Ubiquitin-dependent Turnover of Adenosine Deaminase Acting on RNA 1 (ADAR1) Is Required for Efficient Antiviral Activity of Type I Interferon*

Adenosine deaminase acting on RNA 1 (ADAR1) catalyzes RNA editing of cellular and viral RNAs. Besides RNA editing, ADAR1 has recently been shown to play important roles in maintaining the body balance, including tissue homoeostasis, organ development, and autoimmune regulations, by inhibiting both IFN production and subsequent IFN-activated pathways. Accordingly, the question was raised how IFN signaling induced by viral infections overcomes the inhibitory effect of constitutively expressed ADAR1 (ADAR1-P110) to execute efficient antiviral activity. Here we unexpectedly found that IFN signaling promoted Lys48-linked ubiquitination and degradation of ADAR1-P110. Furthermore, we identified the E3 ligase β transducin repeat-containing protein responsible for IFN-mediated ADAR1-P110 down-regulation. IFN signaling promoted the interaction between β transducin repeat-containing protein and ADAR1-P110 as well as protein turnover of ADAR1-P110. Moreover, we found that both lysine 574 and 576 are essential for ADAR1-P110 ubiquitination. Critically, we demonstrated that down-regulation of ADAR1-P110 is required for IFN signaling to execute efficient antiviral activity during viral infections. These findings renew the understanding of the mechanisms by which IFN signaling acts to achieve antiviral functions and may provide potential targets for IFN-based antiviral therapy.

The IFN family of cytokines plays critical roles in regulating antiviral innate immunity, cell proliferation, and immunomodulatory responses (1,2). So far, three classes of IFNs (type I, II, and III) have been identified. Among them, the type I IFN (IFN-I or IFN␣/␤) family has been extensively studied and well characterized. IFN-I family members trigger their signaling through activation of IFN-I receptors (IFNAR1 and IFNAR2), which is followed by tyrosine phosphorylation and activation of the Janus kinase family (JAK1 and Tyk2). Activated JAK1 and Tyk2 subsequently induce tyrosine phosphorylation of STAT1 and STAT2, which form a transcriptional complex. This complex translocates into the nucleus to induce the expression of interferon-stimulated genes (ISGs) 3 (3). Among these ISGs, the RNA-activated protein kinase (PKR) is well established as an important IFN-inducible antiviral component (4) that is a RNA virus sensor and inhibits viral replication by activating the ␣ subunit of the translation initiation factor (eIF-2␣) to block protein translation (3).
One of the most important functions of IFN-I is the protection of cells from viral infections (2,3). Based on this function, IFN-I has been used to treat some viral infection diseases, including chronic hepatitis C virus infection (5). However, recent studies have demonstrated that IFN-I signaling could play deleterious roles in some normal biological functions, especially when cells undergo excessive activated IFN-I signaling. It has been reported that uncontrolled activated IFN-I signaling contributes to autoimmune diseases, such as systemic lupus erythematosus, rheumatoid arthritis, and Aicardi-Goutières syndrome (6,7). IFN-I also has important effects on lethality in diverse models of sepsis or endotoxic shock (8). In addition, numerous studies have suggested that excessive IFN-I signaling could be involved in impairing normal organismal homoeostasis, including the maintenance of both fetal and adult hematopoietic stem cells (9), intestinal homoeostasis (10), and liver morphological and functional integrity (11).
Based on the complex biological activity of IFN-I, the maintenance of normal organismal functions requires balanced IFN-I signaling. Sufficient IFN-I signaling must be maintained in cells to produce the protective effects for antipathogen responses, and, at the same time, IFN-I signaling must be restricted to prevent harmful effects on tissue homoeostasis. Therefore, some intracellular mechanisms have evolved to limit excessive activated IFN-I signaling. For example, suppressor of cytokine signaling proteins 1 (SOCS1) inhibits cytokineinduced activation of the JAK family (12), protein kinase D2 (PKD2) promotes IFNAR1 ubiquitination and degradation to limit IFN-I signaling (13)(14)(15), and some of ISG-encoded proteins inhibit the recruitment of JAK to IFN-I receptors (16). Recently, adenosine deaminase acting on RNA 1 (ADAR1) has attracted much attention because of its important negative regulatory effect on IFN-I production and subsequent IFN-I-activated pathways (9,(17)(18)(19).
ADAR1 is a double-stranded RNA binding protein that catalyzes RNA editing (converting adenosine to inosine) of cellular and viral RNAs (20). There are two protein size forms of ADAR1 (P110 and P150). ADAR1-P110 is constitutively expressed in ubiquitous types of cells, whereas P150 is only expressed when cells are stimulated by some activators (20,21). The adenosine-to-inosine editing is of broad physiological function. Therefore, ADAR1 executes complex biological activities, including proviral or antiviral effects dependent on different virus types and the interaction between hosts and viruses (20). Despite the complexity, ADAR1 has been shown to promote the infection of numerous types of viruses, including VSV; measles virus; EBV; influenza virus; dengue virus; HIV, type 1; human respiratory syncytial virus; and human herpesviruses 8 (20,21). An increasing number of studies have demonstrated that ADAR1-mediated enhancement of viral infections is mainly through RNA editing or inhibition of IFN activity (including both IFN-I production and IFN-activated PKR antiviral activity) (21).
The inhibitory effect of constitutively expressed ADAR1 on IFN-I signaling maintains the IFN-I signaling balance and tissue homoeostasis. However, it is also of vital importance that the balance of IFN-I signaling should be able to be broken and regulated under some pathological conditions, such as viral infections. Virus-activated IFN-I signaling must act to overcome the ADAR1-mediated inhibitory effect to achieve efficient antiviral activity. However, the action mechanisms still remain unexplored.
Here we discovered that IFN-I signaling down-regulated the level of constitutively expressed ADAR1-P110, which is dependent on Lys 48 -linked polyubiquitination and degradation of ADAR1-P110 mediated by IFN-I signaling. Furthermore, we identified the E3 ligase ␤-TrCP, which interacts with ADAR1-P110 and is responsible for IFN-mediated ADAR1-P110 downregulation. Moreover, we found that both lysine 574 and lysine 576 are essential for ADAR1-P110 ubiquitination. Critically, we demonstrated that down-regulation of ADAR1-P110 is required for IFN signaling to achieve efficient antiviral activity during viral infections. Our findings reveal the action mechanisms by which IFN signaling achieves efficient antiviral functions and may provide potential targets for IFN-based antiviral therapy.

ADAR1-P110 Levels Are Down-regulated in IFN Signaling-
To determine the possible effect of IFN signaling on ADAR1-P110, we first analyzed the change of ADAR1-P110 protein lev-els in cells under conditions of IFN treatment. To this end, HEK293T cells were treated with IFN␣ for various times. Using a specific antibody against ADAR1-P110, we found that endogenous ADAR1-P110 levels were gradually down-regulated with IFN␣ (1000 IU/ml) treatment (Fig. 1A). ADAR1-P110 levels were reduced by around 80% when cells were treated with IFN␣ for 6 h, although its levels went back up slightly with 12-h IFN␣ treatment (Fig. 1A). To further confirm that IFN␣ signaling is capable of promoting the down-regulation of ADAR1-P110, FLAG-ADAR1-P110 was overexpressed in cells that were subsequently treated with IFN␣. The result showed that exogenous FLAG-ADAR1-P110 levels were also substantially down-regulated with IFN␣ treatment (Fig. 1, B and C). When cells were treated with increased amounts of IFN␣, we found that endogenous ADAR1-P110 was downregulated by IFN␣ treatment in a dose-dependent manner (Fig. 1D). Furthermore, IFN-mediated ADAR1-P110 downregulation was significantly inhibited by the proteasome inhibitor MG132 (Fig. 1E). An interesting report showed that ADAR1-P150 can be down-regulated in a caspase-dependent manner (30). To analyze the role of caspase in ADAR1-P110 down-regulation, we utilized a caspase inhibitor, Z-VAD, to block caspase activity. Our data showed that inhibition of caspase activity by Z-VAD did not block IFNinduced ADAR1-P110 down-regulation (Fig. 1F). These data strongly suggest that IFN␣ signaling negatively regulates cellular levels of ADAR1-P110 in a proteasome-dependent manner.
Two lysine residues of ubiquitin, Lys 48 and Lys 63 , have been extensively studied and are most commonly used for the analysis of protein polyubiquitination. We sought to determine whether IFN signaling affects the Lys 48 -linked or Lys 63 -linked ubiquitination of ADAR1-P110. We transfected cells with FLAG-ADAR1-P110 together with HA-tagged Ub-WT or HA-Ub-Lys 48 only (R48K) or HA-Ub-Lys 63 only (R63K). Then cells were treated with IFN␣. We found that IFN␣ treatment remarkably promoted the Lys 48 -linked ubiquitination of ADAR1-P110 ( Fig. 2D) but did not obviously affect IFN␣-induced Lys 63 -linked ubiquitination of ADAR1-P110 (Fig. 2D). Collectively, our data demonstrated that IFN signaling promotes Lys 48 -linked poly-ubiquitination of ADAR1-P110.
␤-TrCP Contributes to ADAR1-P110 Polyubiquitination in IFN Signaling-To find the ubiquitin ligase of ADAR1-P110, we first analyzed the amino acid sequences of ADAR1-P110. We noticed that there is a putative DSG(X) 2ϩn S motif in ADAR1-P110 conserved across various species, including Homo sapi-ens, Mus musculus, and Rattus norvegicus (Fig. 3A). The DSG(X) 2ϩn S motif has been shown to be an important characteristic of substrates for ␤-TrCP E3 ubiquitin ligase, and the serine mutation in this motif of a substrate results in its inability to interact with ␤-TrCP (22). We therefore speculated that ␤-TrCP is a potential ubiquitin ligase of ADAR1-P110. To provide evidence for this hypothesis, the Ser 578 of human ADAR1-P110 was mutated to alanine (SA). Furthermore, both Ser 578 and Ser 585 were mutated to alanine (AA). Cells transfected with FLAG-ADAR1-P110 (WT, SA, or AA) were treated with IFN␣ for 30 min, and then ADAR1-P110 ubiquitination levels were analyzed. We found that the ubiquitination of ADAR1-P110 mutants, especially the AA mutant, was significantly weaker than that of ADAR1-P110-WT (Fig. 3B). To further demonstrate the effect of ␤-TrCP on the specific motif of ADAR1-P110, cells were transfected with FLAG-ADAR1-P110-WT/ SA/AA together with HA-tagged ␤-TrCP (HA-␤-TrCP). As expected, ␤-TrCP induced much higher levels of ADAR1-P110-WT ubiquitination compared with the ubiquitination of ADAR1-P110 mutants (Fig. 3C). These results imply that the interaction between ␤-TrCP and ADAR1-P110 contributes to ADAR1-P110 ubiquitination.
IFN Signaling Promotes the Interaction between ␤-TrCP and ADAR1-P110 -To further understand the effect of ␤-TrCP on ADAR1-P110 ubiquitination, we analyzed the interaction FIGURE 1. IFN treatment induces both endogenous and exogenous ADAR1-P110 down-regulation. A, 293T cells were treated with IFN␣ (1000 IU/ml) for different times. Whole cell extracts were prepared and analyzed by immunoblotting (IB) using the indicated antibodies. ADAR1-P110 levels were quantified by densitometric analyses. B and C, 293T cells transfected with FLAG-ADAR1-P110 (FLAG-ADAR1) were treated with IFN␣ (1000 IU/ml) for different times. Whole cell extracts were subjected to immunoblotting using the indicated antibodies. D, 293T cells were treated with increased amounts of IFN␣ (500, 1000, and 3000 IU/ml) for 6 h. Whole cell extracts were subjected to immunoblotting using the indicated antibodies. E, 293T cells transfected with FLAG-ADAR1-P110 were pretreated with or without MG132 (50 M) for 2 h and then treated with IFN␣ (1000 IU/ml) for 2 h. Whole cell extracts were prepared and analyzed by immunoblotting using the indicated antibodies. F, cells transfected with FLAG-ADAR1-P110 were pretreated with or without Z-VAD (20 M) for 2 h and then treated with IFN␣ (1000 IU/ml) for 6 h. Levels of FLAG-ADAR1-P110 and caspase 3 were analyzed as indicated.
between ␤-TrCP and ADAR1-P110. Cells were co-transfected with FLAG-ADAR1-P110 and HA-␤-TrCP. The result confirmed the interaction between FLAG-ADAR1-P110 and HA-␤-TrCP (Fig. 4A). Also, FLAG-ADAR1-P110 can interact with endogenous ␤-TrCP (Fig. 4B). Moreover, we found that endogenous ␤-TrCP can interact with endogenous ADAR1-P110 (Fig. 4C). Given that the serine mutations in the DSG(X) 2ϩn S motif of protein substrates block the recognition of ␤-TrCP by its substrates, we detected the interaction between ␤-TrCP and ADAR1-P110-SA/AA mutants. The result was consistent with our hypothesis that the DSG(X) 2ϩn S motif of ADAR1-P110 is important for ␤-TrCP binding (Fig.  4D). Next we sought to determine whether IFN signaling upregulates ␤-TrCP levels. Cells were treated with IFN␣ for 2 or 6 h. The result showed that IFN signaling did not obviously affect ␤-TrCP levels within 6 h (Fig. 4E). However, we found that IFN treatment promoted the interaction between ␤-TrCP and ADAR1-P110 (Fig. 4F). All of these data suggest that ␤-TrCP can interact with ADAR1-P110 and that this interaction is promoted by IFN signaling.
␤-TrCP Promotes Protein Turnover of ADAR1-P110 -It has been clarified that Lys 48 -linked ubiquitination of a protein substrate usually leads to its degradation (23). Given that IFN signaling is able to promote the Lys 48 -linked ubiquitination of ADAR1-P110, we speculated that the protein stability of ADAR1-P110 could be regulated by IFN signaling. To address this hypothesis, we conducted cycloheximide (CHX) pulsechase analyses on ADAR1-P110 protein stability. Cells transfected with FLAG-ADAR1-P110 were treated with or without IFN␣ together with the protein synthesis inhibitor CHX. The results showed that IFN␣ treatment accelerated the degradation of ADAR1-P110 (Fig. 5A). Given that ␤-TrCP was able to promote ADAR1-P110 ubiquitination and degradation, we speculated that ␤-TrCP could affect cellular levels of ADAR1-P110. Cells were transfected with increased amounts of ␤-TrCP, and then the level of endogenous ADAR1-P110 was analyzed by immunoblotting. The result showed that overexpression of ␤-TrCP obviously decreased endogenous ADAR1-P110 levels (Fig. 5B). Next, ADAR1-P110 protein stability was analyzed by cycloheximide pulsechase assay. We found that knockdown of ␤-TrCP substantially blocked ADAR1-P110 protein degradation (Fig. 5C). Furthermore, the protein stability of ADAR1-P110-S578A (SA) or ADAR1-P110-S578A,S585A (AA) was analyzed. Compared with ADAR1-P110-WT, the degradation rate of ADAR1-P110-SA and -AA was much lower (Fig. 5D), suggesting that the binding of ␤-TrCP to ADAR1-P110 contributes to the degradation of ADAR1-P110. Taken together, we conclude that ␤-TrCP promotes protein turnover of ADAR1-P110.
Both Lysine 574 and Lysine 576 Are Essential for ADAR1-P110 Ubiquitination-Given that ␤-TrCP promoted the ubiquitination of ADAR1-P110, we next tried to identify the ubiquitin acceptor lysine residues in ADAR1-P110. Previous studies of the substrates of ␤-TrCP ubiquitin ligase showed that the lysine residues for ubiquitin conjugation mostly locate in positions 9 -13 upstream of this DSG(X) 2ϩn S motif in protein substrates (22). Accordingly, we generated the ADAR1-P110-K564R mutant, which harbors a lysine mutation upstream of the DSG(X) 2ϩn S motif of ADAR1-P110. Cells were transfected with either FLAG-ADAR1-P110-WT or -K564R. We found that the ubiquitination level of the ADAR1-P110-K564R mutant was similar as that of ADAR1-P110-WT (Fig. 6A). This result suggests that lysine 564 could not be the ubiquitin acceptor residue of ADAR1-P110.
Down-regulation of ADAR1-P110 Is Required for Efficient Antiviral Activity of Type I IFN-Given that type I IFN promoted the ubiquitination and protein turnover of ADAR1-P110, we sought to determine the pathophysiological significance of ADAR1-P110 down-regulation in IFN signaling. To assess IFN-mediated signaling and antiviral activity, we chose human fibroblast-derived cell line 2fTGH. 2fTGH cells are widely used for IFN studies because of their high sensitivity to IFN-I and excellent ability to produce IFN-I during virus infection (24). VSV is a very sensitive virus model that is usually used to assess IFN antiviral functions (25,26). Firstly, 2fTGH cells were infected with VSV for various times, and then ADAR1-P110 levels were analyzed by immunoblotting. We found that ADAR1-P110 levels were also gradually down-regulated during VSV infections (Fig. 7A), which was consistent with the observation that IFN␣ promoted ADAR1-P110 down-regulation (Fig. 1A). These data suggest that ADAR1-P110 down-regulation occurs during actual viral infections.
Next we analyzed the effect of ADAR1-P110 on IFN-mediated antiviral activity in our infection model, although a previous report has demonstrated that ADAR1 knockdown promoted the effect of IFN on the down-regulation of viral titers (27). Here we pretreated cells with IFN␣ and then infected them with VSV. The VSV-encoded protein VSV-G was analyzed by immunoblotting. As expected, IFN␣ treatment significantly decreased the content of VSV-G (Fig. 7B, lanes 1 and 2). We found that knockdown of endogenous ADAR1-P110 remarkably decreased VSV replication (Fig. 7B, lanes 1 and 3). This result suggests that ADAR1-P110 inhibits cellular antiviral ability in an basal IFN-secreting environment. Furthermore, when exogenous IFN was involved in the antiviral response, we observed that IFN␣-mediated anti-VSV ability was significantly enhanced by knockdown of ADAR1-P110 (Fig. 7B, lane 3 FIGURE 4. IFN signaling promotes the interaction between ␤-TrCP and ADAR1. A, 293T cells were transfected with FLAG-ADRA1-P110 and HA-␤-TrCP together with HA-PKD1. FLAG-ADRA1-P110 proteins were immunoprecipitated (IP) by a FLAG antibody, and then immunoblotting (IB) was performed using the indicated antibodies. WCL, whole cell lysate. B, 293T cells were transfected with FLAG-ADAR1-P110. Immunoprecipitation was performed by a FLAG antibody. Endogenous ␤-TrCP was detected by immunoblotting. C, 293T cells were harvested, and the whole cell lysates were used to analyze the interaction between endogenous ADAR1-P110 and endogenous ␤-TrCP. D, 293T cells were transfected with FLAG-ADAR1-P110 (WT, SA, or AA). FLAG-ADAR1-P110 proteins were immunoprecipitated by a FLAG antibody. ␤-TrCP was detected by immunoblotting. E, 293T cells were treated with IFN␣ (3000 IU/ml) for 0, 2, and 6 h. The level of ␤-TrCP was determined as indicated. F, 293T cells transfected with FLAG-ADRA1-P110 were treated with IFN␣ (1000 IU/ml) for 0, 15, and 30 min. FLAG-ADAR1-P110 was pulled down by M2 affinity gel, and endogenous ␤-TrCP was detected by immunoblotting. NOVEMBER 6). Taken together, our data suggest that endogenous ADAR-P110 inhibits IFN-mediated antiviral activity.

IFN Signaling Promotes ADAR1 Ubiquitination and Degradation
Based on the above results, we speculated that, during viral infections, IFN production and subsequent IFN signaling would encounter the inhibitory effect of ADAR1-P110. In conjunction with our previous data, which showed that IFN signaling actually down-regulated endogenous ADAR1-P110 levels through ubiquitin-dependent degradation, we hypothesized that IFN-mediated ADAR1-P110 down-regulation could be important for IFN antiviral activity. Given that we have demonstrated that the ADAR1-P110-S578A,S585A mutant (AA) impaired ADAR1-P110 ubiquitination and down-regulation (Fig. 3, B and C), we transfected cells with ADAR1-P110-AA, after which cells were infected with VSV-GFP, a VSV construct with the GFP gene. As expected, IFN treatment significantly inhibited VSV infection, as shown by a decreased VSV-GFP signal (Fig. 7C). Overexpression of ADAR1-P110-WT inhibited IFN-mediated antiviral activity (Fig. 7C). Importantly, the inhibitory effect of ADAR1-P110-AA on IFNs-mediated antiviral function was much more significant than that of ADAR1-P110-WT (Fig. 7C). Furthermore, cells were collected for the FACS analyses. The result confirmed that ADAR1-P110-AA had a stronger inhibitory effect on IFN antiviral activity than ADAR1-P110-WT (Fig. 7D, top panel), which can be explained by the enhanced stability of ADAR1-P110-AA under conditions of IFN treatment compared with ADAR1-P110-WT (Fig.  7D, bottom panel). Collectively, we demonstrated that inhibition of ADAR1-P110 down-regulation restricted IFN-mediated antiviral activity.
Then we re-expressed either resistant ADAR1-P110-WT (res-WT) or resistant ADAR1-P110-K574R,K576R (res-2KR) into shADAR1-stable cells. These two resistant expression constructs were generated by mutating the corresponding sites of ADAR1-P110 targeted by ADAR1-P110 shRNAs so that they are unable to be targeted and silenced by ADAR1-P110 shRNAs in shADAR1-stable cells. After that, cells were infected with VSV. As shown in Fig. 7F, re-expression of ADAR1-P110-WT (res-WT) promoted VSV infection in ADAR1 knockdown stable cells. Importantly, re-expression of ADAR-P110 -2KR (res-2KR) more significantly enhanced VSV infection compared with that of ADAR1-P110-WT (Fig. 7F). Consistently, when exogenous IFN was added to cells to inhibit viral infection, we found that the inhibitory effect of ADAR1-P110 -2KR on IFNmediated antiviral activity was much more significant than that of ADAR1-P110-WT (Fig. 7G), indicating that inhibition of ADAR1-P110 ubiquitination and degradation blocked normal antiviral activity mediated by IFN. Taken together, our data suggest that, during viral infection, down-regulation of ADAR1-P110 is required for IFN-I to execute efficient antiviral activity.

Discussion
The common understanding of IFN action is that IFN activity is achieved by activated JAK/STAT signaling pathways (31).
However, activated JAK/STAT signaling could be not enough for IFN signaling to exert efficient antiviral functions. IFN signaling first has to overcome the inhibitory effects mediated by some negative regulatory factors in cells such as ADAR1. ADAR1 has been shown to be a critical suppressor of IFN responses (including IFN production and IFN signaling), which protects cells from the harmful effects of excessively activated IFN signaling and, therefore, provides an important guarantee of tissue homoeostasis (9,10,17). Here we discovered that IFN-I signaling promoted ␤-TrCP-mediated Lys 48 -linked ubiquitination and degradation of constitutively expressed ADAR1 (ADAR1-P110). More importantly, we demonstrated that down-regulation of ADAR1-P110 is very important for efficient IFN-I activity during viral infections (Fig. 7, A-G). Thus, our studies renew the understanding of the mechanisms of IFN-I action and reveal that IFN-I signaling utilizes the E3 ligase ␤-TrCP to down-regulate constitutively expressed ADAR1-P110 to achieve efficient antiviral activity.
As mentioned above, ADAR1 has two forms: P110 and P150. Both ADAR1-P110 and -P150 have RNA-editing activity (20). In addition, ADAR1-P150 has been demonstrated to be an essential negative regulator of the mitochondrial antiviral signaling protein MAVS-mediated RNA-sensing pathway (17). Although ADAR1-P150 is IFN-inducible, the level of ADAR1- P150 during the early stages (a couple of hours) of IFN signaling is very low and hardly detectable. However, ADAR1-P110 levels were obviously down-regulated during several hours (about 2-8 h) of treatment with IFN-I (Fig. 1A). Given that the ADAR1-P150 form does not exist in cells at the initial stage of IFN-I treatment, we speculated that ADAR1-P110 is the pre-dominant form to suppress IFN-I signaling. Therefore, this study focused on constitutively expressed ADAR1-P110. Our data clearly showed that IFN-I signaling induced ADAR1-P110 ubiquitination and promoted protein turnover of ADAR1-P110 (Figs. 2, 3, and 5). It is also interesting but out of the scope of this study to explore the effect of IFN-I signaling on ADAR1-P150, which could be associated with the regulation of virus-sensing pathways. ADAR1 deficiency has recently been shown to be associated with aberrant organ development and severe autoimmune diseases. Mouse embryos with an editing-deficient knockin mutation of ADAR1 died at embryonic day 13.5 (18). Loss of ADAR1 results in increased IFN␣/␤ production and global up-regulation of IFN-inducible transcripts and therefore leads to rapid apoptosis of hematopoietic stem cells (9). Genetic mutations of ADAR1 result in the autoimmune disorder Aicardi-Goutières syndrome, which is associated with excessive activated IFN-I signaling (19). Most recently, it has been reported that ADAR1 isoforms control autoimmunity and multiorgan development (17). All of these studies demonstrated that ADAR1 is an important suppressor of IFN signaling and functions. This conclusion is consistent with many recent reports that showed that ADAR1 promoted the replication of many types of viruses. Here our data demonstrated that overexpression of ADAR1-P110 in cells significantly inhibited IFN-I-mediated antiviral activity (Fig. 7, C-E), and knockdown of ADAR1 promoted cellular antiviral ability mediated by IFN-I (Fig. 7B). Therefore, this study further clarified that constitutively expressed ADAR1-P110 inhibits IFN-I activity and promotes viral infection.
The ADAR1-mediated inhibitory effect on IFN function maintains the balance for organismic homoeostasis. When viral infections occur, the balance need to be broken by virus-induced activated IFN signaling. As shown in our data, IFN signaling rapidly down-regulated ADAR1-P110 levels to overcome the inhibitory effect of ADAR1-P110. An interesting observation is that ADAR1-P110 levels were slowly raised after 12-h treatment with IFN-I (Fig. 1A). It is not surprising. Given that balanced IFN-I signaling is important, lack of ADAR1-P110 in cells could be harmful to cells because of uncontrolled, excessively activated IFN-I signaling. Another interesting result showed that IFN-I signaling promoted the binding of ␤-TrCP to ADAR1-P110 (Fig. 4F). Similarly, a previous study reported that IFN-I signaling promoted the binding of ␤-TrCP to IFNAR1, which resulted in IFNAR1 ubiquitination and degradation (28,29). Therefore, we speculated that IFN-I signaling is able to promote the ␤-TrCP action, although the detailed mechanisms remain unknown. It could be interesting to explore ␤-TrCP action mechanisms in the future. Here the interaction between ␤-TrCP and ADAR1-P110 promotes IFN signaling, whereas the interaction between ␤-TrCP and IFNAR1 restricts the IFN-mediated signaling pathway. This opposite effect of ␤-TrCP on IFN responses could maintain another balance to hold IFN-I signaling in check. However, more evidence is needed to completely address this "balance" hypothesis.
Given that IFN-I signaling leads to ADAR1-P110 down-regulation and that viral infections can induce IFN-I production, we presumed that viral infections can also down-regulate ADAR1-P110. Our data clearly showed that ADAR1-P110 levels were significantly down-regulated during VSV infection (Fig. 7A). Based on this result, we wanted to know what happens to IFN-I antiviral activity when cellular ADAR1-P110 is unable to be ubiquitinated and down-regulated. Thus, we next sought to determine the pathophysiological significance of ADAR1-P110 down-regulation during viral infections. In this study, we have clarified the mechanisms of ADAR1-P110 down-regulation and identified the key serine and lysine residues (including Ser 578 and Ser 585 as well as Lys 574 and Lys 576 ) for ADAR1-P110 ubiquitination and degradation (Figs. 3 and  6). Therefore, we mutated these key residues (AA and 2KR) to limit ubiquitin-dependent down-regulation of ADAR1-P110. Our data demonstrated that ADAR1-P110-AA significantly inhibited IFN-I-mediated antiviral activity (Fig. 7, C and D). Furthermore, we established ADAR1-P110 knockdown cells and then re-expressed either ADAR1-P110-WT or the ADAR1-P110 -2KR mutant. This result showed that re-expression of ADAR-P110 -2KR more significantly inhibited IFN-I signaling and therefore promoted viral infections compared with ADAR1-P110-WT (Fig. 7, F and G). Thus, using this well established model of ADAR1-P110 mutants, we clarified that, during viral infections, inhibition of ADAR1-P110 ubiquitindependent down-regulation significantly restricts IFN-I antiviral efficiency.
This study uncovered the detailed mechanism by which our body breaks the balance of IFN-I signaling to establish efficient antiviral responses by ubiquitin-dependent down-regulation of constitutively expressed ADAR1 during viral infections (Fig. 8). These findings promote a better understanding of the action mechanisms of IFN-I (or other cytokine) activation and therefore may provide potential therapeutic targets for IFN-related diseases.
Cycloheximide Chase Assay-The half-life of FLAG-ADAR1-P110 was determined by CHX pulse-chase assay. Cells were transfected with FLAG-ADAR1-P110. 48 h after transfection, cells were treated with DMSO or CHX (50 g/ml) together with or without IFN␣ for the indicated times. Cells were then collected, and the equal amounts of boiled lysates were analyzed by Western blotting. Each experiment was carried out in triplicate.
RNA Isolation and Real-time PCR-Total RNAs were isolated from cells using TRIzol reagent (Invitrogen). cDNA was synthesized by reverse transcription using oligo(dT) and then subjected to quantitative real-time PCR with VSV RNA primers and ␤-actin primers in the presence of SYBR Green Supermix (Bio-Rad). The primer sequences were as following: VSV (5Ј-ACGGCGTACTTCCAGATGG-3Ј and 5Ј-CTCGGTTCAA-GATCCAGGT-3Ј) and ␤-actin (5Ј-ACCAACTGGGACGA-CATGG AGAAA-3Ј and 5Ј-ATAGCA CAGCCTGGATA-GCAACG-3Ј). The relative expression of the target genes was normalized to ␤-actin mRNA. The results were analyzed from three independent experiments and shown as the average mean Ϯ S.D.
Flow Cytometry Analysis and Immunofluorescence Microscopy-Cells infected with VSV-GFP (MOI ϭ 0.5) were subjected to analyses by either immunofluorescence microscopy or flow cytometry. Briefly, for immunofluorescence microscopy analysis, VSV-GFP viruses were pictured with an upright fluorescence microscope. Magnification was ϫ200. For flow cytometry analysis, cells were collected with cold 1ϫ PBS and acquired in a FACSCalibur (BD Biosciences) equipped with a 488-nm argon laser and a 635-nm red diode laser. Data were analyzed by FlowJo software (FlowJo, Ashland, OR).
Virus and Viral Infection-VSV was a gift from Dr. Chen Wang (Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences). VSV-GFP was a gift from Dr. Chunsheng Dong (Soochow University). The antiviral effect of IFN␣