STING promotes NLRP3 localization in ER and facilitates NLRP3 deubiquitination to activate the inflammasome upon HSV-1 infection

One of the fundamental reactions of the innate immune responses to pathogen infection is the release of pro-inflammatory cytokines, including IL-1β, processed by the NLRP3 inflammasome. The stimulator of interferon genes (STING) has the essential roles in innate immune response against pathogen infections. Here we reveal a distinct mechanism by which STING regulates the NLRP3 inflammasome activation, IL-1β secretion, and inflammatory responses in human cell lines, mice primary cells, and mice. Interestingly, upon HSV-1 infection and cytosolic DNA stimulation, STING binds to NLRP3 and promotes the inflammasome activation through two approaches. First, STING recruits NLRP3 and facilitates NLRP3 localization in the endoplasmic reticulum, thereby facilitating the inflammasome formation. Second, STING interacts with NLRP3 and attenuates K48- and K63-linked polyubiquitination of NLRP3, thereby promoting the inflammasome activation. Collectively, we demonstrate that the cGAS-STING-NLRP3 signaling is essential for host defense against HSV-1 infection.

Introduction The innate immune system detecting pathogens through recognition of molecular patterns is a primary host defense strategy to suppress the infections [1]. Recognition of pathogens stimuli, known as pathogen-associate molecular patterns (PAMPS), is relied on pattern recognition receptors (PRRs). Several families of PRRs have been described, including the Toll-like receptor (TLR) [2], RIG-I-like receptor (RLR) [3], NOD-like receptor (NLR) [4], and C-type lectin receptor (CLR) [5]. The NLRs involved in the assembly of large protein complexes referred to as inflammasomes are emerging as a major route by which the innate immune system responds to pathogen infections [6]. One of the fundamental reactions of the innate immunity is the procession and release of pro-inflammatory cytokines, including interleukine-1β (IL-1β), a pleiotropic cytokine playing crucial roles in inflammatory responses in addition to instructing immune responses [7]. The best-characterized inflammasomes is the NLRP3 inflammasome, which consists of three major components: a cytoplasmic sensor NLRP3 (NACHT, LRR and PYD domains-containing protein 3), an adaptor ASC (apoptosis-associated speck-like protein with CARD domain), and an interleukin-1β-converting enzyme pro-Caspase-1 (cysteinyl aspartate-specific proteinase-1) [6]. NLRP3 and ASC together promote the cleavage of pro-Casp-1 to generate active subunits p20 and p10, which regulate IL-1β maturation [8].
The stimulator of interferon genes (STING) has the essential roles in innate immune response against pathogen infections [9]. Upon binding of cytoplasmic DNA, cyclic GMP-AMP synthase (cGAS) catalyzes the formation of cyclic guanosine monophosphateadenosine monophosphate (cGAMP) by binding to STING. STING subsequently co-localizes with TBK1 and IRF3, leading to induction of type I IFNs, and recruits TRAF6 and TBK1 or TRAF3 and IKKα to activate the NF-κB pathway [10,11]. In human myeloid cells, STING is involved in cytosolic DNA induced-NLRP3 inflammasome activation [12], and in mice BMDMs, STING is required for pathogen-induced inflammasome activation and IL-1β secretion [13,14].

HSV-1 infection induces IL-1β expression and secretion
IFI16 recognize HSV-1 DNA in the nucleus and then exits the nucleus and assembles with ASC and pro-caspase-1 to form an inflammasome complex in HFF cells. And HSV-1 infection can induce NLRP3-ASC interaction in HFF cells [24]. Next, we explored whether HSV-1 infection and HSV120 transfection regulate the inflammasome activation in THP-1 macrophages and LPS-primed mice primary MEFs. In TPA-differentiated THP-1 macrophages, endogenous IL-1β secretion was induced by Nigericin (a positive control for the inflammasome activation) [25] and HSV-1 (Fig 3A and 3B). Consistently, IL-1β maturation and Casp-1 cleavage, as well as pro-IL-1β production were activated upon HSV-1 infection (Fig 3C and

PLOS PATHOGENS
STING activates the NLRP3 inflammasome 3D). The expression of IFI16 was increased after HSV-1 infection (S1 Fig). Notably, IL-1β mRNA was not induced by Nigericin but induced upon HSV-1 infection (Fig 3E and 3F) and HSV-1 ICP27 mRNA was expressed in the infected cells (Fig 3G and 3H). Similarly, in TPAdifferentiated THP-1 macrophages, IL-1β secretion was induced by Nigericin and facilitated by HSV120 (Fig 3I). IL-1β maturation and Casp-1 cleavage and pro-IL-1β production were stimulated by Nigericin and promoted by HSV120 (Fig 3J). IFN-β mRNA expression was not induced by Nigericin but activated by HSV120 (Fig 3K), demonstrating that HSV120 is effective in the cells. Moreover, in LPS-primed mice primary MEFs, endogenous IL-1β secretion was induced by ATP (a positive control), promoted upon HSV-1 infection, and enhanced by HSV120 stimulation (Fig 3L and 3M). HSV-1 ICP27 mRNA was detected in the cells (Fig 3N), suggesting that HSV-1 is replicated. IFN-β mRNA was not induced by ATP but activated by HSV120 in the cells (Fig 3O), demonstrating that HSV120 is effective. Therefore, IL-1β expression and secretion are induced upon HSV-1 infection and cytosolic DNA stimulation.
Moreover, the direct role of NLRP3 in the regulation of HSV-1-induced IL-1β secretion was determined in LPS-primed primary MEFs of C57BL/6 WT pregnant mice and NLRP3 -/-pregnant mice. NLRP3 protein was detected in WT mice LPS-primed primary MEFs, but not in NLRP3 -/mice LPS-primed primary MEFs (Fig 4R), indicating that NLRP3 is knocked out in the null mice. IL-1β secretion was induced by ATP and HSV-1 in WT LPS-primed mice primary MEFs but not in NLRP3 -/mice LPS-primed primary MEFs ( Fig 4S). HSV-1 ICP27 mRNA was expressed in infected cells, indicating that HSV-1 is replicated in the cells ( Fig 4T). Collectively, inhibition, knock-down, and knock-out of the NLRP3 inflammasome components lead to the repression of IL-1β secretion and Casp-1 maturation, therefore the NLRP3 inflammasome is required for HSV-1-induced activation of IL-1β.

STING recruits NLRP3 to the ER to promote the inflammasome formation
STING predominantly resides in the endoplasmic reticulum (ER) to regulate innate immune signaling processes [18]. Here we evaluated whether STING promotes NLRP3 ER localization and activates the inflammasome. In HeLa cells, NLRP3 alone diffusely distributed in the cytoplasm and STING or STING(delTM5) alone located in the ER (as indicated by the ER marker, ER blue), while NLRP3 and STING, but not STING(delTM5), together co-localized and distributed in the ER to form specks (Fig 5A), which is an indication of the NLRP3 inflammasome complex formation [22], suggesting that STING facilitates NLRP3 ER localization and promotes the inflammasome formation. In addition, NLRP3 diffusely distributed in the cytosol of untreated cells, but formed distinct specks upon HSV-1 infection or HSV120 transfection ( Fig  5B). Moreover, NLRP3 diffusely distributed in the cytosol of untreated cells, while upon HSV-1 infection or HSV120 transfection, NLRP3 forms distinct specks in the ER as indicated by Calnexin (ER protein) ( Fig 5C) and ER blue (ER marker) (Fig 5D) In THP-1 macrophages, we also found the improvement of NLRP3 localization in the ER after the HSV-1 infection (Fig 5E), but not in TGN (S2C Fig). Notably, in transfected HeLa cells, NLRP3, STING, and Calnexin were detected in whole cell lysate (WCL) and purified ER fraction, and interestingly, NLRP3 abundance was enhanced by STING in purified ER fraction (Fig 5F). Similarly, in mock-infected THP-1 macrophages and LPS-primed mice primary MEFs, NLRP3, STING, and Calnexin were detected in WCL and purified ER, and NLRP3 abundance was increased in the ER upon HSV-1 infection (Fig 5G and 5H). Therefore, STING, HSV-1, and HSV120 facilitate the NLRP3 inflammasome formation in the ER.
Moreover, the effect of endogenous STING on NLRP3 translocation to the ER was further determined by using shRNA targeting STING (sh-STING). Hela cells stably expressing sh-NC or sh-STING were generated, and then transfected with Flag-NLRP3 and infected with HSV-1 or transfected with HSV120. In the absence of sh-STING, NLRP3 diffusely distributed in the cytosol of untreated cells and formed distinct specks in the ER, as indicated by Calnexin and ER Blue (Fig 5I and 5J, top), upon HSV-1 infection or HSV120-transfection, however, in the presence of sh-STING, NLRP3 failed to form specks upon HSV-1 infection or HSV120 transfection ( Fig 5I and 5J, bottom), indicating that STING knock-down leads to the repression of HSV-1-induced formation of the NLRP3 inflammasome. In addition, NLRP3, Calnexin, and STING were detected in WCL and purified ER fraction of Hela cells (Fig 5K), THP-1 cells ( Fig  5L) and LPS-primed mice primary MEFs (Fig 5M), and notably, NLRP3 level was higher in purified ER fraction upon HSV-1 infection in sh-NC stable cells (Fig 5K-5M, lane 6 vs. 5) as compared with sh-STING stable cells (Fig 5K-5M, lane 8 vs. 7), suggesting that STING knockdown results in the attenuation of NLRP3 localization in the ER upon HSV-1 infection. We also confirmed that STING abundance was down-regulated by sh-STING (Fig 5K-5M). Collectively, STING improves NLRP3 ER localization and promotes the NLRP3 inflammasome formation upon HSV-1 infection and cytosolic DNA stimulation.

STING deubiquitinates NLRP3 to activate the NLRP3 inflammasome
The deubiquitination of NLRP3 is required for the NLRP3 inflammasome activation [30]. We next investigated whether STING plays a role in the deubiquitination of NLRP3, thereby facilitating the inflammasome activation. Interestingly, NLRP3 polyubiquitination catalyzed by HA-UB, HA-UB(K48R), or HA-Ub(K63R) was repressed by STING (Fig 6A and 6B). We also revealed that NLRP3 polyubiquitination catalyzed by HA-UB, HA-UB(K48O) (ubiquitin mutant that only retains a single lysine residue), or HA-UB(K63O) (ubiquitin mutant that only retains a single lysine residue) was suppressed by STING ( Fig 6C). These results reveal that STING decreases K48-and K63-linked polyubiquitination of NLRP3.

PLOS PATHOGENS
STING activates the NLRP3 inflammasome ICP27 mRNA were detected in infected cells, respectively (Fig 7L-7N). Taken together, STING plays specific roles in the NLRP3 inflammasome activation upon DNA virus infection or cytosolic DNA stimulation, but has no effect on the inflammasome activation induced by RNA virus infection or Nigericin stimulation.

NLRP3 is critical for host defense against HSV-1 infection in mice
To gain insights into the biological function of NLRP3 in vivo, we analyzed C57BL/6 NLRP3 +/+ WT mice and C57BL/6 NLRP3 -/deficiency mice. Notably, HSV-1-infected NLRP3 -/mice began to die at 5 days post-infection and all infected mice died at 7 days post-infection, while infected WT mice began to die at 7 days post-infection and 30% WT mice was survival after 11 days post-infection (Fig 8A). The body weights of infected NLRP3 -/mice decreased continuously until died, while the body weights of infected WT mice gradually decreased until 7 days post-infection and then gradually increased (Fig 8B). Thus, NLRP3 deficiency mice are more susceptibility to HSV-1 infection and exhibit early onset of death upon the infection.
Notably, in the mice blood, IL-1β secretion was induced upon HSV-1 infection in WT mice, whereas it was not induced in NLRP3 -/mice (Fig 8C), IL-1β mRNA level was higher in WT mice as compared with NLRP3 -/mice (Fig 8D), IL-6 mRNA and TNF-α mRNA were no significant difference between WT and NLRP3 -/mice (Fig 8E and 8F), and HSV-1 UL30 mRNA was expressed in infected WT and NLRP3 -/mice ( Fig 8G). These results indicate that NLRP3 deficiency leads to the repression of IL-1β expression and secretion in mice. Interestingly, in HSV-1 infected mice lung and brain, IL-1β mRNA and IL-6 mRNA were significantly higher in WT mice as compared with NLRP3 -/mice (Fig 8H, 8I, 8K and 8L), however, the viral titers were much lower in WT mice as compared with NLRP3 -/mice (Fig 8J and 8M), suggesting that NLRP3 deficiency results in the attenuation of IL-1β expression and the promotion of HSV-1 replication in mice lung and brain. Moreover, Hematoxylin and Eosin (H&E) staining showed that more infiltrated neutrophils and mononuclear cells were detected in the lung (Fig 8N, left) and brain (Fig 8O, left) of mock infected WT mice as compared with HSV-1 infected WT mice, and this increase was eliminated in HSV-1 infected WT mice. Immunohistochemistry (IHC) analysis revealed that IL-1β protein level was higher in the lung (Fig 8N, right) and brain (Fig 8O, right) of mock infected WT mice as compared with HSV-1 infected WT mice, and this increase was eliminated in HSV-1 infected HSV-1 infected WT mice, revealing that NLRP3 deficiency mice are more susceptibility to HSV-1 infection and elicit weak inflammatory responses. Collectively, we propose that NLRP3 is essential for host defense against HSV-1 infection by facilitating IL-1β activation (Fig 9).

Fig 9. A proposed model for the regulation of NLRP3 inflammasome activation mediated by the cGAS-STING-NLRP3 pathway.
Left: In the mock infection, all these proteins are in the cytoplasm. NLRP3 contains an N-terminal pyrin domain (PYD), a NACHT-associated domain (NAD) and 7 C-terminal leucine rich repeats (LRR). ASC (the apoptosis-associated speck-like protein with CARD domain) has two domains (PYD and CARD). The effecter protein pro-Casp-1 has the three domains (CARD, p20 and p10). It has been demonstrated that the NLR protein NLRP3, together with the adaptor protein, apoptosis-associated speck-like protein with CARD domain (ASC), promotes the cleavage of the pro-Casp-1 to generate active subunits p20 and p10, which regulate the maturation of IL-1β.STING comprises 5 putative transmembrane (TM) regions, predominantly resides in the endoplasmic reticulum (ER) and regulates innate immune signaling processes. Right: After the infection by HSV-1, everything has changed. Herpes simplex virus type 1 (HSV-1) belongs to the family of Herpesviridae, which causes various mild clinical symptoms. HSV-1 genomic DNA then induce the cGAS-cGAMP-STING pathway for the IFN-I production. Upon binding of cytoplasmic viral DNA, GAS (cyclic GMP-AMP synthase) catalyzed the formation of the second messenger cGAMP, which is an endogenous second messenger by binding to STING and inducing STING's dimerization. After that, HSV-1 infection also induced NLRP3 translocation to Endoplasmic Reticulum (ER) by STING through the interaction between STING and NLRP3. Then STING promotes the deubiquitination of NLRP3 in K48-linked and K63-linked polyubiquitin chains and initiates the activation of NLRP3 inflammasome, promotes the cleavage of the pro-Casp-1 to generate active subunits p20 and p10, which regulate the maturation of IL-1β, and IL-1β then secreted outside the cells. https://doi.org/10.1371/journal.ppat.1008335.g009

PLOS PATHOGENS
STING activates the NLRP3 inflammasome human monocytes and mice BMDMs, but also in human embryonic kidney cells (HEK293T), Hela cells, human leukemic monocytes/macrophages (THP-1), and mice primary mouse embryo fibroblasts (MEFs). More interestingly, our results reveal a distinct mechanism underlying STING-mediated NLRP3 inflammasome activation, and demonstrate for the first time that STING binds to NLRP3 and promotes the inflammasome activation through two approaches. First, STING binds to and improves NLRP3 localization in ER to promote the formation of the NLRP3 inflammasome. Second, STING interacts with NLRP3 and attenuates K48-and K63-linked polyubiquitination of NLRP3 to induce the activation of the NLRP3 inflammasome. Notably, upon HSV-1 infection and HSV120 stimulation, STING binds to NLRP3, promotes the NLRP3-ASC interaction (an indicator of inflammasome complex assembly) [20], facilitates NLRP3-mediated ASC oligomerization (a critical step for inflammasome activation) [21], enhances NLRP3 to form specks (an indicator of inflammasome activation) [22], and enhances IL-1β secretion (a fundamental reaction of the inflammatory responses) [7]. Collectively, the cGAS-STING-NLRP3 pathway plays key roles in the NLRP3 inflammasome activation and IL-1β secretion upon DNA virus infection and cytosolic DNA stimulation.
Moreover, NLRP3 is related to many human diseases. Fibrillar amyloid-β peptide, the major component of Alzheimer's disease brain plaques, facilitates the NLRP3 inflammasome activation [49]. Monosodium urate (MSU) crystals induce the autoinflammatory disease gout and activate the NLRP3 inflammasome [50]. NLRP3, IL-1β, reactive oxygen species (ROS), and TXNIP are implicated in the type 2 diabetes mellitus (T2DM) pathogenesis [51]. Our study gains insights into the biological function of the cGAS-STING-NLRP3 pathway in host defense against HSV-1 infection in mice. NLRP3 deficiency mice are more susceptibility to HSV-1 infection, exhibit early onset of death upon infection, represses IL-1β secretion, and elicits robust inflammatory responses in the tissues. Collectively, these results demonstrate that NLRP3 is essential for host defense against HSV-1 infection by inducting IL-1β expression and secretion.
In conclusion, we reveal a distinct mechanism underlying the regulation of the NLRP3 inflammasome activation upon HSV-1 infection. In this model, STING (the central molecule of the antiviral and inflammatory immune pathways) interacts with NLRP3 (the key component of the inflammasomes), decreases NLRP3 polyubiquitination, improves the localization of NLRP3 in ER, and facilitates the NLRP3 inflammasome activation, thereby inducing IL-1β secretion upon DNA virus infection and cytosolic DNA stimulation.

Materials and methods
Animal study C57BL/6 WT mice were purchased from Hubei Research Center of Laboratory Animals (Wuhan, Hubei, China). C57BL/6 NLRP3 -/mice were kindly provided by Dr. Di Wang of Zhejiang University School of Medicine, China.

Ethics statement
All animal studies were performed in accordance with the principles described by the Animal Welfare Act and the National Institutes of Health Guidelines for the care and use of laboratory animals in biomedical research. All procedures involving mice and experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the College of Life Sciences, Wuhan University.

Cell lines and cultures
African green monkey kidney epithelial (Vero) cells, human cervical carcinoma (Hela) cells, and human embryonic kidney 293T (HEK 293T) cells were purchased from American Type Culture Collection (ATCC) (Manassas, VA, USA). Human acute monocytic leukemia (THP-1) cells were gift from Dr. Jun Cui of State Key Laboratory of Biocontrol, School of Life Sciences, Sun Yat-Sen University, Guangzhou 510275, China. THP-1 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin sulfate. Vero, Hela and HEK293T cells were cultured in Dulbecco modified Eagle medium (DMEM) purchased from Gibco (Grand Island, NY, USA) supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin sulfate. Vero, Hela, HEK293T and THP-1 cells were maintained in an incubator at 37˚C in a humidified atmosphere of 5% CO 2 .

Viruses
Herpes simplex virus 1 (HSV-1) strain and Sendai virus (SeV) strain were gifts from Dr. Bo Zhong of Wuhan University. Zika virus (ZIKV) isolate z16006 (GenBank accession number KU955589.1) was used in this study. Vero cells were maintained at 37˚C in DMEM (GIBCO) supplemented with 10% heat-inactivated FBS with penicillin/streptomycin (GIBCO) (Grand Island, NY, USA). HSV-1 stocks were propagated in Vero cells for 36 h at 0.03 MOI. The infected cells were collected after three times of freezing and thawing in the infected cells and titrated by plaque assay in 12-well plates in Vero cells. The mock-infected was prepared at the same Vero cells without HSV-1 infection, but others are the same procedure.

Plaque assay
Vero cells were cultured in a 12-well plate at a density of 2 × 10 5 cells per well, and infected with 100 μl serially diluted HSV-1 supernatant for 2 h. Then, the cells were washed by PBS and then immediately replenished with plaque medium supplemented with 1% carboxylmethylcelluose. The infected Vero cells were incubated for 2-3 days. After the incubation, plaque medium was removed and cells were fixed and stained with 4% formaldehyde and 0.5% crystal violet.

Enzyme-linked immunosorbent assay (ELISA)
Concentrations of human IL-1β in culture supernatants were measured by ELISA kit (BD Biosciences, San Jose, CA, USA). The mouse IL-1β ELISA Kit was purchased from R&D.

Activated caspase-1 and mature IL-1β measurement
Supernatant of the cultured cells was collected for 1 ml in the cryogenic vials (Corning). The supernatant was centrifuged at 12,000 rpm for 10 min each time by Amicon Ultra (UFC500308) from Millipore for protein concentrate. The concentrated supernatant was mixed with SDS loading buffer for western blotting analysis with antibodies for detection of activated caspase-1 (D5782 1:500; Cell Signaling) or mature IL-1β (Asp116 1:500; Cell Signaling). Adherent cells in each well were lysed with the lyses buffer described below, followed by immunoblot analysis to determine the cellular content of various protein.

Confocal microscopy
HEK293T cells and Hela cells were transfected with plasmids for 24-36 h. Cells were fixed in 4% paraformaldehyde at room temperature for 15 min. After being washed three times with PBS, permeabilized with PBS containing 0.1% Triton X-100 for 5 min, washed three times with PBS, and finally blocked with PBS containing 5% BSA for 1 h. The cells were then incubated with the monoclonal mouse anti-Flag antibody (F3165, Sigma) and Monoclonal rabbit anti-HA (H6908, Sigma) overnight at 4˚C, followed by incubation with FITC-conjugate donkey anti-mouse IgG (Abbkine) and Dylight 649-conjugate donkey anti-rabbit IgG (Abbkine) for 1 h. After washing three times, cells were incubated with DAPI solution for 5 min, and then washed three more times with PBS. Finally, the cells were analyzed using a confocal laser scanning microscope (Fluo View FV1000; Olympus, Tokyo, Japan).

ASC oligomerization detection
HEK293T cells were transfected with plasmids for 24-36 h. Cell lysates were centrifugated at 6000 rpm for 15 min. The supernatants of the lysates were mixed with SDS loading buffer for western blot analysis with antibody against ASC. The pellets of the lysates were washed with PBS for three times and cross-linked using fresh DSS (2 mM, Sigma) at 37˚C for 30 min. The cross-linked pellets were then spanned down and the supernatant was mixed with SDS loading buffer for western blotting analysis.

Statistical analyses
All experiments were reproducible and repeated at least three times with similar results. Parallel samples were analyzed for normal distribution using Kolmogorov-Smirnov tests. Abnormal values were eliminated using a follow-up Grubbs test. Levene's test for equality of variances was performed, which provided information for Student's t-tests to distinguish the equality of means. Means were illustrated using histograms with error bars representing the SD; a P value of <0.05 was considered statistically significant.