also contribute to the accumulation of cyclin B1 protein at early times during infection. Because the level of cyclin B1 during G₁ phase of uninfected cells is regulated by ubiquitination catalyzed by the APC E3 ubiquitin ligase and proteasome-mediated degradation, we examined whether the accumulation of cyclin B1 in HCMV-infected cells early in the infection might be the result of decreased degradation. G₀-synchronized human fibroblasts were infected with HCMV Towne at an MOI of 5 as for the previous experiments. Three hours before harvesting at 18 and 21 h p.i., cultures were treated with the proteasome inhibitor MG132 dissolved in DMSO or with DMSO alone as a control. As shown in Fig. 2B, the addition of proteasome inhibitor to mock-infected cultures had a marked effect on the expression of cyclin B1 at 18 and 21 h p.i. In contrast, the addition of MG132 had a less significant effect on the expression of cyclin B1 in virus-infected cells (Fig. 2B). At 18 h p.i., the inhibition of the proteasome resulted in a moderate increase in the level of cyclin B1 in the infected cells, but by 21 h p.i., MG132 had no effect on cyclin B1 accumulation. These data suggest that the increase in cyclin B1 observed for HCMV-infected cells is the result of both increased transcription and decreased proteasome-mediated degradation.

Viral early gene expression is required for the accumulation of cyclin B1.

We next sought to determine what stage of viral gene expression was necessary for the induction of cyclin B1 expression. To do this, we performed cycloheximide block and release experiments as previously described (38). Briefly, confluence-synchronized cells were infected in the presence of cycloheximide. At 3 h p.i., cells were released from the cycloheximide block, and at various times postrelease, actinomycin D was added to prevent further transcription. All cells were harvested at 18 h p.i. As shown in Fig. 3, the earliest induction of cyclin B1 in the HCMV-infected cells was observed when a 4-h window was allowed between cycloheximide release and the addition of actinomycin D (ActD lane 7-18). At this same time point, we can detect immediate-early (IE) gene expression as well as a very small amount of the HCMV DNA polymerase accessory protein UL44. Cyclin B1 expression was not detected at earlier time points despite the presence of IE proteins, indicating that cyclin B1 is not induced with the kinetics of an IE protein. When a 10-h time interval was allowed between the removal of cycloheximide and the addition of actinomycin D (Fig. 3, ActD lane 13-18), cyclin B1 levels declined in the mock-infected cells but continued to increase in the HCMV-infected cells. Taken together, these data suggest that the induction of cyclin B1 in virus-infected cells occurs at early times during infection and that sustained accumulation likely requires some viral early gene expression.

Cdk1 accumulates in HCMV-infected cells in its active form.

We also examined the expression of Cdk1 in HCMV-infected cells. We found that expression of Cdk1 fluctuated throughout the cell cycle in mock-infected cells (Fig. 4A). At 4 h p.i., there was no detectable Cdk1 in mock or viral samples. By 24 h p.i., Cdk1 was present at high levels in the mock samples. The expression of Cdk1 declined by 96 h p.i. in mock-infected cells as they became confluent, comparable to what was observed for cyclin B1 (Fig. 4A and data not shown). In contrast, the steady-state level of Cdk1 appeared to increase over time in HCMV-infected cell samples. In addition, we detected predominantly the faster migrating form of Cdk1, which represents the active form of Cdk1 that has not been modified by Myt1 or Wee1 (Fig. 4A, compare lane 24 M to lane 24 V). The decrease in Cdk1 Tyr15 phosphorylation in viral samples was confirmed by probing Western blots with an antibody against the phosphorylated Tyr15 epitope on Cdk1 (Fig. 4B). Note that this antibody cross-reacts with a similar epitope on Cdk2, which runs slightly faster than Cdk1 on these gels. We observed that overall the levels of the phosphorylated proteins were low in viral samples at all time points; however, there was some enrichment of the pTyr15 form of Cdk1 in immunoprecipitates of cyclin B1 from virus-infected cells (Fig. 4B and C). These results were consistent with our earlier findings showing that Cdk1/cyclin B1 complexes were active in HCMV-infected cells throughout the infection (24).

To determine the relative activity of the Cdk1/cyclin B1 complexes in the mock- and HCMV-infected cells, we subjected whole-cell lysates to immunoprecipitation with an antibody against cyclin B1 under conditions in which cyclin B1 was completely depleted from the lysate (data not shown). The precipitates were then assessed for the ability to phosphorylate histone H1 in an in vitro kinase assay. Surprisingly, histone H1 kinase assays revealed that the highest levels of cyclin B1-associated kinase activity were not observed for the HCMV-infected cells until the late phases of the infection (Fig. 4C). The active form of Cdk1 was coprecipitated with cyclin B1 in the viral samples at each of the time points tested (denoted by dots in Fig. 4C), and there was an increase in histone H1 kinase activity over time in the viral samples. In mock-infected cells, the levels of cyclin B1-associated kinase activity fluctuated over the cell cycle and peaked at 48 h p.i. At this time point, there was a detectable level of the active form of Cdk1 in immunoprecipitates from mock-infected cells; however, because the relative abundance of the active form of Cdk1 does not correlate with the levels of kinase activity in the mock and viral samples at 48 h p.i., these data suggest that there may be an additional level of regulation that controls cyclin B1-associated kinase activity in HCMV-infected cells.

Down-regulation of the Myt1 kinase pathway.

The detection of the active form of Cdk1 in virus-infected cells prompted us to investigate if the alteration of the Myt1 kinase inhibitory pathway might be one mechanism underlying the activation of Cdk1. Previous work from our laboratory showed that the ERK kinases were up-regulated shortly after infection of cells with HCMV (47). The activity of one of the substrates that are positively regulated by the ERKs, p90^Rsk1, was also increased. This was of particular interest because p90^Rsk1 phosphorylates Myt1 and inhibits its activity (45), consistent with the data shown in Fig. 4. In mock-infected cells, the steady-state level of p90^Rsk1 did not fluctuate significantly throughout the cell cycle (Fig. 5A). In contrast, there was a reproducible decrease in the level of p90^Rsk1 in the viral samples at 8 h p.i. relative to the control samples. The levels of total p90^Rsk1 then increased by 24 h p.i., and the protein accumulated to high levels late in infection. A very different pattern of expression was observed when the Western blots were probed with an antibody specific for p90^Rsk1 phosphorylated at Ser381, which is the active form of the kinase (Fig. 5B) (10). Although the steady-state level of total p90^Rsk1 was lower at 8 h p.i. in the viral sample, the level of active, phospho-Ser381-p90^Rsk1 was higher in the viral lysate than in the mock sample at the same time point. Interestingly, later in the infection, the levels of phospho-Ser381-p90^Rsk1 dropped in the viral samples while the total amount of p90^Rsk1 greatly increased (compare Fig. 5A and B). These results paralleled observations by Rodems and Spector that high levels of p90^Rsk1 activity appeared at early times but were not maintained at late stages of the infection (47).

Based on the above observations, phosphorylation of the Myt1 kinase and its subsequent inactivation should be more efficient in HCMV-infected cells at early times in the infection; however, since the levels of active p90^Rsk1 were similar in the mock and viral samples at 24 h p.i., the inactivation of the Myt1 kinase might be comparable at this time point. To determine if there were any differences in the total amount of Myt1 present, we performed Western blot analyses of the steady-state levels of Myt1. As shown in Fig. 5C, the levels of Myt1 in both the mock- and HCMV-infected cells were very low at 8 h p.i. The levels then increased in the mock and viral samples, but the levels in the infected cells were significantly lower than that in mock-infected controls at 24 h p.i. and later time points. These results suggested that the Myt1 pathway for inhibition of Cdk1 activity was compromised in virus-infected cells as a result of high levels of p90^Rsk1 at early times during the infection and reduced expression of Myt1 throughout the infection.

Alteration of Wee1 expression.

Since inhibitory phosphates can be added to Cdk1 by both Myt1 and Wee1, we also proceeded to determine whether viral down-regulation of Wee1 might be another mechanism contributing to the activation of Cdk1. The expression of this protein was reported to be cell cycle regulated (39, 56), consistent with the results shown in Fig. 5D. The level of Wee1 was below the limit of detection at 6 h p.i. in both mock- and virus-infected cells. In mock samples, the highest levels of Wee1 were detected at 24 h p.i., which coincides with S phase (49). The levels of Wee1 then dropped as the cells became asynchronous at 48 h p.i. and at later time points. These data were consistent with the regulated synthesis and degradation patterns of Wee1 that have previously been reported. Wee1 expression was delayed in HCMV-infected cells, and the protein was not detected in viral samples at 24 h p.i. Thus, the absence of Wee1 coupled with the low levels of Myt1 is consistent with the reduced level of the inactive phosphorylated forms of Cdk1 detected in infected cells at 24 h p.i. After 24 h p.i., we detected forms of Wee1 with altered mobilities in virus-infected cells (Fig. 5D). Two of the slower migrating forms are likely the hyperphosphorylated inactive Wee1; however, because we detected three forms instead of the previously described doublet of Wee1 (56), we do not know whether these forms are inactive. In addition, the small amount of Wee1 coupled with the lack of a high-affinity antibody precluded direct measurement of activity. Nevertheless, these results suggest that the Wee1 inhibitory pathway is down-regulated in HCMV-infected cells within the first 24 h of infection and that this is one mechanism that contributes to the activity of Cdk1/cyclin B1.

The expression of Wee1 is regulated at both the transcriptional and posttranslational levels during the cell cycle. Recent evidence suggests that Wee1 is targeted for degradation by Tome-1, which acts as part of the Skp1-Cullin-F-box protein (SCF)ubiquitin ligase (reviewed in references 30 and 32; also see reference 3). Tome-1 itself is subject to proteasome-mediated degradation and is a substrate for the APC during early G₁ of the cell cycle. The destruction of Tome-1 allows the accumulation of Wee1 in late G₁ phase to ensure that Cdk1 is not activated prematurely. As shown in Fig. 2, cyclin B1 accumulated in infected cells at early times. Since cyclin B1 and Tome-1 are targets of the same E3 ubiquitin ligase, we reasoned that if Tome-1 accumulated, then the SCF E3 ubiquitin ligase might be able to target Wee1 for degradation more efficiently in the virus-infected cells. To address whether the lower level of Wee1 during early phases of the infection was due to decreased gene expression or to enhanced degradation of Wee1, we treated mock- or HCMV-infected cells with the proteasome inhibitor MG132 for 3 h before samples were harvested at 22 h p.i. Cell pellets were processed for Western blotting with an antibody directed against Wee1. Accumulation of Wee1 was observed in the virus-infected cells upon treatment with the proteasome inhibitor, suggesting that early in the infection Wee1 is targeted for degradation. This result indicates that the proteasome is active at early phases of HCMV infection and provides indirect evidence that the APC is less active in infected cells.

HCMV infection leads to accumulation of Cdc25 phosphatases.

The activity of Cdk1/cyclin B complexes is positively regulated by the Cdc25 phosphatases, which remove the inhibitory phosphates from Cdk1 (14, 17, 29, 31). Thus, it was possible that a positive effect of the virus on these phosphatases was also involved in maintaining high levels of active Cdk1 in infected cells. Therefore, we determined the expression of Cdc25B and Cdc25C in mock- and HCMV-infected, confluence-synchronized HFF cells by Western blotting. As shown in Fig. 6, the levels of Cdc25B and Cdc25C fluctuated during the cell cycle in mock-infected cells. Cdc25B expression was barely detectable at 24 h p.i., peaked at 48 h p.i., and then dropped at 72 h p.i. as the mock-infected cells approached confluence (Fig. 6A and data not shown). In contrast, we detected Cdc25B at 24 h p.i. in the virus-infected samples, and the protein accumulated throughout the duration of the experiment. In addition, we observed a slight mobility difference in the HCMV-infected samples.

The pattern of Cdc25C accumulation in mock-infected cells was similar to that observed for Cdc25B in that expression was cell cycle regulated; however, the level of Cdc25C in the viral samples increased over the duration of the time course (Fig. 6B). At 24 and 48 h p.i., the level of Cdc25C in the mock samples was slightly higher than in the viral lysates. At 72 h p.i., we detected more of the phosphatase in the viral sample than in the control sample. These results suggested that at late times postinfection, the maintenance of Cdk1/cyclin B1 complexes could also be attributed to accumulation of the Cdc25 phosphatases that oppose the activity of Myt1 and Wee1.

Subcellular localization of Cdk1 and cyclin B1.

Using immunofluorescent staining, we examined the subcellular localization of both Cdk1 and cyclin B1. The localization of Cdk1 in mock-infected cells is cell cycle regulated. At 48 h p.i., the mock-infected cultures contained cells at various stages of cell division (11, 54). Cells in G₁ and S phase showed diffuse cytoplasmic staining for Cdk1 as well as punctate staining at the centrioles (Fig. 7). We also observed cells in S/G₂ that contained Cdk1 in the nucleus and cytoplasm. In contrast, all cells in HCMV-infected cultures at 48 h p.i. displayed diffuse staining for Cdk1 in the nucleus and cytoplasm and at the centrioles (Fig. 7).

As expected, staining for cyclin B1 in mock-infected cells also showed that localization of this protein was regulated throughout the cell cycle (Fig. 7) (54). At 48 h p.i., cultures contained cells in early G₁, with little or no observable cyclin B1, and cells in late G₁ and S phases, during which cyclin B1 was detected primarily in the cytoplasm. In addition, cells in late G₂ were detected by the accumulation of cyclin B1 in the nucleus. Interestingly, the immunofluorescence data suggest that the cyclin B1 level within individual HCMV-infected cells was lower than that in mock-infected controls in which cyclin B1 was detectable; however, all cells in the infected culture expressed cyclin B1, whereas the protein was detected in only a percentage of the mock-infected cells. The finding that cyclin B1 was uniformly distributed in the cytoplasm and nucleus of virus-infected cells (Fig. 7) was particularly interesting. In addition, intranuclear accumulation of cyclin B1, consistent with viral replication centers, was noted occasionally.

To confirm the data obtained from immunofluorescent staining of HCMV-infected cultures, we also used biochemical fractionation and Western blotting to determine the localization of Cdk1 and cyclin B1 as well as proteins involved in the regulation of the kinase complex. Mock- and virus-infected cells were harvested at 24 and 48 h p.i. and separated into cytoplasmic and nuclear fractions by use of isotonic lysis buffer containing digitonin (55). Fractions corresponding to the soluble cytosolic compartment (C) (Fig. 8A) and to the insoluble and nuclear-associated pellet (N) (Fig. 8A) were collected. Lysates containing equivalent cell numbers were subjected to immunoblotting. As shown in Fig. 8A, more cyclin B1 was detected in the nuclear and insoluble fractions of both mock- and virus-infected cells at 24 and 48 h p.i. A change in the solubility of cyclin B1 was detected in mock samples between 24 and 48 h p.i. such that more cyclin B1 was detected in the soluble fraction at 48 h p.i. than at 24 h p.i. This shift could represent the cell cycle-regulated nuclear localization of cyclin B1 but more likely reflects a change in the association of cyclin B1 with a detergent-resistant compartment (4). A similar shift was observed in virus-infected cells between 24 and 48 h p.i.

As a control for the fractionation procedure, Western blots were probed with antibodies against several cellular and viral proteins whose subcellular localization is well established. As a control for the cytosolic fraction, we used glucose-6-phosphate dehydrogenase (G6PD). We detected approximately equal amounts of G6PD in the soluble cytosolic fractions of mock- and virus-infected cells at both time points (Fig. 8A). No G6PD was present in the insoluble and nuclear pellet fractions. To monitor contamination of cytosolic fractions with nuclear proteins, we probed for an early viral protein, UL44. UL44 was only detected in the insoluble and nuclear pellet fraction of viral samples (Fig. 8A). This was expected since UL44 is tightly associated with viral replication centers (16, 50). We also probed for another set of nuclear viral proteins, IE1 72 and IE2 86 (Fig. 8A). Both IE1 and IE2 were detected almost exclusively in the nuclear fractions, although some IE1 could be extracted with nonionic detergent, consistent with a previous report (50).

To determine the distribution of the active and inactive forms of Cdk1, we probed the fractions with antibodies against total Cdk1 and phosphorylated inactive Cdk1. These analyses revealed that different forms of Cdk1 were localized to specific biochemically defined compartments. At 24 and 48 h p.i., mock and viral samples contained the faster migrating active form of Cdk1 in the soluble cytosolic fraction (Fig. 8B, total Cdk1). To identify the inactive, phosphorylated forms of Cdk1, we probed Western blots with an antibody against the Cdk1 Tyr15-phosphorylated epitope described above. At 24 and 48 h p.i., we detected bands that comigrated with the slower migrating forms of Cdk1 in the soluble and insoluble fractions of the mock-infected samples (Fig. 8B, phospho-Tyr15). In addition, a faster migrating band was detected in the insoluble fraction of the mock sample, and further experiments showed that it corresponded to Cdk2 (Fig. 8B, phospho-Tyr15, and data not shown). In contrast, only a low level of Tyr15-phosphorylated Cdk1 appeared in the soluble and insoluble fractions of the viral samples at these times. Taken together, these data suggest that the majority of Cdk1 in HCMV-infected cells resides in the cytoplasm and is in its active form.

We next examined the distribution of the Cdc25 phosphatases within these biochemically defined compartments. We detected most of the Cdc25B in the soluble cytosolic fractions of mock- and virus-infected cells at 24 and 48 h p.i. (Fig. 8C), although small quantities of Cdc25B did partition into the insoluble and nuclear pellet fractions of viral samples at both time points. Cdc25C was also detected in the soluble cytosolic fractions of both mock- and virus-infected cells at both time points (Fig. 8C). These data indicate that the active form of Cdk1 and the Cdc25 phosphatases reside in the same biochemically defined compartment, which corresponds to the cytosol.