Post-Translational Modifications of cGAS-STING: A Critical Switch for Immune Regulation

Innate immune mechanisms initiate immune responses via pattern-recognition receptors (PRRs). Cyclic GMP-AMP synthase (cGAS), a member of the PRRs, senses diverse pathogenic or endogenous DNA and activates innate immune signaling pathways, including the expression of stimulator of interferon genes (STING), type I interferon, and other inflammatory cytokines, which, in turn, instructs the adaptive immune response development. This groundbreaking discovery has rapidly advanced research on host defense, cancer biology, and autoimmune disorders. Since cGAS/STING has enormous potential in eliciting an innate immune response, understanding its functional regulation is critical. As the most widespread and efficient regulatory mode of the cGAS-STING pathway, post-translational modifications (PTMs), such as the covalent linkage of functional groups to amino acid chains, are generally considered a regulatory mechanism for protein destruction or renewal. In this review, we discuss cGAS-STING signaling transduction and its mechanism in related diseases and focus on the current different regulatory modalities of PTMs in the control of the cGAS-STING-triggered innate immune and inflammatory responses.


Introduction
Cell inherent recognition and defense systems against foreign genetic material encompass an ancient and fundamental feature of living systems. The innate immune response provides a critical first line of defense for host immunity via dynamic and complex interactions among its numerous cellular and molecular components. This gifted trait relies on germline-encoded pattern-recognition receptors (PRRs), which recognize microbial products, trigger signaling pathways, and produce soluble mediators, such as type I interferons (IFNs) and pro-inflammatory cytokines [1]. Type I IFN promotes cellular autonomous defense mechanisms in autocrine and paracrine manners by inducing the expression of IFN-stimulated genes, which in turn instructs and enhances the activation of the adaptive immune system [2]. Cytosolic DNA is a potent activator of the type I IFN response [3,4]. Cyclic GMP-AMP synthase (cGAS), an important cytosolic DNA sensor, binds DNA and activates stimulator of interferon genes (STING), which generates a signaling cascade, leading to the production of type I IFNs and other immune mediators [5,6] (Figure 1). Despite the multiple functions engaged by the cGAS-STING pathway in response to a Figure 1. The cGAS-STING-TBK1 signaling pathway. Various DNA derived from virus, dying tumor cells or nucleus and mitochondria binds to and activates the cytosolic DNA sensor cGAS. cGAS utilizes GTP and ATP to produce the second messenger cGAMP, which directly binds to the ER-localized adaptor protein STING. Stimuli other than cGAMP, CDNs derived from bacteria can also activate STING. Subsequently, STING alters into a more closed conformation and transfers to the Golgi via the ERGIC, where it binds and activates the kinase TBK1 and IKK, which phosphorylate IRF3 and IκBα, respectively. Then, Phosphorylated IRF3 and IκBα dimerize and enter the nucleus, initiating the transcription of Type I IFN, TNF and IL6. These cytokines play a pivotal role in host defense, inflammation and antitumor immunity. Abbreviations: cGAS, cyclic GMP-AMP synthase; STING, stimulator of interferon genes; TBK1, TANK-binding kinase 1; cGAMP, cyclic GMP-AMP; ER, endoplasmic reticulum; CDNs, cyclic dinucleotides; ERGIC, ER-Golgi intermediate compartment; IKK, IκB kinase; IRF3, interferon regulatory factor 3; TNF, tumour necrosis factor. cGAS utilizes GTP and ATP to produce the second messenger cGAMP, which directly binds to the ER-localized adaptor protein STING. Stimuli other than cGAMP, CDNs derived from bacteria can also activate STING. Subsequently, STING alters into a more closed conformation and transfers to the Golgi via the ERGIC, where it binds and activates the kinase TBK1 and IKK, which phosphorylate IRF3 and IκBα, respectively. Then, Phosphorylated IRF3 and IκBα dimerize and enter the nucleus, initiating the transcription of Type I IFN, TNF and IL6. These cytokines play a pivotal role in host defense, inflammation and antitumor immunity. Abbreviations: cGAS, cyclic GMP-AMP synthase; STING, stimulator of interferon genes; TBK1, TANK-binding kinase 1; cGAMP, cyclic GMP-AMP; ER, endoplasmic reticulum; CDNs, cyclic dinucleotides; ERGIC, ER-Golgi intermediate compartment; IKK, IκB kinase; IRF3, interferon regulatory factor 3; TNF, tumour necrosis factor.

DNA Viruses
HSV-1 VC: releases extracellular vesicles; STING interaction; stabilizes STING in HEp-2 cells; enhances cGAMP levels; STING-mediated non-canonical autophagy [27,[50][51][52][53] VE: dampens NF-κB activation; degrades cGAS mRNA; binds to STING and TBK1; restrains cGAS catalyze activity; the deubiquitinase (DUB) activity; mediates cGAS deamidation; inhibits cGAS-DNA phase separation; obstructs TBK1 dimerization; induces TBK1 proteasomal degradation; interacts with karyopherin [53][54][55][56][57][58][59][60][61][62] KSHV VC: detection by the cGAS [ VC: translation inhibition [143] Evidently, cGAS was initially characterized as a classical DNA-binding protein that plays a crucial defensive role in the response of type I IFNs to DNA viruses [68,78]. In addition to cGAS-DNA interactions, many studies have highlighted the importance of cGAS-STING in RNA virus biology and disease pathogenesis, although previous studies have focused on RNA sensors against RNA viruses. It is essential to mention that this potential role does not imply that cGAS can recognize viral RNA or directly activate RNA interactors. The detailed mechanism of cGAS in the detection of RNA viral genomes is not well understood. Recent studies have reported the RNA-binding activity of cGAS, which could largely modulate rather than activate it [152], and more evidence is needed to confirm this. Different RNA viruses may activate the cGAS-STING pathway through multiple pathways. For example, cytoplasmic mtDNA induces leakage of damaged mitochondria, fusion sensing of enveloped viruses [146,153], or self-DNA release into the cytoplasm because of damage to the nuclear envelope [148]. In addition, a recent study has shown that STING can restrict the replication of various RNA viruses at the translational level, which will be described later [143]. By rapidly initiating a type I IFN-dependent immune response and limiting viral replication, the cGAS-STING signaling axis has emerged as a pathway with the potential to treat multiple RNA viruses. To realize this potential, a more detailed elucidation of the actual involvement of cGAS and STING in RNA viral infections is urgently needed. Surprisingly, several RNA virus families have been shown to replicate productively through the non-classical functions of the cGAS-STING pathway, such as autophagic degradation, lipid recognition and T cell responses [36,68,115,154]. Given the evolutionary pressure on viruses, the emergence of a growing number of new schemes demonstrates the importance and complicated network manipulation of cGAS-STING signaling cascades during viral escape, presenting a cat-rat race for survival between viruses and their hosts.
Furthermore, the cGAS-STING pathway is widely involved in various bacterial infections. cGAS and STING detect bacterial genomic DNA and bacterial cyclic dinucleotides (CDNs) present in the cytoplasm, respectively, and thus affect the outcome of intracellular bacterial infection [155][156][157]. CDNs released by many bacteria, such as Listeria monocytogenes [158], Neisseria gonorrhoeae [159], Mycobacteria, Legionella, Shigella, Francisella, group B streptococcus, and Chlamydia [16], are potent stimulators of type I IFN induction as well and are sensed by IFI16 and cGAS to activate the STING pathway [160,161]. Notably, unlike the protective effects against viruses, type I IFN plays a protective or deleterious role in the host depending on the bacterial species and infection mode [157]. Thus, the cGAS-STING signaling pathway is more complex and diverse in bacterial infec- tions. Future studies will further clarify the regulatory rules of cGAS-STING in various bacterial infections.

Cancer
The extensive study of cGAS as a key immune regulation engine during viral invasion and bacterial infection has vastly increased our understanding of innate immunity. The cGAS-STING pathway challenges the traditional pathogen-specific structural model and may ensure a correct immune response through different regulatory mechanisms. Whether this new perspective can participate in the cancer-induced immune effect deserves further research, especially with the increase of tumor immunotherapy, which has become particularly important.
There is an innate advantage of the cancer-activated cGAS-STING pathway, for cancer cells share common characteristics, including genomic instability, mutations or deletions of oncogenes, oxidative stress, and exuberant metabolism [162,163]. In these stressed states, micronucleus and mtDNA are prone to leakage [164]. This unstable dsDNA subsequently initiates the cGAS-sensing cascade and hinders early neoplastic cell progression [165]. Exogenous stimuli such as exposure to ionizing radiation or treatment with chemotherapeutic agents can likewise induce DNA damage, which participates in the accumulation of cytoplasmic DNA in tumor cells and acts as a promoter to induce STING-dependent type I IFN [166,167]. Specifically, the activation of cGAMP can also directly trigger tumor immunity, resulting in robust tumor clearance [168]. Intratumoral injection of 2 ,3 -cGAMP significantly retarded tumor growth and reduced lung metastasis in B16F10 mouse models [169]. Furthermore, nuclear cGAS has been shown to promote lethal accumulation of DNA damage in the cancer genome by inhibiting homologous recombination (HR) DNA repair [170].
I. Enhanced adaptive immunity In the tumor microenvironment, the immune surveillance mechanism of cGAS-STING has been well characterized and is mediated by infiltrating immune cells, such as dendric cells (DCs) and T cells, through type I IFN signaling [171]. Researchers found that tumorderived DNA and cGAMP activate the cGAS-STING pathway in DCs via an unknown mechanism, which in turn performs cross-presentation to DCs and recruits CD8+ T cells for direct, non-spontaneous tumor elimination [170,172]. In addition, type I IFN drives the production of multiple chemokines, such as CXCL9 and CXCL10, which are required for cytotoxic T lymphocyte (CTL) metastasis and infiltration [173]. Accordingly, the regulatory functions of CD8+ T cell activation and differentiation provide significant anti-tumor effects in the STING pathway.
A recent study found that type I IFN in this pathway also interferes with immunosuppression of regulatory T (Treg) cells by downregulating phosphodiesterase 4 (PDE4) and upregulating cyclic adenosine monophosphate (CAMP) [174]. Moreover, in response to various changes and stimuli in the intracellular environment, the cGAS-STING signaling pathway promotes the polarization of tumor-associated macrophages toward the M1 phenotype, thereby secreting inflammatory factors and chemokines to recruit and activate T lymphocytes [175]. Xu et al. further verified through in vitro experiments that STING may regulate the phenotype and anti-tumor effects of macrophages through the IL6R-Jak-STAT-IL24 signaling pathway [176].
Activation of STING in non-malignant cells also has tumor-suppressive effects. Under appropriate conditions, STING signaling can indirectly inhibit melanoma growth by supporting the spontaneous response of natural killer (NK) cells to tumor cells [116,177]. Further studies have shown that the anti-tumor immune function of NK cells depends largely on STING activation in non-malignant cells, which is considered an important component of innate immunity and may serve as a target for immunotherapy of related diseases [177].

II. Senescence and Cell Death
Multiple studies have provided strong evidence that cGAS is an important molecular link between DNA damage, senescence-associated secretory phenotype (SASP) gene expression, and aging [178,179]. Tumor aging may denote a favorable outcome as an endogenous barrier to malignant transformation. cGAS recognizes cytosolic chromatin fragments in senescent cells and promotes the transcription of SASP via the STING pathway, which in turn prevents abnormal proliferation of damaged cells in autocrine and paracrine forms [180], particularly in precancerous stages [181]. Moreover, the differences in the application of the cGAS pathway-induced senescence in precancerous, early, and advanced cancers deserve further investigation.
Through transcriptional activation of apoptotic regulators and IRF3, the cGAS-STING signaling pathway initiates a cascade of responses that support anti-proliferative cell states, including necroptosis, apoptosis, and pyroptosis, which are involved in tumorigenesis and metastasis. The necrotrophic apoptotic capacity of STING is closely related to the induction of pro-apoptotic BH3-only proteins compared with those undergoing senescence [180,182]. Noxa and PUMA are involved in both in vivo and in vitro STINGmediated apoptotic processes [182,183]. Alternatively, in a more direct mode of action, phosphorylated IRF3 can combine with B-cell lymphoma 2 (Bcl-2) related X protein (BAX) and Bcl-2 homologous antagonist/killer (BAK) to form a complex that directly drives downstream non-transcription-dependent apoptosis induction [184,185]. Additionally, STING interacts with splenic tyrosine kinase (Syk) and regulates pyroptosis in colitis-associated colorectal cancer [186].
cGAS/STING also promoted more immediate anti-tumor effects. Cellular autophagy is the key effector activity that drives a robust autophagic cell death program and exerts synergistic anti-tumor effects. Interestingly, in evolutionary terms, autophagy is a primitive function of STING and may have preceded the role of STING in IFN signaling [187]. Chen et al. revealed that these two pathways were completely independent [27]. Despite this, an accurate understanding of the molecular mechanisms of the STING-autophagy-cell death axis is still lacking. The underlying pathways may be involved in the regulation of the ER stress response [188], mitochondrial autophagy [189], or calcium signaling [190].

Autoimmune Diseases
The versatility of the cGAS-STING pathway enables it to play a critical role in detecting and eliminating multiple threats. However, its potency is a double-edged sword, with under-activation leading to an inadequate response to threats, such as cancer or viruses, and its over-activation is associated with the development of inflammatory and autoimmune diseases, such as systemic lupus erythematosus (SLE) [191], Aicardi-Goutières syndrome (AGS), STING-associated vasculopathy with onset in infancy (SAVI) [192], and neurodegenerative diseases [193].
cGAS senses dsDNA regardless of its origin and is therefore unable to distinguish between endogenous and foreign DNA. Deoxyribonucleases (DNases) prevent endogenous DNA accumulation by degrading self-DNA in different compartments under normal physiological conditions. Unfortunately, loss-of-function mutations in DNA exonuclease three-prime repair exonuclease 1 (TREX1) interfere with DNA degradation in apoptotic cells and lead to excessive type I IFN signatures that ultimately develop into SLE or AGS [194,195]. Previous studies have shown that TREX1-deficient mice develop inflammatory myocarditis, lymphoid hyperplasia, vasculitis, and kidney disease [196], which can be fully rescued by knockout of the cGAS gene [197]. Similarly, DNase II is a lysosomal endonuclease responsible for the degradation of DNA from engulfed apoptotic cells. DNase II-deficient mice die intra-embryonically owing to the development of anemia and overproduction of type I IFNs [198]. Intriguingly, further deletion of STING completely rescued embryonic lethality and the autoimmune and inflammatory phenotypes [197]. Furthermore, gain-of-function mutations in STING have been proven to be pathological mediators of the auto-inflammatory diseases, such as SAVI [199,200]. Collectively, these human and mouse genetic studies provide conceptual clues for targeting the cGAS-STING pathway for the treatment of certain human inflammatory diseases.
Self-released DNA that leaks from the nucleus or mitochondria can also function as a cGAS ligand to activate the pathway and cause severe inflammatory responses. A previous study found that self-DNA from apoptotic cells activates cGAS, which is the leading cause of myocardial infarction-associated type I IFN production [201]. In addition to damaged nuclear DNA, cGAS also responds to other self-DNA that is mislocalized to the cytoplasm and drives self-inflammation. In Parkinson's disease, mtDNA dysregulation has been noted, which may also be implicated in cGAS binding and type I IFN induction [193]. Notably, intercellular transfer of cGAMP appears to exacerbate inflammation. A mouse model of alcohol-induced liver disease showed that the transfer of cGAMP as an extracellular immune transmitter increases inflammation and disease severity [202]. Given the complexity of the cGAS-cGAMP-STING signaling process, the role of cGAMP transfer in disease severity remains an area worthy of future exploration.

PTM Networks Regulating the cGAS-STING Pathway
The fine regulatory patterns of innate sensors and downstream signaling molecules are of vital importance in cell fate decisions [203]. As a critical regulatory event, PTMs play an important role in regulating protein activity, localization, stability, and proteinprotein interactions by inducing their covalent linkage to new functional groups, such as phosphates, methyl, and acetate [8]. The complex structure of cGAS/STING makes it a perfect platform for a multitude of covalent modifications, including phosphorylation, ubiquitination, acetylation, methylation, and SUMOylation ( Figure 2). These PTM fine regulatory networks generate a robust yet "tunable" cytokine response that contributes to sustaining immune homeostasis in distinct contexts.

Phosphorylation
Phosphorylation is the most extensively investigated PTM in innate immunity and is inversely regulated by protein kinases and phosphatases. Specifically, protein phosphorylation refers to the addition of phosphate groups to tyrosine, threonine, or serine under the catalysis of phosphorylating kinases. Studies have shown that complex enzyme phos-

Phosphorylation
Phosphorylation is the most extensively investigated PTM in innate immunity and is inversely regulated by protein kinases and phosphatases. Specifically, protein phosphorylation refers to the addition of phosphate groups to tyrosine, threonine, or serine under the catalysis of phosphorylating kinases. Studies have shown that complex enzyme phosphorylation networks are widely used in signaling pathways as the preferred modality for innate immune regulation in vivo [204].
In the cGAS-STING pathway, cGAS (de)phosphorylation has been shown to be essential for regulating cGAS subcellular localization or activity. AKT serine/threonine kinase 1 (AKT1), also known as protein kinase B (PKB), regulates many physiological processes including proliferation, metabolism, growth, and cell survival. Seo et al. demonstrated that cGAS is a substrate for the serine/threonine protein kinase AKT, which mediates phosphorylation at S291 (mouse) or S305 (human), and suppresses the enzymatic activity of cGAS, thereby leading to a negative immune response in the host. In contrast, the AKT inhibitor VIII and the S291A mutant of cGAS promoted DNA-induced IFN-β production and inhibited herpes simplex virus (HSV)-1 infection [205]. Unlike AKT, the mitotic kinase cyclin-dependent kinase 1 (CDK1)-cyclin B complex only phosphorylates the S291 site of mouse cGAS during intermittent mitosis, inhibiting cGAMP synthesis. In particular, when cells exit mitosis, this phosphorylation can be reversed by protein phosphatase 1 (PP1), thereby restoring their DNA-sensing capacity [206]. Another study showed that during mitosis, the N-terminus of cGAS is phosphorylated by mitotic kinases, including Aurora kinase B, which then block chromatin DNA sensing [207]. Oligomerization of chromatin-bound cGAS is inhibited, which is required for cGAS activation. Together, these mechanisms ensure that cGAS is inactive when bound to chromatin during mitosis, potentially contributing to the prevention of autoimmune responses. In murine experiments, Shu et al. reported that another protein serine/threonine phosphatase (PSP) family member, PPP6C, inhibited phosphorylation of mouse S420 (human S435), thereby preventing its binding to GTP and inhibiting cGAMP synthesis [208]. In addition to serine phosphorylation, cGAS is also influenced by tyrosine phosphorylation, which regulates its subcellular localization. B-lymphoid tyrosine kinase promotes cGAS phosphorylation at Y215 in resting cells to control the cytosolic retention [170].
Serine phosphorylation events are also essential for both STING and IRF3 activation. The phosphorylation of STING S358 allows TBK1 to bind STING. Then, TBK1 trans-phosphorylates S366 of the adjacent STING complex, allowing the recruitment, phosphorylation, and dimerization of IRF3 [29][30][31]. Conversely, Konnon et al. reported that phosphorylation of activated STING on identical residues by the autophagy-related serine/threonine protein kinases ULK1 and ULK2 increases STING degradation without IRF3 activation [209]. As can be seen, these apparently contradictory results require further investigation to reconcile the discrepancies. Consistent with this, protein phosphatase Mg2+/Mn2+-dependent 1A (PPM1A) has been reported to target and dephosphorylate STING at S358 [209,210]. This inverse inhibitory phosphorylation plays a positive role in attenuating and terminating the type I IFN signaling pathway in the later stages of DNA virus infection. Additionally, ribosomal protein S6 kinase 1 (S6K1) has also shown to be involved in STING phosphorylation. S6K1 interacts with TBK1 to form a ternary signaling complex that promotes signal transduction along the cGAS-STING-TBK1-IRF3 axis [211]. Subsequent experiments revealed that the phosphorylation of tyrosine residues also contributes to the regulation of STING. Tyrosine-protein phosphatase non-receptor type (PTPN) 1 and 2, which act as MITA/STING-related proteins, mediate MITA/STING Y245 dephosphorylation to facilitate STING degradation by the 20S proteasome and attenuate the intrinsic antiviral response [212].
The kinase activity of TBK1 is critical downstream of STING. Numerous serine (S)/threonine (T) sites on the TBK1 protein can be phosphorylated, acting as negative regulators of risk factor-triggered type I IFN induction. Dual-specificity tyrosine-(Y)phosphorylation-regulated kinase 2 (DYRK2) induces TBK1 S527 phosphorylation and K48-linked ubiquitination in a kinase activity-dependent manner [213]. Furthermore, PPM1B was identified as a TBK1 phosphatase that removes TBK1 autophosphorylation to eliminate downstream IRF3 activation and negatively regulate antiviral responses [214]. Several studies have reported that IRF3 is regulated via phosphorylation modifications. DNA-dependent protein kinase (DNA-PK) functions as a potent regulator of the IRF3 pathway and can phosphorylate IRF3 at Thr135, leading to its retention in the nucleus and delay in its proteolysis process [215]. In line with this, PTEN releases IRF3 and promotes its nuclear translocation by reversing phosphorylation of the IRF3 S97 site [216]. In contrast, protein phosphatase 2A (PP2A) mediates IRF3 dephosphorylation, resulting in the negative regulation of type I IFN production [217]. How phosphorylation and dephosphorylation of IRF3 are precisely balanced to ensure the appropriate type I IFN production remains to be fully resolved.

Ubiquitination
Ubiquitination is an essential PTM that involves covalent attachment of ubiquitin (Ub) molecules to a target protein. The ubiquitination system is dependent on a sequential enzymatic cascade of E1 (Ub-activating enzyme), E2 (Ub-conjugating enzyme), and E3 (Ub ligase). Furthermore, these cellular events can be counteracted by Ub hydrolases/deubiquitinases (DUBs) by removing Ub from the substrate protein. Taken together, the dynamic conversion between ubiquitination and deubiquitination plays a crucial role in the maintenance of cellular homeostasis [218].
Regulation of protein ubiquitination is most commonly dependent on E3 ligase activity. cGAS ubiquitination is also inversely modified by several E3 enzymes or DUBs, and plays critical roles in a wide variety of immunological processes, such as antigen processing. Among them, E3 Ub ligases, RNF185 and TRAF6 are responsible for the polyubiquitination of cGAS, while TRIM56 and TRIM41 monoubiquitinate cGAS. RNF185 specifically catalyzes the formation of the K27-linked polyubiquitin chain at K173 and K384 of cGAS, promoting its enzymatic activity [219]. TRAF6, another E3 ubiquitin ligase, also promotes the polyubiquitination of cGAS to upregulate cGAS-dependent type I IFN signaling [220]. In contrast to TRAF6, TRIM56 induced cGAS monoubiquitination at K335 of the CTD structural domain, leading to a significant increase in cGAS dimerization, DNA-binding activity, and cGAMP production. Congruently, one study suggested that macrophages from Trim56−/− mice produced significantly less IFN-β mRNA when infected with HSV-1 instead with influenza virus [149]. Similar to TRIM56, TRIM41, another TRIM family E3 ubiquitin ligase, is responsible for monoubiquitinating cGAS and upregulating cGAS activity as well as cGAMP synthesis [221]. In addition, cGAS K414 is constitutively K48-linked ubiquitinated and then degraded via the p62-mediated autophagy-lysosomal pathway rather than via the ubiquitin-proteasome system. In this study, the researchers also observed that TRIM14 promotes cGAS enzymatic activity upon viral infection, while recruiting the deubiquitinase USP14. USP14 cleaved the K48-linked polyubiquitin chain of cGAS, thus disrupting the cGAS-p62 interaction and preventing the autophagic degradation of cGAS [222]. Both the deubiquitinating enzymes, USP27x and USP29, interact with cGAS in a similar way and mediate the removal of the K48-linked polyubiquitinated chain to stabilize cGAS [223,224].
To positively regulate STING function, the E3 ubiquitin ligases TRIM56, TRIM32, RNF115, and MUL1 facilitate STING dimerization and its interaction with TBK1 by promoting the conjugation of K-63 linked chains, which are two critical steps in STING-mediated signaling. Specifically, TRIM32 and TRIM56 jointly target STING at K150 for K63-linked polyubiquitination, thereby enhancing the activation of downstream pathways [225,226]. The exact mechanism by which TRIM32 and TRIM56 coordinate STING in this process remains to be elucidated. Furthermore, the mitochondrial E3 ubiquitin protein ligase 1 (MUL1) and RNF115 are responsible for the K63-linked polyubiquitination of STING at K224 and K20, K224, and K289 sites, respectively, to boost signaling from TBK1 to IRF3 [227,228]. Intriguingly, MUL1-mediated ubiquitination of STING is required for STING-IRF3 activation, whereas STING-NF-κB signaling does not [227,228]. Contrary to this, to maintain cellular homeostasis and immune response, cytoplasmic DNA stimulation induces the expression of Myb-like, SWIRM, and MPN domains 1 (MYSM1) protein, which interacts with STING and cleaves STING K63-linked ubiquitination at K150 to limit STING over-activation [229]. Furthermore, the autocrine motility factor receptor (AMFR) interacts with STING and catalyzes K27-linked polyubiquitination at multiple sites (K137, K150, K224, and K236), thereby positively regulating STING-dependent signaling activation. Specifically, the interaction between AMFR and STING is enhanced by another protein, INSIG1, to promote downstream signaling [230]. However, a recent study indicated that the deubiquitinating enzyme USP13 could reverse this modification and prevent TBK1 from recruiting signaling complexes, thereby impairing cellular antiviral responses [231]. The same is true for another deubiquitinating enzyme, USP21 [232]. K63-linked polyubiquitination is involved in signal transduction, whereas K48-linked polyubiquitination is involved in proteasomal degradation. TRIM30α and TRIM29 induce termination of signaling activation by K48-linked polyubiquitination of STING at K275 and K370, respectively, and subsequent degradation [233,234]. A subsequent experiment showed that knockdown and deficiency of TRIM30α resulted in high levels of type I IFN and IL-6 production upon dsDNA stimulation, and mice with TRIM30α-/-were more resistant to HSV-1 infection than wild type mice [235]. Notably, TRIM30α is not present in humans [236]. Similarly, E3 ubiquitin ligases RNF5 and RNF90 were identified as STING-interacting molecules capable of K48-linked ubiquitination at K150 during viral infection, leading to degradation of its proteasome [237,238]. RNF26, in contrast, adds a K11-linked ubiquitin chain at the same lysine to induce STING ubiquitination, which prevents STING from RNF5-mediated premature STING degradation and enhances innate immune signaling [239]. Notably, inactive rhomboid protein 2 (iRhom2) can also recruit the deubiquitinating enzyme EIF3S5 to stabilize STING by removing the K48-linked polyubiquitin chain mediated by RNF5 [240]. Moreover, studies on iRhom2 have furthered our understanding of the intracellular trafficking in STING signal transduction. The important translocation-associated protein, TRAPβ, is activated by iRhom2, which may facilitate STING transport from the ER to the perinuclear membrane region [240]. However, HCMV tegument protein UL82 acts as a negative regulator and interferes with the STING pathway by disrupting the iRhom2mediated STING-TRAPβ translocation complex [89]. Recently, USP13 was also shown to function as a deconjugating enzyme for K27-linked and K33-linked polyubiquitination of STING, although the exact mechanism is not known yet [231]. Two other ubiquitin-specific peptidases, USP18 and USP20, were also found to prevent STING degradation following DNA viral infection [241]. Moreover, CYLD, a deubiquitinase, can stabilize STING protein function by deconjugating K48-linked polyubiquitination, contributing to the positive regulation of the natural immune response [242].
Given the importance of STING signaling for host antiviral responses, many viral proteins have evolved to manipulate STING deubiquitination to evade the immune surveillance system. Both HSV-1 VP1-2 and HTLV-1 Tax inhibit the interaction between TBK1 and STING by causing its deubiquitination [58,243]. Likewise, Epstein-Barr virus (EBV) regulates the expression of TRIM29, which further promotes K48-linked polyubiquitination, thereby interfering with STING-mediated antiviral responses [234]. Analogously, Hepatitis B virus (HBV) has been shown to utilize ubiquitination modifications to negatively regulate STING activity and evade the antiviral type I IFN response [85,244]. Furthermore, innate immune sensing mediated by the STING pathway is also critical for eliciting antitumor immune responses against a wide range of cancers. A recent study identified death-associated protein kinase 3 (DAPK3) as a pivotal regulatory complex that drives tumor-intrinsic immunity and immune surveillance through STING-mediated ubiquitination and phosphorylation [245]. In particular, the detailed mechanisms of PTM crosstalk need to be explored in the future.
Interestingly, in addition to STING, TBK1 is another important target for ubiquitination modification. Several E3 ubiquitin ligases, including Nrdp1, Mib1, Mib2, and RNF128, have been identified to induce K63-polyubiquitin conjugation of TBK1 residues under different stimuli to augment their activity [246]. Furthermore, some E3 ubiquitin ligases, including DTX4, TRIP, and TRIM27, specifically interact with TBK1 and trigger polyubiquitination of the K48 linkage at K251 and K372, causing the degradation of TBK1, thereby modulating its stability [247]. The multi-level modification of IRF3 has been a research hotspot in recent years, and the most common modification is ubiquitination involving K48. Moreover, E3 ubiquitin ligases RBCK1, RAUL, and Ro52 contribute to IRF3 proteasomal degradation and inhibit IRF3-dependent antiviral signaling by mediating the ubiquitination of IRF3 at K48 [248][249][250]. Through this modification, proline isomerase Pin1 has also been shown to affect innate signaling and antiviral defense [251].

SUMOylation and Neddylation
In addition to ubiquitination, another small ubiquitin-like modifier covalently linked to the target protein, small ubiquitin-like modifier (SUMO), has also been characterized. SUMOylation is regulated by specific SUMO E1, E2, and E3 cascades, and desulfurases. Similar to ubiquitination, SUMOylation and its reverse reaction, deSUMOylation, can respond rapidly to changing external and internal stimuli and orchestrate large biological immune effects through dynamic alterations in the biochemical properties of the substrate [252].
Members of the tripartite motif (TRIM) family of proteins are well-known for their roles as E3 ubiquitin ligases. Recently, several studies have demonstrated that TRIM molecules, in addition to triggering ubiquitination, are involved in the SUMOylation modification of cGAS-STING signaling. Early in the immune response, TRIM38 targets cGAS and STING to elicit a biochemical reaction of SUMOylation, thereby preventing K48-linked ubiquitination and degradation [253]. As a contrast, the sentrin-specific protease (SENP) family catalyzes the deSUMOylation process and regulates the dynamic equilibrium of the SUMO conjugation-deconjugation cycle during the later stages of viral infection. For instance, SENP2 removes SUMO modifications from cGAS and IRF3, thus avoiding excessive and potentially harmful immune responses [254,255]. In addition, these modifications were mapped to lysine residues 335, 372, and 382 of cGAS, impairing cGAS-DNA binding, oligomerization, and nucleotidyltransferase activity, ultimately weakening the innate antiviral response [253,256]. To maintain homeostasis of the immune environment, SENP7 relieves this inhibition by catalyzing de-SUMOization of cGAS [256]. Several studies have highlighted the critical role of TBK1-mediated SUMOization in antiviral responses. Mass spectrometry allowed the identification of TBK1 at K694 as a SUMO acceptor site, which could directly modulate the antiviral function of TBK1 by damping interactions with adaptor proteins [257]. Some viral proteins, such as the adenovirus early gene product Gam1, have been identified as substrates for SUMO modification, which can be exploited to intentionally manipulate the SUMO system in vivo or in vitro, thus affecting the antiviral functions [257]. Considering the contribution of TBK1 SUMOylation to its antiviral activity, it will be interesting to screen for more viral proteins interfering with its SUMOylation status in the future.
NEDD8, like ubiquitin and other ubiquitin-like proteins (UBLs), is added to the lysine residue of its target substrate by a three-step enzymatic cascade. Indeed, neddylation is also involved in the innate immune responses mediated by cGAS. Recent studies have proposed that the Ube2m-RNF111 axis, acting as an NEDD8 E2-conjugating enzyme and NEDD8 E3 ligase, is responsible for the neddylation of cGAS [258,259]. Consistently, mice deficient in Ube2m or RNF11 showed less resistance to HSV-1 infection [259]. Therefore, this positive PTM is vital for strengthening of cGAS dimerization and its dsDNA-binding ability. Nevertheless, the exact mechanism by which the neddylation pathway affects the functions and mechanisms of the cGAS-mediated DNA sensing immune pathway remains unclear.

Acetylation
Acetylation of histone lysine residues is a crucial epigenetic modification that affects histone structure and gene expression, and supports a number of cellular regulatory sys-tems, such as regulation of protein function, chromatin structure, and signal transduction. The acetylation of cGAS has recently attracted much attention. To date, researchers have identified several cGAS de/acetylation sites (K21, K47, K50, K56, K62, K63, K82, K83, K198,  K384, K392, K394, and K414) using mass spectrometry [260][261][262]. Among these sites, K47, K56, K62, and K83 of N-terminal structural domain of cGAS are targeted for acetylation by KAT5 to increase its DNA-binding ability and ultimately enhance cGAS immune signaling [260]. Furthermore, KAT2A belongs to another acetyltransferase family and directly exerts its effectiveness via regulating the acetylation level of cGAS, which promotes inflammation amplification in SLE [263]. Considering the complexity of PTM, the same PTM at different sites may result in different cGAS activity outcomes. For instance, acetylation of cGAS K384/K394/K414 may limit its activity by suppressing cGAMP synthesis, thus negatively affecting its role in immune signaling [262]. Moreover, acetylation of K384 and K414 hindered cGAS-dependent apoptosis [261]. Interestingly, under the stimulation of DNA, histone deacetylase 3 (HDAC-3) interacts with cGAS to catalyze its deacetylation and participates in cGAS activation [262,264]. Hence, these results suggest that both KATmediated acetylation and KDAC-mediated deacetylation activate cGAS and modulate the type I IFN response, albeit at different sites. Similarly, acetylation contributes to the regulation of cGAS activity. Another histone deacetylase, HDAC-9, directly deacetylates TBK1 and stimulates its kinase activity [265].

Methylation
Akin to acetylation, methylation is an important epigenetic marker of histone tails and is emerging as a potent regulator of innate immunity via epigenetic modulation of chromatin status and DNA transcription. Several studies have highlighted that some arginine and lysine residues in cGAS-STING can also be methylated. Han et al. reported the presence of cGAS methylation in HEK293T cells and its regulation by protein arginine methyltransferase 5 (PRMT5). Specifically, PRMT5 blocks cGAS-DNA binding capacity by inducing arginine symmetric dimethylation of cGAS at Arg124, and further attenuates cGAS-mediated antiviral innate immunity [266]. Thus, signaling initiation by arginine methylation allows the characterization of direct methylation modifications of the cGAS protein and defines PRMT5 as a negative regulator of cGAS-mediated type I IFN production during viral infection. Emerging evidence suggests that STING expression is also epigenetically suppressed by the histone H3K4 lysine demethylases KDM5B and KDM5C, which obstruct the signal transduction initiated by endogenous DNA and is mediated by the cGAS-STING-TBK1-IRF3 axis [81]. However, the specific regulatory mechanisms underlying such modifications remain to be elucidated.

Glutamylation and Deamidation
A distinct PTM, glutamylation, attaches glutamate side chains to glutamic acid residues in the primary sequence of the target proteins. Recently, glutamylation was identified as a critical regulator of cGAS function that has profound effects on DNA binding and activation [267]. For example, tubulin tyrosine ligase-like (TTLL) family members negatively regulate the inflammatory response triggered by cGAS through various mechanisms. TTLL6 primarily targets cGAS at E286 for polyglutamylation to impair its affinity for DNA, whereas TTLL4 promotes monoglutamylation of the E314 site to hinder its synthase activity [268]. On the contrary, the intracellular carboxypeptidases CCP5 and CCP6 can release the inhibitory effects of glutamylation and recover cGAS activity, suggesting that cGAS function is tightly controlled by cellular glutamylation and de-glutamylation [268]. Remarkably, in both cases, only a part of the cGAS protein was inhibited by glutamylation modification. Therefore, it is imperative to conduct further research to explain how this modification affects the functions of unmodified cGAS.
Deamidation has emerged as a feasible, effective, and attractive method of protein modification. Specifically, it refers to the process of deamidating the side chain of asparagine (Asn) or glutamine (Gln) residues to produce aspartic acid (Asp) or glutamic acid (Glu), respectively, under the catalysis of deamidase, which is involved in various biological functions [269]. Currently, the HSV UL37 protein is a known deamidase that deamidates the Mab21 enzymatic domain of cGAS. Specifically, it reduces the cGAMP-synthesizing activity of cGAS, thereby preventing double-stranded RNA (dsRNA)-induced innate immune activation [59]. Nevertheless, it remains unclear whether the deamidation of cGAS crosstalk with other PTMs.

Palmitoylation
Protein S-palmitoylation, a classical lipid modification process, describes the reversible conjugation of the fatty acid palmitate to the cysteine residues of target proteins. Emerging evidence indicates that palmitoylation is a key regulatory mechanism in the STING pathway [270]. In particular, palmitoylated STING is crucial for its assembly into multimeric complexes and for the recruitment of downstream signaling proteins in the Golgi apparatus [271]. The DHHC family comprises a group of proteins associated with palmitoylation. Activated STING translocates to the Golgi apparatus, where the DHHC family members DHHC3, DHHC7, and DHHC15 act as palmitoyltransferases, covalently modifying STING at C88 and C91 [272]. This palmitoylation reaction is effectively antagonized by the STING C88/91S mutant or the palmitoylation inhibitor 2-bromopalmitate (2-BP), which could be a potential feedback regulator that modulates STING signaling [271]. These results demonstrate that STING undergoes palmitoylation in the Golgi apparatus and that this PTM is indispensable for STING signaling.

Redox and Carbonylation
In addition to the unique nucleophilic nature of cysteine residues, the redox sites on their structures also confer different modifying functions, which realizes the fine regulation of many life activities. Recent studies have verified that oxidative PTMs involving cysteine are intimately linked to cGAS-STING signaling. Sustained cellular reactive oxygen species (ROS) during infection leads to the oxidation of STING at C64 and C148 and the formation of intermolecular disulfide bonds, which are sufficient to reduce the ability of STING to activate IFN-β expression [273,274]. In contrast, cytoglobin may protect STING from oxidation by scavenging ROS to promote IFN-β secretion in chronic liver disease [274].
Carbonylation, an irreversible oxidative protein modification, functions as a deleterious inhibitor of cGAS-STING signal transduction. For example, glutathione peroxidase 4 (GPX4) facilitates STING signaling by maintaining lipid redox homeostasis. Mechanistically, GPX4 inactivation during viral infection increased the cellular levels of the lipid peroxidation metabolite 4-HNE, which promoted STING to undergo carbonylation modification at C88, thus blocking subsequent translocation [275]. Additionally, as mentioned earlier, the Snitroalkylation of STING is also dependent on the C88 site. Therefore, competition between these PTMs at the same residue is crucial for the delicate regulation of STING activity.

Caspase-Mediated Cleavage
Caspases are an evolutionarily conserved family of cysteine-dependent proteases that are actively involved in the execution of cell death, inflammation, and innate immunity [276]. Recent studies have revealed that the caspase family can cooperate with the cGAS-STING signaling pathway to form complex crosstalk. Yet, most caspase family members target key junctions of the cGAS-STING signaling pathway for negative regulation, thereby suppressing innate immunity. For example, during inflammasome activation, caspase-1 leads to cGAS inactivation and reduces cGAMP levels by recruiting and cleaving cGAS at D140/157, ultimately repressing type I IFN production [277]. This experiment further revealed that cGAS may also be cleaved by caspase-4, 5, and 11 in a different manner than caspase-1 under non-classical inflammasome activation conditions [277]. Collectively, these data support the conclusion that all inflammatory caspases are capable of cleaving cGAS; however, the specific mechanism requires further investigation. Another study reported that caspase-3 is also involved in cGAS cleavage. Upon DNA or RNA virus infection, activated caspase-3 strongly cleaves cGAS, mitochondrial antiviral signaling (MAVS), and IRF3 in DNA or RNA virus-infected cells, effectively avoiding cytokine overproduction [278]. An intriguing finding was that caspase-7 is only involved in the inactivation of cGAS in mouse cells, indicating the existence of a speciesdependent immune mechanism. There is increasing evidence that mitochondrial apoptosis is associated with cGAS-STING signaling. As previously mentioned, mtDNA is released during apoptosis, triggering cGAS/STING-dependent type I IFN production. Instead, activated caspase-9 attenuates this response [279,280]. This conclusion is supported by the observation that the genetic deletion of caspase-9, caspase-3/7, or Apaf-1 leads to type I IFN production by dying cells [280]. Therefore, mutual coordination between the mitochondria-caspase-cGAS/STING signaling pathway plays a crucial role in regulating innate immune responses.

Protein-Protein Interaction
The chemical space and functional repertoire of proteins have been dramatically expanded by PTM of various amino acid residues [281]. In addition, the cGAS-STING signaling pathway is regulated by a variety of unknown mechanisms (Table 2), including protein-protein physical binding or blocking of the formation of activation complexes, which affect the enzymatic activity, stability, DNA binding, and translocation of cGAS-STING signaling. These studies highlight the critical role of disordered PTMs in the initiation, progression, and outcomes of various diseases. Based on these known and possibly additional unknown PTMs, the structures of cGAS/STING-related pathway proteins have become more complex, their functions enhanced, their regulation more refined, and their effects more pronounced.

Treatment with PTMs
In recent years, the cGAS-STING pathway has gained clinical importance owing to its great potential for type I IFN production and T cell priming. Several therapeutic strategies targeting the cGAS-STING pathway have been developed and tested using preclinical models [300][301][302]. In particular, the emerging understanding of the different biological mechanisms of PTMs targeting this pathway has also facilitated the development of novel therapeutic strategies. Nitrofuran derivatives, such as C-176, C-178, and H-151, have been identified to robustly antagonize STING palmitization by covalently binding to STING at Cys91 [271]. Moreover, unsaturated nitro-fatty acids (NO2-FAs) as antiinflammatory agents have the ability to engage STING and affect its palmitoylation [303]. More strikingly, aspirin, a non-steroidal anti-inflammatory drug, has been demonstrated to play a broader role by directly modulating cGAS acetylation and impairing cGAS-mediated immune responses. In particular, the effective dose of aspirin required to inhibit cGAS in this experiment fell far below the upper limit of its use in humans [262]. This evidence provides a rationale for targeting the PTM pathway to prevent STING-induced autoimmune diseases. Unfortunately, these antagonists demonstrated low experimental cell activity and augmented cGAS-STING signaling only in mice but not in humans, which may account for the inability to relieve autoimmune symptoms in patients in clinical trials [304]. Together, the development of new STING agonists necessitates the results of the interaction and activation of human STING before clinical trials.

Conclusions and Future Directions
Since the discovery of the cGAS-cGAMP-STING pathway, a series of biochemical, structural, and genetic studies has established the basic framework and mechanism of this DNA-sensing pathway, although cGAS and STING in some species have been inactivated by mutations and premature stop codons [305]. Nonetheless, much remains to be learned regarding the PTM regulatory mechanisms of this pathway and the versatility of different levels of PTM crosstalk in pathophysiological processes.
The discovery and characterization of the cGAS-STING pathway provides a novel framework for understanding the regulatory capacity of dsDNA immune recognition.
Under normal physiological conditions, cGAS can recognize and bind to the invading pathogen DNA, triggering an innate immune reaction. In contrast, aberrant activation of the cGAS-STING pathway by endogenous DNA may also result in autoimmune and inflammatory diseases. Furthermore, the activation of the STING protein as a bridge between innate and adaptive immunity can effectively inhibit tumor progression. Consequently, the regulation of the cGAS-STING pathway and the expression of type I IFN and related inflammatory factors are of great significance for alleviating autoimmune diseases and intervening in the progression of malignant tumors.
As promising as this field is, there are some issues that require additional attention or solutions. Currently, experimental studies on cGAS-STING-related PTMs have mainly focused on antiviral immune responses. However, a covalent STING palmitoylation inhibitor was recently identified to attenuate metastasis during late-stage cancer therapeutic interventions [271]. Therefore, gaining an in-depth understanding of the PTM regulatory network of the cGAS-STING signaling axis would be an exciting advance in the fields of cancer immunology and clinical therapy. Future resources need to be allocated to the investigation of PTM sites in tumor cells to accelerate the understanding of the complex interactions between PTM regulation and the tumor microenvironment, and to develop improved therapeutic options for patients with cancer.
In summary, the regulation of the cGAS-STING pathway by numerous PTMs is a highly relevant and productive area of research that will undoubtedly deliver exciting discoveries in the coming years.