ANXA2 enhances virus replication through negatively regulating cGAS-STING and RLRs-mediated signal pathway

Host nucleic acid receptors can recognize the viral DNA or RNA upon virus infection, which further triggers multiple signal pathways to promote the translocation of the interferon regulatory factor 3 (IRF3) into nucleus and produce type I interferon (IFN), leading to the host antiviral response. Here, we report a novel negative regulator Annexin A2 (ANXA2) that regulates type I IFN production through multiple mechanisms. Ectopic expression of ANXA2 inhibited the production of type I IFN induced by DNA- and RNA viruses and enhanced virus replication, while knockout of ANXA2 expression enhanced the production of type I IFN and inhibited virus replication. Mechanistically, ANXA2 not only disrupted MDA5 recruiting MAVS, but also inhibited the interaction between MAVS and TRAF3 upon RNA virus infection. In addition, ANXA2 impacted the translocation of STING from endoplasmic reticulum to Golgi apparatus upon DNA virus infection. Interestingly, ANXA2 also inhibited IRF3 phosphorylation and nuclear translocation through competing with TANK-binds kinase 1 (TBK1) and inhibitor-κB kinase ε (IKKε) for binding to IRF3. Anxa2 deficiency in vivo increased the production of type I IFN, which resulted in suppression of encephalomyocarditis virus (EMCV) replication. Our findings reveal that ANXA2, as a negative regulator of type I IFN production, plays an important role in regulating the host antiviral responses. Author summary Annexin is a family of evolutionarily conserved multi-gene proteins, which are widely distributed in various tissues and cells of plants and animals. These proteins can reversibly bind to phospholipid membranes and to calcium ions (Ca2+). To date, several studies have confirmed that some members of the Annexin family regulate the antiviral innate immune response. Until now, regulation of the production of type I IFN by ANXA2 is not reported. In this study, ANXA2 were found to strongly inhibit the production of type I IFN, leading to increased virus replication while knockout of ANXA2 expression inhibited virus replication by increasing the amount of IFN. Compared with wild-type littermates, ANXA2 deficiency mice produced more type I IFN to inhibit virus replication. Our results provide methanistic insights into the novel role of ANXA2 in the antiviral innate immune responses.

1 were also higher than that of in the WT HeLa cells (Fig S3D-L). Moreover, the VSV and EMCV 119 genomic RNA copy number and the HSV-1 genomic DNA copy number were significantly lower in the 120 HeLa-Anxa2 -/cells than that of in the HeLa cells (Fig S3M-O). Collectively, these results indicated that 121 the ANXA2 deficiency enhanced the type I IFN production, which resulted in the inhibition of viral 122 replication.

ANXA2 deficiency enhances host antiviral responses in vivo 124
To further define the function of ANXA2 in inhibiting type І IFN production and host antiviral 125 responses in vivo, Anxa2 -/mice and their WT littermates were challenged with EMCV via intraperitoneal 126 injections. We found that the mRNA levels of Ifnβ1 in the heart and brain from Anxa2 -/mice were 127 significantly higher than those from WT mice after infection with EMCV for 48 h (Fig 3A and 3B) and 128 72 h (Fig 3C and 3D). Correspondingly, the protein levels of IFN-β in serum from Anxa2 -/mice were 129 also significantly increased ( Fig 3E). Consistent with these results, the EMCV genomic RNA copy 130 number in the heart were significantly lower in Anxa2 -/mice than that of their WT littermates ( Fig 3F).

131
Fewer signs of severe inflammation were observed in the brain and heart tissues from Anxa2 -/mice than 132 in those from their WT littermates ( Fig 3G). Additionally, brain tissues from WT littermates infected 133 with EMCV exhibited severe glial cell and nerve cell degenerative necrosis, while those from Anxa2 -/-134 mice exhibited few signs of glial cell and nerve cell degeneration (Fig 3G and 3H).

136
To elucidate the underlying molecular mechanisms by which ANXA2 negatively regulates type І 137 IFN production, we first assessed the effect of ANXA2 on the IFN-β promoter activation induced by key 6 138 molecules in the type І IFN signal pathway in HEK293T cells. As shown in Fig S4A- (Fig 4B, C). These results suggest that ANXA2 may inhibit type І IFN production upstream of IRF3 153 phosphorylation.

154
To examine whether ANXA2 inhibits IRF3 phosphorylation, HEK293T cells were transfected with 155 an empty vector or a plasmid expressing HA-tagged ANXA2, respectively. These cells were then mock 156 infected or infected with VSV, EMCV or HSV-1. We found that ectopically expressed ANXA2 157 significantly inhibited the phosphorylation of IRF3 induced by VSV, EMCV or HSV-1 infection ( Fig   158   4D-F). Similarly, ectopically expressed ANXA2 reduced poly(I:C)-or SeV-induced IRF3 159 phosphorylation in a dose-dependent manner (Fig S5A and 5B). Subsequently, primary peritoneal 160 macrophages from ANXA2 -/mice and their wildtype littermates were infected with VSV or HSV-1. As 161 shown in Fig 4G and 4H, the phosphorylation levels of IRF3 in primary peritoneal macrophages from 162 ANXA2 -/mice were higher than those from WT mice infected with VSV or HSV-1, although the total 163 protein levels of IRF3 were not affected. Consistent with these results, the phosphorylation level of IRF3 164 in the HeLa-Anxa2 -/cells was higher than that in the wild type HeLa cells, and overexpression of 165 ANXA2 in HeLa-Anxa2 -/cells restored the inhibitory effects of ANXA2 on SeV-mediated 166 phosphorylation of IRF3 (Fig S5C). 7

167
It has been shown that the phosphorylation of IRF3 results in its translocation to nucleus. Therefore,

168
we further studied the effect of ANXA2 on IRF3 nuclear translocation. As shown in Fig 4I and Fig S5D, 169 the amount of nuclear-translocated IRF3 was significantly reduced following ANXA2 expression upon

170
SeV and EMCV infection. In agreement with these results, Western blot analysis of the protein level of 171 IRF3 in the cytoplasmic and nuclear fractions showed that the amount of nuclear translocation of IRF3 172 induced by VSV infection decreased in a dose-dependent manner with the increasing expression of 173 ANXA2, while knockout of ANXA2 expression significantly enhanced the nuclear translocation of IRF3 174 upon VSV infection (Fig 4J and 4K). Overall, our findings reveal that ANXA2 inhibits the 175 phosphorylation and nuclear translocation of IRF3 induced by DNA virus or RNA virus infection.

177
To identify the target of ANXA2, we examined the interaction between ANXA2 and key molecules that the CARD domain of MDA5 is required for its interaction with ANXA2. Upon EMCV infection,

191
Therefore, we tested the interaction between MDA5 and MAVS and found that ANXA2 inhibited the     ). In addition, we found that interactions existed between endogenous ANXA2 and endogenous 219 STING in HEK293T cells with or without HSV-1 infection (Fig 7B). To map the STING domain required that the TM of STING is dispensable for the interaction between ANXA2 and STING ( Fig 7C). Next,

224
we further explored whether ANXA2 affects the positioning of STING in the Golgi apparatus after   ANXA2 and IRF3. Next, we tried to identify which domain of ANXA2 is required for its inhibition of 249 TBK1-mediated IFN-β production. We found that ANXA2-D 2 , ANXA2-D 3 , ANXA2-D 4 and ANXA2-250 D 5 , but not ANXA2-D 1 or ANXA2-D 6 , inhibited IFN-β-luc reporter activation and the mRNA expression 10 251 levels induced by TBK1 (Fig 8G and 8H).

278
increasing evidence has demonstrated that annexins play important roles in innate immune and 11 279 inflammatory responses [46,47]. It has been reported that in the TLR4/TLR3-TRIF signaling pathway, 280 the C-terminus of ANXA1 directly associates with TBK1 to promote the TLRs-mediated IFN-β 281 production, indicating that ANXA1 plays an important role in regulating host antiviral responses [48].

282
ANXA1 is not only related to IFN production signaling pathways, but can also affect IFN downstream 283 signaling pathway. For example, ANXA1 rapidly promotes IFN-β production and IRF3 activation after    downstream kinases such as TBK1 and IKKε to phosphorylate IRF3, leading to increased production of 304 type I IFN and expression of antiviral genes [5,51]. In this study, we found that ANXA2 not only 305 interacted with MDA5 through CARD domain to inhibit its recruitment of MAVS but also interacted 306 with the linker between CARD and TM of MAVS to inhibit the interaction between MAVS and TRAF3.

307
Upon DNA virus infection, the cGAS senses and binds to the viral DNA and catalyzes formation 12 308 of 2'-3'-cGAMP, an atypical cyclic di-nucleotide second messenger that can be sensed by STING. STING 309 translocates from ER to Golgi, leading to phosphorylation of STING. In this study, we found that 310 ANXA2 interacted with the TM domain of STING and inhibited its localization on Golgiosome and 311 phosphorylation. Activated STING recruited and activatedTBK1 [52]. Therefore, the type I IFN signaling 312 triggered by RNA virus and DNA virus converges on TBK1. Activated TBK1 phosphorylates IRF3 and 313 promotes its dimerization and translocation into the nucleus, where it forms an active transcriptional 314 complex that binds to IFN promoter and triggers the type I IFN genes transcription [53,54].

315
Normally, IRF3 mainly exists in the cytoplasm and can shuttle between cytoplasm and nucleus. The nucleus [57]. Our previous study also showed that DDX19 inhibits TBK1-and IKKε-mediated  [32,59]. In this study, we found that 102-207 aa of ANXA2 is required 334 to inhibit TBK1 induced IFN-β promoter activation, and 102-339 aa of ANXA2 is necessary to interact 335 with IRF3.

336
ANXA2 plays an essential role in the regulation of innate immune response. A report demonstrated 13 337 that ANXA2 plays an anti-inflammatory role in response to injury or viral infection [60]. Another 338 previous study showed that ANXA2 has a role in limiting inflammation by promoting anti-inflammatory 339 signals [61]. In agree with these, the mice lacking ANXA2 showed a lower survival rate when they were 340 infected with bacteria, reflecting a dysregulated inflammatory response [62]. Using the ANXA2 knockout 341 mice as model, we demonstrated that IFN-β production significantly increase in primary peritoneal 342 macrophages from Anxa2 -/mice upon EMCV infection compared to that from wild type mice, which in 343 turn inhibited viral replication. These data suggest that ANXA2 is a negative regulator of IFN-β 344 production and antiviral immune response during viral infection in vivo. Therefore, the virus infection 345 may limit IRF3 activation and IFN-β production by inducing ANXA2 expression, thereby promoting its 346 escape from the host innate immune responses.

347
In summary, we identified a novel function of ANXA2 involved in host antiviral responses. production. Therefore, this study provides a novel target for anti-viral drug design to prevent viral 354 invasion.

408
The cellular supernatants were collected and used to assess for their ability to inhibit vesicular

862
ANXA2 also competes with TBK1 or IKKε to bind to IRF3 to inhibit IRF3 phosphorylation and nuclear 863 translocation.