G1 Phase Cell Cycle Arrest Induced by SARS-CoV 3a
Protein via the Cyclin D3/pRb Pathway
Xiaoling Yuan*, Zhenyu Yao*, Jie Wu, Yusen Zhou, Yajun Shan, Bo Dong, Zhenhu Zhao, Ping Hua,
Jiapei Chen, and Yuwen Cong
Department of Pathophysiology, Beijing Institute of Radiation Medicine, and State Key Laboratory of Pathogen and Biosecurity,
Department of Pathogenic Molecular Biology, Beijing Institute of Microbiology and Epidemiology, Beijing, China
SARS-CoV 3a is a structural protein, mainly localizing to Golgi appa-
ratus and co-localizing with SARS-CoV M in co-transfected cells.
Here we observed that transient expression of 3a inhibited cell
growth and prevented 5-bromodeoxyuridine incorporation, sug- Our results suggested that the SARS-CoV 3a protein plays
gesting that 3a deregulated cell cycle progression. Cell cycle analysis important roles in the SARS-CoV life cycle and virus-
demonstrated that 3a expression was associated with blockage of induced pathogenesis.
cell cycle progression at G1 phase in HEK 293, COS-7, and Vero
cells 24–60 h after transfection. Mutation analysis of 3a revealed
that C-terminal region (176 aa 274 aa), including a potential
calcium ATPase motif, was essential for induction of cell cycle arrest.
Topological analysis showed that 3a predominantly located in Golgi
subgenomic mRNAs may be functionally bicistronic (6). The
apparatus, with its N-terminus residing in the lumen (Nlum) and
C-terminus in the cytosol (Ccyt). Analyzing the cellular proteins
exact roles of these SARS-CoV–speciﬁc accessory proteins are
involving in regulation of cell cycle progression, we demonstrated poorly understood. Previous studies demonstrated that accessory
that 3a expression was correlated with a significant reduction of protein genes of other coronaviruses, which vary in size, se-
cyclin D3 level and phosphorylation of retinoblastoma (Rb) protein quence, and position in the genome, are dispensable for virus
at Ser-795 and Ser-809/811, not with the expression of cyclin D1, replication, at least in cell culture. Still, they may be important
D2, cdk4, and cdk6 in 293 cells. Increases in p53 phosphorylation for virus–host interaction in the in vivo situation. For example,
on Ser-15 were observed in both SARS-CoV M and 3a transfected mutants or deletion of one of these genes, such as the 7b gene
cells, suggesting that it might not correlate with the 3a-induced of feline coronavirus and gene 3 of swine enteric and respiratory
G0/G1 phase arrest. The reduction of cyclin D3 level and phosphory- coronavirus, have been reported to relate to reduced virulence
lation of Rb were further confirmed in SARS-CoV infected Vero cells. and pathogenesis (7, 8), indicating a possible in vivo function.
These results indicate that SARS-CoV 3a protein, through limiting The product of the SARS-CoV 3a gene (CDS: 25252–26074),
the expression of cyclin D3, may inhibit Rb phosphorylation, which also referred to as ORF3, X1, and U274 in other articles, was
in turn leads to a block in the G1 phase of the cell cycle and an identiﬁed separately in the SARS-CoV–infected cells, lung speci-
inhibition of cell proliferation. men from a patient with SARS, and crude virions (9–11). Pre-
Keywords: SARS-CoV 3a; growth inhibition; G1 phase cell cycle arrest; viously, it was reported that 3a protein is located in the Golgi
cyclin D3; pRb apparatus, and the second or third trans-membrane regions are
responsible for the Golgi localization (12). The integral mem-
Severe acute respiratory syndrome (SARS), caused by SARS brane protein 3a interacts with other structural proteins, such
coronavirus (SARS-CoV), is a life-threatening emerging infec- as S, M, and E proteins, as well as nonstructural protein U122,
tious disease originating from Guangdong Province, China (1, 2). in infected or transfected Vero E6 cells (9). Recently published
SARS-CoV, a distant member of Group 2 coronaviruses, has papers showed that the 3a gene product is a structural protein
recently been identiﬁed as the etiologic agent of SARS (3–5). of SARS-CoV (13, 14). Overall, these ﬁndings suggest that the
Analysis of the nucleotide sequence of the SARS-CoV genome 3a protein is an important protein in the viral life cycle. In this
showed that it is nearly 30 kb in length and organized with the study, we ﬁrst present evidence that overexpression of the 3a
gene order that is characteristic for coronavirus [5 -replicase (rep), gene can inhibit cell growth and block cell cycle progression at
Spike (S), Envelope (E), Membrane (M), Nucleocapsid (N)-3 ]. the G1 phase. The domain responsible for these functions was
The genome of SARS-CoV also contains nine genes specifying further identiﬁed through construction of a series of truncated
so-called “accessory proteins” located between S and E, and M mutants of the 3a gene. The mechanism behind the G1 phase
and N genes. All these ORFs are predicted to be expressed from arrest principally involved a decrease in expression of cyclin D3
eight subgenomic mRNAs identiﬁed in SARS-CoV–infected and phosphorylated retinoblastoma (Rb) protein. These results
Vero cells, and remarkably, up to four of the eight SARS-CoV suggested that the 3a protein plays important roles in the SARS-
CoV life cycle and virus-induced pathogenesis.
MATERIALS AND METHODS
(Received in original form September 7, 2005 and in final form December 25, 2006 )
* These authors contributed equally to this work. Cell Culture and Transfection
This work was supported by a grant from the Nature Sciences Foundation of Human embryonic kidney cell lines, HEK 293, and African green mon-
China (30470093). key kidney cell lines, Vero E6 and COS-7, were cultured in Dulbecco’s
Correspondence and requests for reprints should be addressed to Yuwen Cong, modiﬁed Eagle’s medium (DMEM) (Gibco BRL, Grand Island, NY)
Ph.D., Department of Pathophysiology, Beijing Institute of Radiation Medicine, supplemented with 10% FBS. Cultures were incubated at 37 C in a
No. 27 Taiping Road, Beijing, 100850, China. E-mail: email@example.com humidiﬁed environment with 5% CO2. When cell density in a culture
Am J Respir Cell Mol Biol Vol 37. pp 9–19, 2007
plate reached 70% conﬂuence, the cells were transfected with different
Originally Published in Press as DOI: 10.1165/rcmb.2005-0345RC on April 5, 2007 plasmid DNAs using Lipofectamine 2000 (Invitrogen, Carlsbad, CA),
Internet address: www.atsjournals.org following the protocol provided by the manufacturer. Brieﬂy, the total
10 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 37 2007
TABLE 1. PRIMERS USED FOR WILD-TYPE AND TRUNCATED 3a CONSTRUCTS
Construct Name Polarity Sequence
3a/pCMV-myc Sense* 5 -CGGAATTCGGATGGATTTGTTTATGAGATTTTTTAC-3
Anti-sense† 5 -CCGCTCGAGTTACAAAGGCACGCTAGTAGTCG-3
3a D1-222/pCMV-myc Sense 5 -CGGAATTCGGATGGATTTGTTTATGAGATTTTTTAC-3
Anti-sense 5 -CCGCTCGAGTTAGTCTGTAGTAATTTGTGTAG-3
3a D147-274/pCMV-myc Sense 5 -CGGAATTCGGATGGTTTATGATGCCAACTACTTTGTTTGC-3
Anti-sense 5 -CCGCTCGAGTTACAAAGGCACGCTAGTAGTCG-3
3a D176-274/pCMV-myc Sense 5 -CGGAATTCGGATGTCAACACCAAAACTCAAAG-3
Anti-sense 5 -CCGCTCGAGTTACAAAGGCACGCTAGTAGTCG-3
3a D99-274/pCMV-myc Sense 5 -CGGAATTCGGATGGCAGGTATGGAGGCGCAAT-3
Anti-sense 5 -CCGCTCGAGTTACAAAGGCACGCTAGTAGTCG-3
3a D77-274/pCMV-myc Sense 5 -CGGAATTCGGATGTTCCAGTTCATTTGCAATTTA-3
Anti-sense 5 -CCGCTCGAGTTACAAAGGCACGCTAGTAGTCG-3
3a-HA-pcDNA3.1 Sense 5 -CGGAATTCATGGATTTGTTTATGAGATTTTTTAC-3
Anti-sense 5 -AAGGAAAAAAGCGGCCGCTTAAGCGTAATCTGGAACATCGT-3
M-HA-pcDNA3.1 Sense 5 -CGGAATTCATGGCAGACAACGGTACTATTAC-3
Anti-sense 5 -AAGGAAAAAAGCGGCCGCTTAAGCGTAATCTGGAACATCG-3
M/pCMV-myc Sense 5 -CGGAATTCCCATGGCAGACAACGGTACTATTAC-3
Anti-sense 5 -CCGCTCGAGTTACTGTACTAGCAAAGCAATATTG-3
GFP/pCMV-myc Sense 5 -CGGAATTCGGATGGTGAGCAAGGGCGAGGAGC-3
Anti-sense 5 -CCGCTCGAGTTACTTGTACAGCTCGTCCATGCCG-3
Gene sequences are correspond to SARS-CoV (ZJ01).
* Underlined nucleotides represent restriction site before start codon (ATG).
Underlined nucleotides represent restriction site before the stop codon.
amount of DNA transfected into the cells in each well was adjusted to Confocal Microscopy Analysis
2 g/ml by using empty pCMV-myc vector. Cells were incubated with
The cellular localization of the SARS-CoV 3a protein and its mutants
transfection mixtures for 5 h and then replaced with fresh medium.
was studied in transfected cells according to the procedure described
Construction of Expressing Vectors of SARS-CoV 3a and
For 5-bromodeoxyuridine (BrdUrd) incorporation, Vero E6 and
its Mutants COS-7 cells were transfected with 3a/pCMV-myc and M/pCMV-myc.
The 3a gene used for this study was PCR-ampliﬁed from the SARS- At 24 h after transfection, cells on glass cover slips were incubated with
CoV (ZJ01, AY297028) genome using Taq DNA polymerase (NEB). 10 mol/liter BrdUrd for 4 h at 37 C and ﬁxed with 100% methanol
PCR was performed with a forward primer (containing an EcoR I site) at 4 C for 10 min. Incorporated BrdUrd was exposed by treatment with
complementary to the 5 end of the 3a gene and a reverse primer 2 M hydrochloric acid at 37 C for 2 h, followed by neutralization in 0.1
(containing a Xho I site) complementary to the 3 end of the 3a gene M borate buffer (pH 8.5). After washing in PBS, cells were permeabil-
(Table 1). This product was cut with EcoR I and Xho I, and cloned into ized in 0.1% Triton X-100/PBS for 5 min and incubated with anti-
the multiple cloning site (MCS) of the pCMV-myc vector (Clontech), BrdUrd (1:100; Sigma, St. Louis, MO) and anti-myc (1:100; Cell Signal-
producing a 3a/pCMV-myc plasmid. The sequence of the plasmid was ing, Beverly, MA) antibodies for 1 h. Images were viewed and collected
conﬁrmed by sequencing. The serial mutants of 3a gene/pCMV-myc, with a confocal ﬂuorescence microscope connected to a Bio-Rad Radi-
green ﬂuorescent protein (GFP)/pCMV-myc, 3a-hemagglutinin (HA)/ ance 2100 laser scanner (Bio-Rad, Richmond, CA).
pcDNA3.1, SARS-CoV membrane protein (M)/pCMV-myc, and M-
Flow Cytometric Cell Cycle Analysis
HA/pcDNA3.1 constructs were made in a similar fashion, and the
oligonucleotide primers used are listed in Table 1. For ﬂow cytometry, 2 106 transfected cells were ﬁxed overnight with
70% cold ethanol at 4 C. Cells were then permeabilized in 0.1% Triton
Growth Curves and 3-(4,5-Dimethylthiazolyl-2)-2,5-Diphenyl X-100/PBS, incubated with anti-myc antibody (1:100) and ﬂuorescein
Tetrazolium Bromide Assay isothiocyanate–conjugated mouse anti-IgG (1:100) (Santa Cruz Bio-
technology, Santa Cruz, CA), resuspended in propidium iodide (PI, 50
HEK 293 cells seeded in a 24-well plate (Costar, Cambridge, MA) were g/ml) staining solution (containing DNase-free RNase A 20 g/ml)
transfected with 3a/pCMV-myc and pCMV-myc in triplicate. At 12-h for 30 min in the dark, and analyzed immediately by ﬂow cytometry.
intervals after transfection, cells were rinsed with PBS and a single DNA contents and other ﬂuorescence-activated cell sorter data were
cell suspension was prepared by trypsinization. Viable cells, which are analyzed with CellQuest software (Becton Dickinson, San Jose, CA)
resistant to trypan blue staining, were counted with a hemacytometer (15).
chamber. To avoid bias, counting was done blindly by two individuals
for each sample. Cells transfected with at least three independent clones Western Blot Analysis
were analyzed. The transfected (or infected) cells were harvested at the indicated times
For the 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyl tetrazolium bro- after transfection (or infection). Preparation the total cell lysates and
mide (MTT) assay, HEK 293 cells seeded in a 96-well plate (Costar) Western blot analysis were performed according to the procedure
were transfected with different concentrations of 3a/pCMV-myc and described before (16). Brieﬂy, the cell lysates were clariﬁed by centrifu-
pCMV-myc plasmids in triplicate. After 48 h, each well was supple- gation at 12,000 g for 10 min at 4 C. Equal amounts of protein
mented with 20 l of MTT solution (5 mg/ml), and incubated for 3 h. (20 g) were separated by SDS-PAGE and transferred to NC membranes
The medium was removed and 200 l of DMSO were added to each (Osmonics, Inc., Westboro, MA). The membranes were probed with
well. Then the plate was vibrated to dissolve crystals and the absorbance primary antibodies (antibodies against myc, -actin, cyclin D1, cyclin
(O.D.) at 570 nm was measured. The experiments were independently D2, p53, phospho-p53 on Ser-15, Rb, phospho-Rb on Ser-795 and Ser-
repeated three times. 809/811, CDK4, CDK6, cyclin D3, cdc2, and cyclin A; all antibodies
Rapid Communication 11
were purchased from Cell Signaling) and subsequently with horseradish AGT GCG TG-3 ; antisense primer, 5 -GCA ACT CGT CAT ACT
peroxidase–conjugated secondary antibodies. Antibody detection was CCT GCT T-3 ) (299 bp) and -actin (sense primer, 5 -CAC TCT TCC
performed using an enhanced chemiluminescence (ECL) detection kit AGC CTT CCT TCC-3 ; antisense primer, 5 -CGG ACT CGT CAT
(Cell Signaling). ACT CCT GCT T-3 )(388 bp) genes.
When it was necessary to reprobe the membrane with another anti-
body, the membrane was stripped with stripping buffer (2% SDS, 100 Statistical and Densitometric Analysis
mM -mercaptoethanol, 62.5 mM Tris-HCl pH 6.8) at 50 C for 30 min Statistic analysis was performed by using Student’s t test. Data are
and washed with TBST (Tris-HCl 20 mmol/liter, NaCl 140 mmol/liter, reported as the mean and SD. Bands on Western blotting were scanned
Tween 20 0.1%) buffer before use. under a scanner (Microtek, Carson, CA), and the mean density of each
band was analyzed by using Quantity One software (Bio-Rad). The
Infection of Vero E6 Cells with SARS-CoV protein expression and mRNA levels were plotted after normalization
SARS-CoV (BJ01, AY278488) was grown as described previously (17). against actin or corresponding nonphosphorylated total protein signal.
When the cell density in a 60-mm culture ﬂask reached 90% conﬂuence,
the Vero E6 cells were infected with 1 104 TCID50 of SARS-CoV, RESULTS
in a ﬁnal volume of 2 ml of DMEM with 2% FBS for 1 h at 37 C. Then
the cells were washed with PBS, and replaced with complete medium Cell Growth Is Inhibited in Transfected Cells Expressing the
to allow growth. At 6 and 24 h after infection, cells were washed with SARS-CoV 3a Protein
PBS and lysed in 200 l of lysis buffer containing 50 mM Tris-HCl (pH
In this study, the 3a gene of SARS-CoV (ZJ01, AY297028), was
8.0), 150 mM NaCl, 0.5% NP-40 (Sigma), 0.5% sodium deoxycholate,
and 0.005% SDS. A total of 20 l of the lysate was used for Western cloned into the pCMV-myc vector and expressed in HEK 293
blot analysis (described above). cells as described before (12). We have observed that HEK 293
cells transfected with 3a/pCMV-myc grow slower than pCMV-
RT-PCR Assay myc–transfected cells. We thus speculated that expression of 3a
Total RNA was prepared from 3a/pCMV-myc– or pCMV-myc– gene may inhibit cell proliferation. Compared with exponentially
transfected cells. Reverse transcription-PCR (RT-PCR) was performed growing control cells transfected with pCMV-myc, cells express-
using the primers for cyclin D3 (sense primer, 5 -CCT CCT ACT TCC ing the 3a gene showed a signiﬁcant decrease in cell number
Figure 1. Inhibition of cell proliferation by expression of the SARS-CoV 3a protein. (A ) Growth inhibition of protein 3a in the transfected cells.
HEK 293 cells were transfected with equivalent amounts of pCMV-myc (control) or 3a/pCMV-myc. At 12, 24, 36, 48, and 60 h after transfection,
samples were collected and counted using a hand-held counter. Cell counts at each time point were the mean values of three independent
experiments with standard deviation (SD). *P 0.05 versus control. (B ) Effects of different amounts of plasmid DNA on the growth of pCMV-
myc– or 3a/pCMV-myc–transfected cells. One hundred microliters of HEK 293 cells were transfected with different concentrations of pCMV-myc
or 3a/pCMV-myc in 96-well culture plates. After 48 h, each well was supplemented with MTT solution and the absorbance (O.D.) at 570 nm was
measured. The experiments were independently repeated three times, and one experiment was presented with three duplicates at each concentration.
The data shown was the mean of three independent experiments SD. *P 0.05 versus control. (C ) Overexpression of the 3a protein inhibited
DNA replication of COS-7 cells. At 24 h after transfection with 3a/pCMV-myc or M/pCMV-myc, COS-7 cells were incubated with 10 mol/liter
BrdUrd for 4 h and stained with anti-BrdUrd and anti-myc antibodies. Images were viewed and collected under a confocal fluorescence microscope.
The left panels show myc-3a or myc-M positive cells, the middle panels show BrdUrd incorporation cells, and the right panels display overlay of
BrdUrd and myc-3a (myc-M) staining images.
12 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 37 2007
Figure 2. G1 phase cell cycle arrest induced by the SARS-CoV 3a protein. (A ) Expression of the 3a gene–induced G1 phase cell cycle arrest. 3a/
pCMV-myc, M/pCMV-myc, or pCMV-myc was transfected into HEK 293 cells. At 24 h after transfection, samples were collected and stained with
anti-myc antibody and PI. The DNA contents of cells were measured by flow cytometry. The cell cycle profiles of myc-3a (or myc-M) expressed
cells are shown in the middle column and represented as “myc positive,” while “myc negative” in the right column represents the nontransfected
cells in the same population. The experiments were independently repeated three times. (B ) G1 cell cycle arrest induced by the 3a protein in
different cells. COS-7 and Vero E6 cells were transfected with 3a/pCMV-myc for 24 h, followed by staining with anti-myc antibody and propidium
iodide. Cells were analyzed by flow cytometry as above. The histogram showed the percentages of myc-3a–positive and –negative cells at various
phases of cell cycle with means SD for three independent sets of experiments. (C ) G1 cell cycle arrest induced by the 3a protein at different
times after transfection. HEK 293 cells were transfected with 3a/pCMV-myc. At 24, 36, 48, and 60 h after transfection, samples were collected
and analyzed by flow cytometry as before. Myc-3a–positive and –negative cells were shown with different column. The histogram showed the
percentages of cells at various phase of cell cycle with means SD for three independent sets of experiments. (D ) Expression levels of myc-3a
protein at different times after transfection. HEK 293 cells were transfected with equivalent amounts of 3a/pCMV-myc plasmid. At given times
after transfection, cell lysates were prepared and probed with anti-myc or anti-actin antibody.
Rapid Communication 13
after 24 h (Figure 1A). In the 3a/pCMV-myc–transfected cells, 75% of myc-3a–negative COS-7 cells had incorporated BrdUrd,
less than 4.8% of the cells were dead, which indicated that the while almost all myc-3a–expressing cells had little or no BrdUrd
reason for the decrease in cell number was not the cell death, incorporation (only 2.85% positive cells). Similar results were
but may be the cell growth inhibition induced by 3a gene expres- obtained in myc-3a–transfected Vero E6 cells (data not shown).
sion. The above experiments were repeated with MTT assay, a As a control, both myc-M (the membrane protein of SARS-
more sensitive colorimetric test to monitor the cell proliferation. CoV)–positive and –negative cells had similar rates of BrdUrd
As shown in Figure 1B, the growth inhibition of 3a/pCMV- incorporation (63% versus 68%). These data indicate that 3a
myc–transfected HEK 293 cells was signiﬁcantly dependent on expression inhibits cells growth and prevented cell cycle entry
the dose of plasmid used for transfection, whereas the growth into S phase.
of 3a/pCMV-myc–transfected HEK 293 cells was marginally in-
hibited at the highest dose of plasmid (1.0–2.0 g/ml). Based on Induction of G1 Phase Cell Cycle Arrest by the SARS-CoV
its apparent ability to inhibit cell growth, 1.0 g/ml 3a/pCMV- 3a Protein
myc plasmid was used for transfection in our subsequent experi-
ments. To address the mechanism of 3a on cell growth inhibition, Cell cycle deregulation is a common response of host cells to
cell DNA synthesis was further measured by BrdUrd incorpora- many virus infections, and some viral proteins have been shown
tion (15). In view of the propensity of HEK 293 cells easily to to be efﬁcient in inducing cell cycle arrest, such as orf-a of feline
form cell’s dumpling after treatment with 2 M hydrochloric acid, immunodeﬁciency virus and nonstructural protein p28 of murine
COS-7 cells were used in the assay. As shown in Figure 1C, hepatitis coronavirus (MHV) (18, 19). Flow cytometry is a rapid,
Figure 3. Subcellular localization and cell cy-
cle arrest induction of the 3a-truncated
mutants. (A ) Schematic representation of the
3a-truncated mutants. Serial 3a-truncated
mutants were constructed based on bioinfor-
matic analysis of the 3a protein. The boxes
represent the trans-membrane domains re-
ferred to as a, b, and c for the first, second, and
third domain, and the black bar represents
the Yxx motif, respectively. The amino acid
positions for these domains are given below.
The G1 phase cell cycle arrests of these 3a
mutants are shown on the right (Y, yes; N, no).
(B ) Subcellular localization of 3a-truncated
mutants. The plasmids of serial mutants of 3a
(D77–274, D99–274, D176–274, and D1–
222) and pDsRed-Golgi were co-transfected
into HEK 293 cells separately. At 24 h after
transfection, the cells on glass slips were
cells were permeabilized with Triton X-100
and viewed with anti-myc antibody. The nu-
cleus was stained with Hoechst 33342. Co-
localizations of serial 3a-truncated mutants
with Golgi marker were observed under a
confocal fluorescence microscope. (C ) The ef-
fects of 3a-truncated mutants on G1 phase
cell cycle arrest. The plasmids of the serial
3a-truncated mutants were transfected into
HEK 293 cells. At 24 h after transfection, sam-
ples were collected and stained with anti-myc
antibody and PI. The DNA contents of cells
were measured by flow cytometry as before.
The cell cycle profiles of “myc-positive” and
“myc-negative” cells are shown in the middle
and right columns respectively. The experi-
ments were independently repeated three
14 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 37 2007
Figure 3. (Continued).
quantitative, multiparameter cell analysis based on the measure- separately transfected into COS-7 and Vero E6 cells. The trans-
ment of visible and ﬂuorescent light emission. Using the myc- fection efﬁciencies in the two cell types were 18%. As in the
tag to identify the transfected (myc-tag–positive cells) and case of 293 cells, transfection with pCMV-myc had little effect
nontransfected cells (myc-tag–negative cells) in 3a/pCMV- on cell cycle proﬁles of COS-7 and Vero E6 cells (data not shown).
myc–transfected HEK 293 cells (15), the cell cycle of the two In both cell lines, G1 phase arrest induced by 3a gene expression
populations was analyzed, and the transfection efﬁciency of was observed, which was as obvious as that in 293 cells, sug-
3a/pCMV-myc was revealed to be 20.2% by ﬂow cytometry. As gesting that a common mechanism in the different cell lines is
shown in Figure 2A, 67% of the cells positive for 3a gene involved in 3a protein–induced G1 phase arrest (Figure 2B).
expression were in the G1 phase, whereas only 50% of the The G1 phase arrest is a crucial DNA damage checkpoint,
3a-negative cells were in G1 phase at 24 h after transfection. To which acts as an important safeguard for genomic stability. Cells
further conﬁrm the effect of 3a gene expression, HEK 293 cells in G1 phase arrest may go into apoptosis, or recover from the
were transfected with M/pCMV-myc and GFP/pCMV-myc. The G1 phase to enter into the S phase (15, 20). To observe the
transfection efﬁciencies of myc-GFP and myc-M expression plas- denouement of the G1 phase arrest induced by the 3a protein,
mids were found to be 35.7% and 34.0%, respectively. As shown cell cycle analysis was performed in 3a/pCMV-myc–transfected
in Figure 2A, the myc-M– and myc-GFP–positive and –negative HEK 293 cells from 24–60 h after transfection. A signiﬁcant
populations had similar cell cycle progression proﬁles, indicating increase in the percentage of cells in the G1 phase, and concomi-
that the expression of myc-M or myc-GFP had little effect on tant reduction in percentage of cells in S phase, was noticeable
the cell cycle. As controls, HEK 293 control cells and pCMV- in 3a protein–expressing cells between 24 and 60 h after transfec-
myc–transfected HEK 293 cells had similar cell cycle proﬁles, tion relative to control cells, with a maximum induction of G1
with 50% of the cells in the G1 phase (data not shown). To phase arrest being observed at 24 h after transfection (Figure
test whether the G1 phase arrest induced by 3a gene expression 2C). After 24 h, although there was some increases in the number
was cell line speciﬁc or not, pCMV-myc and 3a/pCMV-myc were of cells in S phase, a G1 phase arrest was still noticeable in 3a
Rapid Communication 15
protein–expressing cells, indicating that this arrest was not due
to a cell density effect. Western blotting assay showed that the
expression levels of myc-3a protein in the transfected cells
were similar between 24 and 60 h after transfection (Figure 2D).
The reasons for the increase in the number of cells in S phase
at the later times might be that some cells expressing lower
level of myc-3a could slowly enter into S phase. Sub-G1 phases,
representing one type of cell apoptosis, were observed from 24
to 60 h after transfection, but were not over 6% of analyzed
cells. These data support the idea that the 3a protein is not an
inductor of apoptosis, but of cell cycle arrestor.
Cellular Localization and Induction of G1 Phase Arrest of
Truncated 3a Protein Mutants
To deﬁne the functional domain of the 3a protein involved in
the induction of cell cycle arrest, a series of truncated mutants
were constructed according to the bioinformation analysis re-
ported in other articles (Figure 3A) (4, 5, 12). The 3a protein
is proposed to have three trans-membrane regions located at
residues 34–56, 77–99, and 103–125. Truncated mutants of the
3a gene were cloned into pCMV-myc vector separately. These
mutants were expressed and migrated to the expected molecular
mass. The subcellular localization of these proteins was analyzed
in HEK 293 cells as described before (12). As shown in Figure
3B, mutants D77–274, D99–274, and D1–222, containing one to
three of the predicted trans-membrane domains, co-localized
well with Golgi-DsRed fusion protein in the cytoplasm of co-
transfected cells, which is similar to wild-type 3a protein. While
mutant D176–274 contained no trans-membrane domains, it Figure 4. Topology of the 3a protein in the Golgi apparatus. HEK 293
partly co-localized with the Golgi marker, but mainly localized cells were transfected with 3a/pCMV-myc, 3a-HA/pcDNA3.1, M/pCMV-
to the cytoplasm and to a small extent in the nucleus. myc, or M-HA/pcDNA3.1 plasmids, respectively. At 24 h after transfec-
Furthermore, a cell cycle analysis was performed as before tion, cells on glass slips were collected, permeabilized with Triton X-
in HEK 293 cells. At 24 h after transfection with 3a-mutants/ 100 or digitonin, and then viewed with anti-myc or anti-HA antibody.
The nucleus was stained with Hoechst 33342. Cellular localization of
pCMV-myc and pCMV-myc plasmids, cells were collected and
the 3a and M fusion proteins was analyzed under a confocal fluorescence
analyzed by ﬂow cytometry. The transfection efﬁciencies of myc-
3a, D77–274, D99–274, D176–274, and D1–222 in the samples
analyzed were 18.4, 29.2, 22.1, 18.6, and 41.2%, respectively. As
shown in Figure 3C, when compared with transfection-negative
cells, mutants D77–274, D99–274, D176–274, and 3a all had a ﬁrm the integrity of the Golgi complex (21, 22). As shown in
similar ability to induce G1 phase cell cycle arrest, but D1–222 Figure 4, the myc epitopes of the 3a and M proteins could both
had not. It was concluded that the C-terminal domain of 3a may be detected after Triton X-100 treatment, but not after digitonin
be responsible for its cell cycle arrest induction. Recent articles treatment, indicating a lumenal position of the myc tag. In con-
reported that the C-terminal domain of the 3a protein contains trast, the HA epitopes were detected in both Triton X-100– and
a Yxx motif (160 aa–173 aa), which was important for the digitonin-treated cells, demonstrating a cytosolic localization of
internalization of the 3a protein from the plasma membrane (9), the HA epitope. These data clearly demonstrated that the ex-
and a potential calcium ATPase motif (200 aa–274 aa) (9, 10). pressed 3a protein, as well as the M protein, localize to Golgi
The C-terminally truncated mutant (D1–222), containing the apparatus with a NlumCcyt orientation. Bioinformatics analysis
Yxx motif (160 aa–173 aa), could not induce G1 phase arrest, and subcellular localization analysis of the series of 3a mutants
which further suggested that the potential calcium ATPase motif both indicated that the three trans-membrane regions lie in the
is important for the 3a protein to induce cell cycle arrest. N-terminus between 34 and 143 aa of the 3a protein. As presented
in the above results, the functional domain of the 3a protein for
Topology of the SARS-CoV 3a Protein in the Golgi Apparatus
inducing cell cycle arrest probably localized to its cytoplasmic
The SARS-CoV 3a protein was reported to mainly localize to region. In this conformation, it is proposed that the C-terminal
the Golgi apparatus. To gain more information about the mecha- domain of the 3a protein can easily interact with cytoplasmic
nism by which the 3a protein induces cell cycle arrest, the mem- protein involved in cell cycle regulation.
brane topology of the 3a protein in Golgi apparatus was ana-
lyzed. The SARS-CoV 3a and M proteins were fused to the G1 Phase Arrest Induced by the SARS-CoV 3a Protein via Cyclin
myc and HA tags at their N- and C-terminus, respectively, and D3/pRb Pathway
expressed in HEK 293 cells. The orientation of the HA and myc Rb phosphorylation is a critical step in the G1 to S phase transi-
epitopes was determined by immunoﬂuorescence microscopy tion. Phosphorylation of Rb is regulated primarily by complexes
using HA- or myc-speciﬁc antibodies in permeabilized (Triton of cyclin D associated with CDK4/6 and later by cyclin E associ-
X-100–treated) and semipermeabilized (digitonin-treated) cells. ated with CDK2. To understand the mechanism of 3a-induced
The SARS-CoV M protein, which is proposed to localize to the G0/G1 phase arrest, we ﬁrst examined the phosphorylation status
Golgi apparatus with its N-terminus in the lumen and its C- of Rb in transfected cells by Western blotting analysis. The total
terminus in the cytosol (Nlum-Ccyt), served as a control to con- HEK 293 cells transfected with pCMV-myc, 3a/pCMV-myc, and
16 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 37 2007
Figure 5. The effects of the 3a protein on the phos-
phorylation and/or expression of Rb, cyclins, and
cdks related to G1 cell cycle. (A ) Expression and
phosphorylation of Rb in the transfected cells. HEK
293 cells were transfected with equivalent amounts
of pCMV-myc, M/pCMV-myc, or 3a/pCMV-myc
plasmid, and cells were transfected without plasmid
as control. At 12, 24, 36, and 48 h after transfection,
cell lysates were prepared and probed with anti-
myc, anti-phospho-Rb on Ser-795 or Ser-809/811,
anti-Rb, or anti-actin antibody. The bands of p-Rb
were quantified using densitometric analysis and
plotted after normalization against actin. The histo-
gram shows the means SD for three independent
sets of experiments. (B ) Levels of cyclin D3, CDK4,
and CDK6 in the transfected cells. Cell lysates were
prepared as before and probed with anti-Cyclin D3,
anti-CDK4, anti-CDK6 or anti-actin antibody. Cells
were transfected without plasmid as control. The
bands of cyclin D3 were quantified using densito-
metric analysis and plotted after normalization
against actin. The histogram shows the means
SD for three independent sets of experiments. (C )
Levels of cyclin D3 and Rb in SARS-CoV–infected
cells. At 6 and 24 h after SARS-CoV infection, Vero
E6 cells lysates were prepared as before and probed
with anti-cyclin D3, anti-Rb, or anti-actin antibody.
The bands of cyclin D3 were quantified using densi-
tometric analysis and plotted after normalization
against actin. The histogram shows the means
SD for three independent sets of experiments. (D )
Transcription of cyclin D3 mRNA in transfected cells.
After transfection with equivalent amounts of
pCMV-myc or 3a/pCMV-myc, the total RNA pre-
pared from the HEK 293 cells were used for RT-PCR.
Reactions were performed with primers for -actin
as internal controls and primers for cyclin D3 were
detected at the same time. PCR products were sepa-
rated on 1.2% agarose gel. The bands of cyclin D3
were quantified using densitometric analysis and
plotted after normalization against actin. The histo-
gram shows the means SD for three independent
sets of experiments. (E ) Expression and phosphory-
lation of p53 in the transfected cells. Cell lysates
were prepared as before and probed with anti-p53,
anti-phospho-p53 on Ser-15 or anti-actin antibody.
Cells were transfected without plasmid as control.
The bands of p-p53 were quantified using densito-
metric analysis and plotted after normalization
against the amounts of p53. The histogram shows
the means SD for three independent sets of
M/pCMV-myc plasmids were collected at 24 h after transfection in myc-3a– and myc-M–transfected cells and was more signiﬁcant
and analyzed with antibodies against Rb phosphorylation on in myc-3a–transfected cells, indicating that the expression of 3a
Ser-795 and Ser-807/811. As shown in Figure 5A, myc-3a and inhibits Rb phosphorylation and block cell cycle progression at
myc-M were expressed at the expected molecular mass, and G0/G1 phase. The phosphorylation status of Rb was further
when compared with control and myc-transfected cells, the Rb studied in 3a/pCMV-myc–transfected cells. Expression of myc-
phosphorylation on Ser-795 and Ser-807/811 was down-regulated 3a was observed at all the time points with the highest level at
Rapid Communication 17
Figure 5. (Continued).
24 h after transfection, whereas the phosphorylation of Rb on controls. Furthermore, the level of cyclin D3 was signiﬁcantly
Ser-795 and Ser-807/811 decreased gradually after transfection. decreased from 12–48 h after transfection with 3a/pCMV-myc,
In bisacrylamide cross-linked gels, hyperphosphorylated Rb mi- which was consistent with the changes of Rb phosphorylation
grates slowly, while hypophosphorylated and nonphosphory- (Figure 5B). The expression of cyclin D1 and D2 was not ob-
lated Rb comigrate and appear as a more rapidly migrating band served in HEK 293 transfected with 3a/pCMV-myc, M/pCMV-
(18, 20). Using an anti-Rb antibody, the phosphorylation status myc, and pCMV-myc, respectively (data not shown).
of Rb was further determined. In control pCMV-myc and M/ In addition, cyclin D3 expression and Rb phosphorylation
pCMV-myc–transfected cells, the majority of Rb appeared as a were examined in SARS-CoV–infected Vero E6 cells. In this
slowly migrating band, while in 3a/pCMV-myc–transfected cells, experiment, no cytopathic effects (CPU) were observed at 6 h
the rapidly migrating band was detected from 24 h after transfec- after infection, while at 24 h after infection, the cells showed
tion, indicating hypophosphorylation of Rb. These results sug- 25% CPU. As shown in Figure 5C, the protein level of cyclin D3
gested that the 3a protein arrests cell proliferation by regulating was decreased at 6 h and signiﬁcantly decreased at 24 h after
Rb phosphorylation (Figure 5A). infection. When compared with the mock infection, the rapidly
Rb Ser-795 and Ser-809/811 are targets of active cyclin migrating bands representing the hypophosphorylated and non-
D/CDK4/6 complexes (23). The observed decrease in its phos- phosphorylated forms of Rb appeared at 6 h and signiﬁcantly
phorylation prompted us to examine the expression of cyclin at 24 h after infection. The decrease in cyclin D3 levels and the
D1, D2, D3, cdk4, and cdk6 in 3a/pCMV-myc–transfected cells appearance of the hypophosphorylated and nonphosphorylated
by Western blots. No signiﬁcant changes were observed in the forms of Rb in infected cells were consistent with that in the
levels of cdk4 and cdk6 between pCMV-myc, M/pCMV-my, and transfected cells. In addition, transcription of cyclin D3 mRNA
3a/pCMV-myc. And the expression of cdk4 and cdk6 were not was analyzed by RT-PCR assay. As compared with the control cells,
obviously different from 12–48 h after transfection with 3a/ the transcriptional level of cyclin D3 was decreased obviously in
pCMV-myc (Figure 5B). In contrast, cyclin D3 was decreased 3a/pCMV-myc–transfected cells (Figure 5D). From the above
in both myc-3a– and myc-M–expressed cells and became more data, it was concluded that in 3a/pCMV-myc–transfected cells and
signiﬁcantly so in myc-3a–expressed cells when compared with SARS-CoV–infected cells, expression of cyclin D3 was reduced,
18 AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 37 2007
which, in turn, may lead to phosphorylation of Rb, a block in This suggested that the 3a protein may favor SARS-CoV replica-
G1 phase of cell cycle, and an inhibition of proliferation. tion by inducing cell cycle arrest at the G1 phase, and moreover,
CDK-inhibitors (CKIs) are well known to interfere with cell plays an important role in SARS-CoV–induced pathogenesis.
cycle progression and to cause phase-speciﬁc cycle arrest. These
Conflict of Interest Statement : None of the authors has a financial relationship
inhibitors perturb the phosphorylation process by directly inter- with a commercial entity that has an interest in the subject of this manuscript.
acting with their target proteins, such as cyclins or CDKs. As shown
in Figure 5E, p53 phosphorylation on Ser-15 was up-regulated in Acknowledgments : The authors thank Prof. Eric J. Snijder (Leiden University Medi-
both 3a and M gene–expressed cells and gradually increased in a cal Center, The Netherlands), Prof. Milton Taylor (Indiana University, USA),
Dr. Sherief (Rutgers), and Dr. Gang Li for critical reading of the manuscript; Associated-
time-dependent manner in 3a/pCMV-myc–transfected cells, while Prof. Zhou Tao for the assay of confocal microscopy; and Drs. Liu Hong-Yan,
the protein levels of p53 were unaffected by 3a and M gene Li Su-Yan, and Feng Yan-Bin for the construction of some plasmids. The authors
expression. As no signiﬁcant difference on p53 phosphorylation also thank Dr. Baochang Fan (Indiana Univetsity) for some help.
was observed between 3a and M gene–expressed cells, and the
levels of p21Cip and p27Kip, members of the Cip/Kip subfamily, References
were unchanged by 3a or M expression (data not shown), it is 1. Poutanen SM, Low DE, Henry B, Finkelstein S, Rose D, Green K, Tellier
proposed that p53 phosphorylation might be induced by higher R, Draker R, Adachi D, Ayers M, et al. Identiﬁcation of severe acute
expression of foreign proteins at the Golgi apparatus and not respiratory syndrome in Canada. N Engl J Med 2003;348:1995–2005.
associated with 3a-induced G1 phase arrest. 2. Tsang KW, Ho PL, Ooi GC, Yee WK, Wang T, Chan-Yeung M, Lam
WK, Seto WH, Yam LY, Cheung TM, et al. A cluster of cases of
According to current concepts, the cell cycle commitment severe acute respiratory syndrome in Hong Kong. N Engl J Med 2003;
after restriction point passage requires the sustained stimulation 348:1977–1985.
by mitogens of the synthesis of labile D-type cyclins, which 3. Drosten C, Gunther S, Preiser W, van der Werf S, Brodt HR, Becker
associate with cyclin-dependent kinase (CDK) 4/6 to phosphory- S, Rabenau H, Panning M, Kolesnikova L, Fouchier RA, et al. Identi-
late Rb family proteins and sequester the CDK inhibitor such ﬁcation of a novel coronavirus in patients with severe acute respiratory
as p21(WAF1) and p27kip1. Cyclin D3 is expressed in nearly syndrome. N Engl J Med 2003;348:1967–1976.
4. Marra MA, Jones SJ, Astell CR, Holt RA, Brooks-Wilson A, Butterﬁeld
all proliferating cells and has shown the most broad expression
YS, Khattra J, Asano JK, Barber SA, Chan SY, et al. The Genome
pattern of the three D-type cyclins (cyclin D1, D2, and D3). Lin sequence of the SAS-associated coronavirus. Science 2003;300:1399–
and coworkers reported that the Cdk6–cyclin D3 complex is 1404.
unique among the D cyclin and kinase combinations in the ability 5. Rota PA, Oberste MS, Monroe SS, Nix WA, Campagnoli R, Icenogle
to promote the cell cycle start, evading the inhibition by JP, Penaranda S, Bankamp B, Maher K, Chen MH, et al. Characteriza-
p27(KIP1) and p21(CIP1) with a resemblance to viral cyclin- tion of a novel coronavirus associated with severe acute respiratory
bound Cdk6 (24). PTEN tumor suppressor gene, PKA activation, syndrome. Science 2003;300:1394–1399.
6. Snijder EJ, Bredenbeek PJ, Dobbe JC, Thiel V, Ziebuhr J, Poon LL,
and some anti-tumor agents such as glucocorticoids were re- Guan Y, Rozanov M, Spaan WJ, Gorbalenya AE. Unique and con-
ported to have arrested cell cycle progression in G1 phase by served features of genome and proteome of SARS-coronavirus, an
decreasing cyclin D3 mRNA levels and/or by inducing its protea- early split-off from the coronavirus Group 2 lineage. J Mol Biol 2003;
somal degradation. Furthermore, enforced expression of cyclin 331:991–1004.
D3 abrogated the PTEN-induced cell cycle arrest, while silencing 7. Herrewegh AA, Vennema H, Horzinek MC, Rottier PJ, de Groot RJ.
cyclin D3 by RNA interference further inhibited S phase entry, The molecular genetics of feline coronaviruses: comparative sequence
analysis of the ORF7a/7b transcription unit of different biotypes. Virol-
indicating a key role for cyclin D3 repression in these events
(25, 26). The decreases in the protein level of cyclin D3 were 8. Paul PS, Vaughn EM, Halbur PG. Pathogenicity and sequence analysis
observed in the measles virus (MV) and MHV-infected cells and studies suggest potential role of gene 3 in virulence of swine enteric
may play some roles for the virus in inducing G1 cell cycle arrest and respiratory coronaviruses. Adv Exp Med Biol 1997;412:317–321.
(20, 27). In this article, we ﬁrst reported that the expression of 9. Tan YJ, Teng E, Shen S, Tan TH, Goh PY, Fielding BC, Ooi EE, Tan
SARS-CoV 3a gene arrested cell cycle progression in G1 phase HC, Lim SG, Hong W. A novel severe acute respiratory syndrome
by a signiﬁcant down-modulation of cyclin D3. It would be inter- coronavirus protein, U274, is transported to the cell surface and under-
goes endocytosis. J Virol 2004;78:6723–6734.
esting to deﬁne the pathway for 3a protein to decrease the 10. Yu CJ, Chen YC, Hsiao CH, Kuo TC, Chang SC, Lu CY, Wei WC, Lee
expression of cyclin D3. CH, Huang LM, Chang MF, et al. Identiﬁcation of a novel protein
The role of cell cycle arrest induced by the 3a protein was 3a from severe acute respiratory syndrome coronavirus. FEBS Lett
not determined in the life cycle of SARS-CoV, but was proposed 2004;565:111–116.
from recent studies. Infection of MHV, a member of coronavirus 11. Zeng R, Yang RF, Shi MD, Jiang MR, Xie YH, Ruan HQ, Jiang XS,
family, was recently reported to result in inhibition of host cellu- Shi L, Zhou H, Zhang L, et al. Characterization of the 3a protein of
SARS-associated coronavirus in infected vero E6 cells and SARS
lar DNA synthesis and accumulation of cells in G1 phase in
patients. J Mol Biol 2004;341:271–279.
activating DBT and 17Cl-1 cells through inducing cyclin D2 12. Yuan X, Li J, Shan Y, Yang Z, Zhao Z, Chen B, Yao Z, Dong B, Wang
and cyclin E degradation (28). The expression of nonstructural S, Chen J, et al. Subcellular localization and membrane association of
protein p28 of MHV was reported to induce G1 phase arrest in SARS-CoV 3a protein. Virus Res 2005;109:191–202.
transfected cells and might be responsible for MHV to induce 13. Ito N, Mossel EC, Narayanan K, Popov VL, Huang C, Inoue T, Peters CJ,
cell cycle arrest (18). Increasing data proposed that cell cycle Makino S. Severe acute respiratory syndrome coronavirus 3a protein is
arrest in the G1 phase might favor coronavirus replication and a viral structural protein. J Virol 2005;79:3182–3186.
14. Shen S, Lin PS, Chao YC, Zhang A, Yang X, Lim SG, Hong W, Tan
exacerbate virus-induced pathogenicity, especially in some as- YJ. The severe acute respiratory syndrome coronavirus 3a is a novel
pect, for example increasing amounts of ribonucleotide pools structural protein. Biochem Biophys Res Commun 2005;330:286–292.
for efﬁcient coronavirus RNA synthesis, preventing the induc- 15. Deng LW, Chiu I, Strominger JL. MLL 5 protein forms intranuclear
tion and execution of early cell death in infected cells, assisting foci, and overexpression inhibits cell cycle progression. Proc Natl Acad
in efﬁcient coronavirus assembly, beneﬁting cap-dependent Sci USA 2004;101:757–762.
translation of coronavirus proteins, and decreasing the killing 16. Yuan X, Cong Y, Hao J, Shan Y, Zhao Z, Wang S, Chen J. Regulation
of LIP level and ROS formation through interaction of H-ferritin with
efﬁciency of coronavirus-infected cells by cytotoxic T cells (20, G-CSF receptor. J Mol Biol 2004;339:131–144.
28). In this article, we reported that expression of 3a could 17. Zhang Y, Li T, Fu L, Yu C, Li Y, Xu X, Wang Y, Ning H, Zhang S,
signiﬁcantly inhibit cell growth and induce cell G1 phase arrest Chen W, et al. Silencing SARS-CoV Spike protein expression in cul-
in different types of transfected cells and infected Vero E6 cells. tured cells by RNA interference. FEBS Lett 2004;27:141–146.
Rapid Communication 19
18. Chen CJ, Sugiyama K, Kubo H, Huang C, Makino S. Murine coronavirus 24. Lin J, Jinno S, Okayama H. Cdk6-cyclin D3 complex evades inhibition
nonstructural protein p28 arrest cell cycle in G0/G1 phase. J Virol by inhibitor proteins and uniquely controls cell’s proliferation compe-
2004;78:10410–10419. tence. Oncogene 2001;20:2000–2009.
19. Gemeniano MC, Sawai ET, Sparger EE. Feline immunodeﬁciency virus 25. Zhu X, Kwon CH, Schlosshauer PW, Ellenson LH, Baker SJ. PTEN
Orf-A localizes to the nucleus and induces cell cycle arrest. Virology induces G(1) cell cycle arrest and decreases cyclin D3 levels in endome-
2004;325:167–174. trial carcinoma cells. Cancer Res 2001;61:4569–4575.
20. Chen CJ, Makino S. Murine coronavirus replication induces cell cycle 26. Ausserlechner MJ, Obexer P, Bock G, Geley S, Koﬂer R. Cyclin D3
arrest in G0/G1 phase. J Virol 2004;78:5658–5669.
and c-MYC control glucocorticoid-induced cell cycle arrest but not
21. Locker JK, Rose JK, Horzinek MC, Rottier PJ. Membrane assembly of
apoptosis in lymphoblastic leukemia cells. Cell Death Differ 2004;
the triple-spanning coronavirus M protein. Individual transmembrane
domains show preferred orientation. J Biol Chem 1992;267:21911– 11:165–174.
21918. 27. Naniche D, Reed SI, Oldstone MB. Cell cycle arrest during measles
22. Maeda J, Repass JF, Maeda A, Makino S. Membrane topology of coro- virus infection: a G0-like block leads to suppression of retinoblastoma
navirus E protein. Virology 2001;281:163–169. protein expression. J Virol 1999;73:1894–1901.
23. Connell-Crowley L, Harper JW, Goodrich DW. Cyclin D1/Cdk4 regu- 28. Chen CJ, Makino S. Murine coronavirus-induced apoptosis in 17Cl-1
lates retinoblastoma protein-mediated cell cycle arrest by site-speciﬁc cells involves a mitochondria-mediated pathway and its downstream
phosphorylation. Mol Biol Cell 1997;8:287–301. caspase-8 activation and bid cleavage. Virology 2002;302:321–332.