Biochimica et Biophysica Acta 1863 (2016) 3148–3159

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamcr

Yersinia YopJ negatively regulates IRF3-mediated antibacterial response

through disruption of STING-mediated cytosolic DNA signaling
Ye Cao a,b,c, Kai Guan c, Xiang He c, Congwen Wei c, Zirui Zheng c, Yanhong Zhang c, Shengli Ma a,b,c,
Hui Zhong c,⁎, Wei Shi a,b,⁎⁎
a
Key Laboratory for Molecular Enzymology & Engineering, the Ministry of Education, Jilin University, Changchun 130012, China
b
College of Life Science, Jilin University, Changchun 130012, China
c
State Key Laboratory of Pathogen and Biosecurity, Beijing Institute of Biotechnology, Beijing 100850, China

a r t i c l e i n f o a b s t r a c t

Article history: The Yersinia outer protein J (YopJ) plays a pivotal role in evading the host immune response and establishes a per-
Received 7 June 2016 sistent infection in host cells after bacterial infection. YopJ is a cysteine protease and can act as a deubiquitinating
Received in revised form 30 September 2016 enzyme that deubiquitinates several targets in multiple signaling pathways. Stimulator of interferon genes
Accepted 7 October 2016
(STING) is a critical adapter for the induction of interferon regulatory factor 3 (IRF3) phosphorylation and subse-
Available online 11 October 2016
quent production of the cytokines in response to nucleic acids in the cytoplasm. Our studies demonstrate that
Keywords:
YopJ targets STING to inhibit IRF3 signaling. Specially, YopJ interacts with STING to block its ER-to-Golgi traffic
Yop J and remove its K63-linked ubiquitination chains. Deubiquited STING perturbs the formation of STING-TBK1 com-
STING plex and the activation of IRF3. The 172th cysteine of YopJ mediated STING deubiquitination and IRF3 signaling
Deubiquitination inhibition. Consequently, mice infected with WT and ΔYopJ/YopJ bacteria induced lower levels of IRF3 and
IRF3 IFN-β, decreased inflammation and reduced staining of STING as compared to ΔYopJ and ΔYopJ/YopJ C172A
IFN-β strains infection. The data herein reveal a previously unrecognized mechanism by which YopJ modulates innate
immune signaling.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction

The host innate immune system provides a critical first line of

defense against infectious agents, including pathogenic bacteria [1].
Abbreviations: BMDMs, bone marrow-derived macrophages; c-di-GMP, cyclic When pathogenic bacteria invade hosts, the pathogen associated mole-
diguanylate monophosphate; DMEM, Dulbecco's modified Eagle's medium; ER, endoplas- cule patterns (PAMPs) can be recognized by pattern recognition recep-
mic reticulum; FBS, fetal bovine serum; GST, Glutathione S-transferase; HEK293T, human
tors (PRRs) and then trigger the rapid induction of cytokines to
embryonic kidney; IB, immunoblotting; IHC, immunohistochemical; K63, lysine 63;
MAPK, mitogen-activated protein kinase; MAP2Ks, mitogen-activated protein kinase ki- eliminate the pathogens [2].
nases; MAVS, mitochondrial antiviral signaling protein; M-CSF, macrophage colony- Several independent groups have identified STING as a central and
stimulating factor; NF-κB, nuclear factor κB; PAMPs, the pathogen associated molecule multifaceted mediator of the innate immune response [3]. Later studies
patterns; PBS, phosphate-buffered saline; PRRs, pattern recognition receptors; RIG-I, show that STING acts as not only an adaptor, but also a direct sensor
retinoic acid-induced gene I; SD, standard deviations; sictrl, siRNA control; siSTING,
STING siRNA oligos; STING, stimulator of interferon genes; SUMO, small ubiquitin related
of cytosolic DNA and cyclic diguanylate monophosphate (c-di-GMP)
modifiers; TAK1, TGFβ-activated kinase 1; TBK1, TANK-binding kinase 1; TBST, Tris- [4–6]. Cytosolic detection of pathogen-derived DNA and c-di-GMP are
buffered saline and tween 20; TMAs, tissue microarrays; TNFα, tumor necrosis factor α; the major mechanisms of bacterial-induced interferon production.
T3SS, type III secretion system; WT Y. pestis, wild type Yersinia pestis; YopE, Yersinia STING is ubiquitinated during the invasion of certain nucleic acids
outer protein E; YopM, Yersinia outer protein M (YopM); YopJ, Yersinia outer protein J;
and then separated from the endoplasmic reticulum (ER) to form
YopJ C172A, the 172th cysteine residue of YopJ was replaced by alanine; ΔYopJ Y. pestis,
Y. pestis lacking YopJ; Yops, Yersinia outer proteins. punctate complexes with TBK1 at the Golgi apparatus [7–10]. The
⁎ Corresponding author. STING-TBK1 complexes further activate the phosphorylation of
⁎⁎ Correspondence to: W. Shi, Key Laboratory for Molecular Enzymology & Engineering, IRF3, and the phosphorylated IRF3 forms dimers to activate tran-
the Ministry of Education, Jilin University, Changchun 130012, China. scription of downstream genes in the nucleus [11–14]. Therefore,
E-mail addresses: caoye1987122@163.com (Y. Cao), geminigk@aliyun.com (K. Guan),
hexiang_spring@yahoo.com (X. He), weicw@yahoo.com (C. Wei), zirui_zheng@163.com
STING functions as both an adaptor and a sensor, and has been dem-
(Z. Zheng), yanhongzhang333@163.com (Y. Zhang), shengli0128@163.com (S. Ma), onstrated to facilitate downstream signal transmission to IRF3 and
towall@yahoo.com (H. Zhong), shiw@jlu.edu.cn (W. Shi). nuclear factor κB (NF-κB).

http://dx.doi.org/10.1016/j.bbamcr.2016.10.004
0167-4889/© 2016 Elsevier B.V. All rights reserved.
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3149

Many Gram-negative bacterial pathogenic agents have acquired 2.3. Bone marrow-derived macrophages (BMDMs) isolation
strategies to escape host immune system by injecting effector pro-
teins into host cells by using a type III secretion system (T3SS). BMDMs were isolated from C57BL/6 mice, cultured in DMEM con-
Y. pestis injects at least 6 kinds of Yersinia outer proteins (Yops) taining 20 ng/ml macrophage colony-stimulating factor (M-CSF,
into the infected macrophages, resulting in the inhibition of the in- eBiosciences, CA, USA) for 6 days, and then seeded into 6-well plates
herent immune signaling, which ultimately allows the bacteria to es- 24 h before infection.
tablish a systemic infection [15,16]. Through subverting the innate
immune response, YopJ nullifies the activation of pro-inflammatory 2.4. Bacterial genome DNA extracting and bacterial infection
cytokines, and subsequently induces macrophage apoptosis to pre-
vent CD8+ T-cell responses and contributes to successful infection WT Y. pestis was cultured in Hiss Agar (CM1102, LAND BRIDGE) and
[17]. YopJ, but not the catalytically inactive YopJ C172A, inhibits the the genome DNA was extracted using the TIANamp Bacteria DNA Kit
mitogen-activated protein kinase (MAPK) and the NF-κB signaling (DP302, TIANGEN) with the A260/A280 of DNA solution between 1.75
pathways by acting as both an acetyltransferase and deubiquitinase and 1.85. Both the stimulation of Y. pestis genome DNA (Y. pestis DNA)
[18–20]. Although NF-κB is found to be one of the host target mole- and poly (dA: dT) (P0883, Sigma–Aldrich) were at the concentration of
cule of the YopJ, Y. pestis could affect host immune system at multiple 0.2 μg/ml for 6 h. RAW264.7 macrophages were infected with Y. pestis
levels. In this study, we show that T3SS effector YopJ in Y. pestis tar- at a multiplicity of infection (MOI) of 20. The infection was initiated by
gets STING and inhibits cytosolic DNA antibacterial signaling, reveal- centrifuging the plate at 700 ×g for 5 min. After incubation for 1 h at
ing a previously unrecognized mechanism by which YopJ modulates 37 °C, the plates were washed three times with 1× phosphate-buffered
innate immune signaling. saline (PBS), transferred into fresh medium containing gentamicin
(100 μg/ml) to kill the extracellular bacteria. Samples were subjected to
immunoblotting (IB). Normalization of the levels of α-Tubulin was used
2. Materials and methods to measure the mean intensity for quantifying the density of each band.

2.1. Strains, plasmids and siRNA

2.5. Dual luciferase reporter assays
Y. pestis biovar Microtus str. 91001 was used as wild type in this study
HEK293T cells were transfected with reporter plasmids (1 μg per
and ΔYopJ Y. pestis was constructed based on the wild one. The
plate, pGL3-IRF3-luc/UAS-luc or pGL3-IFN-β-luc), Renilla luciferase
pACYC184 vector was used to construct the YopJ or YopJ C172A (the
constructs (pRL-TK vector, 0.02 μg per plate, E2241, Promega), and var-
172th cysteine residue was replaced by alanine) overexpressing strains
ious combinations of expression plasmids using jetPRIME® reagent
(ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis). The mammalian expression
after being seeded in 12-well plates for 24 h. Indicated empty vectors
plasmids of retinoic acid-induced gene I (RIG-I), mitochondrial antiviral
were used to equalize the total amounts of plasmids. The reporter activ-
signaling protein (MAVS), STING, TBK1, Yersinia outer protein E (YopE),
ity of prepared cell extracts was detected using the Dual-Luciferase®
Yersinia outer protein M (YopM), YopJ, ubiquitin, K48-ubiquitin, K63-
Reporter Assay System (E1910, Promega). Results were presented as
ubiquitin encoding Flag-, Myc-, HA-fusion proteins were constructed
fold relative to the activity of uninfected or unstimulated cells. Data
by inserting PCR-amplified fragments into pcDNA3-Flag, pCMV-Myc,
were representative of three independent experiments.
pCMV-HA vacant vectors respectively and the 172th nucleotide of the
YopJ gene was mutated from cysteine to alanine by using a Quik-
Change Mutagenesis Kit (Stratagene) to generate the Myc-YopJ C172A 2.6. Immunoprecipitation and IB analysis
plasmids (Supplementary Table S1). pCDH-EF1-Flag-STING was con-
structed by inserting PCR-amplified fragments of STING into pCDH- Cell lysates were prepared in lysis buffer (50 mM Tris-HCl pH 7.5,
EF1-MCS-T2 A–puro vacant vector. The reporter constructs pGL3-IRF3- 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 10 mM
luc/UAS-luc (IRF-luciferase reporter plasmids), pGL3-IFN-β-luc have sodium fluoride, protease inhibitors cocktail, 5% Nonidet P-40).
been described earlier [21–23]. Glutathione S-transferase (GST) fusion Anti-Flag M2 Affinity Gel (A2220, Sigma–Aldrich) or anti-mouse
proteins were generated by expression in pGEX4T-1-based vectors IgG conjugated beads (42472, Sigma-Aldrich) were used for immu-
(Amersham Biosciences Biotech) in Escherichia coli BL21 (DE3). Enzyme noprecipitation of soluble proteins. An aliquot of the total lysates
analysis and DNA sequencing were performed to verify all kinds of (5%, v/v) was included as a control. IB was performed with anti-
plasmids. The fluctuation of A260/A280 of all plasmids solution was be- Acetylated-Lysine (9441; Cell Signaling Technology), anti-HA
tween 1.75 and 1.85. siRNA (20 nM, GenePharma) was transfected by (H9658; Sigma-Aldrich), HRP-conjugated anti-Flag (A8592; Sigma-
jetPRIME® reagent (114-01, Polyplus) following the manufactory's in- Aldrich), anti-IRF3 (2241-1; Epitomics), anti-IRF3 Phospho pS386
structions and the sequences of siRNA were showed in Supplementary (2562-1; Epitomics), anti-MAVS (sc-166583; Santa Cruz biotechnol-
Table S2. Bacterial infection was performed 48 h later after the transfec- ogy), anti-Myc (A5598; Sigma-Aldrich), anti-RIGI (sc-98911; Santa
tion of siRNA. Cruz biotechnology), anti-TBK1 (3296–1; Epitomics), anti-ubiquitin
(sc-166,553, Santa Cruz Biotechnology), anti-ubiquitin K63-linkage
specific (05-1308; HERCK MILIIPORE), anti-ubiquitin K48-linkage
2.2. Cell culture and transfections specific (6919-s; Epitomics), anti-STING (ab92605; abcam) or anti-
α-Tubulin (T6074; Sigma-Aldrich) antibody, respectively. Films (X-
Human embryonic kidney (HEK293T) and the murine macro- OMAT BT film, Carestream) were used to visualize the antigen-antibody
phage cell line RAW264.7 were grown in Dulbecco's modified Eagle's complexes by chemiluminescence (NEL104001EA, PerkinElmer Life
medium (DMEM, 10-013, Corning cellgro) supplemented with 10% heat- Sciences).
inactivated fetal bovine serum (FBS, 10,100-147, Gibco), 2 mM L-gluta-
mine, 100 U/ml penicillin, and 100 mg/ml streptomycin. Thp-1 cells 2.7. GST pull-down assays
were grown in RPMI Medium 1640 (31,800-022, Gibco) with 10% heat-
inactivated FBS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 mg/ml DE3 was used to express the GST and GST fusion proteins. The subse-
streptomycin and treated by TPA (0.1 μg/ml, 723,134, Sigma) for 48 h be- quent purification of both kinds of proteins was according to the protocol
fore using. Transient transfections were performed with jetPRIME® re- (28-9523-61, Amersham Biosciences Biotech). Cell lysates were incubat-
agent following the manufactory's instructions. ed with 5 μg purified GST or GST fusion proteins bound to glutathione
3150 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159

beads at 4 °C for 2 h. The absorbates were washed by lysis buffer and IB The interaction between exogenous STING and endogenous TBK1 was
was performed with the total lysates (5%, v/v) as a control. determined by the intensity and distribution of the red spots under fluo-
rescent microscopy.
2.8. Immunofluorescence
2.13. Histopathological scoring
Cells were washed three times with 1 × PBS. For Flag-STING, Myc-
YopJ, or Golgin 97 visualization, cells were fixed with 2% paraformalde- Liver and spleen tissues were fixed in 10% phosphate-buffered for-
hyde for 15 min at 37 °C and permeabilized for 5 min with 0.2% Triton X- malin, dehydrated by a tissue processor and embedded in paraffin. Mi-
100. Samples were blocked with 5% skim milk in Tris-buffered saline crotome (5 μm) was used to section tissues and then hematoxylin and
and tween 20 (TBST) for 1 h. Anti-Flag (1:150, SAB4301135, Sigma- eosin (HE) were stained by an autostainer. Histological scoring of
Aldrich), anti-Myc (1:150, A5598, Sigma-Aldrich), and anti-Golgin 97 coded slides was assessed by a single pathologist, blinded to the source
(1:150, sc-74632, Santa Cruz) antibodies were used to detect the Flag- of the slides, based on a modified system.
STING, Myc-YopJ and Golgin 97 proteins, respectively. Staining was vi-
sualized with secondary antibodies conjugated with Alexa Fluor 405 2.14. Immunohistochemical (IHC) staining
(ab175649, abcam), 594 (SA00006-7, Proteintech) or 488 (SA00003-3,
Proteintech) and the images were captured with a digital camera IHC staining for STING was performed on the paraffin-embedded tis-
under a confocal microscope (Zeiss LSM510). To see the colocation of sue microarray. The tissue microarrays (TMAs) were performed on 5 μm
STING and Golgi more intuitively, we interchanged the colors of STING thick sections. A graded alcohol series was used for tissue slides de-
and YopJ in the analysis software. paraffinizing and rehydrating. Antigen retrieval was performed by plac-
ing the slides in 10 mM sodium citrate buffer (pH 6.0) and maintained at
2.9. Cycloheximide chase assay a sub-boiling temperature for 10 min. The slides were immersed in 10%
normal goat serum in 1× PBS for 30 min to block non-specific staining,
HEK293T cells were cotransfected with Flag-STING and Myc-vector followed by primary antibody overnight at 4 °C in a humidified cham-
or Myc-YopJ, treated with cycloheximide (50 μM) for the indicated ber. 1 × PBS was used for slides washing. Then, secondary antibody
time points. Samples were analyzed by IB with anti-Flag and anti-Myc was used for 30 min at room temperature.
antibody. The half-life of the STING protein was calculated as the time
required for degradation of 50% of the protein. 2.15. Statistical analysis

2.10. RNA analysis Two-tailed Student's t-test was used to statistically analyze the dif-
ference between two groups. All data points were the average of tripli-
Total RNA was extracted from samples using TRNzol-A + Regent cates, with error bars representing standard deviations (SD). All data
(DP421, TIANGEN), and first-strand cDNA was generated from total were representative of results from at least three independent experi-
RNA using TransScript One-Step gDNA Removal and cDNA Synthesis ments. ⁎⁎P b 0.01 or ##P b 0.01 was considered statistically significant.
SuperMix (AT311, TRANSGENE). Quantitative RT-PCR was performed
using AceQ qPCR SYBR Green Master Mix (Q121-02, VazymeTM) in trip- 3. Results
licate and analyzed on an ABI Prism 7700 analyzer (Applied Biosystems,
Foster, CA, USA). All real-time values were normalized to GAPDH. The 3.1. YopJ inhibits IRF3 activity induced by cytosolic DNA pathway
primer sequences were showed in Supplementary Table S3 [24].
Due to the importance of IRF3 in host defense against bacterial infec-
2.11. Ubiquitination assays tion [25], we initiated our study by examining the extent of cytosolic
DNA activated IRF3 activation in THP-1 cells. c-di-GMP can be directly
HEK293T cells were cotransfected with Flag-STING, Myc-YopJ or recognized by STING to activate TBK1-IRF3 signaling. To this end, cells
Myc-vector together with HA-ubiquitin, HA-K48 ubiquitin or HA-K63 carrying pGL3-IRF3–luc/UAS-luc and pRL-TK vector were transfected
ubiquitin. For detection of endogenous ubiquitinated-STING following together with Y. pestis DNA or c-di-GMP. As was shown, the IRF3 activity
WT, ΔYopJ, ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis infection, cells was increased by cytosolic Y. pestis DNA and c-di-GMP (Fig. 1A). Similar
were infected at MOI = 20. Cells were treated with MG132 for 8 h be- enhancement of IFN-β luciferase activity by Y. pestis DNA or c-di-GMP
fore harvesting cells. Immunoprecipitated were performed with anti- was also observed in THP-1 cells (Fig. 1A). Since STING expression is
Flag M2 Affinity Gel or Protein A/G plus-Agarose together with anti- very low in HEK293T cells (Data not shown), we reconstitute respon-
STING antibody. The ubiquitination signal was detected using anti-HA, siveness of HEK293T cells to c-di-GMP by Flag-STING overexpression.
anti-ubiquitin, anti-ubiquitin K63-linkage specific or anti-ubiquitin As we can see, Y. pestis DNA or c-di-GMP transfection also increased
K48-linkage specific antibody, respectively. IRF3 and IFN-β reporters in HEK293T cells (Fig.1B). These data sug-
gested that Y. pestis infection can trigger cytosolic DNA sensing pathway
2.12. In situ PLA assays by its genome DNA or its second messenger.
We next would like to determine if Yops were capable of blocking
RAW264.7 macrophages were transiently transfected with Flag- DNA-mediated induction of IRF3 activation. Expression constructs for
STING, cells were infected with different strains of Y. pestis after 24 h. YopE, YopJ and YopM were generated and cotransfected into THP-1
The infected cells were treated according to the protocol of Duolink In cells respectively together with Y. pestis genome or c-di-GMP. As
Situ PLA® Probe Anti-Rabbit MINUS (DUO92005, OLINK BIOSCIENCE). shown in Fig. 1C, the induction of IRF3 activation by Y. pestis DNA or c-

Fig. 1. YopJ inhibits IRF3 activity induced by cytosolic DNA pathway. (A, B) THP-1 (A) or HEK293T cells (B) were transfected with pGL3-IRF3-luc/UAS-luc or pGL3-IFN-β-luc together with
pRL-TK vector and stimulated by Y. pestis genome or c-di-GMP for 6 h before harvesting. Luciferase activity was measured and normalized for transfection efficiency. (C, E) THP-1 (C) or
HEK293T cells (E) were transfected with pGL3-IRF3-luc/UAS-luc together with indicated plasmids and stimulated by Y. pestis genome or c-di-GMP for 6 h before harvesting. Luciferase
activity was measured and normalized for transfection efficiency. (D, F) THP-1 (D) or HEK293T cells (F) were transfected with pGL3-IRF3-luc/UAS-luc, pRL-TK vector together with in-
creasing amount of Myc-YopJ and stimulated by Y. pestis genome or c-di-GMP for 6 h before harvesting. Luciferase activity was measured and normalized for transfection efficiency.
Whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control. Cell-based studies were performed at least three independent times
with comparable results. Data represent means ± the standard errors of the means. Student's t-test was used for statistical analysis: **P b 0.01; ##P b 0.01, versus different control
respectively.
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3151

di-GMP was only inhibited by YopJ. YopJ inhibited IRF3 induction by YopJ was also observed in HEK293T cells (Fig. 1E and F), suggesting
Y. pestis DNA or c-di-GMP in a dose-dependent manner (Fig. 1D). Simi- that YopJ is able to inhibit IRF3 activity induced by cytosolic DNA
lar repression of Y. pestis DNA-induced activation of IRF3 reporters by pathway.

A B
THP-1 HEK293T+Flag-STING
250 150
Ctrl ** Ctrl
Relative activation (fold)

Relative activation (fold)

200 Y.pestis DNA Y.pestis DNA
## **
c-di-GMP 100 c-di-GMP
150
##
100
50
** ##
50 ** ##

0 0
IRF3-luc IFN-β-luc IRF3-luc IFNβ-luc

C D
THP-1 THP-1
80 100
Ctrl Ctrl
Y.pestis DNA 80 Y.pestis DNA
IRF3 activation (fold)

IRF3 activation (fold)

60 **
c-di-GMP c-di-GMP
60
40 ## **
40
##
20 ** **
## 20 ##

0 0
Myc-vector Myc-YopE Myc-YopJ Myc-YopM Myc-YopJ (µg) 0 0.2 0.4 0.8

(kDa) (kDa)

55- 35- IB: Myc

35- IB: Myc
25- 55- IB: α-Tubulin

55- IB: α-Tubulin

E F
HEK293T+Flag-STING HEK293T+Flag-STING
30 40
Ctrl Ctrl
Y.pestis DNA
IRF3avtivation (fold)

Y.pestis DNA
IRF3 activation (fold)

30
20 c-di-GMP c-di-GMP

20
** ##
10
**
10
## ##
** ** ##
0 0
Myc-vector Myc-YopE Myc-YopJ Myc-YopM 0 0.2 0.4 0.8
Myc-YopJ (µg)
(kDa) (kDa)
35-
IB: Flag 35- IB: Flag

55-
35- IB: Myc
35- IB: Myc
25- 55- IB: α-Tubulin

55- IB: α-Tubulin

3152 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159

A B IP C GST-Pulldown

GST-YopJ
Flag IgG Lysates

Lysates
IP
+ - - - - - - - +

GST
Flag-YopE
(kDa)
Flag Lysates - + - - + - - + - Flag-YopJ
IgG

- - + - - - + - - Flag-YopM 35-
Flag-STING
+ - + - + Flag-STING (kDa) + + + + + + + + + Myc-STING IB: Flag
(kDa) + + + + + Myc-YopJ
100-
35- IB: Myc
35- IB: Myc 70-
55- GST-YopJ
55-
35- IB: Flag
35- IB: Flag

25- 35-

GST
25-
Coommassie blue staining

D YopJ STING Golgin97 YopJ + STING STING + Golgin97 Merge

Myc-vector

-c-di-GMP
Myc-YopJ
Myc-vector

+c-di-GMP
Myc-YopJ

Fig. 2. YopJ associates with STING. (A) HEK293T cells were cotransfected with Myc-YopJ and Flag-vector or Flag-STING expression plasmids respectively. Anti-Flag or IgG
immunoprecipitates and whole cell lysates were analyzed by IB with anti-Myc or anti-Flag antibody. (B) Myc-STING expression plasmids were transfected into HEK293T cells together
with Flag-vector, Flag-YopE Flag-YopJ or YopM. Anti-Flag or IgG immunoprecipitates and whole cell lysates were analyzed by IB with anti-Myc or anti-Flag antibody. (C) HEK293T cells
were transfected with Flag-STING expressing plasmids. The GST-YopJ fusion protein absorbates from cell lysates were analyzed by IB with anti-Flag antibody (top panel). Loading of
the GST proteins was assessed by coomassie blue staining (bottom panel). (D) Representative confocal immunofluorescent images in HEK293T cells transfected with the indicated
plasmids in the presence or absence of c-di-GMP. Original magnification ×100. Scale bars, 10 μm. Cell-based studies were performed at least three independent times with comparable
results.

3.2. YopJ associates with STING this possibility, Flag-STING was cotransfected together with Myc-YopJ.
IB analysis of anti-Flag immunoprecipitates with an anti-Myc antibody
Due to the importance of STING in cytosolic DNA innate immune sig- showed a significant association between Myc-YopJ and Flag-STING
naling, a possible link between YopJ and STING were proposed. To test (Fig. 2A). Specially, YopE and YopM failed to associate with STING
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3153

A B Myc-STING C

Flag-vector

Myc-YopM
Y.pestis

Myc-YopE
Myc-YopJ
Flag-STING
(kDa) WT ΔYopJ
(kDa) 0 0.2 0.4 0.6 0.8 Myc-YopJ (µg) 130-
(kDa) IB: RIG-I
100-
35- IB: Flag 35- IB: Myc 70-
55- IB: MAVS
35- IB: Myc 55-
35- 35- IB: STING
IB: Flag
55- IB: α-Tubulin 25- 100-
IB: TBK1
70-

55- IB: α-Tubulin 55- IB: α-Tubulin

D E Flag-STING
WT ΔYopJ
Myc-vector Myc-YopJ
(kDa) 0 2 4 6 8 0 2 4 6 8 Y.pestis (hpi)
(kDa) 0 2 4 6 8 0 2 4 6 8 CHX (hrs)
IB: IRF3-p
55- 35- IB: Flag

55- IB: IRF3 35- IB: Myc

55- IB: α-Tubulin 55-

IB: α-Tubulin

F G
25 30 Flag-vector
Flag-vector
Flag-STING
Flag-STING
20
IFN-β activation (fold)
IRF3 activation (fold)

20
15

10
10
**
5
**

0 0
Myc-vector Myc-YopE Myc-YopJ Myc-YopM Myc-vector Myc-YopE Myc-YopJ Myc-YopM

(kDa) (kDa)
35- IB: Flag 35- IB: Flag

55- 55-
35- IB: Myc 35- IB: Myc
25- 25-

55- IB: α-Tubulin 55- IB: α-Tubulin

Fig. 3. YopJ blocks IRF3 signaling by inducing STING for degradation. (A) HEK293T cells were cotransfected with increasing amount of Myc-YopJ expression plasmids together with Flag-
STING, and whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control. (B) Myc-STING were cotransfected with Flag-vector, Flag-
YopE, Flag-YopJ or Flag-YopM expression plasmids respectively, and whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control.
(C) RAW264.7 macrophages were infected with WT or ΔYopJ Y. pestis (MOI = 20). Whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal
loading control. (D) RAW264.7 macrophages were infected with WT or ΔYopJ Y. pestis (MOI = 20) for 0, 2, 4, 6, 8 h respectively, and whole cell lysates were analyzed by IB with the
indicated antibodies. α-Tubulin was used as equal loading control. (E) HEK293T cells were cotransfected with Flag-STING and Myc-vector or Myc-YopJ respectively. After 24 h, cells
were treated with cycloheximide (50 μM) for 0, 2, 4, 6, 8 h respectively, and whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading
control. (F, G) HEK293T cells were cotransfected with pGL3-IRF3-luc/UAS-luc (F) or pGL3-IFN-β-luc (G) together with pRL-TK vector and the indicated plasmids. Luciferase activity
was measured 24 h later and normalized for transfection efficiency. Whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control.
Cell-based studies were performed at least three independent times with comparable results. Data represent means ± the standard errors of the means. Student's t-test was used for
statistical analysis: **P b 0.01.
3154 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159

(Fig. 2B). To demonstrate the interaction of YopJ and STING in vitro, ly- It has been known that STING relocalizes from ER to Golgi after acti-
sates from HEK293T cells expressing Flag-STING were incubated with vation [9,10,13]. We thus wanted to explore the effect of YopJ on STING
GST or GST-YopJ fusion protein. Analysis of the absorbates by IB analysis translocation in response to c-di-GMP treatment. Immunofluorescence
with anti-Flag antibody showed that STING bound GST-YopJ, but not results showed that YopJ colocalized with STING even in the absence
GST (Fig. 2C). of c-di-GMP transfection (Fig. 2D). As expected, more colocalization of

A B D
Flag-STING 30
210 **
Myc-C172A
Myc-vector
Myc-vector

uninfect **
Myc-YopJ

200

IRF3 activation (fold)

IFN-β expression (fold)

WT
20 190
(kDa) ΔYopJ

55- IB:IRF3-p 180 ΔYopJ/YopJ

50
ΔYopJ/C172A
10 40
55- IB:IRF3 ** 30
20
35- IB: Flag 10
0 0
Myc- Myc- Myc- Myc-
35- IB: Myc vector vector YopJ C172A

Flag-STING
55- IB: α-Tubulin (kDa)

35- IB: Flag

35- IB: Myc E

20
55- IB: α-Tubulin

15
C
PLA signals

WT ΔYopJ ΔYopJ/YopJ ΔYopJ /C172A

(kDa) 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 Y.pestis (hpi) 10

35- IB: STING
5
55- IB: α-Tubulin ** **
0
uninfect WT ΔYopJ ΔYopJ ΔYopJ
F /YopJ /C172A

uninfect WT ΔYopJ ΔYopJ/YopJ ΔYopJ /C172A Y.pestis

STING+TBK1

DAPI

Merge
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3155

STING and Golgin97 was identified in response to c-di-GMP stimulation. As YopJ targeted STING for degradation, protein levels of STING
However, in cells with overexpressed YopJ, the colocalization of STING were checked after infection with different strains of Y. pestis at the
with Golgin97 was significantly reduced (Fig. 2D). These results suggest indicated time points. Consistently, the protein level of STING was
that YopJ blocks STING translocation from ER to Golgi after activation reduced significantly by WT and ΔYopJ/YopJ Y. pestis, while ΔYopJ
through their physical association. and ΔYopJ/YopJ C172A Y. pestis had a compromised activity on
STING down-regulation (Fig. 4C). To demonstrate that the STING
3.3. YopJ blocks IRF3 signaling by inducing STING for degradation signaling pathway was a critical way for YopJ induced interferon
response, we infected RAW264.7 macrophages with different strains
To further explore the mechanism for the inhibitory effect of YopJ of Y. pestis. The results showed that ΔYopJ and ΔYopJ/YopJ C172A
on IRF3 antiviral signaling, we examined the effect of YopJ on STING Y. pestis could activate the IFN-β expression more strikingly than
abundance. Immunofluorescence results showed that the staining WT and ΔYopJ/YopJ Y. pestis did (Fig. 4D).
intensity of STING was much weaker in cells with overexpressed Previous studies indicated that the activated STING forms dimers to
YopJ (Fig. 2D). In addition, a striking reduction in STING expression assemble with TBK1, leading to IRF3 activation and the final induction of
with overexpressed YopJ was found in a dose dependent manner type I interferon [28]. To further examine the mechanism of the inhibi-
(Fig. 3A). A study in 2007 reported that YopJ influenced the expres- tory effect of YopJ in innate immune signaling under physiological con-
sion of CMV-based plasmids [26]. We then constructed STING into ditions, we explored the effect of YopJ on STING-TBK1 interaction.
an EF1 promoter plasmid pCDH-EF1-MCS-T2A-Puro and found YopJ Immunoprecipitation experiment indicated that STING-TBK1 interac-
could still reduce STING expression (Supplementary Fig. S1A and tion was reduced strikingly by YopJ (Supplementary Fig. S2). Then
B). Specially, YopE or YopM expression did not change STING level STING-TBK1 complexes were visualized by an in situ proximity ligation
(Fig. 3B). Quantitative RT-PCR revealed no change in endogenous assay (PLA). Small spots of STING-TBK1 complex in RAW264.7 macro-
STING mRNA level with increased YopJ expression, suggesting that phages infected with different strains of Y. pestis were observed. Results
YopJ down-regulates STING by posttranscriptional modification showed that ΔYopJ and ΔYopJ/YopJ C172A Y. pestis-infected RAW264.7
(Supplementary Fig. S1C). macrophages displayed significantly increasing in abundance of STING-
We then assessed whether YopJ was able to mediate the degradation TBK1 complexes as compared to those from WT and ΔYopJ/YopJ
of endogenous STING under physiological conditions. RAW264.7 mac- Y. pestis-infected RAW264.7 macrophages (Fig. 4E and F). These results
rophages infected with WT Y. pestis or ΔYopJ Y. pestis showed that indicate the importance of YopJ deubiquitinating protease activity in
STING level was higher in cells infected with ΔYopJ Y. pestis as compared down-regulating STING activity.
with that of WT Y. pestis, whereas the protein levels of RIG-I, MAVS and
TBK1 were not affected by YopJ deletion (Fig. 3C). To further delineate 3.5. YopJ deubiquitinates STING
the mechanism for the YopJ-mediated STING degradation, IRF3 phos-
phorylation level was analyzed following Y. pestis infection. As illustrat- It has been reported that the removal of K63-linked ubiquitin chains
ed in Fig. 3D, ΔYopJ Y. pestis infection strongly stimulated the signals for from TRAF6 by YopJ inactivates the NF-κB signaling pathway [20]. Since
IRF3 activation than WT Y. pestis infection. STING ubiquitination is indispensable for its function in IFN-β signaling,
We then investigated the half-life of STING in cells with overexpressed we then checked the effect of YopJ on STING ubiquitination modifica-
YopJ. As was shown in Fig. 3E, the approximated half-life of STING was tion. As was shown in Fig. 5A, ΔYopJ Y. pestis induced much more obvi-
considerably reduced by YopJ introduction. In reporter assays, YopJ, but ous ubiquitination of STING than WT Y. pestis did, indicating that YopJ
not YopE or YopM, dramatically suppressed STING-induced activation of might act as a deubiquitinating enzyme. Furthermore, the
the IRF3 and IFN-β reporters activation (Fig. 3F and G). Taken together, ubiquitination modification of STING was decreased by ectopic YopJ ex-
these results indicate that YopJ targets STING for degradation to block pression in response to c-di-GMP or poly (dA: dT) stimulation (Fig. 5B
the phosphorylation and activation of IRF3 signaling. and C). Actually, basal STING ubiquitination level was sharply decreased
by YopJ introduction (Fig. 5B and C). Furthermore, STING ubiquitination
3.4. YopJ 172th cysteine is the active site for blocking STING-mediated induced by c-di-GMP or poly (dA: dT) was nearly unchanged by YopJ
signaling C172A (Fig. 5B and C). Since YopJ is both a deubiquitinating enzyme
and an acetyltransferase, we next explored whether the acetyl-
Previous reports showed that YopJ functions by deubiquitinating transferase activity of YopJ had any effect on STING acetylation.
various signal transduction proteins [20,26,27]. A catalytic cysteine mu- Flag-STING was cotransfected with Myc-YopJ or Myc-YopJ C172A
tant of YopJ (YopJ C172A) is unable to inhibit signaling pathways, which plasmids and immunoprecipitation was performed to check if
prompted us to evaluate the contribution of YopJ protease activity in STING was been acetylated. The validity of Acetylated-Lysine
blocking STING activity [20]. Transfection of HEK293T cells with plas- antibody was shown in Supplementary Fig. S3A. As was shown in
mids encoding Flag-STING together with WT YopJ or YopJ C172A re- Supplementary Fig. S3B, neither YopJ nor YopJ C172A could acety-
vealed that the levels both of STING protein abundance and IRF3 late STING. Therefore, YopJ triggers STING deubiquitination by its
phosphorylation were diminished considerably by WT YopJ, but not Cys-172 protease activity.
YopJ C172A (Fig. 4A). Consistently, ΔYopJ and ΔYopJ/YopJ C172A By using K48 and K63 ubiquitin antibodies in ubiquitination assay,
Y. pestis also had a compromised activity on inhibiting STING-induced only anti-K63 ubiquitin antibody could react with immunoprecipitated
IRF3 promoter activation compared to WT and ΔYopJ/YopJ Y. pestis STING in macrophages after Y. pestis infection (Fig. 5D and Supplemen-
(Fig. 4B). tary Fig. S3C). The endogenous ubiquitination assay showed that the

Fig. 4. YopJ 172th cysteine is the active site for blocking STING mediated signaling. (A) Flag-STING was cotransfected with Myc-vector, Myc-YopJ or Myc-YopJ C172A expression plasmids
into HEK293T cells respectively, and whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control. (B) pGL3-IRF3-luc/UAS-luc, pRL-TK
vector were cotransfected with Myc-vector, Myc-YopJ or Myc-YopJ C172A expression plasmids respectively. Luciferase activity was measured 24 h later and normalized for transfection
efficiency. Whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading control. (C) RAW264.7 macrophages were infected with WT, ΔYopJ,
ΔYopJ/YopJ, ΔYopJ/YopJ C172A Y. pestis (MOI = 20) for 0, 2, 4, 6, 8 h respectively. Whole cell lysates were analyzed by IB with the indicated antibodies. α-Tubulin was used as equal loading
control. (D) RAW264.7 macrophages were infected with WT, ΔYopJ, ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis (MOI = 20) respectively. RNA was extracted and mRNA level of IFN-β was
analyzed by quantitative RT-PCR. The data were normalized to the expression of the GAPDH reference gene. (E) Quantification of PLA signals per cell in (F), presented relative to which of
control cells treated with solvent. (F) In situ PLA of STING-TBK1 complexes in RAW264.7 macrophages infected with different strains of Y. pestis (MOI = 20). STING-TBK1 complexes were
red spots. Original magnification ×100. Scale bars, 10 μm. Cell-based studies were performed at least three independent times with comparable results. Data represent means ± the
standard errors of the means. Student's t-test was used for statistical analysis. ⁎⁎P b 0.01.
3156 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159

A B Flag-STING
C Flag-STING
Flag-STING -c-di-GMP +c-di-GMP -poly (dA:dT) +poly (dA:dT)
uninfect

- + + + - + + + HA-Ub
ΔYopJ

- + + + - + + + HA-Ub
WT

(kDa) Y.pestis - - + - - - + - Myc-YopJ

- - + - - - + - Myc-YopJ
130- (kDa) - - - + - - - + Myc-YopJ-C172A - - - + - - - + Myc-YopJ-C172A
(kDa)
100- 100- 100-
70- 70-
IB: Ub 70-
55- 55- 55- IB: HA
IB: HA
35-
35- 35-
25-
35- IB: Flag
35- IB: Flag
IP: Flag 35- IB: Flag
Input: IP: Flag
IP: Flag Input:
35- IB: Flag Input:
35- IB: Myc
35- IB: Myc

E Flag-STING F Flag-STING

-c-di-GMP +c-di-GMP -poly (dA:dT) +poly (dA:dT)

- + - + - + - + - + - + Myc-YopJ - + - + - + - + - + - + Myc-YopJ
+ + - - - - + + - - - - HA-Ub + + - - - - + + - - - - HA-Ub
- - + + - - - - + + - - HA-Ub K48 - - + + - - - - + + - - HA-Ub K48
- - - - + + - - - - + + HA-Ub K63 - - - - + + - - - - + + HA-Ub K63
(kDa) (kDa)

100- 100-
70- 70-
55- 55- IB: HA
IB: HA

35- 35-
25- 25-

35- IB: Flag

35- IB: Flag
IP: Flag
IP: Flag Input:
Input:
35- IB: Myc
35- IB: Myc
ΔYopJ /C172A

D G
ΔYopJ/YopJ
uninfect

ΔYopJ

Flag-STING
WT

(kDa) Y.pestis
DMSO MG132 NH4Cl
130- (kDa) - + - + - + Myc-YopJ
100-
35- IB: Flag
70-
IB: K63-Ub
55- 35- IB: Myc

35- IB: α-Tubulin

35-

35- IB: STING

IP: STING
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3157

K63-linked ubiquitination modification of STING was significantly demonstrated that K63-linked ubiquitination signals of STING were dra-
decreased in WT and ΔYopJ/YopJ Y. pestis-infected macrophages com- matically reduced in the presence of YopJ, which might contribute to
pared to ΔYopJ/YopJ C172A Y. pestis-infected macrophages, while the destabilizing STING. We propose that the deubiquitination and degrada-
K48-linked ubiquitination was barely detected (Fig. 5D and Supplemen- tion of STING during Y. pestis infection is a new bacterial strategy to
tary Fig. S3C). Using an ubiquitin mutant with only one lysine at position modulate the host antibacterial responses.
48 or 63 available for conjugation (HA-K48 ubiquitin or HA-K63 ubiqui- Recent studies identify that pathogenic bacteria can evade the
tin), we found that c-di-GMP or poly (dA: dT) transfection led to in- host immune system during the course of infection through
creased K63-linked but not K48-linked ubiquitination of STING, which targeting at various proteins in multiple signaling pathways. Differ-
was significantly abrogated by YopJ (Fig. 5E and F), indicating YopJ me- ent groups have already demonstrated that YopJ blocks the MAPK
diates K63-linked deubiquitination of STING. Since YopJ-mediated deg- and NF-κB signaling pathways by acting as both deubiquitinase and
radation of STING was partially inhibited by the proteasome inhibitor acetyltransferase. In 2000, it was reported that YopJ is an ubiquitin-
MG132, but not lysosome inhibitor NH4Cl (Fig. 5G), we conclude that like protein protease by which small ubiquitin related modifiers
the YopJ-induced K63-linked deubiquitination may direct STING to (SUMO) of unidentified conjugated proteins were cleaved [27]. Sub-
proteasome-mediated degradation. sequently, YopJ was proved to act as a deubiquitinase targeting at
TRAF proteins to inhibit NF-κB signaling pathways [20,26]. In 2006
3.6. YopJ contributes to bacterial virulence by regulating STING-mediated and 2012, YopJ was regarded as a serine/threonine acetyltransferase
antibacterial response targeting both at mitogen-activated protein kinase kinases (MAP2Ks)
and TGFβ-activated kinase 1 (TAK1) [18,19]. Our data show that the
To further delineate the role of YopJ mediated STING down- deubiquitinase activity of YopJ also targets at a scaffolding protein,
regulation in Y. pestis virulence, BMDMs were infected with different STING. Therefore, YopJ modulates various inflammatory signaling path-
strains of Y. pestis at the indicated time points. The mean numbers of ways at multiple levels with different mechanisms.
bacteria recovered from WT and ΔYopJ/YopJ Y. pestis-infected BMDMs There are various mechanisms for mammalian cells to limit bacterial
were considerably higher than those from ΔYopJ and ΔYopJ/YopJ growth inside endosome and lysosome compartments [29]. Under
C172A Y. pestis-infected BMDMs (Fig. 6A). After down-regulation of physiological surroundings, most bacterial colonization in host cells
STING, the mean number of bacteria recovered from ΔYopJ Y. pestis- can be restrained by innate immune system [2,30]. Previous studies in-
infected siSTING cells was much higher than ΔYopJ Y. pestis-infected dicated that STING plays a pivotal role both in the process of ubiquitin
siCtrl cells, while the mean number of bacteria recovered from WT colocalization of specific bacterial subpopulation and the initiating of
Y. pestis-infected siSTING cells were barely increased (Fig. 6B), indicat- autophagy targeting [31]. The limited bacterial killing by macrophages
ing a major function of YopJ in STING degradation is required for viru- relies on the delivery of this population to lysosome. In the process of
lence. To establish the role of YopJ in Y. pestis infection, mice were keeping the host cytosol from bacterial colonizing, TBK1 is also neces-
challenged by intravenous injection of WT, ΔYopJ, ΔYopJ/YopJ or sary for activating autophagy by recognizing the ubiquitin-coated bac-
ΔYopJ/YopJ C172A Y. pestis. Firstly, HE staining was performed to exam- teria. Because YopJ associates with STING, inhibits STING translocation
ine disease pathology of formalin-fixed liver and spleen. Total patholog- from ER to Golgi and suppresses STING-TBK1 interaction, it is reason-
ical severity scoring indicated a significantly increased inflammation, able to propose that this bacterial protein interferes with cell signaling
tissue injury and necrosis in ΔYopJ or ΔYopJ/YopJ C172A bacteria infect- initiated from TLRs, STING, TBK1, or autophagy pathways during
ed tissues, while WT and ΔYopJ/YopJ strains generally caused lower Y. pestis infection. We also demonstrated that YopJ, but not YopJ
levels of inflammation and necrosis (Fig. 6C). These observations sug- C172A, could inhibit the activation of IFN-β production both in macro-
gest that YopJ is essential for Y. pestis virulence. IHC staining was then phages and under physiological conditions in mice. The K48-linked
performed to examine the STING staining in situ. Consistently, signifi- ubiquitination of STING couldn't be detected after infection with differ-
cantly increased staining of STING in the inflammatory lesions of the ent strains of Y. pestis. K48-linked ubiquitination of STING has been
liver and spleen from ΔYopJ and ΔYopJ/YopJ C172A–infected mice identified in response to viral infection in several papers [32,33]. Failure
was identified as compared to which of WT and ΔYopJ/YopJ-infected to detect K48-linked ubiquitination in our experiments may due to its
mice (Fig. 6D). These results demonstrate that YopJ contributes to bac- low level in response to bacterial infection. As YopJ induced STING deg-
terial virulence by negatively regulating STING-mediated antibacterial radation can be partially rescued by MG132 treatment, we propose that
response. YopJ-induced K63-linked deubiquitination may direct STING to
proteasome-mediated degradation, which needs further investigation.
4. Discussion It should be noted that the protein stability was increased via K63-
linked ubiquitination [34]. In conclusion, we show that YopJ could in-
YopJ has been reported to be a deubiquitinating enzyme that nega- hibit the activation of IFN-β production in macrophages by binding to
tively regulates NF-κB signaling via removing K63-linked ubiquitination STING and removing its K63-linked ubiquitination.
conjugates from critical proteins, such as TRAF2, TRAF6, and IKK [20]. In YopP in Y. enterocolitica, a 94% identical of YopJ in Y. pseudotuberculosis
this study, we investigated the mechanisms of the IRF3 antagonism and Y. pestis, shows greater ability for secretion, translocation, and conse-
imposed by YopJ in a deubiquitinase catalytic dependent manner. We quent cytotoxicity than YopJ [35,36]. But both expression of YopP and
showed that YopJ, but not YopJ C172A, blocked the assembly of complete absence of YopJ in Y. pseudotuberculosis resulted in attenu-
STING-TBK1 complexes and IRF3 activation. Importantly, we have ation of virulence, which shows that reasonable levels of YopJ-

Fig. 5. YopJ deubiquitinates STING. (A) THP-1 cells were transfected with Flag-STING expression plasmids. After infecting with WT or ΔYopJ Y. pestis (MOI = 20). Cells were treated with
MG132 for 8 h before harvesting cells. Anti-Flag immunoprecipitates and whole cell lysates were analyzed by IB with the indicated antibodies. (B\ \C) HEK293T cells were cotransfected
with Flag-STING, HA-ubiquitin together with Myc-vector, Myc-YopJ or Myc-YopJ C172A expression plasmids respectively in the presence or absence of c-di-GMP (B) or poly (dA: dT)
(C) stimulation. Cells were treated with MG132 for 8 h before harvesting cells. Anti-Flag immunoprecipitates and whole cell lysates were analyzed by IB with the indicated antibodies.
(D) RAW264.7 macrophages were infected with WT, ΔYopJ, ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis (MOI = 20) respectively. Cells were treated with MG132 for 8 h before
harvesting cells. Anti-STING immunoprecipitates were analyzed by IB with the indicated antibodies. (E, F) HEK293T cells were cotransfected with Flag-STING and Myc-vector or Myc-
YopJ together with HA-ubiquitin, HA-K48 or HA-K63 ubiquitin expression plasmids respectively in the presence or absence of c-di-GMP (E) or poly (dA: dT) (F) stimulation. Cells
were treated with MG132 for 8 h before harvesting cells. Anti-Flag immunoprecipitates whole cell lysates were analyzed by IB with indicated antibodies. (G) HEK293T cells were
cotransfected with Flag-STING and Myc-vector or Myc-YopJ and treated with DMSO, MG132 or NH4Cl respectively. Whole cell lysates were analyzed by IB with the indicated
antibodies. α-Tubulin was used as equal loading control. Cell-based studies were performed at least three independent times with comparable results.
3158 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159

A 25 B 25
WT WT+siCtrl
ΔYopJ ΔYopJ+siCtrl

Y. pestis proliferation (fold)

Y. pestis proliferation (fold)

20 20
ΔYopJ/YopJ WT+siSTING
15 ΔYopJ /C172A 15 ΔYopJ+siSTING

10 10

5 ** **
5
**
0 0
1 6 12 24 1 6 12 24
Y. pestis (hpi) Y. pestis (hpi)

C uninfect WT ΔYopJ ΔYopJ/YopJ ΔYopJ /C172A Y.pestis

Liver

Spleen

D uninfect WT ΔYopJ ΔYopJ/YopJ ΔYopJ /C172A Y.pestis

Liver

Spleen

Fig. 6. YopJ contributes to bacterial virulence by regulating STING-mediated antibacterial response. (A) BMDMs were infected with WT, ΔYopJ, ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis
(MOI = 20) respectively for the indicated time points. Then BMDMs were collected using 5% deoxysodium cholate, smeared to solid Hiss Agar. The quantity of Y. pestis clones were
estimated after 36 h. (B) BMDMs were transfected with siCtrl or siSTING. After 48 h, they were infected with WT or ΔYopJ Y. pestis (MOI = 20) respectively for the indicated time points.
Then BMDMs were collected using 5% deoxysodium cholate, smeared to solid Hiss Agar. The quantity of Y. pestis clones were estimated after 36 h. (C) Liver or spleen from mice infected
with WT, ΔYopJ, ΔYopJ/YopJ or ΔYopJ/YopJ C172A Y. pestis (MOI = 20) respectively for 3 days were subjected to immunohistochemical staining of STING. All images are representative and
taken at magnification ×200. Data shown are representative from two independent experiments (n = 5 mice per strain). (D) Liver, or spleen from mice infected with WT, ΔYopJ, ΔYopJ/
YopJ or ΔYopJ/YopJ C172A Y. pestis respectively for 3 days were subjected to histopathological analysis by HE staining (MOI = 20). Each micrograph shows an example of the pathology
caused by each strain. All images are representative and were taken at magnification ×100. Data shown are representative from two independent experiments (n = 5 mice per strain).

mediated cytotoxicity are indispensable for optimized virulence for of a systemic infection in mice [38,39]. Here we found that the mean
specific pathogenic species [37]. Other studies also showed that numbers of bacteria recovered from ΔYopJ and ΔYopJ/YopJ C172A
this virulence characteristic of YopJ is essential for the establishment Y. pestis-infected cells were dramatically reduced compared to
 Y. Cao et al. / Biochimica et Biophysica Acta 1863 (2016) 3148–3159 3159

those from WT and ΔYopJ/YopJ Y. pestis-infected cells, indicating the [15] J.E. Trosky, A.D. Liverman, K. Orth, Yersinia outer proteins: Yops, Cell. Microbiol. 10
(2008) 557–565.
importance of YopJ for Y. pestis to establish a systemic infection. To- [16] G.I. Viboud, J.B. Bliska, Yersinia outer proteins: role in modulation of host cell signal-
gether, our work show that the interaction between YopJ and STING ing responses and pathogenesis, Annu. Rev. Microbiol. 59 (2005) 69–89.
not only helps us to understand the mechanism of pestis to evade the [17] K. Trulzsch, G. Geginat, T. Sporleder, K. Ruckdeschel, R. Hoffmann, J. Heesemann, H.
Russmann, Yersinia outer protein P inhibits CD8 T cell priming in the mouse infec-
host immunity, but also suggests that YopJ may form a possible tar- tion model, J. Immunol. 174 (2005) 4244–4251.
get for pestis treatment. [18] S. Mukherjee, G. Keitany, Y. Li, Y. Wang, H.L. Ball, E.J. Goldsmith, K. Orth, Yersinia
Supplementary data to this article can be found online at http://dx. YopJ acetylates and inhibits kinase activation by blocking phosphorylation, Science
312 (2006) 1211–1214.
doi.org/10.1016/j.bbamcr.2016.10.004. [19] N. Paquette, J. Conlon, C. Sweet, F. Rus, L. Wilson, A. Pereira, C.V. Rosadini, N.
Goutagny, A.N. Weber, W.S. Lane, S.A. Shaffer, S. Maniatis, K.A. Fitzgerald, L. Stuart,
Funding N. Silverman, Serine/threonine acetylation of TGFbeta-activated kinase (TAK1) by
Yersinia pestis YopJ inhibits innate immune signaling, Proc. Natl. Acad. Sci. U. S. A.
109 (2012) 12710–12715.
This work was supported in part by the Basic Research Program of [20] H. Zhou, D.M. Monack, N. Kayagaki, I. Wertz, J. Yin, B. Wolf, V.M. Dixit, Yersinia vir-
China (2012CB518900), the National Natural Science Foundation of ulence factor YopJ acts as a deubiquitinase to inhibit NF-kappa B activation, J. Exp.
Med. 202 (2005) 1327–1332.
China (31170029, 31207911 and 31270800), the China Postdoctoral [21] K. Guan, Z. Zheng, T. Song, X. He, C. Xu, Y. Zhang, S. Ma, Y. Wang, Q. Xu, Y. Cao, J. Li, X.
Science Foundation Grant (2013M542507). The funders had no role in Yang, X. Ge, C. Wei, H. Zhong, MAVS regulates apoptotic cell death by decreasing
study design, data collection and analysis, decision to publish, or prepa- K48-linked ubiquitination of voltage-dependent anion channel 1, Mol. Cell. Biol.
33 (2013) 3137–3149.
ration of the manuscript. [22] T. ASSong, C. Wei, Z. Zheng, Y. Xu, X. Cheng, Y. Yuan, K. Guan, Y. Zhang, Q. Ma, W.
Shi, H. Zhong, c-Abl tyrosine kinase interacts with MAVS and regulates innate im-
mune response, FEBS Lett. 584 (2010) 33–38.
[23] C. Wei, Y. Cao, X. Yang, Z. Zheng, K. Guan, Q. Wang, Y. Tai, Y. Zhang, S. Ma, Y. Cao, X.
Conflicts of interest Ge, C. Xu, J. Li, H. Yan, Y. Ling, T. Song, L. Zhu, B. Zhang, Q. Xu, C. Hu, X.W. Bian, X. He,
H. Zhong, Elevated expression of TANK-binding kinase 1 enhances tamoxifen resis-
The authors do not have any potential conflicts of interest to declare. tance in breast cancer, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) E601–E610.
[24] Z. Zheng, C. Wei, K. Guan, Y. Yuan, Y. Zhang, S. Ma, Y. Cao, F. Wang, H. Zhong, X. He,
The authors declare that they have no competing interests. Bacterial E3 ubiquitin ligase IpaH4.5 of Shigella flexneri targets TBK1 to dampen the
host antibacterial response, J. Immunol. 196 (2016) 1199–1208.
Transparency document [25] A.A. Patel, H. Lee-Lewis, J. Hughes-Hanks, C.A. Lewis, D.M. Anderson, Opposing roles
for interferon regulatory factor-3 (IRF-3) and type I interferon signaling during
plague, PLoS Pathog. 8 (2012), e1002817.
The Transparency document associated with this article can be [26] C.R. Sweet, J. Conlon, D.T. Golenbock, J. Goguen, N. Silverman, YopJ targets TRAF pro-
found, in online version. teins to inhibit TLR-mediated NF-kappaB, MAPK and IRF3 signal transduction, Cell.
Microbiol. 9 (2007) 2700–2715.
[27] K. Orth, Z. Xu, M.B. Mudgett, Z.Q. Bao, L.E. Palmer, J.B. Bliska, W.F. Mangel, B.
References Staskawicz, J.E. Dixon, Disruption of signaling by Yersinia effector YopJ, a
ubiquitin-like protein protease, Science 290 (2000) 1594–1597.
[1] S. Uematsu, K. Fujimoto, The innate immune system in the intestine, Microbiol. [28] W. Sun, Y. Li, L. Chen, H. Chen, F. You, X. Zhou, Y. Zhou, Z. Zhai, D. Chen, Z. Jiang, ERIS,
Immunol. 54 (2010) 645–657. an endoplasmic reticulum IFN stimulator, activates innate immune signaling
[2] S. Akira, S. Uematsu, O. Takeuchi, Pathogen recognition and innate immunity, Cell through dimerization, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 8653–8658.
124 (2006) 783–801. [29] T.S. Gomez, J.A. Gorman, A.A. de Narvajas, A.O. Koenig, D.D. Billadeau, Trafficking de-
[3] Y. Ran, H.B. Shu, Y.Y. Wang, MITA/STING: a central and multifaceted mediator in in- fects in WASH-knockout fibroblasts originate from collapsed endosomal and lyso-
nate immune response, Cytokine Growth Factor Rev. 25 (2014) 631–639. somal networks, Mol. Biol. Cell 23 (2012) 3215–3228.
[4] T. Abe, A. Harashima, T. Xia, H. Konno, K. Konno, A. Morales, J. Ahn, D. Gutman, G.N. [30] Y. Kumagai, S. Akira, Identification and functions of pattern-recognition receptors, J.
Barber, STING recognition of cytoplasmic DNA instigates cellular defense, Mol. Cell Allergy Clin. Immunol. 125 (2010) 985–992.
50 (2013) 5–15. [31] C. Pujol, K.A. Klein, G.A. Romanov, L.E. Palmer, C. Cirota, Z. Zhao, J.B. Bliska, Yersinia
[5] Q. Yin, Y. Tian, V. Kabaleeswaran, X. Jiang, D. Tu, M.J. Eck, Z.J. Chen, H. Wu, Cyclic di- pestis can reside in autophagosomes and avoid xenophagy in murine macrophages
GMP sensing via the innate immune signaling protein STING, Mol. Cell 46 (2012) by preventing vacuole acidification, Infect. Immun. 77 (2009) 2251–2261.
735–745. [32] B. Zhong, L. Zhang, C. Lei, Y. Li, A.P. Mao, Y. Yang, Y.Y. Wang, X.L. Zhang, H.B. Shu, The
[6] D.L. Burdette, K.M. Monroe, K. Sotelo-Troha, J.S. Iwig, B. Eckert, M. Hyodo, Y. ubiquitin ligase RNF5 regulates antiviral responses by mediating degradation of the
Hayakawa, R.E. Vance, STING is a direct innate immune sensor of cyclic di-GMP, Na- adaptor protein MITA, Immunity 30 (2009) 397–407.
ture 478 (2011) 515–518. [33] Y. Wang, Q. Lian, B. Yang, S. Yan, H. Zhou, L. He, G. Lin, Z. Lian, Z. Jiang, B. Sun,
[7] T. Tsuchida, J. Zou, T. Saitoh, H. Kumar, T. Abe, Y. Matsuura, T. Kawai, S. Akira, The TRIM30alpha is a negative-feedback regulator of the intracellular DNA and DNA
ubiquitin ligase TRIM56 regulates innate immune responses to intracellular virus-triggered response by targeting STING, PLoS Pathog. 11 (2015), e1005012.
double-stranded DNA, Immunity 33 (2010) 765–776. [34] K.S. Choi, H.J. Choi, J.K. Lee, S. Im, H. Zhang, Y. Jeong, J.A. Park, I.K. Lee, Y.M. Kim, Y.G.
[8] J. Zhang, M.M. Hu, Y.Y. Wang, H.B. Shu, TRIM32 protein modulates type I interferon Kwon, The endothelial E3 ligase HECW2 promotes endothelial cell junctions by in-
induction and cellular antiviral response by targeting MITA/STING protein for K63- creasing AMOTL1 protein stability via K63-linked ubiquitination, Cell. Signal. 28
linked ubiquitination, J. Biol. Chem. 287 (2012) 28646–28655. (2016) 1642–1651.
[9] N. Dobbs, N. Burnaevskiy, D. Chen, V.K. Gonugunta, N.M. Alto, N. Yan, STING activa- [35] B.W. Wren, The yersiniae–a model genus to study the rapid evolution of bacterial
tion by translocation from the ER is associated with infection and autoinflammatory pathogens, Nat. Rev. Microbiol. 1 (2003) 55–64.
disease, Cell Host Microbe 18 (2015) 157–168. [36] S.A. Handley, P.H. Dube, P.A. Revell, V.L. Miller, Characterization of oral Yersinia
[10] K. Mukai, H. Konno, T. Akiba, T. Uemura, S. Waguri, T. Kobayashi, G.N. Barber, H. Arai, enterocolitica infection in three different strains of inbred mice, Infect. Immun. 72
T. Taguchi, Activation of STING requires palmitoylation at the Golgi, Nat. Commun. 7 (2004) 1645–1656.
(2016) 11932. [37] I.E. Brodsky, R. Medzhitov, Reduced secretion of YopJ by Yersinia limits in vivo cell
[11] H. Ishikawa, G.N. Barber, STING is an endoplasmic reticulum adaptor that facilitates death but enhances bacterial virulence, PLoS Pathog. 4 (2008), e1000067.
innate immune signalling, Nature 455 (2008) 674–678. [38] D.M. Monack, J. Mecsas, D. Bouley, S. Falkow, Yersinia-induced apoptosis in vivo aids
[12] B. Zhong, Y. Yang, S. Li, Y.Y. Wang, Y. Li, F. Diao, C. Lei, X. He, L. Zhang, P. Tien, H.B. in the establishment of a systemic infection of mice, J. Exp. Med. 188 (1998)
Shu, The adaptor protein MITA links virus-sensing receptors to IRF3 transcription 2127–2137.
factor activation, Immunity 29 (2008) 538–550. [39] N. Lemaitre, F. Sebbane, D. Long, B.J. Hinnebusch, Yersinia pestis YopJ suppresses
[13] H. Ishikawa, Z. Ma, G.N. Barber, STING regulates intracellular DNA-mediated, type I tumor necrosis factor alpha induction and contributes to apoptosis of immune
interferon-dependent innate immunity, Nature 461 (2009) 788–792. cells in the lymph node but is not required for virulence in a rat model of bubonic
[14] Y. Tanaka, Z.J. Chen, STING specifies IRF3 phosphorylation by TBK1 in the cytosolic plague, Infect. Immun. 74 (2006) 5126–5131.
DNA signaling pathway, Sci. Signal. 5 (2012) (ra20).