Angiotensin-Converting Enzyme 2 Potentiates SARS-CoV-2 Infection by Antagonizing Type I Interferon Induction and Its Down-Stream Signaling Pathway

ABSTRACT The innate interferon (IFN) response constitutes the first line of host defense against viral infections. It has been shown that IFN-I/III treatment could effectively contain severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) replication in vitro. However, how SARS-CoV-2 survives through the innate antiviral mechanism remains to be explored. Our study uncovered that human angiotensin-converting enzyme 2 (ACE2), identified as a primary receptor for SARS-CoV-2 entry, can disturb the IFN-I signaling pathway during SARS-CoV-2 infection in human lung cells. We identified that ACE2 was significantly upregulated by SARS-CoV-2 and Sendai virus (SeV) infection, and exogenous expression of ACE2 suppressed IFN-I production in a dose-dependent manner. Mechanistically, ACE2 disrupted poly (I:C)-mediated inhibition of SARS-CoV2 replication by antagonizing IFN-I production by blocking IRF3 phosphorylation and nuclear translocation. Moreover, ACE2 quenched the IFN-mediated antiviral immune response by degrading endogenous STAT2 protein, inhibiting STAT2 phosphorylation and nuclear translocation. Interestingly, IFN-inducible short ACE2 (dACE2 or MIRb-ACE2) can also be induced by virus infection and inhibits the IFN signaling. Thus, our findings provide mechanistic insight into the distinctive role of ACE2 in promoting SARS-CoV-2 infection and enlighten us that the development of interventional strategies might be further optimized to interrupt ACE2-mediated suppression of IFN-I and its signaling pathway. IMPORTANCE Efficient antiviral immune responses against SARS-CoV-2 infection play a key role in controlling the coronavirus diseases 2019 (COVID-19) caused by this virus. Although SARS-CoV-2 has developed strategies to counteract the IFN-I signaling through the virus-derived proteins, our knowledge of how SARS-CoV-2 survives through the innate antiviral mechanism remains poor. We herein discovered the distinctive role of ACE2 as a restraining factor of the IFN-I signaling in facilitating SARS-CoV-2 infection in human lung cells. Both full-length ACE2 and truncated dACE2 can antagonize IFN-mediated antiviral response. These findings are key to understanding the counteraction between SARS-CoV-2 pathogenicity and the host antiviral defenses.

constitutes the first line of host defense against viral infections. It has been shown that IFN-I/III treatment could effectively contain SARS-CoV-2 replication in vitro (1) and in vivo (2). Therefore, it is rationalized that SARS-CoV-2 has developed strategies to counteract the IFN-I signaling. Compared with SARS-CoV and Middle East respiratory coronavirus (MERS-CoV), SARS-CoV-2 nsp1 and nsp6 proteins suppressed IFN-I signaling more efficiently (3). The SARS-CoV-2 nsp12 protein attenuates IFN-I production by inhibiting IRF3 nuclear translocation (4), nsp14 abolishes the induction of interferon-stimulated genes (ISGs) via a translational shutdown (5), nucleocapsid protein represses retinoic acid-inducible gene I (RIG-I)-mediated IFN-b production (6), and membrane glycoprotein M antagonizes the mitochondrial antiviral signaling protein (MAVS)-mediated innate antiviral response (7). SARS-CoV-2 is an enveloped, positive-sense, single-stranded RNA beta-coronavirus. The entry of SARS-CoV-2 into its target cells, such as lung alveolar or bronchial cells, is reported to depend on its receptor angiotensin-converting enzyme 2 (ACE2) (8,9). However, a recent report has demonstrated that SARS-CoV-2 could efficiently enter a human alveolar basal epithelial carcinoma cell line, A549 cells, which expressed a low level of ACE2, but its replication was abolished (10). In contrast, the exogenous constitutive expression of human ACE2 in wild type (WT) A549 cells confers SARS-CoV-2 to successful replication upon entering these cells (11). Interestingly, we observed that ACE2 expression was upregulated by SARS-CoV-2 (Fig. S1A) and Sendai virus (SeV) (Fig. S1B) infection in WT A549 cells, which is following the observation that the ACE2 expression was evidenced during SARS-CoV-2 infection of the respiratory epithelia in COVID-19 (12) (Fig. S1C, adapted from Nawijn and Timens [13]). From this perspective, we speculate that there may exist an ACE2-mediated mechanism antagonizing host antiviral signaling to facilitate SARS-CoV-2 replication.

RESULTS
To address the role of ACE2 in regulating host antiviral signaling, we first generated a stable cell line of A549 expressing the full-length of human ACE2 using a lentiviral vector designated ACE2-A549 (Fig. S1D). RT-qPCR and Western immunoblotting analyses revealed a significant elevation of ACE2 mRNA and ACE2 protein levels in ACE2-A549 cells compared to WT A549 cells (Fig. 1A). Consistent with previous studies, WT A549 cells were not susceptible to SARS-CoV-2. However, viral replication was significantly enhanced in ACE2-A549 cells at 48 h after authentic SARS-CoV-2 infection ( Fig. 1B and C). We next examined the viral attachment and entry efficacy of SARS-CoV-2 in WT and ACE2-A549. SARS-CoV-2 binding and endocytosis were observed in WT A549 cells, with a moderate increase in ACE2-expressing cells (Fig. 1D). A similar observation was also obtained in NCI-H1299 cells (Fig. S1E). The representative amplification curves of RT-qPCR were shown in Fig. S1F, revealing a quite minor difference in Ct value between WT and ACE2-expressing cells. These data indicate that the robust SARS-CoV-2 replication in ACE2-A549 cells is unlikely to be merely determined by ACE2-mediated endocytosis enhancement.
It has been demonstrated that a host factor, such as receptor tyrosine kinase, AXL, promotes RNA virus replication by antagonizing type I interferon signaling (14,15). Thus, we deciphered whether ACE2 can regulate the IFN-I signaling pathway. We first examined the innate cytokine responses during the early stage of SARS-CoV-2 infection in A549 cells. IFN-a2b, IFN-b, and IFN-l3 were significantly upregulated in WT A549 cells at 6 h after SARS-CoV-2 infection (Fig. 1E). Conversely, IFN-a2b and IFN-b were marginally detected in mock-infected or SARS-CoV-2 infected ACE2-A549 cells. In addition, we hardly observed a significant increase of IFN-l3 mRNA in response to SARS-CoV-2 infection compared with that in mock-infected ACE2-A549 cells (Fig. 1E). These observations suggest that ACE2 may be able to significantly regulate endogenous type I/III IFN responses during SARS-CoV-2 infection. A significant upregulation of interferon-induced gene expressions (ISGs) expression, such as MX1 and ISG15, was consistently identified in WT A549 cells but not in ACE2-A549 cells at 6 h after SARS-CoV-2 infection (Fig. 1F). Interestingly, following a previous study (16), we detected the In addition, we further tested those cells after being transfected or infected with poly (I:C) or SeV, respectively, both of which are often used to stimulate antiviral signaling pathways. Upon poly (I:C) transfection and SeV infection, the induction of IFN-a2b, IFN-b, and IFN-l3 mRNA (Fig. 1G) and protein secretion (Fig. 1H) was dramatically impaired in ACE2-A549 cells, as well as the production of ISGs, including MX1 and ISG15 expression (Fig. S1H). Of note, IL-1b and IL-6 were markedly induced in ACE2-A549 cells compared to WT A549 cells at the baseline and 10 h after poly (I:C) treatment (Fig. S1I). Genetic ablation of ACE2 enhanced the antiviral innate immune response without or with poly (I:C) treatment ( Fig. S1J and K). To investigate whether ACE2 also regulated the IFN-I response signaling, WT or ACE2-A549 cells were treated with recombinant human IFN-a for 4 and 10 h. We found that MX1 and ISG15 mRNA expression was significantly suppressed in ACE2-A549 cells compared to WT A549 cells (Fig. 1I). Thus, ACE2 may play a regulatory role in both the production and downstream responses of the IFN-I signaling pathway. To exclude the possible bias from a mixed population of cells and one cell line, we thereby established low passage single cellderived clones (clone#1, clone#5, and clone#6) and treated these three clonal cell lines with poly(I:C), SeV, or IFNa. RNA-Seq (clone#6) and RT-qPCR (clone#1, clone#5, and clone#6) analyses revealed that IFN induction ( We found that the inhibitory effects of ACE2 on IFN production and IFN responses were dose-dependent, and ACE2 already exerted a significant inhibitory effect on the IFN-I signaling at a concentration as low as 0.1 mg/mL ( Fig. 1L and M). Furthermore, we sorted two groups of A549 cells with low or high ACE2 protein from ACE2-expressing A549 cells (fresh . Cells with high ACE2 expression exhibited more apparent inhibitory effects on IFNb production than cells with minor ACE2 (Fig. 1N). These data suggest a significant correlation between the elevated amounts of ACE2 protein and the inhibitory effects on IFN-I signaling in both U-251 MG and A549 cells. Overall, these data explain clinical observations on the dramatically skewed responses from IFN to inflammatory responses (12,17,18). It has been shown that ACE2 is a human interferon-stimulated gene in the airway epithelial cells (19), which is reported to be a novel, transcriptionally independent truncated isoform of ACE2 (dACE2 or MIRb-ACE2) and could be induced by IFNs and viruses (20)(21)(22). We next tested if the IFN-induced truncated protein is also induced in our system and if it also shows inhibitory potential. Using custom-designed assays, we detected the expression of full-ACE2, dACE2, and total-ACE2 in A549 cells at baseline or after 10 h of stimulating with SeV, poly (I:C), and IFNa. Following previous studies (19)(20)(21)(22), dACE2 was significantly induced by IFNa and can also be upregulated by poly (I:C) and SeV stimulation (Fig. 1O). However, full-ACE2 was induced by poly (I:C) and SeV, but not IFNa stimulation (Fig. 1O), suggesting that full-ACE2 can be induced by an RNA virus directly. Alternatively, one plausible explanation is that poly (I:C) and SeV stimulation have the potential to induce IFN production and thereby stimulate dACE2 production rapidly. This chime with the data obtained from SARS-COV-2 infection, in which full-ACE2 was induced at the early stage (6 h) of infection (Fig. 1P). Ng et al. (22) also reported that dACE2 was induced at a late stage (24 h) of SARS-COV-2 infection in Calu-3 cells. These results sparked our interest in evaluating the role of dACE2 in the regulation of IFN signaling. Thus, we further exogenous expressed dACE2 in A549 cells and established a dACE2-A549 cell line, pLV-linker-GFP-A549 cells, as a negative control (NC) cells. Compared to full-ACE2-A549 cells, expressing dACE2 can also inhibit IFNb and MX1 production after SeV infection (Fig. 1Q) or IFNa stimulation (Fig. 1R). Altogether, these results confirm that both full-length and truncated dACE2 can regulate the IFN signaling during viral infection.
We then explored the underlying mechanism by which ACE2 is a host regulatory factor for type I IFN responses. The dsRNA, which is generated during coronavirus genome replication and transcription, could be recognized by RIG-I and/or melanoma differentiation gene 5 (MDA5) in the cytoplasm (23,24). Then IFN-a/b was activated in the axis of the MAVS-TBK1/IKKi-IRF3/7 signaling pathway, which prompted us to test whether ACE2 modulates TBK1 and IRF3 phosphorylation, which are two key steps in the IFN induction pathway. We found that overexpression of ACE2 significantly downregulated endogenous and exogenous IRF3 protein and suppressed IRF3 phosphorylation ( Fig. 2A and S2A) but did not affect IRF3 mRNA expression in ACE2-expressing A549 cells treated with poly (I:C) or SeV (Fig. S2B). NC-293T cells and ACE2-expressing 293T cells exhibited comparable IRF3 levels after IRF3 transfection (Fig. S2C). In contrast, endogenous TBK1, TBK1 phosphorylation, and the phosphorylation of JNK, P38, and ERK (Fig. S2D) were partially enhanced. Further immunofluorescence and Western immunoblotting analyses showed that in mock-treated A549 cells, IRF3 was distributed in the cytoplasm with or without ACE2 overexpressing (Fig. 2B, row 1). However, poly (I:C) and SeV-induced IRF3 nuclear translocation was severely impaired in cells overexpressing ACE2 (Fig. 2B, rows 2 and 3, and Fig. 2C). We then cotransfected IRF3-5D (a constitutively active IRF3 mutant) with IFN-b promoter and determined the activation of IFN-b promoter in NC-293T cells and ACE2-expressing 293T cells. Luciferase reporter assays showed that overexpression of ACE2 significantly inhibited IRF3-5D-triggered IFN-b promoter activation (Fig. S2E). These observations emphasized the critical role of ACE2 in targeting the downstream of the TBK1 signaling pathway and suppressing IFN production by reducing IRF3 protein and blocking IRF3 phosphorylation and nuclear translocation. Next, we investigated whether ACE2 could suppress poly (I:C)-mediated inhibition of SARS-CoV-2 replication. ACE2-A549 and Calu-3 cells were pretreated with 5 mg/mL of poly (I:C) for 10 h to trigger IFN induction, followed by SARS-CoV-2 infection for an additional 48 h. Viral replication was significantly reduced after poly (I:C) treatment in Calu-3 cells (Fig. 2D). By contrast, pretreating with poly (I:C) did not affect SARS-CoV-2 replication in ACE2-A549 cells (Fig. 2E). Collectively, the results demonstrate that ACE2 is likely to abort poly (I:C)-mediated inhibition of SARS-CoV-2 replication by antagonizing IFN induction.
To further investigate the underlying mechanisms by which ACE2 regulates the IFN response, we stimulated WT A549 and ACE2-A549 cells with IFN-a. WB and RT-qPCR analyses revealed that mRNA level (Fig. S2F), endogenous protein, and phosphorylation level of STAT2 (Fig. 2F) were significantly depressed in the indicated time points, whereas STAT1 phosphorylation was slightly reduced, and IRF9 was not inhibited (Fig. 2F). In addition, the nuclear translocation of STAT2 was completely blocked in ACE-A549 cells (Fig. 2G), indicating that ACE2 was likely to suppress IFN-a-triggered ISGs activation by degrading endogenous STAT2, blocking STAT2 phosphorylation and translocation. Consistent with this observation, exogenous expression of ACE2 aborted IFN-a-mediated inhibition of SARS-CoV-2 ( Fig. 2H and I) and Zika virus (Fig. 2J) infection. Of note, dACE2 could also antagonize the antiviral effects of IFNa (Fig. 2J). These data further confirmed that both full-ACE2 and dACE2 could play a critical role in antagonizing IFN-I-mediated antiviral responses. Interestingly, we noticed that ACE2 was highly expressed in the nucleus after stimulating with poly (I:C) (Fig. 2C), IFNa (Fig. 2G), or SeV (Fig. S2G). Moreover, RNA-Seq analysis revealed that ACE2 expression inhibited the transcription of plentiful host genes after poly (I:C) or IFNa stimulation compared to non-ACE2 expressing NC-cells ( Fig. 1J and K). Thus, it is speculated that ACE2 nuclear translocation is likely to participate in transcriptional regulation, and it will be important to address the mechanistic insights into the regulating effects in future studies.

DISCUSSION
Altogether, we uncovered that human ACE2, which was identified as a primary receptor for SARS-CoV-2 entry, can abort the IFN-I induction and the antiviral effects of the IFN-I response signaling pathway during SARS-CoV-2 infection (Fig. 2K). Moreover, both full-length and truncated ACE2 can suppress the IFN-I signaling. Our data provide mechanistic insight into the distinctive role of ACE2 in promoting SARS-CoV-2 infection in human lung cells and enlightens us that the development of interventional strategies might be further optimized to interrupt ACE2-mediated suppression on IFN-I and its signaling pathway. The detailed molecular basis of the interaction between ACE2 and IFN-related signaling pathways remains to be explored in future studies.

ACE2 Antagonizes IFN Signaling Pathway mSphere
Statistical analyses. Western blotting and immunofluorescence data were obtained from at least three repeated experiments. At least 3,000 fluorescent cells were imaged and counted with a flow-type tissue cytometer to quantify infected cells. More than three replicates were established per sample. The data were analyzed using Prism 7.0 software (GraphPad, USA) and are presented as the means 6 s.e.m. Statistical significance between the two groups was determined by an unpaired two-tailed Student's t test. Multiple group comparisons were performed using a two-way analysis of variance (ANOVA). Differences were considered to be significant for P , 0.05 (indicated with an asterisk [*]).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. FIG S1, TIF  J.X. and J.C. conceived and designed the experiments. J.C. and J.L. performed the experiments and analyzed the data. H.P. and D.B.F. prepared the authentic SARS-CoV-2 and performed the infection assay. Z.L.C., C.S.Z., P.Z., X.Z. and S.Z. contributed to data analysis and interpretation. J.C. and J.X. drafted the manuscript. All authors contributed to the editing of the manuscript.All data supporting the findings of this study are available within the article and SI Appendix. All correspondence and requests for materials should be addressed to J.X. (xujianqing@fudan.edu.cn).We declare that we have no competing financial interests.