Deactylation by SIRT1 enables liquid–liquid phase separation of IRF3/IRF7 in innate antiviral immunity

Qin, Ziran; Fang, Xiuwu; Sun, Wenhuan; Ma, Zhenyu; Dai, Tong; Wang, Shuai; Zong, Zhi; Huang, Huizhe; Ru, Heng; Lu, Huasong; Yang, Bing; Lin, Shixian; Zhou, Fangfang; Zhang, Long

doi:10.1038/s41590-022-01269-0

Download PDF

Article
Published: 25 July 2022

Deactylation by SIRT1 enables liquid–liquid phase separation of IRF3/IRF7 in innate antiviral immunity

Nature Immunology volume 23, pages 1193–1207 (2022)Cite this article

12k Accesses
24 Citations
6 Altmetric
Metrics details

Subjects

Abstract

Innate antiviral immunity deteriorates with aging but how this occurs is not entirely clear. Here we identified SIRT1-mediated DNA-binding domain (DBD) deacetylation as a critical step for IRF3/7 activation that is inhibited during aging. Viral-stimulated IRF3 underwent liquid–liquid phase separation (LLPS) with interferon (IFN)-stimulated response element DNA and compartmentalized IRF7 in the nucleus, thereby stimulating type I IFN (IFN-I) expression. SIRT1 deficiency resulted in IRF3/IRF7 hyperacetylation in the DBD, which inhibited LLPS and innate immunity, resulting in increased viral load and mortality in mice. By developing a genetic code expansion orthogonal system, we demonstrated the presence of an acetyl moiety at specific IRF3/IRF7 DBD site/s abolish IRF3/IRF7 LLPS and IFN-I induction. SIRT1 agonists rescued SIRT1 activity in aged mice, restored IFN signaling and thus antagonized viral replication. These findings not only identify a mechanism by which SIRT1 regulates IFN production by affecting IRF3/IRF7 LLPS, but also provide information on the drivers of innate immunosenescence.

Nuclear mRNA decay: regulatory networks that control gene expression

Article 18 April 2024

Endogenous aldehyde-induced DNA–protein crosslinks are resolved by transcription-coupled repair

Article Open access 10 April 2024

Immune microniches shape intestinal Treg function

Article Open access 03 April 2024

Main

In humans and mice, pathogen-associated molecular patterns are detected by germline-encoded pattern-recognition receptors (PRRs), such as retinoic acid-inducible gene I (RIG-I)-like receptors, Toll-like receptors, cytosolic DNA sensors and nucleotide-binding oligomerization domain-like receptors, serving as the first line of defense against pathogen invasion^1,2. After recognition of foreign nucleic acids, PRRs recruit adaptors to activate the IFN regulatory factors IRF3 and IRF7 via carboxyl terminus phosphorylation by the serine/threonine protein kinase TBK1 and IKKɛ^3,4. To eliminate infection and avoid harmful immune pathology, a precise and appropriate induction of type I IFNs, including IFN-β and IFN-α, needs to be achieved by activated IRF3 and/or IRF7 via binding and transactivating IFN-stimulated response elements (ISREs) in IFN-I promoters^5,6. These type I IFNs are the best-defined cytokines, known for an ability to induce an antiviral state in virus-infected cells and non-infected bystander cells by inducing a transcriptional program that interferes with multiple stages of viral replication^7,8.

In both influenza and COVID-19 pandemics, very low and delayed IFN-I induction and much higher rate of hospitalization and death occurred in elderly patients when compared to juvenile or other younger individuals^{9,10,11,12,13}. These observations suggest that innate antiviral immunity undergoes deterioration with aging; however, the underlying mechanism for age-associated degeneration of IFN signaling and related innate immunosenescence is not clear.

Here, we show that IRF3/IRF7 is transcriptionally active in a phase-separated state and reveal a mechanism for the role of IRF3/IRF7 LLPS-mediated ISRE binding, in which DBD deacetylation is identified as a necessary step. This study also provides a therapeutic method to improve IRF3/IRF7 LLPS by using SIRT1 agonists, which might reverse innate immunosenescence in elderly individuals.

Results

IRF3 undergoes LLPS after viral infection

From its N to C terminus, the IRF3 protein contains a well-conserved DBD, an IRF-associated domain (IAD) that facilitates dimerization and an inhibitory domain (ID) that keeps IRFs in an inactive monomeric state until activated by C-terminal phosphorylations^14,15,16. Investigation with IUPred and PONDR (using an online program http://pondr.com/) predicted that there is a largely unfolded intrinsically disordered region (IDR) between the N and C termini of IRF3 (Extended Data Fig. 1a). Given that proteins with extensive IDRs that are also involved in multivalent protein interactions tend to undergo LLPS^17,18, we investigated whether IRF3 has the ability to undergo phase separation. We purified recombinantly expressed green fluorescent protein (GFP)–IRF3 fusion proteins (Extended Data Fig. 1b). GFP–IRF3 formed spherical droplets with an aspect ratio (maximal diameter to minimal diameter) close to 1 (Fig. 1a–d), indicating a high degree of circularity. In addition, the abundance of GFP droplets increased, from barely detectable small foci to regular droplets and large droplet clusters, with increasing protein concentration, salt concentrations within a pH of 5.0–8.0 and to some extent temperatures (Fig. 1a–d). Consistent with this, droplet formation was substantially inhibited by 5% 1,6-hexanediol (Hex), a compound that putatively disrupts weak hydrophobic interactions (Extended Data Fig. 1c) and was completely abolished by treatment with heat or proteinase K (Extended Data Fig. 1d,e). Time-lapse microscopy found that the droplet size of AF488-labeled IRF3 increased over time (Fig. 1e), ruling out the possibility that the LLPS puncta were artificially formed by the GFP tag. After selective fluorescence recovery after photobleaching (FRAP) of the center part of the droplets, fluorescence recovery was observed on a rapid time scale (Fig. 1f) and more efficiently at 37 °C or 39.6 °C than at 16 °C (Fig. 1g), indicating an efficient diffusion of the IRF3 liquid droplets at physiological temperature.

**Fig. 1: IRF3 undergoes liquid–liquid phase separation.**

A spherical shape and the ability to fuse and recover from photobleaching are some of the features of a liquid-like phase-separated structure^17,18,19. To test whether IRF3 also undergoes LLPS in intact cells, we ectopically expressed GFP–IRF3 in U2OS cells and found that GFP–IRF3 formed discrete puncta in the nucleus after infection with Sendai virus (SeV) (Fig. 1h,i), indicating the enrichment of GFP–IRF3 in vivo. In line with the in vitro results, live-cell imaging showed that the IRF3-enriched nuclear condensates readily fused into larger structures over time (Fig. 1h) and recovered quickly following FRAP (Fig. 1i). Three-dimensional (3D) reconstruction also revealed the spherical appearance of GFP–IRF3 condensates in cells (Extended Data Fig. 1g). Notably, in the absence of stimulation, GFP–IRF3 was largely dispersed in the cytoplasm and the SeV-induced condensation of nuclear IRF3 could also be disrupted by 5% 1,6-hexanediol (Fig. 1j), suggesting that the transcriptionally activated IRF3 is in a phase-separated state.

To unambiguously determine the effect of C-terminal phosphorylation on IRF3 LLPS, we utilized genetic code expansion orthogonal systems^20,21 to incorporate phospho-serine in recombinant IRF3 and to produce Ser386-phosphorylated or Ser396-phosphorylated IRF3 (IRF3^Ser386-Pho or IRF3^Ser396-Pho, respectively). Antibodies specific for Ser386-phosphorylated or Ser396-phosphorylated IRF3 could detect purified IRF3^Ser386-Pho or IRF3^Ser396-Pho, respectively, but not IRF3^Non-Pho (Extended Data Fig. 1g). In vitro droplet formation assays consistently showed that IRF3^Ser386-Pho and IRF3^Ser396-Pho were substantially more efficient than IRF3^Non-Pho in condensing and forming liquid-like droplets (Fig. 1k). Of note, endogenous IRF3 also exhibited the formation of nuclear puncta in both L929 and HeLa cells in response to viral stimulation (Fig. 1l and Extended Data Fig. 1h). To further verify the subcellular location of endogenously phosphorylated IRF3, we performed a proximity ligation assay (PLA)²², which enables the detection of protein modification or aggregation in situ with high specificity and sensitivity²³. We obtained punctate PLA signals from SeV-infected cells that localized exclusively in the nucleus (whereas no IRF3 PLA signal was detected in non-infected cells; Fig. 1m). Together, these results indicate that IRF3 undergoes LLPS and forms nuclear condensates upon activation by viral infection.

DBD, IDR and IAD are necessary for IRF3 LLPS

To identify the domains in IRF3 that are required for phase separation, a series of truncated IRF3 labeled with GFP were bacterially expressed and purified (Extended Data Fig. 1b). Although the removal of the IDR domain (d-IDR) showed an effect on limiting IRF3 LLPS, deletion of the IAD domain (d-IAD) considerably reduced but did not eliminate droplet formation. Deletion of the DBD domains (d-DBDs) abolished IRF3 LLPS (Fig. 2a). Similar to the C-terminal phosphorylation (Fig. 1k), removal of the ID domain (d-ID) substantially enhanced IRF3 droplet formation (Fig. 2a). Thus, the d-DBD and, to a lesser extent, the d-IAD and d-IDR are required for IRF3 phase separation in vitro.

**Fig. 2: DBD, IAD and IDR are necessary for IRF3 LLPS.**

Activated IRF3 dimers recognize ISRE upstream of IFN-I and ISGs⁵. Thus, we evaluated whether ISRE DNA was involved in GFP–IRF3 droplet formation. Purified wild-type (WT) IRF3, but not d-DBD, could condense to liquid-like droplets with Cy3-labeled ISRE DNA. Under the same dosage and conditions, these droplets fused by IRF3 and ISRE DNA were substantially larger in size and showed stronger fluorescence intensity than the droplets condensed by IRF3 alone (Fig. 2b). These observations suggest that IRF3 can phase separate with ISRE DNA and this ability is likely mediated by DBD. To confirm this, we incubated Cy3-ISRE DNA with GFP–IRF3 WT or GFP–IRF3 d-DBD under a matrix of concentration in vitro to generate a phase diagram, which clearly demonstrated the requirement of IRF3 DBD for an efficient condensation of IRF3 with ISRE DNA (Fig. 2b). In comparison to the DNA-free protein droplets, IRF3 fused with ISRE DNA exhibited higher phase reversibility and faster recovery rates after photobleaching (Fig. 2c,d). Thus, the IRF3 association with ISRE DNA promotes LLPS.

Next, we compared the phase separation properties of IRF3 WT with IRF3 truncations in cells. In response to viral infection, the GFP–IRF3 WT rapidly condensed into discrete nuclear puncta, whereas GFP–IRF3 d-DBD remained evenly distributed in the cytoplasm, similar to the localization of GFP–IRF3 WT before viral stimulation (Fig. 2e and Extended Data Fig. 1i). Similar to GFP–IRF3 d-DBD, GFP–IRF3 d-IDR and GFP–IRF3 d-IAD were diffusely distributed in the cytoplasm with small dots, if any, scattered in the cytosol, whereas GFP–IRF3 d-ID condensed as nuclear puncta even more efficiently than GFP–IRF3 WT (Fig. 2e and Extended Data Fig. 1i). In line with these findings, transfection of IRF3 WT and IRF3 d-ID, but not the d-IDR, d-IAD or d-DBD, restored SeV-induced IFNB1 expression in IRF3-deficient (IRF3 knockout (KO)) HEK293T cells (Extended Data Fig. 1j), suggesting that the ability of IRF3 to form a phase separation cluster in the nucleus is linked to its transcriptional activity. To further ascertain whether each domain plays a role in IRF3 phase separation, we used an optogenetic platform ‘optoDroplet’ to design optoIDR constructs in which different domains from IRF3 were fused to mCherry and the photolyase homology region of Arabidopsis thaliana Cry2, a light-sensitive protein that self-associates upon blue light exposure to facilitate LLPS²⁴. Consistent with previous reports^25,26,27, mCherry–Cry2(-) alone showed little clustering upon blue light activation (Fig. 2f). Notably, fusing DBD, IDR or IAD of IRF3 to mCherry–Cry2 led to rapid blue light-dependent cluster assembly and speck formation in most cells (Fig. 2f). Therefore, DBD, IDR and IAD are necessary for IRF3 to perform efficient LLPS in cells.

IRF3 compartmentalizes IRF7 in LLPS condensates

Next, we sought to identify the components in the IRF3 condensates. Given that components involved in phase separation do not necessarily have direct and strong interactions with each other, we employed a method genetically encoded chemical crosslinking of proteins coupled with mass spectrometry (GECX–MS)²⁸ to identify IRF3-binding proteins in live cells (Fig. 3a). This assay identified IRF7, another key inducer of IFN-I signaling^5,6, with multiple unique peptides, as a strong component of IRF3 condensates (Fig. 3a). As the closest family member, IRF7 shares very high homology with IRF3 and is predicted to have IDR (Extended Data Fig. 2a). Thus, we tested whether IRF7 has LLPS properties.

Purified mCherry-conjugated IRF7 (Extended Data Fig. 2b) also formed microdroplets in solution (Fig. 3b), which was inhibited by 5% 1,6-hexanediol (Hex) and disrupted by heat or proteinase K (Extended Data Fig. 2c–e). More numerous and larger droplets were observed in high salt concentrations, within a pH range of 5.0–8.0 and at high temperatures (Fig. 3b,c and Extended Data Fig. 2f,g). The droplet size of IRF7s labeled with AF594 increased with time (Fig. 3d and Extended Data Fig. 2i), indicating fusion and that the puncta were not artificially formed by the mCherry tag. The FRAP assay revealed fluorescence recovery after bleaching (Extended Data Fig. 2h), indicating the diffusion within IRF7 droplets. Similar to IRF3, deletion of d-DBD, d-IDR or d-IAD inhibited, whereas deletion of the d-ID considerably enhanced IRF7 droplet formation (Fig. 3e and Extended Data Fig. 2j). Moreover, the purified mCherry–IRF7 WT, but not d-DBD, was found to phase separate with FITC-labeled ISRE DNA (Fig. 3f). Thus, IRF7 also undergoes LLPS with ISRE DNA.

We next evaluated whether IRF7 phase separated with IRF3. Upon incubation, the fluorescently labeled IRF3 and IRF7 rapidly formed micrometer-sized liquid droplets and the small liquid droplets quickly fused into larger ones, accompanied by increased fluorescence intensity and a larger equivalent diameter (Fig. 3g and Extended Data Fig. 2k). Time-lapse imaging revealed that IRF7 readily fused with phase-separated IRF3 and colocalized to all IRF3 droplets, while IRF7 d-IAD barely fused into the IRF3 liquid droplets (Fig. 3g), demonstrating that the IRF7 d-IAD is needed for the condensation of IRF7 with IRF3. Following photobleaching, the condensates of IRF7 fused with IRF3 were recovered more efficiently than those of IRF7 alone (Fig. 3h). Thus, fusion with IRF3 enhanced the IRF7 LLPS in vitro.

In cells, PLA revealed that the endogenous IRF3 and IRF7 together yielded discrete punctate nuclear staining immediately after the viral infection, which was not detected in non-infected cells (Fig. 3i). When tracked by live-cell imaging, the puncta condensed by GFP–IRF3 and mCherry–IRF7 appeared in the nucleus in less than 3 h after SeV infection, with GFP–IRF3 puncta forming first and mCherry–IRF7 merging into the puncta as a function of time (Fig. 3j). Considering that IRF7 is constitutively expressed in pDC and monocytes, whereas in most other cell types IRF7 is induced by innate IFN priming, this result suggests that IRF3 might be able to prime IRF7 condensation in cells. To verify this, we compared IRF7 condensation in control and IRF3-deficient cells. We found that, in response to SeV infection, endogenously expressed IRF3 and IRF7 could rapidly colocalize as nuclear puncta in control WT cells, whereas in the IRF3 KO cells, IRF7 was evenly distributed in the nucleus and failed to condense as LLPS puncta (Fig. 3k and Extended Data Fig. 2l). Together, these results suggest that IRF3 promotes IRF7 LLPS in vivo and compartmentalizes IRF7 to the nuclear puncta.

SIRT1 is required for IRF3 LLPS and transcription in cells

After stably expressing GFP–IRF3 in IRF3 KO cells, we noticed that the IRF3 condensation appeared 3 h after SeV infection and peaked at 8 h (Fig. 4a). The timing of initial induction and peak for IFNB1 mRNA expression was largely consistent with that for IRF3 condensation but delayed by approximately 2 h (Fig. 4a). Moreover, the IRF3 nuclear condensates were also enriched for MED1, a component of the super-enhancer marker^25,29 (Fig. 4b), suggesting that the IRF3 LLPS puncta are likely to concentrate at the transcription apparatus. In contrast, we detected H3K9me3, a transcriptionally repressive mark and found that signals from this mark were completely separated from IRF3 condensates (Fig. 4b). These results support the idea that IRF3 forms LLPS condensates with ISRE, which are enriched with key transcription factors and co-activators to enable IFN-I gene expression.

**Fig. 4: SIRT1 is required for IRF3 LLPS and innate antiviral immune responses.**

To identify critical component(s) of the IRF3 LLPS puncta in the nucleus, we treated GFP–IRF3 stable cells with a library of 216 epigenetic compounds that target key protein modifications and gene transcription (Fig. 4c). Of note, cells pre-treated with SIRT1 inhibitor EX527 lost Ifnb1 induction upon viral infection (Fig. 4d) and the nuclear condensation of IRF3 induced by either RNA virus SeV or the DNA virus herpes simplex virus type 1 (HSV-1) was also significantly repressed compared to the control (Fig. 4e and Extended Data Fig. 3a). In line with this, PLA and immunofluorescence staining showed that pre-treatment with EX527 also prevented IRF3 from compartmentalizing IRF7 (Fig. 4f and Extended Data Fig. 3b) and enriching transcriptional co-activator MED1 (Fig. 4g). To consolidate the function of EX527 in the innate antiviral response, reporter plasmids containing the promoter of IFN-β, PRD I–III (containing the IRF3-binding site of the Ifnb1 promoter) or the promoter of IFN-α were transfected into HEK293T cells. After SeV stimulation, a significant downregulation of the activities of these reporters were observed (Extended Data Fig. 3c). EX527 also severely reduced mRNA expression of Ifnb1 and Ifna upon infection with vesicular stomatitis virus (VSV) or HSV-1 (Extended Data Fig. 3d), indicating that the enzymatic activity of SIRT1 was critical in viral-induced IFN-I expression.

To explore the function of SIRT1 in the innate antiviral response, we transfected four independent small interfering RNAs (siRNAs) that target human SIRT1 (sh-SIRT1 no. 1 to sh-SIRT1 no. 4) into HEK293T cells and identified no. 1 and no. 2 sh-SIRT1 as the efficient ones (Extended Data Fig. 3e). Depletion of SIRT1 with sh-SIRT1 no. 1 or sh-SIRT1 no. 2 resulted in significant downregulation of IFN-β, PRD I–III and IFN-α-Luc promoter activity and expression of IFNB1 and IFNA upon induction by SeV, relative to their expression in cells transfected with control shRNA (Extended Data Fig. 3f,g). As a negative control, sh-SIRT1 no. 3 and sh-SIRT1 no. 4 showed no effect (data not shown). In RAW264.7 mouse macrophages, SeV-induced expression of Ifnb1 and Ifna mRNA was also decreased by Sirt1 depletion (Extended Data Fig. 3h). After challenge with VSV, Ifnb1 and Ifna expression was inhibited and the VSV-specific mRNA and titers were increased in Sirt1-depleted cells (Fig. 4h). Consistent with this, the replication of VSV-expressing GFP (GFP–VSV) was also increased by SIRT1 depletion (Fig. 4i). We next investigated whether SIRT1 overexpression had any effect on IFN-I signaling. Ectopic expression of WT SIRT1, but not its deacetylase-inactive mutant H363Y (hereafter denoted as HY), elevated SeV-induced luciferase activity of the IFN-β/IFN-α promoter or PRD I–III reporter (Extended Data Fig. 4a). In both RAW264.7 and HEK293T cells, overexpression of SIRT1 WT, but not HY, stimulated the expression of IFN-β- and IFN-α-encoding messenger RNAs induced by SeV or the synthetic RNA duplex poly(I:C) (Extended Data Fig. 4b,c). Notably, the rather potently stimulated expression of IFNB1 and IFNA by the constitutively active IRF3 (called ‘IRF3-5D’ here) or IRF7 could still be further enhanced by SIRT1 WT, but not by SIRT1 HY (Extended Data Fig. 4d), suggesting that SIRT1 might act downstream of IRF3/IRF7 phosphorylation. Fluorescence microscopy showed that SIRT1 WT, but not the HY mutant inhibited the replication of VSV–GFP (Extended Data Fig. 4e). Upon VSV infection, Ifnb1 and Ifna expression were activated and the abundance of VSV-specific mRNA, VSV-G proteins and VSV titers were reduced by SIRT1 WT but not by SIRT1 HY (Fig. 4j and Extended Data Fig. 4f). Similar results were obtained in cells transfected with SIRT1 WT or HY and infected with HSV-1 (Fig. 4k). Together, these data suggest that SIRT1 positively regulates IFN-I signaling induced by RNA or DNA viruses.

Myeloid SIRT1 deficiency causes innate immune evasion in mice

To demonstrate the role of SIRT1 in regulating IRF3 LLPS and downstream innate antiviral immunity, we isolated mouse embryonic fibroblasts (MEFs) from Sirt1 KO (Sirt1^−/−) mice³⁰. SeV-induced nuclear condensates of endogenous IRF3 and IRF7 were clearly observed in control MEFs, whereas they were significantly reduced in Sirt1^−/− MEFs (Fig. 5a). The Sirt1^−/− mouse is embryonic lethal. To investigate the function of SIRT1 in innate antiviral immunity in vivo, we generated Sirt1 conditional KO mice (Lyz2-Cre⁺Sirt1^f/f), which undergo deletion of loxP-flanked Sirt1 alleles (Sirt1^f/f) specifically in myeloid cells via Cre recombinase expressed under control of myeloid cell-specific promoter Lyz2 (Lyz2-Cre) (Extended Data Fig. 5a). Lyz2-Cre⁺Sirt1^f/f mice were viable and normal in size and had no gross physiological or behavioral abnormalities. Primary peritoneal macrophages were then prepared from Lyz2-Cre⁻Sirt1^f/f and Lyz2-Cre⁺Sirt1^f/f mice and stimulated with SeV, RNA mimics, VSV or HSV-1. The expression of Ifnb1, Ifna and their downstream chemokine-encoding genes Cxcl10 and Ccl5 were significantly downregulated in Lyz2-Cre⁺Sirt1^f/f macrophages stimulated with SeV, 5′-triphosphorylated RNA or poly (I:C) compared to that in Lyz2-Cre⁻Sirt1^f/f macrophages (Extended Data Fig. 5b–d). Consistent with this, SeV- or HSV-1-inducedproduction of IFN-β and IFN-α proteins in Lyz2-Cre⁺Sirt1^f/f peritoneal macrophages was much less than that in Lyz2-Cre⁻Sirt1^f/f macrophages (Fig. 5b). After infection with VSV, the level of Ifnb1 and Ifna mRNA was much lower and the copy number and replication of VSV were much higher in Lyz2-Cre⁺Sirt1^f/f peritoneal macrophages than in Lyz2-Cre⁻Sirt1^f/f peritoneal macrophages (Extended Data Fig. 6a). Similar results were obtained when HSV-1 was used as the stimulus (Extended Data Fig. 6b). We also isolated bone-marrow-derived macrophages (BMDMs) from Lyz2-Cre⁺Sirt1^f/f and Lyz2-Cre⁻Sirt1^f/f mice and assayed their antiviral responses together with the Sirt1^+/+ and Sirt1^−/− MEFs in vitro; qPCR analysis of Ifnb1 and Ifna expression in these cells showed similar results with peritoneal macrophages (Extended Data Fig. 6c,d). VSV-G proteins and VSV-expressing GFP were also higher in Sirt1^−/− MEFs than in controls (Fig. 5c).

**Fig. 5: SIRT1 deficiency abolishes IRF3/IRF7 LLPS and leads to innate immune evasion.**

To further elucidate the function of SIRT1 in vivo, we challenged Lyz2-Cre⁻Sirt1^f/f and Lyz2-Cre⁺Sirt1^f/f mice with VSV. As expected, the expression of Ifnb1 and Ifna mRNA in the lungs, liver and spleen and the concentration of IFN-β and IFN-α in the serum of Lyz2-Cre⁺Sirt1^f/f mice were much lower than those in Lyz2-Cre⁻Sirt1^f/f mice (Fig. 5d,e). Consistent with that, higher VSV-specific mRNA, VSV-G expression and VSV titers in the lungs, liver and spleen together with more severe lung injury and higher mortality were observed and in Lyz2-Cre⁺Sirt1^f/f mice than in Lyz2-Cre⁻Sirt1^f/f mice (Fig. 5f–j). We also compared the antiviral responses of Lyz2-Cre⁻Sirt1^f/f and Lyz2-Cre⁺Sirt1^f/f mice to infection with HSV-1 and got consistent and similar results (Fig. 5k–m). These in vivo data indicate that SIRT1 is an important activator of innate immune responses to both RNA and DNA viruses.

SIRT1 directly deacetylates IRF3/IRF7 in the DBD

The findings described above raised the question of how SIRT1 regulates IRF3/IRF7 LLPS and subsequent innate antiviral immunity. Considering that the enzymatic activity of SIRT1 is crucial and IRF3/IRF7 can be acetylated^31,32,33, we investigated whether IRF3/IRF7 could be regulated via deacetylation. Indeed, we noticed that endogenous IRF3 and IRF7 seemed to be highly acetylated in Sirt1-deficient macrophages and MEFs (Fig. 6a). Using acetyl lysine antibody, we confirmed that acetylated IRF3 and IRF7 were enriched in macrophages and MEFs upon EX527-mediated SIRT1 inhibition (Extended Data Fig. 7a). We thus extracted Flag-tagged IRF3 and IRF7 from the EX527 pre-treated cells and subjected them to mass spectrometry. This enabled us to identify Lys39 and Lys77 as specific acetylation residues of IRF3 and Lys45 and Lys92 as specific acetylation residues of IRF7 (Extended Data Fig. 7b). Sequence alignment revealed that both acetylation sites identified from IRF3 or IRF7 are conserved in their orthologs and located in rather similar motifs of their DBD (Fig. 6b). To predict the involvement of these site-specific acetylations in regulating IRF3/IRF7 LLPS, we performed PONDR score analysis of IRF3/IRF7 WT and mutants by replacing lysine with glutamine (Gln), which mimics the acetylation state of lysine and tested the effects of these mutations on IRF3/IRF7. Compared to IRF3 WT and IRF7 WT, IRF3-K39 and 77Q (IRF3-2KQ) and IRF7-K45 and 92Q (IRF7-2KQ) were apparently less disordered at the mutated sites (Fig. 6c), suggesting that deacetylation of these residues would likely promote IRF3/IRF7 LLPS.

To test whether SIRT1 deacetylates IRF3 and IRF7, hyperacetylated IRF3 and IRF7 were purified from Sirt1-deficient cells and then incubated with the prokaryotic purified SIRT1 WT or SIRT1 HY proteins (Extended Data Fig. 7c). As expected, SIRT1 WT, but not SIRT HY, efficiently removed IRF3 and IRF7 acetylation in the presence, but not in the absence, of NAD⁺ (Fig. 6d), indicating that SIRT1 directly deacetylates both IRF3 and IRF7 in vitro. In peritoneal macrophages and BMDMs, endogenous SIRT1 was found to associate with IRF3 and IRF7 and their interactions were stimulated by viral infection (Fig. 6e). To confirm that SIRT1 can deacetylate IRF3 and IRF7 on specific lysine residues, we generated antibodies for Lys39- or Lys77-acetylated IRF3 and Lys45- or Lys92-acetylated IRF7. Immunoprecipitation with these antibodies showed that cells transfected with SIRT1 WT, but not SIRT1 HY, exhibited undetectable acetyl modifications of IRF3 or IRF7 on these lysine residues (Fig. 6f). Consistent with above findings, IRF3 acetylation on Lys39 and Lys77 and IRF7 acetylation on Lys45 and Lys92 was much more abundant in Lyz2-Cre⁺Sirt1^f/f peritoneal macrophages than in Lyz2-Cre⁻Sirt1^f/f macrophages (Fig. 6g). Of note, SeV infection promoted deacetylation of both IRF3 and IRF7 in Lyz2-Cre⁻Sirt1^f/f cells but not in Lyz2-Cre⁺Sirt1^f/f cells (Fig. 6g), showing that Sirt1 is critically required. Considering that the identified acetylated lysines are all located in the DBD of IRF3/IRF7, which is a key region responsible for IRF3/IRF7–ISRE association, acetylation and acetyl-mimicking mutations at these sites likely affect the interactions between IRF3/IRF7 DBD and ISRE, thereby interfering with IRF3/IRF7 function.

In support of this hypothesis, we purified IRF3-2KQ and IRF7-2KQ (Extended Data Fig. 7d). These two mutants showed severely reduced binding affinity for ISRE DNA in microscale thermophoresis (MST) binding affinity quantification assays in vitro (Fig. 6h). To test whether these acetyl-mimicking mutations induce changes of phase separation in IRF3 and IRF7, we incubated fluorescently labeled IRF3/IRF7-2KQ with ISRE DNA and then compared the micrometer-sized liquid droplets formation with IRF3/IRF7 WT. Both IRF3-2KQ and IRF7-2KQ were inefficient for LLPS (Fig. 6i,k). Following bleaching, binding of either IRF3-2KQ or IRF7-2KQ to ISRE DNA was recovered much less efficiently compared to that in their WT counterparts (Fig. 6j,l and Extended Data Fig. 7e,f). Thus, the IRF3/IRF7 acetyl-mimicking mutation in DBD abrogates LLPS with ISRE DNA.

To predict the functional consequences of DBD acetylation in vivo, we transfected IRF3 KO cells to re-express IRF3 WT or mutants. Compared to IRF3 WT, either IRF3-K39Q, IRF3-K77Q single mutant or the IRF3-2KQ double mutant did not show transcriptional activity upon viral infection (Fig. 6m and Extended Data Fig. 7g). We also generated single or double KQ mutations based on IRF3-5D and found that transfection of any of these constructs failed to induce IFNB1 and ISG56 (Fig. 6m and Extended Data Fig. 7g). Moreover, we found a similar phenomenon in IRF7 (the acetylation simulation of a single site in its DBD), either K45Q or K92Q blocked the transcriptional activity of IRF7 and even IRF7-6D, a constitutively active IRF7 (Fig. 6m and Extended Data Fig. 7g). These results suggest that DBD acetylation at a specific single residue is sufficient to block the transactivation of IRF3 and IRF7.

Site-specific acetylation of IRF3/IRF7 DBD abrogates LLPS

To unambiguously determine the effect of acetylation on IRF3/IRF7 DBD, we employed a genetic code expansion orthogonal system to incorporate acetyl lysine into recombinant IRF3 and IRF7 to create single or double site-specific fully acetylated proteins in vitro (Fig. 7a). As expected, antibodies specific to Lys39- or Lys77-acetylated IRF3 could detect IRF3^Lys39-Ac and IRF3^Lys77-Ac, respectively and they recognized IRF3^{Lys39, 77-Ac (dual-Ac)}. Antibodies for Lys45- or Lys92-acetylated IRF7 could detect IRF7^Lys45-Ac and IRF7^Lys92-Ac, respectively and they recognized IRF7^{Lys45, 92-Ac (Dual-Ac)} (Fig. 7a). We then performed MST assays to compare the ISRE DNA-binding activities of these proteins and to evaluate any difference in the activity and specificity. We found that DBD-acetylated IRF3/IRF7, at either single or double sites, was much less efficient than IRF3/IRF7^Non-Ac in binding to ISRE DNA in vitro, indicating that DBD-acetylated IRF3/IRF7 is indeed less active (Fig. 7b). To test whether acetylation alters the phase separation of IRF3/IRF7 in vitro, we incubated fluorescently labeled IRF3/IRF7 with ISRE DNA. IRF3^Non-Ac and ISRE DNA rapidly formed larger liquid droplets with much higher fluorescence intensity and reversibility than IRF3^Lys39-Ac, IRF3^Lys77-Ac and IRF3^{Lys39, 77-Ac (dual-Ac)} did (Fig. 7c). Similar results were obtained when IRF7^Non-Ac was compared to IRF7^Lys45-Ac, IRF7^Lys92-Ac and IRF7^{Lys45, 92-Ac (Dual-Ac)} (Fig. 7d). Therefore, the presence of an acetyl moiety at a specific lysine residue of IRF3/IRF7 DBD abrogates IRF3/IRF7 LLPS.

Compared to IRF3^Non-Ac, IRF3^Lys39-Ac and IRF3^Lys77-Ac showed enhanced binding affinity for SIRT1 protein in MST assays and the enhancement was close to ~1,000-fold in IRF3^{Lys39, 77-Ac (dual-Ac)}. Similarly, the binding affinity of IRF7^Lys45-Ac, IRF7^Lys92-Ac and IRF7^{Lys45, 92-Ac (dual-Ac)} to SIRT1 was ~50, ~500 and ~3,500 times higher, respectively, than that of IRF7^Non-Ac (Fig. 7e). In the presence of NAD⁺, all purified IRF3/IRF7 proteins that were acetylated at a single site or double sites could be efficiently deacetylated by SIRT1 in vitro at a similar speed (Fig. 7f). Upon in vitro incubation with SIRT1 and NAD⁺, the small-sized, less-abundant liquid droplets fused by IRF3/IRF7^dual-Ac and ISRE DNA could be rescued to a scale similar to IRF3^Non-Ac in number, fluorescence intensity and diameter (Fig. 7g). Together, these results suggest that SIRT1 recognizes acetylated IRF3/IRF7 and promotes LLPS by catalyzing deacetylating reactions.

To verify the influence of DBD acetylation on IRF3/IRF7 LLPS and cellular IFN-I signaling, we developed an orthogonal system in vivo (Fig. 7h). Using this system, IRF3 KO cells were engineered to express IRF3 WT, IRF3^K39-Ac, IRF3^K77-Ac, IRF3^{Ks39, 77-Ac (dual-Ac)}, IRF7 WT, IRF7^Ks45-Ac, IRF7^K92-Ac and IRF7^{K45, 92-Ac (dual-Ac)}, respectively (Fig. 7h). Upon SeV infection, the IRF3 WT condensed as discrete nuclear puncta, whereas the IRF3^K39-Ac, IRF3^K77-Ac and IRF3^{Ks39, 77-Ac (dual-Ac)} presented a diffuse distribution in the nucleus (Fig. 7i). Similar results were obtained by comparing IRF7 WT and site-specific acetylated IRF7 proteins (Fig. 7i), showing that a single site-specific acetylation in IRF3/IRF7 DBD is indeed sufficient to block phase separation. As assessed by chromatin immunoprecipitation (ChIP) and qPCR analysis, acetylation of IRF3 or IRF7 at either a single or double specific sites impaired their ability to bind with the targeted ISRE sequence in IFNB1 or IFNA promoter and to trigger their mRNA expression upon viral infection (Fig. 7j). Consistent with this, the replication of GFP–VSV was strongly repressed in IRF3 or IRF7 WT-expressed cells, whereas the engineered expression of acetylated IRF3 or IRF7 barely affected GFP–VSV replication (Fig. 7k). Together, these results suggest that the addition of an acetyl moiety at specific DBD sites abolishes phase separation of IRF3/IRF7 and subsequent induction of IFN-I.

Reinforcing SIRT1 activity rescues innate immunosenescence

Among patients with influenza, elderly individuals showed a significantly lower cellular SIRT1 enzymatic activity and lower serum IFN-β than non-aged (Fig. 8a); SIRT1 activity showed significant positive correlation with IFN-β production (Fig. 8b). Compared to the non-senescent cells, the senescent showed increased DBD acetylation of IRF3/IRF7 (Lys39- and Lys77-acetylated IRF3, Lys45- and Lys92-acetylated IRF7) together with reduced SIRT1 activity and IFN-I responses (Fig. 8c). Moreover, peritoneal macrophages from aged mice also showed significant lower cellular SIRT1 enzymatic activity and lower IFN-β induction than those from non-aged mice (Extended Data Fig. 8a–c); in those mouse cells, SIRT1 activity were also positively correlated with IFN-β production upon infection with either VSV or HSV-1 (Extended Data Fig. 8b,c). Compared to peritoneal macrophages from non-aged mice, the endogenous IRF3/IRF7 DBD acetylations were much more abundant in macrophages from aged mice and they were more resistant to SeV-induced deacetylations (Extended Data Fig. 8d). SeV-induced nuclear condensates of endogenous IRF3 and IRF7 were clearly observed in primary mouse pulmonary fibroblasts from non-aged mice, whereas they were sharply reduced in cells from aged mice (Extended Data Fig. 8e). After challenge with VSV, the aged mice showed inhibited IFN-β expression in the organs and serum (Extended Data Fig. 8f,g) and enhanced viral replication, lung injury and mortality compared to non-aged mice (Extended Data Fig. 8h–k). HSV-1 infection also led to similar results as for VSV in aged and non-aged mice (Extended Data Fig. 8l–p). These in vitro and in vivo data indicate that loss of SIRT1 activity in aged host causes innate immunosenescence by interfering with IRF3/IRF7 LLPS and IFN signaling.

To investigate whether restoring SIRT1 activity can elevate innate antiviral immunity in the host, we used a natural phytoalexin compound, resveratrol (SRT501) and a chemically synthetic compound SRT2183 (refs. ^34,35). Pre-treatment with SRT501 or SRT2183 strongly inhibited the expression of Lys39- and Lys77-acetylated IRF3, Lys45- and Lys92- acetylated IRF7 in WT cells, but not in SIRT1-deficient cells (Fig. 8d and Extended Data Fig. 9a), indicating that those SIRT1 activating compounds (STACs) indeed promote deacetylation of IRF3 and IRF7 via SIRT1. As expected, the nuclear condensation of IRF3 and IRF7 induced by viral infection can be further enhanced by SRT501 or SRT2183 in WT cells, but not in SIRT1-deficient cells (Fig. 8e) and the SeV-induced binding of IRF3 and IRF7 to the ISRE of IFN-I promoters and the expression of Ifnb1 and Ifna was also significantly elevated by SRT501 or SRT2183 in Sirt1^+/+ MEFs, but not in Sirt1^−/− MEFs (Extended Data Fig. 9b and Fig. 8f), confirming the innate immunity-enhancing effect of these STACs.

To investigate whether these STACs can prevent innate immunosenescence in vivo, aged mice were pre-injected with or without SRT501 or SRT2183 followed with VSV infection (Fig. 8g). The mRNA expression of Ifnb1 and Ifna in the spleen, liver and lungs and the concentrations of IFN-β and IFN-α in the serum of SRT501- or SRT2183-treated aged mice were all significantly higher than those in the controls (Extended Data Figs. 9c and 8g). Consistent with these results, SRT501 or SRT2183 treatment considerably reduced the VSV-specific mRNA levels, VSV-specific protein VSV-G expression and VSV titers in those organs of aged mice (Fig. 8h and Extended Data Fig. 9d). Notably, the aged mice that were pre-treated with SRT501 or SRT2183 showed lesser damage in lungs and seemed to be less sensitive to VSV infection than those pre-treated with PBS (Fig. 8i and Extended Data Fig. 9e). Immunofluorescence confirmed that pre-treatment with SRT501 or SRT2183 enhanced nuclear condensation of IRF3 and IRF7 in VSV-infected liver tissues (Extended Data Fig. 9f). Notably, SRT501 and SRT2183 could also rescue innate antiviral immune responses and inhibit viral load in aged mice when infected with HSV-1 (Fig. 8j–l). Together, the results from the in vitro and in vivo analyses demonstrated that SRT501 and SRT2183 can rescue IRF3/IRF7 LLPS and thus prevent innate immunosenescence in aged hosts (Fig. 8m).

Discussion

Increasing studies have shown evidence that LLPS has an important function in human health and disease³⁶. Phase-separated multi-molecular assemblies provide a general regulatory mechanism to compartmentalize biochemical reactions within cells^19,37. Here phase separation enables IRF3 to activate transcription in an efficient and specific manner. Under natural conditions without crowding agents, IRF3 forms liquid-like droplets with ISRE DNA both in vitro and in vivo, which function as hubs for IRF3 to compartmentalize its partner IRF7 and co-activator MED1, to activate the transcription of IFN-I and ISG genes. Thus, we have identified LLPS as a notable mechanism by which IRF3 multivalently accesses DNA and efficiently engages the transcriptional machinery to stimulate antiviral gene expression.

LLPS is emerging as a key mechanism for transcriptional regulation²⁵ but little is known about how phase-separated transcriptional condensates are controlled. In a screen searching for determinants of IRF3 nuclear puncta, we found that the deacetylase activity of SIRT1 is crucial. Mice deficient in Sirt1 exhibited abolished IRF3/IRF7 puncta formation and severely reduced antiviral immunity. In Sirt1-deficient cells, IRF3 and IRF7 were highly acetylated at two conserved lysine residues in their DBD, which in turn abrogated IRF3/IRF7 LLPS with ISRE DNA and impaired the transcriptional induction of IFN-I. Using a genetic code expansion orthogonal system, we incorporated acetyl lysine into IRF3/IRF7 DBD to create single and double site-specific fully acetylated proteins; compared to IRF3/IRF7 WT, those acetylated counterparts poorly bound to ISRE DNA and were much less efficient in forming LLPS. By developing the orthogonal system in vivo, we re-expressed the site-specific fully acetylated IRF3/IRF7 proteins in IRF3 KO cells. Single site-specific acetylation in DBD sufficiently blocked IRF3/IRF7 LLPS, IFN-I promoter binding and transcription, thereby promoting viral replication. Thus, the transcriptional activation by IRF3/IRF7 LLPS could be completely prohibited by addition of acetyl moiety/moieties to specific IRF3/IRF7 DBD residue(s) and these activities were restored after the acetyl moiety/moieties were removed by deacetylase SIRT1.

Age is an important risk of viral disease but the underlying molecular mechanisms remain elusive. In this study, we uncovered DBD deacetylation as a previously unknown prerequisite step for IRF3/IRF7 transactivation. Sirt1 is known to delay aging and extend lifespan in mice³⁸ and stimulating SIRT1 deacetylase rescues cell survival, mitochondrial function and premature aging^39,40,41. Based on the findings of the current study, the age-related decline in SIRT1 activity could enhance IRF3/IRF7 DBD acetylation, thus compromising innate antiviral immunity in aged individuals and serving as an important cause of innate immunosenescence. In line with this, we found that the cellular SIRT1 activity correlated well with the serum IFN-β level in patients with influenza and that SIRT1 activity was significantly reduced in macrophages from aged individuals. In this study, we also demonstrated that restoration of SIRT1 activity with agonists can rescue IRF3/IRF7 LLPS and IFN-I production and thus prevents innate immunosenescence.

In summary, we have disclosed the biological function and precise mechanism of IRF3/IRF7 LLPS in innate immune regulation and have described, at the molecular level, how aging is associated with declined innate antiviral immunity. By elucidating a new deacetylation-mediated control mechanism of IRF3/IRF7 LLPS, we have revealed a previously unknown interplay between SIRT1 activity and antiviral innate immune responses. These mechanistic studies might explain the diminished innate antiviral immunity frequently found in aged patients with viral infectious diseases. We envision that the findings described here will pave way for further translational studies toward improving the antiviral capabilities of the elderly and alleviate the current and future medical crises of an aging society.

Methods

Mice

Lyz-Cre⁺Sirt1^f/f (008041) mice from Shanghai Model Organisms Center were generated by CRISPR/Cas9-mediated gene editing. Sirt1^+/− mice were kindly provided by C.-X. Deng (National Institutes of Health). The Sirt1 gene was mutated by deleting exons 5 and 6, which encode a part of the catalytic domain. All mice including the aged (14–20 months old) were on a C57BL/6 background. Male and female mice were 8 weeks old and all experiments used littermate controls. Mice were maintained under specific-pathogen-free conditions in the animal facility of Soochow University. The animal room has a controlled temperature (18–23 °C), humidity (40–60%) and a 12-h light/12-h dark cycle. Experimental/control animals were bred separately. Animal experiments involving infectious virus were performed in the physical containment level 2 laboratory of Soochow University. The Institutional Committee for Animal Welfare of the Soochow University approved and oversaw this study. All animal studies were conducted in accordance with Soochow University guidelines.

Cells and reagents

HEK293T, HeLa, COS7, U2OS, A549 and Vero-E6 cells were obtained from American Type Culture Collection and RAW264.7 cells were kindly provided by Stem Cell Bank, Chinese Academy of Sciences. Sirt1^+/+ and Sirt1^−/− MEF cells derived from embryos at embryonic day 11 were grown in DMEM/Ham’s F-12 supplemented with 10% fetal bovine serum and antibiotic–antimycotic solution. Peritoneal macrophages were collected from mice 4 d after injection of thioglycolate (BD) and were cultured in DMEM supplemented with 10% FBS. BMDMs were isolated from the tibia and femur. Cells were cultured in DMEM with 10% FBS, glutamine and 30% L929 supernatant at 37 °C for 7 d. DCs were isolated from the tibia and femur. Cells were cultured in 1640 with 10% FBS, glutamine and 10 ng ml⁻¹ GM-CSF, 100 ng ml⁻¹ FLT3L at 37 °C for 7 d. DCs can be separated by flow cytometry using CD11c and PDCA1 markers. All cells were cultured at 37 °C under 5% CO₂ in DMEM supplemented with 10% FBS, 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin. Poly (I:C) and 5′-triphosphorylated RNA were bought from Invivogen. MG132 and EX527 were obtained from MCE. SRT501 and SRT2183 were purchased from TargetMol.

Real-time RT–PCR

Total RNA was prepared using the RNAiso Plus (Takara). A total of 1 μg of RNA was reverse-transcribed using the HiScript II Q RT SuperMix for qPCR (+gDNA wiper) (Vazyme). Real-time PCR was conducted with ChamQ Universal SYBR qPCR Master Mix (Vazyme) using a StepOne Plus real-time PCR system (Applied Bioscience). Quantitation of all target gene expression was normalized to the control genes 18sRNA for mouse genes or 18SRNA for human genes. qPCR primers were as follows:

Murine Ifnb1 forward: 5′-AAAGATCAACCTCACCTACAGGG: 5′,

Murine Ifnb1 reverse: 5′-CAACAATAGTCTCATTCCACCCAG-3′,

Murine Cxcl10 forward: 5′-GTCCTAATTGCCCTTGGTCTTCT-3′,

Murine Cxcl10 reverse: 5′-TCGCACCTCCACATAGCTTACAG-3′,

Murine Ccl5 forward: 5′-GCTTCCCTGTCATTGCTTGCTCT-3′,

Murine Ccl5 reverse: 5′-CTGATTTCTTGGGTTTGCTGTGC-3′,

Murine 18s forward: 5′-CGCGGTTCTATTTTGTTGGT-3′,

Murine 18s reverse: 5′-AGTCGGCATCGTTTATGGTC-3′,

Murine Sirt1 forward: 5′-TCAGTGTCATGGTTCCTTTGC-3′,

Murine Sirt1 reverse: 5′-TGGCTTCATGATGGCAAGTG-3′,

Murine Ifna forward: 5′-TGCATGGAATACAACCCTCCTA-3′,

Murine Ifna reverse: 5′-CAGACTTCTGCTCTGACCACCTC-3′,

VSV forward: 5′-ACGGCGTACTTCCAGATGG-3′,

VSV reverse: 5′-CTCGGTTCAAGATCCAGGT-3′,

Human IFNB1 forward: 5′-CCAACAAGTGTCTCCTCCAAAT-3′,

Human IFNB1 reverse: 5′-AATCTCCTCAGGGATGTCAAAGT-3′,

Human SIRT1 forward: 5′-TCCTGGACAATTCCAGCCAT-3′,

Human SIRT1 reverse: 5′-CCCGCAACCTGTTCCAGCGT-3′,

Human 18S forward: 5′-CGGCTACCACATCCAAGGAA-3′,

Human 18S reverse: 5′-GCTGGAATTACCGCGGCT-3′,

HSV-1 forward: 5′-GGTCGCCCTGTCGCCTTA-3′,

HSV-1 reverse: 5′-GGTCGCCATGTTTCCCGT-3′,

Human CXCL10 forward: 5′-TTTGCTGCCTTATCTTTCTGACT-3′,

Human CXCL10 reverse: 5′-ATTGTAGCAATGATCTCAACACG-3′,

Human ISG56 forward: 5′-GCTTTCAAATCCCTTCCGCTAT-3′,

Human ISG56 reverse: 5′-CTTGGCCCGTTCATAATTTTTTC-3′,

Human IFNA forward: 5′-TGAGACCCACAGCCTGGATAA-3′,

Human IFNA reverse: 5′-AACTGGTTGCCATCAAACTCCT-3′.

Immunoprecipitation and immunoblot analysis

Cells were lysed with 1 ml lysis buffer (20 mM Tris-HCl, pH 7.4, 2 mM EDTA, 25 mM NaF and 1% Triton X-100) containing protease inhibitors (Sigma) for 10 min at 4 °C. After centrifugation at 12 × 10³g for 15 min, protein concentrations were measured and equal amounts of lysates were used for IP. IP was performed with anti-FLAG M2 beads (Sigma, A2220) for 1 h at 4 °C or with various antibodies (identified below) and protein A-Sepharose (GE Healthcare Bio-Sciences AB) for 3 h at 4 °C. Thereafter, the precipitants were washed three times with washing buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate and 0.1% SDS) and the immunocomplexes were eluted with sample buffer containing 1% SDS for 5 min at 95 °C. The immunoprecipitated proteins were separated thereafter by SDS–PAGE.

IB analysis was performed with specific antibodies (identified below) and secondary anti-mouse or anti-rabbit antibodies conjugated to horseradish peroxidase (HRP) (identified below). Visualization was achieved with chemiluminescence.

Antibodies used for IP, IB analysis and immunofluorescence (IF) are listed in the Reporting Summary.

Immunofluorescence and confocal microscopy

HeLa, COS7, U2OS or MEF cells grown on glass coverslips were transfected with plasmids or treated with various stimuli as indicated in the figures. Cells were then washed with PBS and fixed with 4% paraformaldehyde in PBS for 20 min, permeabilized with 0.2% Triton X-100 and blocked with 3% bovine serum albumin. Then the cells were stained with the appropriate antibodies (as listed in the ‘Immunoprecipitation and immunoblot analyses’ section and indicated in the figures), followed by incubation with fluorescent dye-conjugated secondary antibodies. The nuclei were counterstained with DAPI (Sigma-Aldrich). The tissue paraffin sections were blocked by 3% BSA and 0.1% Tween-20 in PBS for 1 h after re-hydration and antigen retrieval and then the sections were stained with indicated antibodies (identified above) in 3% BSA, followed by incubation with fluorescent dye-conjugated secondary antibodies (identified above) in 3% BSA. The nuclei were counterstained with DAPI in 3% BSA. Images were recorded by Zeiss LSM880 confocal microscope system (ZEN 2.3 (blue edition)) using ×100 NA APO lens or Zeiss Axio imager D2 microscope system with ApoTome using a ×40 NA APO len and analyzed with Zeiss confocal software.

Purification of acetylated or phosphorylated recombinant proteins

After optimizing the constructs and methods, Escherichia coli strain, DH10B, was transformed with plasmids aminoacyl tRNA synthetase pSupAR-tRNA-chRS42 and pBAD/Myc-Strep C vector carrying the ORF of target protein with an amber codon at the desired site for lysine acetylation. The cells were first grown overnight in LB medium supplemented with 100 μg ml⁻¹ ampicillin and 50 mg ml⁻¹ chloramphenicol (Amp⁺Cm⁺) at 37 °C. A 10-ml volume of the overnight cultured bacteria culture was then subcultured in 200 ml LB Amp⁺Cm⁺. When the OD600 reached 0.6–0.7, 20 mM nicotinamide (NAM) and 10 mM acetyl lysine were added and 30 min later, when the OD600 reached 0.8–1.0, protein expression was induced at 30 °C for 18–20 h by adding 2 mM l-arabinose. Cells were collected after induction and washed with ice-cold PBS containing 20 mM NAM. The proteins were purified with Strep-tag affinity chromatography column according to the manufacturer’s protocol.

To obtain IRF3 proteins that were fully phosphorylated at positions Ser386 or Ser396, that is, IRF3 complementary DNA in which Ser386(TCC) or Ser396(TCC) is replaced with amber codon (UAG), we cloned IRF3 into the pQLink vector. pEVOL-SEP plasmids bearing SEP tRNA and SEP synthetase were created by Gibson assembly, in which SEP tRNA is driven by the E. coli proK promoter and SEP synthetase is driven by an arabinose inducible promoter. Plasmids were transformed into C321-competent cells. The cells were first grown overnight in LB medium supplemented with 100 μg ml⁻¹ ampicillin, 50 mg ml⁻¹ kanamycin and 50 mg ml⁻¹ chloramphenicol (Amp⁺Kan⁺Cm⁺) at 37 °C. A 10-ml volume of the overnight cultured bacteria culture was then subcultured in 200 ml LB Amp⁺Kan⁺Cm⁺ for further culturing. When the OD600 reached 0.6–0.7, unnatural amino acids (SEP) at a final concentration of 1 mM were added before induction. l-arabinose (0.02% final) and isopropyl β-d-1-thiogalactopyranoside (0.2 mM final) were then added to induce SEP synthase and IRF3 expression for 20 h at 16 °C. Subsequent purification steps were similar to those previously described for other proteins.

Intracellular expression of site-specific fully acetylated proteins

The eukaryotic orthogonal system used the ribozyme that is a chimera derived from two archaeal ribozymes to increase the insertion efficiency of acetyl lysine and four repeated tRNA sequences under the U6 promoter to strengthen the binding of tRNA to the amber codon. The pcDNA vector plasmid expressing IRF3/IRF7 protein (specifically mutated to TAG at K39/K77/K45/K92 sites) and chimeric ribozyme plasmid were transfected into cells at a ratio of 3 to 1 (similar to prokaryotic systems, specific ribozymes recognize amber codons and insert acetyl lysine). After 8 h, the medium was refreshed and 4 mM acetyl lysine was added. Cells after transfection were cultured for a total of 48 h. At 12 h before collection, 20 mM nicotinamide (NAM) was added to reduce degradation of acetylated proteins. Intracellular expressed acetylated proteins were confirmed with site-specific antibodies and for further analysis.

Mass spectrometry

SDS–PAGE gels were minimally stained with Coomassie brilliant blue, cut into six molecular weight ranges based on heavy chain IgG bands and digested with trypsin. Immunocomplexes were identified on a Thermo Fisher LTQ (majority) or Velos-Orbitrap mass spectrometer. Spectral data were then searched against the human protein RefSeq database in BioWorks or the Proteome Discoverer Suites using either SeQuest (for LTQ data) or Mascot (Orbitrap data) software. The IP/MS results were transferred into a FileMaker-based relational database generated in-house, where protein identification numbers (protein GIs) were converted to GeneID identifiers according to the NCBI ‘gene accession’.

Genetically encoded chemical crosslinking mass spectrometry

GECX–MS was performed on the intracellular IRF3 protein. In this method, an unnatural amino acid (Uaa) with proximity-enabled bioreactivity is genetically incorporated into IRF3 as a bait that reacts with the target residue (such as Cys) of the interacting proteins only when the two proteins co-condensate and place the Uaa in proximity to the target residue, thereby covalently crosslinking the interacting protein with IRF3 for mass spectrometry identification. In general, Uaa is genetically integrated into the Flag–IRF3 vector plasmid and transfected into 293T cells. Cells were cultured with chemical crosslinking Uaa BprY, which contains a proximity-enabled reactive group, alkyl bromide. Expressed Flag–IRF3 protein was then immunoprecipitated with a-Flag–M2 resin (Sigma) followed by elution with Flag peptide (Sigma, 1 mg ml⁻¹ in 50 mM HEPES, pH 7.5, 100 mM NaCl, 0.1% NP40 and 5% glycerol). After affinity purification of the target protein, the co-purified crosslinked protein is digested by protease and analyzed with mass spectrometry to reveal the protein identity.

Labeling proteins with fluorescent dyes

Purified proteins were labeled with Alexa Fluor 488 or Alexa Fluor 594 carboxylic acid (succinimidyl ester) according to the manufacturer’s protocols (Thermo Fisher Scientific).

In vitro phase separation assay

Purified proteins were diluted to various concentrations in buffer containing 20 mM Tris-HCl, pH 7.0 and 1 mM dithiothreitol with the indicated salt concentrations. Protein solution (5 μl) was loaded onto a glass slide, covered with a coverslip and imaged using an LSM880 confocal microscope system (Zeiss). Phase separation of recombinant eGFP–IRF3 or mCherry–IRF7 protein with Cy3-labeled DNA or FITC-labeled DNA was performed in physiological buffer (20 mM Tris-HCl, pH 7.0, 15 mM NaCl, 135 mM KCl, 5 mM phosphate, 1.5 mM MgCl₂ and 1 mg ml⁻¹ BSA). The mixture was transferred to 96-well plates (Corning) coated with 20 mg ml⁻¹ BSA (Sigma) and imaged. Double-stranded DNA oligonucleotides were generated by annealing sense and anti-sense single-stranded DNA oligonucleotides in annealing buffer (20 mM Tris-HCl, pH 7.5 and 50 mM NaCl) ramping down from 95 °C to 25 °C at 1 °C min⁻¹

Cy3 or FITC-labeled single-stranded DNA: 5′-ACATAGGAAAACTGAAAGGGAGAAGTGAAAGTG-3′

Cy3-5 × ISRE DNA: 5′-TAGTTTCACTTTCCCTAGTTTCACTTTCCCTAGTTTCACTTTCCCTAGTTTCACTTTCCCTAGTTTCACTTTCCC-3′

In vitro FRAP assays

FRAP experiments were performed on a Zeiss LSM880 confocal microscope system. For FRAP of proteins or DNA, spots of ~2-μm diameter in ~10-μm droplets were photobleached with 20% laser power for 1 s using 488-nm and 595-nm lasers. Time-lapse images were acquired within about a 2-min time course after bleaching with a 1-s interval. Fluorescence intensities of regions of interest (ROIs) were corrected by unbleached control regions and then normalized to pre-bleached intensities of the ROIs. The corrected and normalized data were fitted to the single exponential model using GraphPad Prism 8.

Cellular FRAP assays

Cellular FRAP experiments were performed on a Zeiss LSM880 confocal microscope system at 37 °C in a live-cell imaging chamber. HeLa cells were transfected with plasmids or treated with various stimuli as indicated in figures and then grown on chambered cover glass until they reached the desired density. Target proteins was fully or partially photobleached with 20% laser power for 2 s using a 488 nm laser. Time-lapse images were acquired over a 1-min time course after bleaching. Images were processed by ImageJ and FRAP data were fitted to a single exponential model by GraphPad Prism 8.

Chromatin immunoprecipitation assay

Cells grown to 100% confluence were crosslinked with 1% formaldehyde for 15 min at 37 °C. Crosslinking reactions were quenched with 0.125 M glycine for 5 min at room temperature. Cells were rinsed twice with PBS and collected with a silicon scraper. Cells were resuspended, lysed (1% SDS, 50 mM Tris-HCl, pH 8.0, 5 mM EDTA and protease inhibitors) and sonicated to obtain DNA fragments of ~300–500 bp in length on average. Samples were then centrifuged at 14 × 10³ r.p.m. for 10 min. The supernatant was diluted (20 mM Tris-HCl, pH 8.0, 2 mM EDTA, 1% Trion X-100, 150 mM NaCl and protease inhibitors) and pre-absorbed with 50 μl of protein A beads (Zymed). Supernatants were then incubated with 10 μg antibodies (as indicated) overnight at 4 °C. The immunocomplexes were collected with 100 μl of protein A beads in a 3-h co-incubation. Then, the mixture was washed sequentially with TSE I (0.1% SDS, 20 mM Tris-HCl, pH 8.0, 2 mM EDTA, 1% Triton X-100, 150 mM NaCl and protease inhibitors), TSE II (0.1% SDS, 20 mM Tris-HCl, pH 8.0, 2 mM EDTA, 1% Triton X-100, 500 mM NaCl and protease inhibitors), LiCl buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.25 mM LiCl, 0.1% NP40 and 1% deoxycholate sodium) and TE (10 mM Tris-HCl, pH 8.0 and 1 mM EDTA, pH 8.0). The bound immunocomplex was eluted by incubating with 400 μl of fresh elution buffer (25 mM Tris-HCl, pH 8.0, 10 mM EDTA and 0.5% SDS) at 65 °C for 15 min with occasional vortexing. The crosslinking was reversed in an overnight incubation at 65 °C. Whole-cell extract (WCE) DNA (fraction reserved from the sonication step) was also subjected to crosslinking reversal. Immunoprecipitated DNA and WCE DNA were then purified by treating with RNase A, proteinase K and multiple phenol:chloroform:isoamyl alcohol extractions. Immunoprecipitated DNA was analyzed with absolute qRT–PCR and the amplification product was expressed as a percentage of the input (WCE) for each condition. The primers used to amplify the IRF3/IRF7-binding region of the ISRE-IFNB1/IFNA promoters were the following: ISRE-IFNB1 promoter, forward: 5′-TCACAGTTTGTAAATCTTTTTCCCT-3′ and reverse: 5′-CAGTGTCGCCTACTACCTGT-3′,

ISRE-IFNA promoter, forward: 5′-TGAGACCCACAGCCTGGATAA-3′ and reverse: 5′-AACTGGTTGCCATCAAACTCCT-3′.

Primers used for the control region were the following:

forward: 5′-TTGAGATGGAGTCTTGCTCTG-3′,

reverse: 5′-GTGAAACCCTGTCCCTACTAAA-3′.

Deacetylation assays

Deacetylase assays were conducted in buffer A (100 mM NaCl, 50 mM Tris-Cl, pH 7.4, 5 mM MgCl₂ and 2 mM β-mercaptoethanol) supplemented with 1 mM NAD⁺ (AppliChem) at 23 °C. Then, 12 µM site-specifically lysine-acetylated IRF3 or IRF7 proteins were incubated with 0.14 µM Sirt1 for the indicated time. Samples taken at indicated time points were heated for 5 min at 95 °C to stop the reaction and then were separated by SDS–PAGE and IB.

Microscale thermophoresis

Binding affinities between target proteins and fluorescently labeled ligand proteins or ligand nucleic acids were measured in PBST binding buffer (PBS containing 0.05% Tween) using a MONOLITH NT.115 system (NanoTemper Technologies). Briefly, 10 µl of target protein was mixed with 10 µl of twofold serially diluted ligand at room temperature and measurements were repeated on independent protein preparations to ensure reproducibility. The data were analyzed by plotting peptide concentrations against liquid-induced fluorescence changes (change in raw fluorescence on y axis). Curve fitting was performed by using Prism 8 (GraphPad Software) and the given K_d values were calculated with 95% confidence levels.

Enzyme-linked immunosorbent assay

The concentrations of IFN-β in culture supernatants and serum were measured by ELISA kits (DY8234-05 and DY814-05, R&D Systems). The concentrations of IFN-α in serum was measured by ELISA kits (BMS6027TWO, Thermo Fisher Scientific).

Viral infection and plaque assay

Mouse macrophages or other cells (2 × 10⁵) were plated 24 h before infection. Cells were infected with VSV (0.1 MOI), HSV-1 (1 MOI) or SeV (100 hemagglutination units per ml) for various times, as indicated in the figures. VSV/HSV-1 plaque assay and VSV/HSV-1 replication were determined by a standard median tissue culture infectious dose assay on permissive Vero cell monolayers in 96-well plates with a series of tenfold-diluted samples. After 1 h of infection, the plates were incubated for 48 h. The medium was then removed and the cells were fixed with 4% paraformaldehyde for 15 min and stained with 1% crystal violet for 30 min before plaque counting.

Viral infection in vivo

For in vivo viral infection studies, 8-week-old control and mutant mice were infected with VSV (2 × 10⁹ p.f.u. per mouse) by intraperitoneal injection and HSV-1 2 × 10⁹ p.f.u. per mouse) by tail intravenous injection. At 12 h for VSV or 24 h for HSV-1 after infection, we collected the blood from the orbital sinus for ELISA and obtained the lungs, spleen and liver (brain was also needed for HSV-1) from each mouse for analysis of RNA, protein and viral titers. To measure the VSV titers in the lungs, spleen and liver and HSV-1 titers in the brain, snap-frozen tissues were weighed and homogenized three times (60 s each) in MEM. After homogenization, the suspensions were centrifuged at 1,620g for 30 min and the supernatants were used for plaque assays on monolayers of Vero cells seeded in 96-well plates with a series of tenfold-diluted samples. For survival experiments, mice were monitored for survival after infection with VSV or HSV-1.

Lung histology

Lungs from control or virus-infected mice were dissected, fixed in 10% phosphate-buffered formalin, embedded in paraffin, sectioned, stained with hematoxylin and eosin solution and examined by light microscopy for histological changes.

Statistical analysis

Data are shown as mean ± s.d. and statistical analyses were performed with a two-tailed unpaired Student’s t-test or as indicated in the figure legends. P valus of 0.05 or less were considered statistically significant. The exact value of n, representing the number of mice in the experiments, is indicated in the figure legends. For mouse survival studies, Kaplan–Meier survival curves were generated and analyzed for statistical significance with GraphPad Prism 7.0 and 8.0. Pilot studies were used for estimation of the sample size to ensure adequate power. There was no exclusion of data points or mice. No randomization or blinding was used.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The de-identified datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. Source data are provided with this paper.

References

Bowie, A. G. & Unterholzner, L. Viral evasion and subversion of pattern-recognition receptor signalling. Nat. Rev. Immunol. 8, 911–922 (2008).
Article CAS Google Scholar
Wu, J. & Chen, Z. J. Innate immune sensing and signaling of cytosolic nucleic acids. Annu. Rev. Immunol. 32, 461–488 (2014).
Article CAS Google Scholar
Seth, R. B., Sun, L., Ea, C. K. & Chen, Z. J. Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NF-κB and IRF 3. Cell 122, 669–682 (2005).
Article CAS Google Scholar
Ishikawa, H. & Barber, G. N. STING is an endoplasmic reticulum adaptor that facilitates innate immune signalling. Nature 455, 674–678 (2008).
Article CAS Google Scholar
Sato, M. et al. Distinct and essential roles of transcription factors IRF-3 and IRF-7 in response to viruses for IFN-α/β gene induction. Immunity 13, 539–548 (2000).
Article CAS Google Scholar
Honda, K. et al. IRF-7 is the master regulator of type-I interferon-dependent immune responses. Nature 434, 772–777 (2005).
Article CAS Google Scholar
Yan, N. & Chen, Z. J. Intrinsic antiviral immunity. Nat. Immunol. 13, 214–222 (2012).
Article CAS Google Scholar
McNab, F., Mayer-Barber, K., Sher, A., Wack, A. & O’Garra, A. Type I interferons in infectious disease. Nat. Rev. Immunol. 15, 87–103 (2015).
Article CAS Google Scholar
Park, A. & Iwasaki, A. Type I and type III interferons - induction, signaling, evasion, and application to combat COVID-19. Cell Host Microbe 27, 870–878 (2020).
Article CAS Google Scholar
Bartleson, J. M. et al. SARS-CoV-2, COVID-19 and the ageing immune system. Nat. Aging 1, 769–782 (2021).
Article Google Scholar
Solana, R. et al. Innate immunosenescence: effect of aging on cells and receptors of the innate immune system in humans. Semin. Immunol. 24, 331–341 (2012).
Article CAS Google Scholar
Piroth, L. et al. Comparison of the characteristics, morbidity, and mortality of COVID-19 and seasonal influenza: a nationwide, population-based retrospective cohort study. Lancet Respir. Med. 9, 251–259 (2021).
Article CAS Google Scholar
Ioannidis, J., Axfors, C. & Contopoulos-Ioannidis, D. G. Population-level COVID-19 mortality risk for non-elderly individuals overall and for non-elderly individuals without underlying diseases in pandemic epicenters. Environ. Res. 188, 109890 (2020).
Article CAS Google Scholar
Lin, R., Mamane, Y. & Hiscott, J. Structural and functional analysis of interferon regulatory factor 3: localization of the transactivation and autoinhibitory domains. Mol. Cell. Biol. 19, 2465–2474 (1999).
Article CAS Google Scholar
Qin, B. Y. et al. Crystal structure of IRF-3 reveals mechanism of autoinhibition and virus-induced phosphoactivation. Nat. Struct. Biol. 10, 913–921 (2003).
Article CAS Google Scholar
Takahasi, K. et al. X-ray crystal structure of IRF-3 and its functional implications. Nat. Struct. Biol. 10, 922–927 (2003).
Article CAS Google Scholar
Hyman, A. A., Weber, C. A. & Julicher, F. Liquid–liquid phase separation in biology. Annu. Rev. Cell Dev. Biol. 30, 39–58 (2014).
Article CAS Google Scholar
Boeynaems, S. et al. Protein phase separation: a new phase in cell biology. Trends Cell Biol. 28, 420–435 (2018).
Article CAS Google Scholar
Alberti, S., Gladfelter, A. & Mittag, T. Considerations and challenges in studying liquid–liquid phase separation and biomolecular condensates. Cell 176, 419–434 (2019).
Article CAS Google Scholar
Neumann, H., Peak-Chew, S. Y. & Chin, J. W. Genetically encoding N(epsilon)-acetyllysine in recombinant proteins. Nat. Chem. Biol. 4, 232–234 (2008).
Article CAS Google Scholar
Ryu, Y. & Schultz, P. G. Efficient incorporation of unnatural amino acids into proteins in Escherichia coli. Nat. Methods 3, 263–265 (2006).
Article CAS Google Scholar
Soderberg, O. et al. Characterizing proteins and their interactions in cells and tissues using the in situ proximity ligation assay. Methods 45, 227–232 (2008).
Article Google Scholar
Soderberg, O. et al. Direct observation of individual endogenous protein complexes in situ by proximity ligation. Nat. Methods 3, 995–1000 (2006).
Article Google Scholar
Shin, Y. et al. Spatiotemporal control of intracellular phase transitions using light-activated optoDroplets. Cell 168, 159–171 (2017).
Article CAS Google Scholar
Sabari, B.R. et al. Coactivator condensation at super-enhancers links phase separation and gene control. Science https://doi.org/10.1126/science.aar3958 (2018).
Shen, C. et al. Phase separation drives RNA virus-induced activation of the NLRP6 inflammasome. Cell 184, 5759–5774 (2021).
Article CAS Google Scholar
Shi, B. et al. UTX condensation underlies its tumour-suppressive activity. Nature 597, 726–731 (2021).
Article CAS Google Scholar
Yang, B. et al. Spontaneous and specific chemical cross-linking in live cells to capture and identify protein interactions. Nat. Commun. 8, 2240 (2017).
Article Google Scholar
Lu, Y. et al. Phase separation of TAZ compartmentalizes the transcription machinery to promote gene expression. Nat. Cell Biol. 22, 453–464 (2020).
Article CAS Google Scholar
Wang, R. H. et al. Impaired DNA damage response, genome instability, and tumorigenesis in SIRT1 mutant mice. Cancer Cell 14, 312–323 (2008).
Article CAS Google Scholar
Caillaud, A. et al. Acetylation of interferon regulatory factor-7 by p300/CREB-binding protein (CBP)-associated factor (PCAF) impairs its DNA binding. J. Biol. Chem. 277, 49417–49421 (2002).
Article CAS Google Scholar
Huai, W. et al. KAT8 selectively inhibits antiviral immunity by acetylating IRF3. J. Exp. Med. 216, 772–785 (2019).
Article CAS Google Scholar
Li, M. et al. Grass carp (Ctenopharyngodon idella) KAT8 inhibits IFN 1 response through acetylating IRF3/IRF7. Front. Immunol. 12, 808159 (2021).
Article CAS Google Scholar
Jang, M. et al. Cancer chemopreventive activity of resveratrol, a natural product derived from grapes. Science 275, 218–220 (1997).
Article CAS Google Scholar
Milne, J. C. et al. Small molecule activators of SIRT1 as therapeutics for the treatment of type 2 diabetes. Nature 450, 712–716 (2007).
Article CAS Google Scholar
Wang, B. et al. Liquid–liquid phase separation in human health and diseases. Signal Transduct. Target Ther. 6, 290 (2021).
Article CAS Google Scholar
Hnisz, D., Shrinivas, K., Young, R. A., Chakraborty, A. K. & Sharp, P. A. A phase separation model for transcriptional control. Cell 169, 13–23 (2017).
Article CAS Google Scholar
Satoh, A. et al. Sirt1 extends life span and delays aging in mice through the regulation of Nk2 homeobox 1 in the DMH and LH. Cell Metab. 18, 416–430 (2013).
Article CAS Google Scholar
Cohen, H. Y. et al. Calorie restriction promotes mammalian cell survival by inducing the SIRT1 deacetylase. Science 305, 390–392 (2004).
Article CAS Google Scholar
Lagouge, M. et al. Resveratrol improves mitochondrial function and protects against metabolic disease by activating SIRT1 and PGC-1α. Cell 127, 1109–1122 (2006).
Article CAS Google Scholar
Scheibye-Knudsen, M. et al. A high-fat diet and NAD(+) activate Sirt1 to rescue premature aging in Cockayne syndrome. Cell Metab. 20, 840–855 (2014).
Article CAS Google Scholar

Download references

Acknowledgements

The current work was supported by a special program from the Ministry of Science and Technology of China (2021YFA1101000), the Chinese National Natural Science Funds (U20A20393, U20A201376, 32125016, 31701234, 91753139, 31925013, 31671457, 31870902, 32070907, 32100699 and 31871405), the China National Postdoctoral Program for Innovative Talents (BX2021208), the China Postdoctoral Science Foundation (2021M692350), the Zhejiang Natural Science Fund (LD19C070001) and Jiangsu National Science Foundation (19KJA550003).

Author information

These authors contributed equally: Ziran Qin, Xiuwu Fang.

Authors and Affiliations

International Biomed-X Research Center, Second Affiliated Hospital of Zhejiang University School of Medicine, Zhejiang University, Hangzhou, China
Ziran Qin & Long Zhang
Institutes of Biology and Medical Science, Soochow University, Suzhou, China
Ziran Qin, Xiuwu Fang, Wenhuan Sun, Zhenyu Ma, Tong Dai, Shuai Wang & Fangfang Zhou
MOE Laboratory of Biosystems Homeostasis & Protection and Innovation Center for Cell Signaling Network, Life Sciences Institute, Zhejiang University, Hangzhou, China
Zhi Zong, Heng Ru, Huasong Lu, Bing Yang, Shixian Lin & Long Zhang
Faculty of Basic Medical Sciences, Chonqing Medical University, Chongqing, China
Huizhe Huang
Department of Pharmaceutical Chemistry and the Cardiovascular Research Institute, University of California, San Francisco, CA, USA
Bing Yang
Cancer Center, Zhejiang University, Hangzhou, China
Long Zhang

Authors

Ziran Qin
View author publications
You can also search for this author in PubMed Google Scholar
Xiuwu Fang
View author publications
You can also search for this author in PubMed Google Scholar
Wenhuan Sun
View author publications
You can also search for this author in PubMed Google Scholar
Zhenyu Ma
View author publications
You can also search for this author in PubMed Google Scholar
Tong Dai
View author publications
You can also search for this author in PubMed Google Scholar
Shuai Wang
View author publications
You can also search for this author in PubMed Google Scholar
Zhi Zong
View author publications
You can also search for this author in PubMed Google Scholar
Huizhe Huang
View author publications
You can also search for this author in PubMed Google Scholar
Heng Ru
View author publications
You can also search for this author in PubMed Google Scholar
Huasong Lu
View author publications
You can also search for this author in PubMed Google Scholar
Bing Yang
View author publications
You can also search for this author in PubMed Google Scholar
Shixian Lin
View author publications
You can also search for this author in PubMed Google Scholar
Fangfang Zhou
View author publications
You can also search for this author in PubMed Google Scholar
Long Zhang
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Z.Q., X.F., T.D., W.S., Z.M. and S.W. designed the experiments and analyzed the data. Z.Q., F.C. and W.S. performed the experiments. Z.Z. designed the cartoon for the working model. B.Y. performed the mass spectrometry analysis. H.H., H.L., X.H. and L.Z. provided valuable discussion. L.Z. and F.Z. wrote the manuscript.

Corresponding authors

Correspondence to Fangfang Zhou or Long Zhang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Source data

Supplementary information

Reporting Summary

Rights and permissions

Reprints and permissions

About this article

Cite this article

Qin, Z., Fang, X., Sun, W. et al. Deactylation by SIRT1 enables liquid–liquid phase separation of IRF3/IRF7 in innate antiviral immunity. Nat Immunol 23, 1193–1207 (2022). https://doi.org/10.1038/s41590-022-01269-0

Download citation

Received: 11 January 2022
Accepted: 13 June 2022
Published: 25 July 2022
Issue Date: August 2022
DOI: https://doi.org/10.1038/s41590-022-01269-0

This article is cited by

Crosstalk between protein post-translational modifications and phase separation
- Yang Liu
- Wenjuan Feng
- Bin Wu
Cell Communication and Signaling (2024)
The SIRT1/Nrf2 signaling pathway mediates the anti-pulmonary fibrosis effect of liquiritigenin
- Qingzhong Hua
- Lu Ren
Chinese Medicine (2024)
MAVS deSUMOylation by SENP1 inhibits its aggregation and antagonizes IRF3 activation
- Tong Dai
- Lei Zhang
- Fangfang Zhou
Nature Structural & Molecular Biology (2023)
Phase separation in cGAS-STING signaling
- Quanjin Li
- Pu Gao
Frontiers of Medicine (2023)
Phase separation in immune regulation and immune-related diseases
- Ning Huang
- Hao Dong
- Bin Shao
Journal of Molecular Medicine (2022)