Many DNA viruses usurp host cell cycle regulation for their own replication advantage (reviewed in reference
56). Small DNA tumor viruses, such as simian virus 40 (
13,
22), adenovirus (
18,
36), and human papillomavirus (
75,
79), encode proteins that promote cells to enter the S phase. In contrast, herpesviruses, a group of large DNA viruses that encode their own DNA polymerases, generally block cell cycle progression at the G
0/G
1 phase during lytic infection cycles (reviewed in reference
25). Studies of cell cycle dysregulation induced by RNA viruses have been relatively limited. That reovirus infection causes the inhibition of cellular DNA synthesis has been known for some time (
12,
20,
30), but not until more recently was it shown that reovirus-induced inhibition of cell proliferation results from G
2/M cell cycle arrest that is mediated by the viral σ1s nonstructural protein (
59). Human immunodeficiency virus type 1 infection also induces cell cycle arrest in the G
2/M phase (
32), and the expression of the accessory gene product Vpr alone is sufficient for inducing G
2/M cell cycle arrest (
60,
62). Vpr-mediated cell cycle arrest apparently favors virus replication, since the long terminal repeat is most active and the expression of the viral genome is optimal in the G
2 phase of the cell cycle (
27). Cell cycle perturbations have also been seen in cells infected with the paramyxovirus simian virus 5 (
44), measles virus (
49,
50,
53), and coxsackievirus (
48).
Cyclins and cyclin-dependent kinases (Cdks) form complexes and play important regulatory roles in controlling cell cycle progression (reviewed in references
55 and
58). G
1-phase progression requires the activities of cyclin D-Cdk4/6 complexes, and cyclin E-Cdk2 activity is necessary for the G
1/S-phase transition. These G
1 cyclin-Cdk complexes regulate the cell cycle through phosphorylation of the retinoblastoma protein (pRb) p107 and pRB family proteins and p130. In the quiescent G
0 phase, pRb is nonphosphorylated, while during the G
1 phase, it is sequentially hypophosphorylated by cyclin D-Cdk4/6 complexes in early G
1 and hyperphosphorylated by the cyclin E-Cdk2 complex in late G
1 (
47). It remains hyperphosphorylated in the S, G
2, and M phases of cycling cells (
14). When pRb binds to the E2F family of transcription factors, it functions as a transcriptional repressor, and its hyperphosphorylation in late G
1 results in inactivation, release of E2F, and transcription of genes important for DNA synthesis (reviewed in references
17 and
31).
The activities of G
1 cyclin-Cdk complexes are regulated by cellular Cdk inhibitors (CKIs) (reviewed in reference
65). INK4 family CKIs bind to Cdk4 and Cdk6, thus blocking cyclin D-Cdk4/6 activities (reviewed in references
63 and
64). CKIs of the Cip/Kip family, including p21
Cip1, p27
Kip1, and p57
Kip2, are potent inhibitors of cyclin E- and A-dependent Cdk2 (reviewed in reference
51). The regulation of G
1 cyclin quantities is also important for proper cell cycle progression. Cyclin D1 expression is induced through the RAS-RAF-MEK-ERK pathway upon mitogenic stimulation (
1,
24,
38,
43,
77), and cyclin D1 undergoes ubiquitin-dependent proteolysis when it is phosphorylated by glycogen synthase kinase 3β (
15). The presence of growth factors maintains D-type cyclins at relatively constant levels throughout the cell cycle. The amounts of cyclin E in actively growing cells are maximal at the G
1/S-phase transition and low in other cell cycle phases (
6,
26). The synthesis of cyclin E occurs in the late G
1 phase and is transcriptionally controlled by E2F transcription factors. Rapid turnover of cyclin E in the S phase is mediated by phosphorylation-dependent ubiquitination and subsequent proteolysis (
9,
52,
80).
Coronaviruses are enveloped RNA viruses that cause gastrointestinal and upper respiratory tract illnesses in animals and humans (
57,
78). Recently, a novel coronavirus was shown to be the etiologic agent for the emerging infectious disease severe acute respiratory syndrome (
16,
42). Mouse hepatitis virus (MHV), a prototypic coronavirus, causes various diseases, including hepatitis, enteritis, and encephalitis, in rodents (
10,
78). After MHV infection, MHV RNA synthesis takes place in the cytoplasm. MHV particles, which contain three envelope proteins, S, M, and E, and an internal helical nucleocapsid, which consists of N protein and genomic RNA, bud from internal cellular membranes (
40,
45,
74). Extensive morphological, physiological, and biological changes occur in cells infected with MHV (
2,
3,
33,
66,
70,
71), and MHV infection may induce apoptotic cell death in certain cultured cell lines at a time late in infection (
2,
4,
7).
In the present study, we examined the effect of MHV replication on cell cycle progression. We found that MHV infection in asynchronously growing cells led to the inhibition of host DNA synthesis and the accumulation of infected cells in the G0/G1 phase. When serum-starved, quiescent cells were infected with MHV, they failed to enter the S phase after serum stimulation. Further analyses suggested that a reduction in the amounts of Cdk4, Cdk6, and G1 cyclins in infected cells resulted in the accumulation of hypophosphorylated and/or nonphosphorylated pRb, leading to the arrest of cell cycle progression in the G0/G1 phase.
MATERIALS AND METHODS
Virus and cells.
Plaque-cloned MHV type 2 (MHV-2) was used throughout this study. Mouse fibroblast 17Cl-1 cells (
69) and astrocytoma DBT cells (
34) were maintained as previously described (
7).
Cell cycle analysis by flow cytometry.
Nuclear DNA content was measured by using propidium iodide staining and fluorescence-activated cell sorting (FACS) analysis as described previously (
82). Briefly, adherent cells were treated with trypsin, washed with phosphate-buffered saline (PBS), resuspended in low-salt stain buffer (3% polyethylene glycol 8000, 50 μg of propidium iodide/ml, 0.1% Triton X-100, 4 mM sodium citrate, 10 μg of RNase A/ml), and incubated at 37°C for 20 min. An equal volume of high-salt stain buffer (3% polyethylene glycol 8000, 50 μg of propidium iodide/ml, 0.1% Triton X-100, 400 mM sodium chloride) then was added to the cell suspension. Propidium iodide-stained nuclei were stored at 4°C overnight before FACS analysis (FACScan; Becton Dickinson), and at least 15,000 nuclei were counted for each sample. Data analysis was performed by using ModFit LT version 2.0 (Verity Software House).
Measurement of cellular DNA synthesis.
17Cl-1 cells or DBT cells plated in 96-well plates at approximately 50% confluence were mock infected or infected with MHV-2 at a multiplicity of infection (MOI) of 20. Cells were labeled continuously with 1 μCi of [3H]thymidine (Amersham)/well from 4 to 11 h postinfection (p.i.) and harvested onto glass fiber filters (Packard) with a cell harvester (Packard). Total [3H]thymidine incorporation into the cells was determined by scintillation counting (Beckman LS 6000IC).
Synchronization of cells.
Subconfluent cultures of 17Cl-1 cells were synchronized in the G0 phase by using serum deprivation. Approximately 4 × 105 cells were plated in a 60-mm plate and maintained in medium containing 0.5% fetal calf serum (FCS) for 48 h. Synchronized cells were mock infected or infected with MHV-2 at an MOI of 10. After 1 h of virus adsorption, cells were treated with medium containing 10% FCS and harvested at various times p.i. for cell cycle and Western blot analyses.
Total cell lysate preparation.
Infected and mock-infected cells were collected at various times after MHV-2 inoculation and washed once with PBS. Cells were lysed directly in sodium dodecyl sulfate (SDS) sample buffer (60 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.01% bromophenol blue), boiled for 10 min, and passed through a 23-gauge needle several times to shear the DNA.
Western blot analysis.
Whole-cell lysates separated by SDS-polyacrylamide gel electrophoresis were transferred to polyvinylidene difluoride membranes (Amersham). The membranes were blocked in blocking solution (0.05% Tween 20 and 5% nonfat dry milk in PBS), incubated with primary and secondary antibodies diluted in blocking solution for 1 h each, and developed with an enhanced chemiluminescence kit (Amersham). The following mouse monoclonal antibodies were used: anti-pRb (G3-245; BD Pharmingen); anti-cyclin D1 (Ab-3; Oncogene); and anti-Cdk2, anti-Cdk4, and anti-cyclin D3 (BD Biosciences). The following rabbit polyclonal antibodies were used: anti-p21 (C-19), anti-cyclin E (M-20), anti-Cdk6 (C-21), anti-cyclin D2 (M-20), and anti-p16 (M-156) (Santa Cruz Biotechnology); anti-p27 (2552; Cell Signaling); and anti-phospho-pRb (Ser807 and Ser811) (9308; Cell Signaling). Actin was detected with goat antiactin polyclonal antibody (I-19; Santa Cruz Biotechnology). Horseradish peroxidase-conjugated goat anti-mouse and anti-rabbit immunoglobulin G antibodies and donkey anti-goat immunoglobulin G antibody (Santa Cruz Biotechnology) were used as secondary antibodies.
Cdk2 kinase assay.
Cells were lysed in lysis-immunoprecipitation buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% NP-40, 50 mM NaF, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, protease inhibitor cocktail [P8340; Sigma], phosphatase inhibitor cocktails [P2850 and P5726; Sigma]). Five hundred micrograms of protein lysate from each sample was immunoprecipitated with 2 μg of anti-Cdk2 antibody (M-2; Santa Cruz Biotechnology) and protein A beads. The immunocomplexes were washed twice with lysis-immunoprecipitation buffer and twice with kinase buffer (25 mM Tris-HCl [pH 7.5], 5 mM β-glycerophosphate, 2 mM dithiothreitol, 0.1 mM Na3VO4, 10 mM MgCl2; Cell Signaling) and then incubated in kinase buffer containing 200 μM ATP and 1 μg of a fusion protein containing maltose binding protein and the C-terminal region (701 to 928) of pRb (Rb-C) (Cell Signaling) at 30°C for 30 min. The reaction was stopped by adding SDS sample buffer and boiling the samples for 5 min. Proteins were separated on SDS-8% polyacrylamide gels and visualized by Western blot analysis with an anti-phospho-pRb (Ser807 and Ser811) antibody.
Statistical and densitometric analyses.
Statistical analysis was performed by using Student's t test. Data are reported as the mean and standard error (SE). A P value of <0.05 was considered significant. Bands on Western blots were scanned, and the mean density of each band was analyzed by using TotalLab software (Ultra · Lum Inc., Claremont, Calif.); prior titration experiments had confirmed that image densities were linearly proportional to protein masses. Each protein signal was normalized against the actin signal in each sample before comparisons for fold changes.
DISCUSSION
In this study, we investigated the effect of MHV infection on host cell cycle progression. Analysis of [
3H]thymidine incorporation demonstrated that MHV infection resulted in the inhibition of cellular DNA synthesis, and this effect required active MHV replication. FACS analyses demonstrated that MHV infection in actively growing cells caused an increase in the percentage of cells in the G
0/G
1 phase and that MHV infection in quiescent G
0-phase cells significantly prevented the cells from entering the S phase after mitogenic stimulation. Consistent with the cell cycle profile data, MHV replication inhibited pRb hyperphosphorylation, which is an essential step for E2F activation and S-phase progression. All of these data indicated that MHV replication arrested cell cycle progression in the G
0/G
1 phase. A decrease in Cdk2 kinase activity was seen in MHV-infected DBT and 17Cl-1 cells, while the amounts of the CKIs p21
Cip1, p27
Kip1, and p16
INK4a did not change in infected cells. These data indicated that MHV-induced inhibition of Cdk2 activity (Fig.
5) and pRb hyperphosphorylation (Fig.
4) was not caused by the activation of these CKIs. MHV replication in asynchronous cultures, however, resulted in reduced amounts of G
1 cyclins in both DBT and 17Cl-1 cells as well as in decreases in Cdk4 and Cdk6 levels in 17Cl-1 cells. When quiescent 17Cl-1 cells were infected with MHV, they failed to accumulate Cdk4, Cdk6, cyclin D1, and cyclin D3 after serum stimulation, contrary to the increased accumulation of these G
1 cyclins and Cdks in mock-infected cells. In mock-infected cells, the levels of cyclins D2 and E remained unchanged at 18 h p.i. (17 h after serum stimulation), while MHV infection induced a reduction in the levels of these two cyclins. A straightforward interpretation of all of these data is that the formation of only limited amounts of G
1 cyclin-Cdk complexes led to reduced Cdk activities and insufficient pRb hyperphosphorylation, resulting in an inhibition or delay of cell cycle progression in the G
0/G
1 phase in MHV-infected cells. Because most of our biochemical studies were focused on proteins that are known to be involved in cell cycle progression in the G
0/G
1 phase, our studies did not rule out the possibility that MHV replication also affected other stages of the cell cycle progression. Further studies are required to characterize the effect of MHV replication on other cell cycle stages.
We have not determined exactly which point of progression in the cell cycle becomes inhibited in the G
0/G
1 phase of MHV-infected cells, but we can speculate on where the host cell cycle is arrested based on our analyses of various G
1 regulatory proteins. Actively growing cells go through repeated cycles of the G
1/S/G
2/M phases, and when the environment is deprived of growth factors, cells enter the quiescent G
0 phase. The majority of uninfected cells showing 2N DNA content in FACS analysis therefore probably represented G
1 cells in actively growing cultures (Fig.
2A) and G
0 cells in serum-starved cultures (Fig.
3A). In the G
0 phase, pRb is nonphosphorylated; then it is sequentially hypophosphorylated by cyclin D-Cdk4/6 complexes in early G
1 and hyperphosphorylated by the cyclin E-Cdk2 complex in late G
1 (
47). The loss of hyperphosphorylated pRb (Fig.
4A and B) and the reduction in Cdk2 activity (Fig.
5) after MHV infection in cycling cells indicated that infected cells failed to enter the late G
1 phase. The reduction in the amounts of Cdk4, Cdk6, and D-type cyclins (Fig.
7 and
8) in MHV-infected cells most likely caused the suppression of Cdk4/6 activities. Taken together, these results indicate that MHV-infected cells were most likely arrested in the early G
1 phase. MHV infection of 17Cl-1 cells synchronized in the G
0 phase resulted in a very limited increase in the amounts of Cdk4, Cdk6, and cyclins D1 and D3 and a decrease in the amount of cyclin D2 after serum stimulation (Fig.
9), indicating very low Cdk4/6 activities in these cells. Accordingly, cell cycle progression from G
0 to G
1 was most likely blocked in cells synchronized in the G
0 phase, and the cells probably remained in a G
0-like state. A previous report on measles virus-induced cell cycle arrest in T cells examined the amount of rRNA as a method for discrimination between G
0 cells with fewer ribosomes and G
1 cells with a higher level of ribosomes (
53). Unfortunately, this experimental approach was not suitable for determining the exact point of MHV-induced cell cycle arrest, because MHV replication induces severe 28S rRNA degradation (
3).
MHV replication caused a reduction in the amounts of G
1 cyclins (Fig.
8 and
9). The lower level of cyclin E might have resulted in reduced cyclin E-Cdk2 activity, and the lower levels of D-type cyclins most likely resulted in reduced cyclin D-Cdk4/6 activities. What, then, is the mechanism of reduction of the amounts of G
1 cyclins in MHV-infected cells? Because the amounts of cyclins can be regulated by their synthesis and degradation, MHV replication could affect G
1 cyclin levels through both mechanisms. DNA microarray analyses with several MHV-permissive cell lines revealed a slight decrease in cyclin D1 mRNA levels in MHV-infected cells (C. J. Chen and S. Makino, unpublished data), suggesting that MHV infection could affect cyclin mRNA transcriptional activity or stability. MHV replication might also affect cyclin translation, because host protein synthesis is suppressed in MHV-infected cells (
3,
33,
66,
70,
71). Furthermore, MHV infection may promote cyclin D2 and E degradation; the amounts of cyclins D2 and E increased slightly after serum stimulation in quiescent 17Cl-1 cells but decreased after MHV infection in quiescent 17Cl-1 cells (Fig.
9). Decreased expression of cyclins and Cdks appears to be a common mechanism by which several viruses disrupt G
1 cell cycle progression, as demonstrated for the cell cycle arrest induced by herpes simplex virus type 1 (
19,
67), coxsackievirus (
48), and measles virus (
53). For coxsackievirus, virus replication induces cell cycle arrest in part through an increase in the ubiquitin-dependent proteolysis of cyclin D1 (
48).
The expression of transmissible gastroenteritis virus N protein results in a higher percentage of cells undergoing cell division, suggesting a cell cycle delay or arrest in the G
2/M phase (
81). Disrupted cytokinesis is also observed in cells expressing IBV N protein and cells infected with IBV (
8). We did not detect an increase in the G
2/M-phase population in MHV-infected asynchronous cultures (Fig.
2), indicating that MHV N protein does not have an effect on cytokinesis or that its putative effect on cytokinesis is masked by other MHV-induced functions in infected cells.
What is the biological significance of MHV-induced cell cycle arrest? One possibility is that cell cycle arrest in the G
0/G
1 phase provides increased amounts of ribonucleotide pools for efficient MHV RNA synthesis; ribonucleotides are the precursors for synthesizing deoxyribonucleotides, and a reduction in cellular DNA synthesis most likely increases the levels of ribonucleotide pools in cells. Cell cycle arrest may also benefit MHV replication in some other ways. MHV replication in cultured cells generally results in cell death, including apoptotic cell death (
2,
4,
7). The onset of caspase activation and apoptosis occurs very late p.i., when the highest level of MHV production has been achieved (
7), yet how MHV manages to accomplish its maximum replication prior to cell death is unknown. Accumulated data from other laboratories imply that cross talk exists between cell cycle signaling and apoptosis signaling; apoptosis follows cell cycle arrest in some systems (
28,
68), but in others, the induction of apoptosis appears to require progression through the cell cycle (
83). It is possible that cell cycle arrest in MHV-infected cells prevents the induction and execution of early cell death in infected cells. Cell cycle arrest may also assist in efficient MHV assembly, which occurs in the intermediate compartment between the endoplasmic reticulum and the Golgi complex (
40,
74) and most likely requires proper intracellular membrane structures, whereas most membrane trafficking steps are disrupted during mitosis (
46,
76). Indeed, a one-step growth curve for MHV-2 in DBT cells shows that exponential virus production occurs from 4 to 10 h p.i. and that the highest virus titer is maintained from 12 to 24 h p.i (
35); the most efficient virus production occurs when infected cells are arrested in the G
0/G
1 phase, indicating that the MHV-induced cell cycle arrest may assist in efficient MHV assembly. Furthermore, cell cycle arrest may be beneficial to MHV protein synthesis. Cap-dependent translation is reduced during mitosis due to the impaired function of cap-binding protein (
5). Because all MHV mRNAs are 5′ capped and the translation of all MHV proteins, except for E protein (
73), is cap dependent, arresting cells in the G
0/G
1 phase to prevent cells from entering mitosis should be beneficial for the cap-dependent translation of MHV proteins. Finally, MHV-induced cell cycle arrest potentially has additional important biological significance for virus-induced pathogenicity. It has been reported that noncycling cells are less likely to be killed by cytotoxic T cells (
54); hence, MHV-infected cells that are arrested in the G
0/G
1 phase may not be killed efficiently by cytotoxic T cells.