Activation of STAT6 by STING is critical for antiviral innate immunity.
ABSTRACT STAT6 plays a prominent role in adaptive immunity by transducing signals from extracellular cytokines. We now show that STAT6 is required for innate immune signaling in response to virus infection. Viruses or cytoplasmic nucleic acids trigger STING (also named MITA/ERIS) to recruit STAT6 to the endoplasmic reticulum, leading to STAT6 phosphorylation on Ser(407) by TBK1 and Tyr(641), independent of JAKs. Phosphorylated STAT6 then dimerizes and translocates to the nucleus to induce specific target genes responsible for immune cell homing. Virus-induced STAT6 activation is detected in all cell-types tested, in contrast to the cell-type specific role of STAT6 in cytokine signaling, and Stat6(-/-) mice are susceptible to virus infection. Thus, STAT6 mediates immune signaling in response to both cytokines at the plasma membrane, and virus infection at the endoplasmic reticulum.
- [show abstract] [hide abstract]
ABSTRACT: Interleukin (IL)-4 plays an important role in the differentiation of naive T helper (Th) cells into Th2. Mast cells can produce a significant amount of IL-4 and have been proposed to play a major role in the induction of Th2 responses. Recently, it has been reported that mast cells have a distinct IL-15 receptor system different from that of T or natural killer cells. In the present study, we demonstrated that IL-15 induced IL-4 production from a mouse mast cell line, MC/9, and bone marrow-derived mast cells. IL-4 mRNA expression was increased by IL-15, suggesting that IL-15 promotes IL-4 expression at the transcriptional level. In these mast cells, signal transducer and activator of transcription (STAT) 6 were rapidly tyrosine-phosphorylated in response to IL-15. In MC/9 cells, the expression of a C-terminally truncated dominant negative form of STAT6 significantly suppressed the IL-4 mRNA up-regulation by IL-15, suggesting that STAT6 activation is essential for the IL-15-mediated IL-4 production. Additionally, tyrosine phosphorylation of Tyk2 was rapidly increased by IL-15 treatment in this cell line. Altogether, our results suggest that IL-15 plays an important role in stimulating early IL-4 production in mast cells that may be responsible for the initiation of Th2 response.Journal of Biological Chemistry 09/2000; 275(38):29331-29337. · 4.65 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: IFN-alpha/beta-mediated functions promote production of MIP-1alpha (or CCL3) by mediating the recruitment of MIP-1alpha-producing macrophages to the liver during early infection with murine CMV. These responses are essential for induction of NK cell inflammation and IFN-gamma delivery to support effective control of local infection. Nevertheless, it remains to be established if additional chemokine functions are regulated by IFN-alpha/beta and/or play intermediary roles in supporting macrophage trafficking. The chemokine MCP-1 (or CCL2) plays a distinctive role in the recruitment of macrophages by predominantly stimulating the CCR2 chemokine receptor. Here, we examine the roles of MCP-1 and CCR2 during murine CMV infection in liver. MCP-1 production preceded that of MIP-1alpha during infection and was dependent on IFN-alpha/beta effects for induction. Resident F4/80(+) liver leukocytes were identified as primary IFN-alpha/beta responders and major producers of MCP-1. Moreover, MCP-1 deficiency was associated with a dramatic reduction in the accumulation of macrophages and NK cells, as well as decreased production of MIP-1alpha and IFN-gamma in liver. These responses were also markedly impaired in mice with a targeted disruption of CCR2. Furthermore, MCP-1- and CCR2-deficient mice exhibited increased viral titers and elevated expression of the liver enzyme alanine aminotransferase in serum. These mice also had widespread virus-induced liver pathology and succumbed to infection. Collectively, these results establish MCP-1 and CCR2 interactions as factors promoting early liver inflammatory responses and define a mechanism for innate cytokines in regulation of chemokine functions critical for effective localized antiviral defenses.The Journal of Immunology 03/2005; 174(3):1549-56. · 5.52 Impact Factor
Activation of STAT6 by STING Is Critical
for Antiviral Innate Immunity
Huihui Chen,1,2,6Hui Sun,1,2,6Fuping You,1,2,6Wenxiang Sun,1,2Xiang Zhou,1,2Lu Chen,1,2Jing Yang,1,2Yutao Wang,1,2
Hong Tang,1,2Yukun Guan,1,2Weiwei Xia,1Jun Gu,1Hiroki Ishikawa,4Delia Gutman,4Glen Barber,4Zhihai Qin,5
and Zhengfan Jiang1,2,3,*
1State Key Laboratory of Protein and Plant Gene Research
2Key Laboratory of Cell Proliferation and Differentiation of the Ministry of Education
School of Life Sciences, Peking University, Beijing, China
3Peking University-Tsinghua University Joint Center for Life Sciences, Beijing, China
4Department of Medicine and Sylvester Comprehensive Cancer Center, University of Miami School of Medicine, Miami, FL 33136, USA
5State Key Laboratory of Biomacromolecules, Institute of Biophysics, Beijing, China
6These authors contributed equally to this work
STAT6 plays a prominent role in adaptive immunity
by transducing signals from extracellular cytokines.
We now show that STAT6 is required for innate
immune signaling in response to virus infection.
Viruses or cytoplasmic nucleic acids trigger STING
(also named MITA/ERIS) to recruit STAT6 to the
endoplasmic reticulum, leading to STAT6 phosphor-
ylation on Ser407by TBK1 and Tyr641, independent of
JAKs. Phosphorylated STAT6 then dimerizes and
translocates to the nucleus to induce specific target
genes responsible for immune cell homing. Virus-
induced STAT6 activation is detected in all cell-types
tested, in contrast to the cell-type specific role of
STAT6 in cytokine signaling, and Stat6–/–mice are
susceptible to virus infection. Thus, STAT6 mediates
immune signaling in response to both cytokines at
the plasma membrane, and virus infection at the
Innate immunity is the first line of defense against microbial
infection. Recognition of pathogens is mainly mediated by
pattern recognition receptors (PRRs), including Toll-like recep-
tors (TLRs), RIG-I-like receptors (RLRs) and NOD-like receptors
(NLRs) (Takeuchi and Akira, 2010), that trigger signal cascades
to upregulate the expression of various cytokines. In the case
of viral infection, endosomal TLRs and cytoplasmic RLRs detect
viral DNAs or RNAs and induce the production of type I IFN,
which are potent inhibitors of viral replication (Gitlin et al., 2006;
Kato et al., 2005, 2006). RLRs, including RIG-I and Mda5, are
sensors of viral RNAs in the cytoplasm; in response to viral infec-
tion, RLRs associate with the adaptor protein MAVS/Cardif/
IPS-1/VISA (Kawai et al., 2005; Meylan et al., 2005; Seth et al.,
2005; Xu et al., 2005), an integral membrane protein that func-
tions on both mitochondria and peroxisomes through distinct
mechanisms (Dixit et al., 2010); the RLR/MAVS complex facili-
tates TBK1/IKKε-mediated activation of IRF3/7 and NF-kB,
which lead to the induction of type I IFNs. Besides viral RNA,
cytoplasmic double-stranded DNA (dsDNA) also induces type I
IFNs, but the exact identity of the receptor in this situation is
currently not fully established (Ishii et al., 2006; Stetson and
Medzhitov, 2006). A recently identified adaptor protein, endo-
plasmic reticulum IFN stimulator (STING, also named MITA/
ERIS) (Ishikawa and Barber, 2008; Sun et al., 2009; Zhong
et al., 2008) exhibits a vital role in dsDNA signaling (Ishikawa
through STING (IFI16 [Unterholzner et al., 2010]) or via the RIG-I–
MAVS axis (involving RNA polymerase III mediated transcription
of cytoplasmic DNA [Ablasser et al., 2009; Chiu et al., 2009]),
and both pathways ultimately result in the recruitment and acti-
vation of TBK1, which in turn activates IRF3/7 and NF-kB.
Many cytokines, including type I IFNs, exert their effects
through the canonical JAK (Janus kinase)-STAT (signal trans-
ducers and activators of transcription) pathway (Levy and Dar-
nell, 2002). Specifically, IL-4 and IL-13 activate STAT6 (Takeda
et al., 1996) resulting in T helper cells 2 (Th2) polarization (Aki-
moto et al., 1998; Hebenstreit et al., 2006; Shimoda et al.,
1996). IL-4 induces the phosphorylation of IL-4 receptor, which
in turn recruits cytosolic STAT6 by its SH2 domain; the recruited
STAT6 is phosphorylated on tyrosine 641 (Y641) by JAK1, which
results in the dimerization and nuclear translocation of STAT6
to activate target genes (Mikita et al., 1996, 1998). Several cyto-
kines, including IL-3/15, IFN-a and platelet-derived growth
factor (PDGF-BB), activate STAT6 in different cell types (Bula-
nova et al., 2003; Masuda et al., 2000; Quelle et al., 1995), and
induce over 150 diverse targets, many of which are involved in
Th2-associated processes (Elo et al., 2010; Wei et al., 2010). A
thorough understanding of biological consequences of STAT6
signaling awaits additional studies.
It is known that NF-kB, AP-1 and IRFs are responsible for the
induction of many IFN-stimulated genes (ISGs), however, the
role of STAT6 in anti-viral response is unclear. Here we report
a STAT6-dependent antiviral innate immune signaling event
436 Cell 147, 436–446, October 14, 2011 ª2011 Elsevier Inc.
that leads to the induction of chemokines, including CCL2,
CCL20, and CCL26, and these chemokines recruit immune cells
to combat viral infection. More importantly, virus induces STAT6
activationindependently of JAK,but instead relies on STINGand
TBK1, as well as MAVS in the case of RNA virus. The physiolog-
ical significance of the novel pathway is reflected by a higher
susceptibility of Stat6–/–mice to viral infections; moreover, unlike
other cell type-specific STAT6 signaling pathways, virus-
induced STAT6 activation is ubiquitously detected, implying
a fundamental requirement of this mechanism in the defense
against viral infections.
STAT6 Interacts with STING in Response to Virus
Using C-terminal STING (aa 178–379) as bait in the yeast 2-
hybrid screen, we identified an STING-STAT6 interaction and
confirmed it in 293 cells by coimmunoprecipitation (coIP) (Fig-
ure S1A available online). Specifically, the DNA-binding domain
(DBD) of STAT6 and STING C terminus (aa 317–379) were
required for this interaction (Figures S1B–S1D). We next exam-
ined this interaction at endogenous protein levels. Analysis
with confocal microscope showed a dispersed pattern of
STAT6 in the cytosol of unstimulated HeLa cells; upon infection
with Sendai virus (SeV, an RNA virus), STAT6 redistributes to the
perinuclear regions, colocalizes with STING, and eventually
translocates into the nucleus (Figure 1A). CoIP analyses also re-
vealed an inducible interaction of endogenous STAT6 with
STING, as well as MAVS and TBK1, in SeV-infected primary
MEFs, 2fTGH and THP-1 cells (Figure 1B). Consistent with these
observations, endogenous STAT6 co-fractionate with STING in
HeLa cell lysates after Herpes simplex virus 1 (HSV-1, a DNA
virus) infection (Figure 1C). Similar results were obtained from
SeV-infected HeLa cells, with an additional location to a mixed
fraction containing MAVS-resident mitochondria-associated
ER membrane (MAM) and also MAVS-resident peroxisomal
membrane (Dixit et al., 2010; Ishikawa et al., 2009; Zhong
et al., 2008) (Figure 1D). These data demonstrate that STAT6
interacts with STING during virus infection.
STAT6 Is Activated upon Virus Infection
293 cells lack a functional endogenous STAT6 but express the
other components of the IL-4 signaling pathway (Mikita et al.,
1996, 1998). Taking advantage of this property, we first estab-
lished a 293 cell-line stably expressing Flag-STAT6 (293-
STAT6) and confirmed its normal responsiveness to IL-4/13
with intact Y641phosphorylation (data not shown). Virus infection
resulted in the nuclear translocation of STAT6, suggesting
that STAT6 may serve as a transcriptional activator under this
situation. To confirm this hypothesis, we assessed Y641phos-
phorylation of STAT6, since it is required for STAT6 activation
in response to cytokines. We found that STAT6 was indeed
phosphorylated on Y641in SeV-infected and poly (I:C)/poly
dAdT-transfected cells, and this STAT6 phosphorylation takes
place prior to the phosphorylation of IRF3 and other STATs
(Figures 2A, 2D, and 2H, and Figures S2B and S2H). A STAT6-
responsive luciferase reporter (E3-Luc) (Yuan et al., 2006) was
activated in 293-STAT6 cells upon virus infection and poly
(I:C)/poly dAdT transfection, whereas a nonresponsive control
reporter (mutated at the STAT6-binding site, E3-Luc-M) was
not affected (Figure 2C and Figure S2A). By contrast, neither
reporter was activated in 293 cells, indicating a transactivation
function of STAT6 in response to virus. These findings imply
a previously unknown pathway of STAT6 activation in response
to viral infection and cytoplasmic dsRNA/DNA.
STAT6 can be activated by several cytokines. To clarify
a potential role of cytokines in STAT6 activation during viral chal-
lenges, we first monitored cytokine production in virus-infected
cells. Neither IL-4 nor IL-13 was induced by virus (Figure 2B
SeV 0 7 14
0 7 14 0 7 14 0 7 14 (h)
7 14 (h)0
0 2 6 (h)
Figure 1. Virus-Induced STAT6-STING Interaction
and STAT6 Translocation
(A) STAT6 translocates and colocalizes with STING after
virus infection. Confocal microscopy of endogenous
STING (red), STAT6 (green) and the merge in HeLa cells
infected with Sendai virus (SeV) for the indicated hours.
Nuclei were stained with DAPI. All images are represen-
tative of at least three independent experiments in which
>95% of the cells displayed similar staining. Scale bars
represent 10 mm.
(B) STAT6 interacts with STING and TBK1 after virus
infection. Primary MEFs, 2fTGH and THP-1 cells were in-
fected with SeV for the indicated hours. Cell lysates were
STAT6, TBK1 and MAVS antibodies. WCL, whole cell
(C and D) Virus infection induces STAT6 translocation.
Western blot analyses of fractionated HeLa cells infected
with Herpes simplex virus 1 (HSV-1) (C) or SeV (D) for the
indicated hours. Cyt, cytosolic; ER, endoplasmic retic-
ulum-rich; Nuc, nuclear; and Mit, mitochondrial; were re-
vealed by Caspase 3, Calnexin, Histone H3, and COX IV,
See also Figure S1.
Cell 147, 436–446, October 14, 2011 ª2011 Elsevier Inc. 437
and Figure S2C), thus excluding their involvement in STAT6 acti-
vation after virus infection. Strikingly, other cytokines including
type I IFNs, IL-8 and STAT6-induced genes (CCL2 and CCL20,
see below), displayed similar kinetics post infection. Therefore,
CCL2/20 is unlikely regulated by cytokines like type I IFNs or
IL-8. In fact, when media of SeV-infected 293-STAT6 cells
were used to treat naive 293-STAT6 cells, STAT6 phosphoryla-
tion was only detected in virus-infected but not media-treated
cells, whereas phosphorylation of STAT1/2/3/5 was detected
in media-treated cells (Figure 2D and data not shown), excluding
any STAT6-activating cytokines in the media within these time
points. Furthermore, STAT6 phosphorylation was intact upon
SeV and poly (I:C) stimulation when production of cytokines
including IL-8 and type I IFNs was inhibited by cycloheximide
(CHX) pretreatment (Figures 2E and 2F). These data collectively
indicate a cytokine –independent pathway of STAT6 activation
upon virus infection.
Next we used 2fTGH and its derivative cell lines (Kumar et al.,
1997) (each deficiency in a single key component of the JAK–
STAT pathway, Figure S2D)to test if any of these known proteins
in JAK-STAT pathway would be required for virus-induced
STAT6 activation. Notably, U4A cells with JAK1 deficiency did
Figure 2. Virus-Induced STAT6 Activation
(A) SeV induces STAT6 phosphorylation. 293-STAT6 cells were infected with SeV for the indicated hours. Phosphorylation of the indicated proteins was analyzed
by western blot. (Top panel: membrane was probed with a-P-STAT6, developed, and reprobed with a-P-IRF3.)
(B) Kinetics of cytokine induction by virus. Supernatants of cells in (A) and prolonged infected-cells as indicated were subject to ELISA or type I IFN bioassay.
Asterisk indicates levels that were not detectable.
(C) Virus infection activates STAT6. 293-STAT6 cells transfected with an E3-Luc (STAT6-responsive reporter) or E3-Luc-M (STAT6-nonresponsive reporter) were
assay (fold induction).
(D) Virus-induced STAT6 activation is independent of cytokines in culture media. Naive 293-STAT6 cells were incubated with supernatants from SeV-infected
293-STAT6 cells for indicated times. Both infected (upper panels) and media-treated cells (lower panels) were analyzed for phosphorylation of the indicated
proteins by western blot. Lysate from IL-4 treated cells was used as a positive control for STAT6 phosphorylation detection.
(E) Virus-induced STAT6 phosphorylation does not require protein synthesis. Mock or cycloheximide (CHX, 5 mg/ml) pretreated (for 2 hr) 293-STAT6 cells were
infected with SeV or transfected with poly (I:C) for 10 hr. STAT6 phosphorylation was analyzed by western blot.
(F) Left, ELISA analyses of IL-8 production in supernatants of cells treated in (E) for 24 hr. Right, IFN-b-Luc reporter assay for the induction of IFN-b in the same
(G) JAK1 and IFNAR2 are not required for STAT6 activation by virus. U4A (JAK1–/–) and U5A (IFNAR2–/–) cells transfected with an E3-Luc were treated with IL-13
for 12 hr or transfected with poly (I:C) for 24 hr. STAT6 activation was analyzed using luciferase assay (fold induction).
(H) JAK1 and IFNAR2 are not required for STAT6 phosphorylation by virus. 2fTGH, U4A and U5A cells were infected with SeV for the indicated hours.
Phosphorylation of the indicated proteins was analyzed by western blot. IFNs, 2 hr of type I IFNs (500 unit/ml IFN-a & 500 unit/ml IFN-b) treatment.
(I) Y641phosphorylation is a prerequisite for STAT6 activation. 293 cells transfected with an E3-Luc and wild-type (WT) or mutant STAT6 were infected with SeV,
transfected with poly (I:C) for 24 hr or treated with IL-4 for 12 hr. Luciferase activity was analyzed as fold induction (upper panel). Vec, empty vector; F641Y, the
reversed Y641F. Expression of STAT6 was analyzed by western blot (lower panel).
Data are means ± SEM. See also Figure S2.
438 Cell 147, 436–446, October 14, 2011 ª2011 Elsevier Inc.
not respond to IL-4/13 (Figure 2G and Figures S2D–S2G), while
U5A cells deficient in IFNAR2 were completely defective in IFN-
a/b response (Figure S2F). Surprisingly, STAT6 from all these
cells responded normally to virus infection (Figures 2G and 2H
is indispensable. Besides, cytokine effects were re-examined
using these cells. Treatment with a high concentration of IFNs
and other STAT6-responding cytokines including IL-3/15 and
PDGF-BB did not result in STAT6 phosphorylation in 2fTGH or
U5A cells (Figure 2H, Figure S2F, and data not shown), support-
ing our previous conclusion that virus-induced STAT6 activation
Although virus-induced STAT6 activation appears to be quite
different from that induced by IL-4/13, Y641phosphorylation is
essential for both activation pathways since mutation of this
residue totally abolished its response to virus or IL-4/13 (Fig-
ure 2I). The SH2 domain of STATs is essential not only for both
receptor-binding and dimerization, but is also required for DNA
binding by this family of proteins. To investigate the role of
SH2 domain in virus-induced STAT6 activation, we searched
for potential residue(s) in this domain that might be important.
We found that L551A mutant lost the ability to respond to virus,
albeit with normal response to IL-4/13, and the defect was fully
rescued by the reversion mutation (Figures S2I–S2L). Further
experiments showed that the L551A mutation abrogated Y641
phosphorylation and STAT6 homo-dimerization in response to
by IL-4 (Figures S2M-S2O). Collectively these data suggest that
STAT6 is differentially activated by virus and IL-4/13.
Specific Target Genes
(A) Microarray analysis of poly (I:C)-induced genes.
Expression levels in 293 cells were arbitrarily set to 1
(green). Heat map of genes most strongly upregulated by
poly (I:C) in both cells (top) and only in 293-STAT6 cells
(bottom). Stars indicate genes confirmed by quantitative–
PCR, western blot and/or ELISA.
(B) Quantitative–PCR analyses of U4A and U5A cells
transfected with poly (I:C) or treated with IL-4 for the indi-
cated hours. Data were normalized to the relative expres-
sion of the HPRT1 reference gene.
(C) CCL2 and CCL20 are upregulated in the absence of
JAK1 or IFNAR2. ELISA analyses of 2fTGH (2f), U4A and
U5A cells transfected with poly (I:C) (Trans pIC), treated
with poly (I:C) (Add pIC) or infected with SeV for 24 hr.
(D) CCL2 promoter contains a functional STAT6 site that is
responsive to virus infection. A schematic presentation of
STAT6 binding site mutants is shown (upper). 293-STAT6
cells transfected with the indicated promoter-reporters
were infected with SeV or HSV-1. Luciferase activity was
analyzed as fold induction.
Data are means ± SEM. See also Figure S3.
3. Virus-Activated STAT6Regulates
Virus-Activated STAT6 Regulates
a Specific Set of Target Genes
Using DNA microarrays, we compared mRNAs
that are significantly induced in mock or poly
(I:C) transfected 293 and 293-STAT6 cells.
Among 30, 968 genes examined, poly (I:C) trans-
fection induced the expression of numerous ISGs in both cells,
including OAS1, CCL5 and IFIT1/3, and a set of genes only in
293-STAT6 cells (Figure 3A), suggesting that these genes
are specifically regulated by STAT6; notably among the
STAT6-regulated genes are the chemokines CCL2, CCL20,
and CCL26, which are responsible for the recruitment of immune
cells to sites of infection. The microarray data were validated
by either quantitative–PCR or ELISA (Figures 3B and 3C and
Figure S3), respectively.
The transcriptome of poly (I:C) stimulated 293-STAT6 cells
displayed substantial difference from that of IL-4/13 activated
cells. In fact, while CCL11 was only upregulated by IL-4 treat-
ment, CCL26 could be induced by both IL-4 and virus. Consis-
tently, both SeV and poly (I:C) were able to induce CCL2 and
CCL20 in U4A and U5A cells (Figures 3B and 3C). Furthermore,
inspection of CCL2 promoter sequence revealed one typical and
two putative STAT6 binding sites (–1129 to –1120, –585 to –576,
and –293 to –285 relative to the transcriptional start site, respec-
tively). SeV and HSV-1 infection indeed activated a luciferase
reporter driven by DNA segment containing these sites, but not
the one with mutation in the first STAT6 binding site (Figure 3D),
suggesting that CCL2 promoter harbors a functional STAT6
binding site that is responsive to virus infection.
STING Mediates STAT6 Activation by Virus
Next, we sought to investigate the molecular mechanism of
STAT6 signaling in response to virus. The translocation and
interaction of STAT6 with STING after virus infection (Figure 1)
raised the possibility that STING is involved in STAT6 activation.
Cell 147, 436–446, October 14, 2011 ª2011 Elsevier Inc. 439
To address this possibility, we first assessed the effect of RNAi
knockdown of STING. Suppression of STING expression almost
completely abolished STAT6 activation in response to SeV
infection, but had little effects on IL-4 treatment (Figures S4A–
S4C). Meanwhile, the induced interaction between MAVS and
STAT6 (Figure S4D) and translocation of STAT6 from cytosol to
ER fraction and later to nucleus was barely detectable (Fig-
ure S4E). Further evidence for a critical role of STING in virus-
derived STAT6 activation showed that Sting–/–MEFs lost
CCL2/20 induction in response to virus and transfected genomic
DNA, although comparable amounts of IL-6 and type I IFNs
This result also demonstrated that STING is dispensable for
RLR-mediated type I IFN production. As a control, WT and
STING-reconstituted Sting?/?MEFs showed normal response,
highlighting the vital role of STING in virus-induced STAT6
activation. STAT6 from Sting?/?MEFs retained in the cytosol
after virus infection (Figure 4B). Exogenous human STAT6
(hSTAT6) was not phosphorylated on Y641(Figure 4C), nor did
it dimerize (Figure 4D) in Sting?/?MEFs after virus infection.
These data as a whole indicated that STING is required for
virus-induced STAT6 activation.
MAVS Is Required for STAT6 Activation by RNA Virus
Since STAT6 was also localized to MAVS-resident MAM and
peroxisomes (Figure 1D), we speculated that MAVS might take
a part in STAT6 signaling after virus infection. Indeed, MAVS,
RIG-I–N, and Mda5-N (where N denotes N-terminal CARD
module) strongly activated STAT6 (Figure S5A), consistent with
a role for STAT6 in RNA virus infection. Yeast 2-hybrid assays
showed that STAT6 interacted only with STING but not with
MAVS or TBK1, each of which could bind STING (Figure S5B).
This result suggested that STING may act as a platform that
assembles the STAT6 signal complex which includes MAVS in
the case of RNA virus stimulation.
0 7 14
0 7 14
0 714 (h)
0 12 16 1216
IL-6 (pg/ml) CCL20 (pg/ml)
0 12 16 1216
Figure 4. STING Is Required for Virus-Induced
(A) ELISA analyses of cytokine production and type I IFN
bioassay in WT, Sting–/–and Sting–/–-hSTING MEFs
(STING–/–MEFs reconstituted withhuman STING) infected
DNA from Calf thymus (CT gDNA)/Listeria Monocytogenes
(LM DNA)/E. coli.
(B) Western blot analyses of fractionated WT and Sting–/–
MEFs infected with SeV for the indicated hours. Cyt,
cytosol; Nuc, nucleus.
(C) Sting–/–and Sting–/–-hSTING MEFs expressing human
STAT6 were infected with the indicated viruses or treated
with IL-4. STAT6 phosphorylation was analyzed by
(D) Cells in (C) were analyzed by native-PAGE and western
blot for dimerization of STAT6 and IRF3.
Data are means ± SEM. See also Figure S4.
To verify this point, we examined virus-
induced cytokine production in cells derived
from Mavs?/?mice. SeV could only induce
CCL2/20 secretion in bone marrow-derived
macrophages (BMDMs) from WT, but not Mavs?/?mice; on
the contrary, Mavs?/?cells responded normally to HSV-1 (Fig-
ures 5A and Figure S5C). Similarly, CCL2/20 production was
severely diminished in sera or in organs (lungs and livers) of
Mavs?/?mice intravenously infected with SeV compared to
their heterozygous littermates (Figure 5B). SeV infection, but
not HSV-1 infection, resulted in more severe lung pathology,
with massive infiltration of monocytes, in WT relative to Mavs?/?
mice (Figure 5C). In addition, SeV-induced STAT6 nucleus trans-
location was completely abolished in Mavs?/?cells (data not
shown). The requirement for MAVS in STAT6 signaling was
confirmed by that virus-induced CCL2/20 production was fully
restored in human MAVS-reconstituted Mavs?/?MEFs (Fig-
ure 5D). These data demonstrate that MAVS is required for
RNA virus-induced STAT6 activation.
ical STAT6 pathway, and we observed normal CCL11 induction
in relevant knockout MEFs in response to IL-4 except Stat6?/?
MEFs (Figure 5E), suggesting that the canonical STAT6 signaling
pathway is intact and is independent of STING or MAVS.
TBK1 Is Required for STAT6 Phosphorylation
Our previous results suggest that individual JAK deficiency has
little effect on virus-triggered STAT6 phosphorylation, but it
remains to be tested whether simultaneous lack of two or more
JAKs would inhibit the process. Using Jak inhibitor Ruxolitinib
(INCB 018424) and CP690550, which show specific inhibition
of JAK1/2 and JAK3/JAK2, respectively, we found that the inhib-
itors had little impact on virus-induced STAT6 phosphorylation
in U1A cells that is deficient of Tyk2 protein (Figure 6A). Consis-
tently, none of the JAKs was phosphorylated in virus-infected
U5A cells while STAT6 phosphorylation persisted (Figure 6B),
ingly, TBK1, IKKε, and IKKb overexpression led to obvious shift
in the mobility of STAT6 on SDS-PAGE (Figure 6C), whereas
440 Cell 147, 436–446, October 14, 2011 ª2011 Elsevier Inc.