| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Pathology, State University of New York at Stony Brook, New York, 11794, USA
2Correspondence: Department of Pathology, Academic Tower B, SUNY at Stony Brook, Nicolls Road, Stony Brook, NY 1794–8691, USA. E-mail: nreich{at}path.som.sunysb.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: IRF-3 • transcription • defense • gene expression
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B) and the dsRNA-activated factor (DRAF1)
(1
2
3
4)
. One set of genes induced in response to viral
dsRNA is that encoding type I interferons (IFNs) (5
, 6)
.
Newly synthesized IFNs are secreted from the cell and function to
promote resistance to viral infection, control cellular proliferation,
and activate cells of the immune system. IFNs bind to cell surface
receptors and transmit a signal to the nucleus by activation of Janus
tyrosine kinases (JAKs) and a latent transcription factor, the
IFN-stimulated gene factor 3 (ISGF3) (7
8
9
10)
. ISGF3 is
composed of members of the signal transducer and activator of
transcription factor (STAT) family, STAT1 and STAT2, and a member of
the interferon regulatory factor (IRF) family, p48/IRF-9. ISGF3 binds
to a DNA sequence, designated the IFN-stimulated response element
(ISRE), in the promoters of responsive genes and induces the expression
of IFN-stimulated genes (ISGs). The products of the ISGs mediate the
biological effects of IFN.
The infected cell may gain little protective benefit from autocrine IFN
due to the time required to produce and then respond to IFN. It is now
clear that the infected cell induces a distinct subset of ISGs in
direct response to virus infection or dsRNA, independent of IFN
signaling (2
, 3
, 11)
. We identified a specific cellular
transcription factor activated in response to infection that binds to a
DNA target site containing the ISRE and mediates gene induction. This
factor was designated the dsRNA-activated factor, DRAF1
(2
3
4)
. DRAF1 is now known to be a multimeric
transcription factor composed of interferon regulatory factor 3 (IRF-3)
and the transcriptional coactivators CREB binding protein (CBP) or p300
(4
, 12
13
14
15)
. IRF-3 resides in the cytoplasm of uninfected
cells, and after viral infection it accumulates in the nucleus in
association with CBP or p300 to form DRAF1. Association with CBP or
p300 is dependent on specific serine-threonine phosphorylation of IRF-3
by an activated protein kinase that remains to be identified.
IRF-3 is a member of a family of transcription factors known as IRFs
(16
17
18)
. The IRF family is defined by their shared
homology within an amino-terminal DNA binding domain. Through this DNA
binding domain, IRF family members recognize a DNA sequence contained
within an inner core of the ISRE. IRF-3 in complex with CBP/p300 as the
DRAF1 transcription factor recognizes a DNA target site that includes
the ISRE as well as several adjacent adenine residues (2
, 3)
. For this reason DRAF1 activates a distinct subset of genes
induced by IFN.
Infection by many viruses and the biological effect of viral dsRNA
signaling within a cell culminates in programmed cell death, or
apoptosis (19
, 20)
. However, the mechanism whereby viral
dsRNA triggers apoptosis is not completely understood. Apoptosis is a
genetically programmed pathway of cell death that is mediated in two
major execution programs initiated downstream of a death signal
(21
22
23)
. One of these programs is the caspase protease
cascade; the other is mitochondrial dysfunction mediated by the BCL
family of proteins. The decision to undergo apoptosis in many instances
is regulated by the balance of proapoptotic proteins and antiapoptotic
proteins present in the cell. In this report, we address whether
activation of IRF-3/DRAF1 and subsequent induction of gene expression
in response to viral dsRNA play a role in the promotion of apoptosis.
The induction of apoptosis in virus-infected cells may serve to
increase host survival by limiting the spread of infectious virions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. IFNAR1+/+ MEFs and
IFNAR1o/o MEFs were gifts from Paul J. Hertzog (Monash University, Australia) (25)
. EB1 cells
were a gift of Philip Shaw (CHUV Lausanne, Switzerland)
(26)
. Synthetic DNA and RNA polymers, poly (dI)-poly (dC),
poly (I)-poly (C), poly(A), and poly C were purchased from Pharmacia
Biotech (Piscataway, N.J.). Antibodies derived against IRF-3 were
described (4)
. Antibodies to the HA epitope (F-7), FLAG
epitope (OctAH6), CBP (A-22), p300 (N-15), and p53 (DO-1) were
purchased from Santa Cruz Biotech, Inc. (Santa Cruz, Calif.). Antibody
to the T7 epitope, was purchased from Novagen (Madison, Wis.). Antibody
to CrmA (A71–1) and caspase 3 was purchased from PharMingen (San
Diego, Calif.). Antibody to adenovirus 2 E1B 19 kDa protein (Ab-2) was
purchased from Calbiochem (San Diego, Calif.). Polyclonal rabbit serum
against PKR was a gift from Michael Mathews (New Jersey Medical School,
UMDNJ, Newark). Etoposide was purchased from Clontech (Palo Alto,
Calif.). Pfu DNA polymerase was purchased from Stratagene (San Diego,
Calif.) and Taq DNA polymerase was purchased from
Perkin-Elmer (Norwalk, Conn.).
Plasmid constructs
Human IRF-3 cDNA was cloned into pcDNA3 containing a T7 epitope
tag (T7-IRF-3) at the amino terminus using a polymerase chain reaction
(PCR). The cDNA encoding the constitutively active IRF-3 5D was a gift
from John Hiscott (McGill University, Montreal, Canada) and was cloned
into the pcDNA3-T7 vector (13)
. The expression plasmid
T7-
N60IRF-3 was constructed using PCR to amplify the IRF-3 region
encoding 60–427 aa. The plasmid, T7-DBDIRF-3, was constructed in a
similar fashion and encodes the DNA binding domain of IRF-3 (1–118
aa). The plasmid expressing IRF-3 with the serine 396 to aspartic acid
mutation was constructed using the Quick Change mutagenesis kit
(Stratagene). The EGFP gene was purchased from Clontech. The green
fluorescent protein (GFP)-IRF-3 construct was generated by PCR cloning
EGFP upstream of IRF-3 (27)
. PKR expression plasmid was a
gift from Michael Katze (University of Washington, Seattle)
(28)
. The CrmA cDNA was a gift from Yuri Lazebnik (Cold
Spring Harbor Laboratory, N.Y.) and was subcloned into pcDNA3.
Adenovirus E1B 19 K protein cDNA was a gift from Eileen White (Rutgers
University, New Brunswick, N.J.) (29)
was subcloned into
pcDNA3.1(-). The expression plasmids encoding FLAG-tagged p53 mutant
(F320–393) and p53 were gifts from Ute Moll (SUNY Stony Brook, New
York) (30)
.
Transfections and infections
Cells were transfected with calcium phosphate-DNA coprecipitates
with 15 µg of total plasmid DNA per 60 mm plate, 7.5 to 30 µg of
dsRNA (poly I:poly (C), or ssRNA (poly(A) or poly C). A luciferase
reporter gene assay was used containing a minimal promoter
(31)
with one copy of the ISG15 ISRE sequence (2
3
4
, 32)
. Luciferase activity was normalized to ß-galactosidase
activity by using the pSV-ß-Gal gene as a cotransfection control
(Promega, Madison, Wis.). Newcastle disease virus (NDV)
(NJ-LaSota-1946), a gift from Dr. Paula M. Pitha-Rowe (The Johns
Hopkins University, Baltimore, Md.), was propagated in embryonated hen
eggs and titers were determined by hemagglutination assay. Infections
were performed at 100 hemagglutination units/ml or with allantoic fluid
from mock-infected hen eggs (3)
.
Apoptosis assays
Nuclei were visualized after cell fixation in 4% formaldehyde,
permeabilization with 0.5% Triton-X 100/phosphate-buffered saline
(PBS), and DNA staining with 0.01 to 1.0 µg/ml of 4',
6-diamidino-2-phenylindole (DAPI) (Sigma, St. Louis, Mo.). The ApoAlert
Annexin V-EGFP apoptosis detection kit was purchased from Clontech.
Colorimetric protease assay kits for caspase activity were purchased
from PanVera Corporation (Madison, Wis.). Assays were performed with
substrates for caspase-8 (IETD-pNA) and caspase-3
(DEVD-pNA), and included controls with specific peptide
inhibitors for caspase-8 (IETD-FMK) or caspase-3 (DEVD-FMK).
Immunoprecipitation and Immunoblot
Whole cell extracts were prepared by lysis in buffer containing
50 mM Tris (pH 7.6), 400 mM NaCl, 0.5% Nonidet P-40, 5 mM K-EDTA, 1 mM
EGTA, 10 mM Naphosphate, 50 mM NaFl, and protease inhibitors.
Immunoprecipitations were performed with excess antibody for 2–4 h at
4°C. Immunocomplexes were collected on protein-G agarose (GibcoBRL,
Grand Island, N.Y.); proteins were separated by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and
electroblotted to Immobilon (Millipore). Immunoblotting with primary
antibody was followed by secondary antibodies (Amersham) and enhanced
chemiluminescence (DuPont NEN, Boston, Mass.).
Immunofluorescence
Cells were fixed in 4% formaldehyde and permeabilized in 0.5%
Triton-X 100, blocked in 3% bovine serum albumin/PBS-Tween-20 (0.1%),
and incubated with either 0.1 µg/ml of the T7 antibody, FLAG
antibody, or control antibody in blocking solution. Secondary
antibodies conjugated to Rhodamine were used (Jackson
ImmunoResearch Laboratories, Inc., West Grove, Pa.).
Protein-DNA binding assay
Electrophoretic mobility shift assays (EMSA) were performed as
described with a DNA probe containing the ISG15 ISRE (2
3
4
, 32)
. Proteins were synthesized in vitro with a
coupled transcription/translation system (Promega).
Northern blot hybridization
Total cellular RNA was extracted, separated by formaldehyde
denaturing gel electrophoresis, and transferred to Nytran (Schleicher &
Schuell, Keene, N.H.). mRNA was detected by hybridization to
radiolabeled DNA probes generated by Random prime-it II (Stratagene) of
DNA fragments corresponding to the human ISG54 gene (EcoRI
fragment of second exon) (gift from David Levy, New York University,
N.Y.) (33)
or the glyceraldehyde-3-dehydrogenase (GAPDH)
gene (gift from Ken Marcu, SUNY Stony Brook).
Images are presented with the use of Adobe Photoshop 4.0.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Apoptosis appears to play a role
in native immunity, but the molecular signaling events by which cells
react to virus and viral dsRNA remain unclear. We investigated whether
the induction of gene expression by the dsRNA-activated transcription
factor DRAF1 was involved in events leading to apoptosis. If DRAF1
plays a significant role in mediating the induction of apoptosis,
apoptosis should proceed in the absence of IFN autocrine action. To
address this, IFN-responsive cell lines (HeLa, HT1080) and
IFN-unresponsive cell lines (U3A, HEC-1B) were tested for an apoptotic
response to NDV. NDV and other members of the paramyxovirus family are
known to evoke cellular apoptosis (35
, 36)
. U3A cells are
deficient in STAT1 protein, and HEC-1B cells are defective in binding
IFN (24
, 37)
. Therefore, both cell lines are defective in
IFN signaling.
Cells were either mock infected or infected with NDV and evaluated for
morphological signs of apoptosis. Adherent cells were fixed and stained
with the fluorescent DNA intercalating agent DAPI to visualize nuclear
morphology. In all cases, NDV infection led to cell death characterized
by several hallmarks of apoptosis, including morphological changes of
rounding and refraction, loss of adherence, and development of pyknotic
nuclei (i.e., nuclear condensation and fragmentation). The results
indicate that cells respond to NDV with apoptosis either in the
presence or absence of IFN signaling (Fig. 1A
). This IFN-independent apoptosis was not specific to
immortalized cell lines as it was observed using primary murine embryo
fibroblasts (MEFs) derived from either wild-type (wt) mice or mice
deficient in the type I IFN receptor chain IFNAR1 (25
, 38)
. Both IFNAR1+/+ MEFs and
o/o MEFs underwent cell death in response to NDV
and showed apoptotic nuclear morphology (Fig. 1B
).
|
Viral dsRNA is produced during infection, and treatment of cells with
dsRNA in the absence of infection has been shown to induce apoptosis
(39
, 40)
. However, other viral products can elicit
apoptosis, such as the adenoviral E1A protein (41)
. We
addressed whether dsRNA, an activator of IRF-3/DRAF-1, was sufficient
to induce apoptosis in the absence of IFN signaling. Transfection of
cells with dsRNA, poly (I):poly (C), was found to induce apoptosis in
wt MEFs and in IFNAR1-deficient MEFs as observed by nuclear morphology
(Fig. 1B
). Apoptosis was not induced in mock-transfected
cells or by transfection with single-stranded RNA, poly (A), or poly
(C), or with dsDNA, poly (dI)-poly (dC) (data not shown). dsRNA-induced
apoptosis was also observed in the human cell lines HeLa, HEC-1B, and
U3A. Results of relative apoptosis induced by dsRNA in comparison to
control ssRNA are displayed for HEC-1B and U3A cells (Fig. 1C
). These results indicate that apoptosis induced in
response to virus infection and dsRNA can proceed in the absence of IFN
action.
Apoptosis depicted by nuclear morphology was confirmed in a second
manner. The annexin-V protein binds to phosphatidylserine (PS) and is
used to detect PS translocation from the inner face of the plasma
membrane to the outer surface during apoptosis. The annexin-V binding
assay has been widely used as a measure for apoptosis
(42)
. HEC-1B cells and IFNAR1o/o
MEFs undergoing cell death in response to NDV infection were positive
for annexin-V-EGFP binding (Fig. 1D
). Collectively, the
results indicate that NDV or dsRNA directly induce apoptosis in both
human and mouse cells, in both immortalized cell lines and primary
cells, independent of IFN signaling.
Typically, apoptosis proceeds through a complex cascade of protease
activation within the cell and subsequent proteolysis of cellular
components (21
, 22)
. Since the proteases responsible for
executing programmed cell death are the caspases, we examined whether
caspases were activated during NDV-infection of IFN nonresponsive
cells. The activity of caspase-8, an initiator caspase was first
assessed. Lysates were prepared from cells either mock infected or
infected with NDV, and caspase-8 activity was measured using a specific
peptide substrate in a colorimetric protease assay (Fig. 2A
). Caspase-8 was activated in both IFN responsive and
unresponsive cells. A similar result was found for that of caspase-3
(Fig. 2B
). Caspase-3 activation lies downstream of initiator
caspases, and falls into a class referred to as the executioners.
Activation of caspase-3 proceeds via cleavage of the 32 kDa proenzyme
into 17 kDa and 12 kDa polypeptides that heterodimerize to form the
active enzyme. The activation of caspase-3 by NDV infection was
confirmed by detection of the 17 kDa cleaved subunit. HEC-1B, U3A, and
HeLa cells were either mock infected, or infected for 20 h, or
40 h and proteins in cell lysates were analyzed by SDS-PAGE and
subsequent immunoblotting (Fig. 2C
). NDV infection led to
the appearance of the 17 kDa caspase-3 fragment and concomitant
decrease in the level of the 32 kDa proenzyme form in all cell lines
tested. We also tested the effect of broad-spectrum caspase inhibitor,
z-VAD-FMK, on the appearance of cellular apoptosis during NDV infection
(36
, 43)
. Inclusion of z-VAD-FMK during NDV infection
completely blocked the increased incidence of pyknotic nuclei, although
a delayed cytopathic effect was still apparent (data not shown). Taken
together, these results confirm that NDV infection induces apoptosis in
a caspase-dependent and IFN-independent manner.
|
Activation of IRF-3/DRAF1 promotes apoptosis
Since viral dsRNA appears to promote apoptosis and to activate
IRF-3/DRAF1, a direct role of IRF-3/DRAF1 in the apoptotic response was
further investigated. The effects of expression plasmids encoding IRF-3
or mutant forms of IRF-3 were tested (Fig. 3A
). To demonstrate intrinsic DNA binding properties, an
in vitro transcription/translation system was used to
generate wt or mutant forms of IRF-3. The in vitro
translated products were analyzed by an EMSA with a radiolabeled ISRE
probe (Fig. 3B
). Prior to infection, wild-type IRF-3 in its
unmodified form lacks intrinsic DNA binding ability (4)
.
It was reported previously that a constitutively active mutant of IRF-3
could be produced by substitution of five carboxyl serine or threonine
residues with the phosphomimetic aspartic acid (IRF-3 5D)
(13)
. In vivo, this mutation promotes IRF-3
association with CBP or p300 in the nucleus in the absence of viral
infection and induces gene transcription. In vitro, the
constitutively active IRF-3 5D protein can bind the ISRE target site.
The modification appears to alter protein conformation and promote DNA
binding. A T7 epitope tag positioned upstream of the DNA binding domain
(T7-IRF-3 5D) does not impair its ability to function. A single serine
396 to aspartic acid mutation within the IRF-3 carboxyl terminus is not
sufficient to allow the protein to bind DNA. A truncated protein
encoding just the amino-terminal DNA binding domain of IRF-3,
T7-DBDIRF-3, is competent to bind the ISRE. The T7-DBDIRF-3 protein
contains the nuclear localization signal of IRF-3 and localizes to the
nucleus in vivo constitutively (12
, 27)
. An
amino-terminal deletion mutation in IRF-3, T7-N60IRF-3, lacks a
portion of the DNA binding domain and cannot recognize DNA. However,
in vivo this protein can be phosphorylated in response to
viral infection and accumulate in the nucleus in association with
CBP/p300 (12, data not shown).
|
Expression plasmids encoding T7-IRF-3, T7-IRF-3 5D, T7-DBDIRF-3,
or T7-N60IRF-3, were transfected into IFN unresponsive HEC-1B cells
to evaluate their effect. At 48 h post-transfection, adherent
cells were fixed and stained with DAPI to detect nuclear morphology and
coordinate expression of IRF-3 was detected by immunofluorescence with
T7 antibodies. The percent apoptosis in the culture was quantified by
the number of cells positive for the T7 epitope that had pyknotic
nuclei detected with DAPI (Fig. 3C
). Expression of
T7-N60IRF-3 or T7-DBDIRF-3 elicited low apoptotic indices similar to
a vector control. However, in response to T7-IRF-3 5D, 
45% of the
cells demonstrated apoptotic nuclei, and expression of the wt T7-IRF-3
protein demonstrated approximately 35% apoptosis. This latter
result with wt T7-IRF-3 was unexpected since, in uninfected cells,
IRF-3 exists in a latent inactive state in the cytoplasm of the cell
(4
, 12
13
14
15)
. It implied that wt T7-IRF-3 was activated
during the transient expression assay. To investigate whether the
transfected T7-IRF-3 protein was activated, a coimmunoprecipitation
assay was performed to determine whether it associated with CBP/p300
(Fig. 3D
). HEC-1B cells were transiently transfected with
vector, T7-IRF-3, or T7-IRF-3 5D expression plasmids and proteins were
immunoprecipitated from lysates with control antibody, IRF-3 antibody,
or a mixture of CBP and p300 antibodies. Transfected IRF-3 proteins
were detected by immunoblot with a T7 antibody (upper panel) or an
IRF-3 antibody to detect both the endogenous IRF-3 and transfected
proteins (lower panel). The results demonstrated that transient
transfection of IRF-3 resulted in activation of a portion of the
expressed protein as measured by its association with CBP/p300. The
proportion of wt T7-IRF-3 protein that associated with CBP/p300 was
considerably less than that of the active IRF-3 5D protein, but the
level appears sufficient to induce cellular apoptosis. The increase in
apoptosis by IRF-3 or IRF-3 5D overexpression was also observed in the
U3A cell line (data not shown).
To ensure that the cell death mediated by T7-IRF-3 5D was indeed due to
apoptosis, the effects of the characterized apoptosis inhibitors
adenovirus E1B 19K and cowpox virus protein CrmA were examined in
cotransfection assays. The E1B 19K protein is a viral homologue of the
BCL-2 family of antiapoptotic proteins and has been shown to inhibit
apoptosis induced during adenovirus infection and by several other
apoptotic stimuli (41
, 44)
. CrmA is a cowpox viral protein
that binds to and inhibits members of the caspase family of proteases,
and thus inhibits apoptosis induced by a variety of agents
(45)
. HEC-1B cells were cotransfected with an expression
plasmid encoding GFP and either control vector or expression plasmids
encoding T7-IRF-3 5D alone or with E1B 19K or CrmA. Cells expressing
GFP were identified by fluorescent microscopy and used as a
cotransfection indicator. At 48 h post-transfection, the percent
apoptosis in the culture was quantified by the number of cells positive
for GFP fluorescence that had pyknotic nuclei detected with DAPI
staining (Fig. 4
). Using GFP as a cotransfection indicator generated a slightly lower
apoptotic index, but best represented the effects of coexpression of
different untagged plasmids. Coexpression of either E1B 19K or CrmA was
able to inhibit the incidence of pyknotic nuclei and apoptotic cell
death mediated by T7-IRF-3 5D expression. Protein expression of E1B 19K
and CrmA was confirmed by immunoblot (data not shown). This inhibitory
effect of E1B 19K and CrmA was not due to a nonspecific inhibition of
T7-IRF-3 5D expression (Fig. 4
, lower panel). As a positive control,
the effect of a characterized proapoptotic protein, the dsRNA-activated
protein kinase (PKR), was also evaluated in this cotransfection assay
(Fig. 4)
(19
, 46)
. PKR showed a similar level of apoptosis
induction quantified by pyknotic nuclei in cells coexpressing GFP.
|
Activation of gene expression by IRF-3/DRAF1 is requisite for
apoptosis
Since a fraction of wt IRF-3 was activated during transient
transfection to form DRAF1, it should elicit a specific effect on gene
transcription. This was confirmed by cotransfection of IRF-3 with an
ISRE-luciferase reporter gene assay (Fig. 5A
). HEC-1B cells were cotransfected with the reporter gene
and vector alone, T7-DBDIRF-3, or T7-IRF-3. Transfection of IRF-3
resulted in approximately a fourfold increase in specific gene
expression. Since IRF-3 activation appears to promote apoptosis, a
logical hypothesis for the mechanism by which this occurs is via the
activation of IRF-3 target genes. To test this hypothesis, a negative
interference approach was used to inhibit IRF-3-mediated gene
expression. The DBDIRF-3 protein containing only 1–118 aa binds the
ISRE target site, but not CBP/p300 and has been shown to constitutively
localize to the nucleus in vivo (Fig. 3)
(12
, 27)
. Through competitive binding to DNA, DBDIRF-3 should act as
a dominant negative interfering mutant of IRF-3. Coexpression of
T7-DBDIRF-3 with T7-IRF-3 was found to inhibit IRF-3-mediated
activation of the ISRE reporter gene in a dose-dependent manner (Fig. 5A
). The repressive effect of T7-DBDIRF-3 was not due to a
nonspecific repression in the level of T7-IRF-3 protein (lower panel).
|
To determine whether IRF-3-mediated apoptosis was dependent on
activation of target gene expression, T7-IRF-3 was cotransfected with a
negative interfering mutant, T7-DBDIRF-3, and an apoptosis assay was
performed (Fig. 5B
). A GFP expression plasmid was used as a
marker for the cotransfection. Cells were fixed and stained with DAPI
to visualize nuclear morphology. The expression of T7-IRF-3 elicited a
threefold increase in apoptotic cells, and this increase in apoptosis
was completely inhibited by the coexpression of T7-DBDIRF-3. As
positive controls for the inhibition of apoptosis, IRF-3 was
coexpressed with the apoptosis inhibitors E1B19K or CrmA. The results
indicate that IRF-3-mediated apoptosis is dependent on the
transcriptional activation of target genes containing the ISRE.
The effect of a dominant interfering mutant of IRF-3 was also tested
during the cellular response to NDV infection. ISRE-mediated gene
expression is induced during viral infection of HEC-1B cells (2
, 3)
. To test the effect of an IRF-3 interfering mutant, cells
were transfected with the ISRE-luciferase reporter gene and vector
plasmid or plasmids encoding T7-DBDIRF-3 or T7-N60IRF-3, and
subsequently mock-infected or infected with NDV (Fig. 6A
). NDV infection led to 20-fold induction of the reporter
gene, whereas coexpression of T7-DBDIRF-3 or T7-N60IRF-3
dramatically inhibited ISRE-mediated gene expression. Although
T7-N60IRF-3 cannot bind DNA, it may inhibit IRF-3 by sequestration
of CBP/p300 (4
, 12
, 13
, 27)
.
|
To address whether gene induction mediated by IRF-3/DRAF1 is necessary
for apoptosis in response to NDV infection, a similar transient
transfection-apoptosis assay was performed in combination with virus
infection (Fig. 6B
). HEC-1B cells were cotransfected with
expression plasmids for GFP and either empty vector, T7-IRF-3,
T7-DBDIRF-3, E1B 19K, or CrmA. Subsequently cells were either
mock-infected or infected with NDV, and the adherent cells were fixed
and stained with DAPI to visualize nuclear morphology. The percent of
apoptosis was expressed as GFP-positive cells with pyknotic nuclei.
Cells with vector alone responded to NDV infection with 35%
apoptosis. Cotransfection of IRF-3 did not significantly enhance the
apoptosis, indicating a surplus of endogenous IRF-3 protein. Expression
of either of the antiapoptotic proteins E1B 19K or CrmA inhibited the
level of NDV-induced apoptosis. More significantly, expression of the
negative interfering IRF-3 mutant T7-DBDIRF-3 inhibited cellular
apoptosis in response to viral infection. Expression of T7-DBDIRF-3,
E1B19K, or CrmA inhibited the appearance of pyknotic nuclei in response
to NDV infection, but did not completely inhibit the cytopathic effect
of the virus. These results demonstrate that expression of cellular
target genes by IRF-3/DRAF1 is required for apoptosis in response to
viral infection.
Effect of etoposide and p53 on IRF-3/DRAF1 induced apoptosis
It is clear that the induction of gene expression by IRF-3/DRAF1
in response to viral dsRNA is critical in mediating a cellular
apoptotic response, but only a few genes are known to be induced
directly by IRF-3/DRAF1. One of these genes, ISG54, can clearly be
shown to be induced by either dsRNA or by the activation of IRF-3 in
cells unresponsive to autocrine IFN (3
, 11
, 33)
(Fig. 7A
). To follow expression of the endogenous ISG54 gene, cells
were transfected with either dsRNA as poly (I):poly (C) or with the
gene encoding the constitutively active IRF-3 5D. Northern blot
analysis readily detects expression of endogenous ISG54 mRNA in
response to dsRNA or IRF-3 5D.
|
To investigate a possible involvement of IRF-3/DRAF1 in the cellular
response to DNA-damaging agents that induce apoptosis, the effect of
etoposide on expression of the ISG54 gene was assessed. The plant
derivative etoposide is a cancer therapy drug used to induce apoptosis.
Etoposide inhibits the DNA unwinding enzyme topoisomerase II, and this
leads to an increase in DNA strand breaks and protein levels of the p53
tumor suppressor (47
48
49)
. U3A cells or HT1080 cells were
untreated, treated with etoposide, or infected with NDV. RNA was
prepared and analyzed by Northern blot hybridization to detect ISG54
mRNA or GAPDH mRNA as a control. The ISG54 gene was found to be induced
in response to either NDV or etoposide (Fig. 7B
). A
reduction in GAPDH mRNA was observed during viral infection that may
indicate the activation of a latent RNase, and for this reason the
ISG54 induction may be underestimated (50)
. Induction of
ISG54 in response to virus or etoposide is independent of autocrine IFN
action since U3A cells lack STAT1 and cannot respond to IFNs. To
determine whether IRF-3/DRAF1 could be responsible for gene induction
in response to etoposide, activation of IRF-3 was evaluated. Activated
IRF-3 is known to relocalize from the cytoplasm to the nucleus in
response to dsRNA or viral infection (4
, 12
13
14
15
, 27)
. To
detect IRF-3/DRAF1 activation, cells were transfected with GFP-IRF-3
expressing a GFP-IRF-3 fusion protein (27)
. Cells were
either untreated or treated with etoposide, and GFP-IRF-3 protein was
visualized by fluorescence microscopy (Fig. 7C
). The results
indicated that IRF-3 was activated in response to the apoptotic agent
etoposide since GFP-IRF-3 accumulated in the nucleus.
Etoposide leads to activation of the p53 tumor suppressor, and p53
activation is known to be able to trigger apoptosis (47
, 49
, 51
52
53
54
55)
. For this reason, the involvement of p53 was
investigated in the IRF-3-mediated cellular response. It was first
necessary to demonstrate that the IFN-unresponsive cells were sensitive
to the apoptotic effects of p53. This was demonstrated by
overexpression of p53 in HEC-1B cells (Fig. 8A
). To control for the involvement of p53, a dominant
negative approach was used with a p53 mutant encoding a
carboxyl-terminal domain of p53 (F320–393) (30)
. This
truncated form of p53 forms nonfunctional hetero-oligomers with
endogenous p53 and blocks p53 dependent action (56
, 57)
.
Cells were transfected with plasmids expressing GFP with p53 or
dominant negative (DN) p53, or with both p53 and DNp53. Cells were
evaluated for GFP expression and apoptotic nuclei with DAPI staining.
p53 overexpression clearly induced apoptosis in the HEC-1B
cultures, and coexpression of DNp53 inhibited the apoptosis.
|
To evaluate the effect of increased p53 levels and concomitant
apoptosis in the absence of DNA damage or a transfection, a cell line
that expresses wt p53 under the control of the metallothionein promoter
was used (55)
. If p53 indirectly or directly stimulates
IRF-3/DRAF1 activity, DRAF1 target genes should be induced. Cells were
treated with zinc sulfate to induce high levels of p53 synthesis, and
expression of endogenous ISG54 mRNA was measured. A Northern blot of
RNA isolated from untreated or treated cells did not reveal induction
of ISG54, although p53 protein levels increased (Fig. 8B, C
). This result suggests that p53-mediated apoptosis does not
require IRF-3/DRAF1, whereas the effects of etoposide stimulate signals
to activate both p53 and IRF-3/DRAF1 and thereby induce ISG54 mRNA
expression.
Although p53-mediated apoptosis did not appear to require the action of
IRF-3/DRAF1, it remained possible that IRF-3/DRAF1-mediated apoptosis
required the action of p53. To determine the involvement of p53, a
dominant negative approach was used with the p53 mutant encoding the
carboxyl-terminal domain of p53 (30)
. Cells were
transfected with plasmids expressing GFP and the active IRF-3 5D in the
absence or presence of cotransfected DNp53 and evaluated for apoptotic
nuclei with DAPI staining. Expression of p53 was detected by
immunofluorescence with antibodies to a FLAG epitope tag. Apoptosis
induced by IRF-3 5D was found to be unaffected by DNp53 since there was
no alteration in the apoptosis induced in the absence or presence of
cotransfected DNp53 (Fig. 8D
). The results suggest that
apoptosis mediated by IRF-3/DRAF1 is independent of the action of p53.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 6)
. IFNs are
secreted, bind to cell surface receptors, and elicit biological changes
that confer resistance to viral infection. It is also clear that the
infected cell has an intrinsic response mechanism that results in
apoptosis. Therefore, a delicate temporal balance must exist of
antiapoptotic signals to allow production and secretion of protective
cytokines and proapoptotic signals to initiate cell suicide.
We demonstrate in this report that NDV infection and viral dsRNA
stimulate apoptosis in the absence of IFN action since it occurs in
cells that cannot respond to IFN. Apoptosis was measured by the
development of pyknotic nuclei, phosphatidyl serine externalization,
and activation of caspase proteases. An earlier report suggested that
the apoptotic response to virus was dependent on the action of IFNs,
although IFN treatment by itself was not sufficient to induce apoptosis
(58)
. Although it remains to be determined how different
viruses or other factors influence cellular apoptosis, our results
strongly suggest that this response is activated independent of the IFN
pathway.
Since we earlier identified DRAF1 as a transcription factor that
is activated in response to viral infection or dsRNA transfection, we
hypothesized that DRAF1 could be a mediator of cellular apoptosis
(2
3
4)
. The IRF-3 component of DRAF1 is phosphorylated
during viral infection, and this modification allows IRF-3 to bind to
CBP or p300 and accumulate in the nucleus as DRAF1 (4
, 12
13
14
15
, 27)
. To investigate the possibility that DRAF1 induces
transcription of genes that initiate apoptosis, we tested the effect of
overexpression of wtIRF-3 and mutations of IRF-3 in the absence of
viral infection in cells that lack a response to IFN.
Transient overexpression of IRF-3 or a constitutively active mutant,
IRF-3 5D, led to an enhancement of apoptosis during a transient
transfection. Apoptosis was measured by the appearance of pyknotic
nuclei at a specific time and therefore does not include cells that
have lifted from the surface or have not yet progressed to terminal
stages of apoptosis. Even within this window of time, a significant
level of apoptosis was visible in the cultures expressing IRF-3 5D, and
was also apparent with wt IRF-3 (Fig. 3)
. It appears that introduction
of plasmid DNA can lead to activation of signaling pathways sensitive
to dsRNA and the phosphorylation and activation of wt IRF-3. Apoptosis
induced by activated IRF-3 or by NDV infection can be inhibited by the
well-characterized antiapoptotic proteins adenovirus E1B 19K and cowpox
CrmA. More significantly, the apoptotic response can be inhibited by
coexpression of a mutant of IRF-3 that blocks specific gene
transcription by DRAF1. The dominant negative DNA binding domain
(DBDIRF-3) localizes to the nucleus and appears to occupy target sites
without the ability to activate gene expression (12
, 27)
.
The p53 tumor suppressor occupies a prominent role in the apoptotic
response of the cell to stress signals such as oncoproteins, hypoxia,
and DNA damage (51
52
53
54
55)
. To determine the involvement of
p53 in the apoptotic response to viral infection and IRF-3/DRAF1
activation, we first evaluated a genotoxic stress inducer, etoposide,
and the induction of high levels of p53. Genotoxic stress inducers have
recently been reported to activate IRF-3, which can be demonstrated by
visualizing the localization of a GFP-IRF-3 fusion protein with
fluorescent microscopy (Fig. 7)
(59)
. The induction of
apoptosis by etoposide was found to result in the activation of
IRF-3/DRAF1 and specific expression of an endogenous responsive gene
ISG54. Etoposide can therefore elicit stress signals that lead to the
induction/activation of p53 and to the activation of IRF-3/DRAF1. p53
induction in the absence of other stress inducers did not result in
ISG54 gene expression, indicating that p53 does not activate
IRF-3/DRAF1 or the set of genes known to be responsive to IRF-3/DRAF1.
In addition, a dominant negative p53 mutant did not alter apoptosis
induced by IRF-3/DRAF1, indicating independence of the action of p53.
Two other IRF transcription factors have been demonstrated to play a
role in promoting apoptosis by different agents (60
61
62)
.
IRF-1 has been implicated in oncogene-induced apoptosis and DNA
damage-induced apoptosis in mature lymphocytes. However IRF-1
activation was not sufficient to induce apoptosis on its own. IRF-8,
also known as interferon consensus sequence binding protein (ICSBP),
appears to be involved in apoptosis since myeloid cells from mice
deficient in IRF-8/ICSBP show a decrease in both spontaneous apoptosis
during development and in that induced by DNA damage. Our study
presents evidence for a mechanism by which viral dsRNA leads to
apoptosis of infected cells with the activation of the IRF-3/DRAF1
transcription factor. The study clearly indicates that apoptosis in
response to dsRNA or NDV infection is independent of the action of IFN
as it occurs in various cells deficient in the IFN signal pathway.
This report presents evidence demonstrating that apoptosis induced in response to the paramyxovirus, NDV, or dsRNA proceeds independent of interferon or p53 action. Activation of the IRF-3/DRAF1 transcription factor by dsRNA leads to the induction of a subset of cellular genes whose products stimulate apoptosis. This apoptotic response to viral infection has apparently evolved as a mechanism to limit the dissemination of virus.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication April 24, 2000.
Revision received July 31, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
T7) to visualize the transfected
proteins (upper panel) and subsequently with a murine polyclonal
antibody to IRF-3 (
