Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has emerged as a global public health threat, causing over 700 million coronavirus disease 2019 (COVID-19) cases and 7 million deaths (Carabelli et al., 2023; Markov et al., 2023). SARS-CoV-2 encodes four structural proteins, 16 non-structural proteins (Nsps) that assist in viral replication and transcription, and a series of accessory proteins (ORFs) associated with immune evasion (V’kovski et al., 2021; Wang et al., 2020). Some viruses such as Ebola virus (Valle et al., 2021), dengue virus (Jung et al., 2018), and reovirus (Furuichi et al., 1975) encode 2’-O methyltransferases (2’-O-MTases) homologous to those found in eukaryotes. This allows them to mimic the host 5’ cap structure on their RNA, hindering recognition by the innate immune system (Daffis et al., 2010; Züst et al., 2011). Similarly, SARS-CoV-2 nsp16, in complex with nsp10, functions as a 2’-O-MTase, methylating the first nucleotide of a capped viral RNA strand at the 2′-O position, converting the 5’ cap of RNA from ‘cap-0’ to ‘cap-1’ (Benoni et al., 2021; Park et al., 2022). This modification shields SARS-CoV-2 from host antiviral responses such as MDA5 recognition and IFIT1 restriction (Bergant et al., 2022; Russ et al., 2022). Importantly, nsp16 inhibits global host mRNA splicing, which reduces host protein and mRNA levels (Banerjee et al., 2020). Additionally, nsp16 enhances SARS-CoV-2 cell entry by promoting TMPRSS2 expression (Han et al., 2023), highlighting its multifaceted role as a virulence factor. Moreover, nsp16-deficient strains exhibit reduced pathogenicity and induce a strong immune response, suggesting their potential as a live-attenuated vaccine candidate (Ye et al., 2022). Given nsp16’s critical functions, drugs targeting nsp16 or the nsp16-nsp10 complex have been developed (Klima et al., 2022; Nguyen et al., 2022). Therefore, understanding host factors that interact with nsp16 is crucial for uncovering novel therapeutic strategies.

The ubiquitin-proteasome system (UPS) is an important pathway for the targeted degradation of intracellular proteins and is referred to as “the molecular kiss of death” (Dongdem & Wezena, 2021; Park et al., 2020). Substrate ubiquitination requires a cascade of three ligases: E1, E2, and E3 (Zheng & Shabek, 2017). Some E3 ubiquitin ligases function as monomers, while others form multi-subunit E3 complexes. Regardless of their form, all E3 ligases share the same functional specificity in recognizing substrates and determining the specificity of the UPS. In some complexes, an additional protein may be required as a substrate receptor to recognize the substrate (Iconomou & Saunders, 2016; Yang et al., 2021). E3 ligases are indispensable for direct binding to substrates and determining the specificity of the UPS (Li et al., 2021; Wang et al., 2022). Researchers have identified over 600 human genome-encoded E3 ligases that operate in diverse cellular environments, respond to a wide range of cellular signals, and regulate a vast array of protein substrates (Dongdem & Wezena, 2021; Garcia-Barcena et al., 2020). Dysregulation of E3 ligase activity has been implicated in various human diseases, making them promising drug targets (Humphreys et al., 2021). E3 ligases are classified into three classes based on conserved domains and ubiquitin (Ub) transfer mechanisms: Really Interesting New Gene (RING) E3s, homologous to the E6AP carboxyl terminus (HECT) E3s, and RING-between-RING (RBR) E3s (Garcia-Barcena et al., 2020; Morreale & Walden, 2016). RING E3 ligases directly transfer Ub from the E2-Ub complex to the substrate (Deshaies & Joazeiro, 2009), whereas HECT-type E3s transfer Ub to their own catalytic cystine residue before attaching it to the substrate (Huibregtse et al., 1995). RBR E3s combine the characteristics of both RING and HECT types (Walden & Rittinger, 2018). RBR and RING E3s share the RING-binding domains, but RBR family members can generate thioester intermediates with Ub, as is the case with HECT-type E3s (Garcia-Barcena et al., 2020).

To investigate the relationship between SARS-CoV-2 and host factors, we focused on the interaction between nsp16 and the UPS. In this study, we present a novel regulatory mechanism targeting nsp16 via the host UPS. We demonstrate, for the first time, that nsp16 is ubiquitinated and degraded by the host proteasome. Interestingly, two E3 ligases from distinct families, the RING-type MARCHF7 and the HECT-type UBR5, independently recognise and target nsp16 for degradation in the cytoplasm and nucleus. Since nsp16 possesses 2’-O-MTase activity, its enzymatic function converts the viral RNA 5’ cap structure from cap-0 to cap-1. This dual targeting of nsp16 by MARCHF7 and UBR5 leads to its degradation, disrupting its function and ultimately demonstrating effective antiviral activity against SARS-CoV-2. Our findings identify novel therapeutic targets for developing effective strategies to prevent SARS-CoV-2 and treat COVID-19.

Results

Ubiquitination and proteasomal degradation of SARS-CoV-2 nsp16 by the UPS

Ubiquitination plays an important role in SARS-CoV-2 infection and pathogenesis (Gao et al., 2022; Guo et al., 2021; Li et al., 2023; Zhang et al., 2021; Zhang et al., 2024; Zhang et al., 2023). To explore whether the non-structural proteins of SARS-CoV-2 are regulated by the UPS, we examined the effect of the proteasome inhibitor MG132 on the protein expression of all 16 non-structural proteins. MG132 treatment markedly increased the abundance of nsp8, nsp11, and nsp16 proteins in HEK293T cells (Fig. 1A), suggesting their potential degradation by the UPS. It has been reported that nsp8 undergoes UPS-mediated degradation via TRIM22-mediated ubiquitination (Fan et al., 2024). In this study, we focused on nsp16 for further investigation. To confirm the role of the UPS in nsp16 degradation, we employed additional proteasome inhibitors, Bortezomib and Carfilzomib. Unlike the lysosomal inhibitors Bafilomycin A1 and NH₄Cl and the autophagy-lysosomal inhibitor vinblastine, these proteasome inhibitors enhanced nsp16 stability (Fig. 1B). The impact of the USP on the half-life of nsp16 was investigated. To this end, nsp16-expressing cells were treated with a protein synthesis inhibitor, cycloheximide, and the kinetics of nsp16 decay were examined. We observed that the half-life of nsp16 in MG132-treated samples was significantly longer than that in MG132-untreated samples (15 h versus 2 h) (Fig. 1C and 1D).

Figure 1.

To identify proteins interacting with nsp16, we performed co-immunoprecipitation (Co-IP) followed by mass spectrometry (MS) analysis, comparing nsp16-expressing cells treated with and without MG132 (Fig. 1E). KEGG pathway and Gene Ontology analyses of the MS data (Figure 1—figure supplement 1A). Biological process enrichment analysis showed that blocking nsp16 degradation resulted in an increase in the number of proteins that interact with it, including 43 proteins involved in virus-associated processes and 22 proteins involved in regulating mRNA stability. Five proteins were related to the regulation of the cellular antiviral defence. Additionally, interactions were associated with mRNA splicing and RNA methylation (Banerjee et al., 2020), consistent with nsp16’s role as a 2’-O-MTase. Furthermore, SARS-CoV-2 uses nsp16 to disrupt host mRNA splicing to facilitate its infection (Park et al., 2022; Russ et al., 2022; Züst et al., 2011). These findings corroborate the reliability of our MS data.

As expected, we found that several nsp16-binding proteins were related to ubiquitination and degradation pathways (Figure 1 — figure supplement 1B). Four deubiquitinases (DUBs), the E2 ligase UBE2D3, and 14 proteasomal enzymes were identified. Of these, six E3 ligases, UBR5, MARCHF7, HECTD1, TRIM32, MYCBP2, and TRIM21, were chosen for further investigation.

E3 ligases UBR5 and MARCHF7 independently mediate nsp16 degradation

To identify the E3 ligases responsible for nsp16 degradation, we designed small interfering RNA (siRNA) targeting the six candidate E3 ligases and evaluated their effects on nsp16 abundance. Knockdown of UBR5 and MARCHF7 stabilised the protein levels of nsp16 (Fig. 2A). Furthermore, we generated stable cell lines with UBR5 or MARCHF7 knockdowns to further investigate their effects on nsp16 stability and confirm silencing efficiency (Fig. 2B). To determine whether UBR5 and MARCHF7 cooperate in nsp16 degradation, we knocked down UBR5 or MARCHF7 using siRNA in MARCHF7- or UBR5-knockdown cells, respectively. Knockdown of UBR5 in MARCHF7-knockdown cells, and vice versa, further enhanced nsp16 stability. This indicates that UBR5 and MARCHF7 act independently to degrade nsp16 (Fig. 2C). We further validated the independent function of these E3 ligases by examining nsp16 degradation. MARCHF7 overexpression induced nsp16 degradation in both MARCHF7-knockdown and MARCHF7/UBR5-double-knockdown cells. Similarly, UBR5 overexpression degraded nsp16 even in the presence of MARCHF7 knockdown. Knockdown efficiencies were confirmed for all experiments (Figure 2 — figure supplement 1A and 1B). Notably, overexpression of wild-type MARCHF7 or UBR5, but not mutants lacking the functional RING domain (1-542 aa) or the HECT domain inactivated mutant, disrupts the stability of nsp16 in cells with MARCHF7 or UBR5 knockdown (Fig. 2D and 2E). These findings demonstrate that UBR5 and MARCHF7 independently ubiquitinate nsp16 via their HECT or RING domains, targeting it for proteasomal degradation.

Figure 2.

UBR5 and MARCHF7 mediate distinct Ub linkages on nsp16

We first confirmed the ubiquitination of nsp16, both in the presence and absence of exogenous Ub (Fig. 3A-B). To further investigate the regulation by E3 ligases, we examined the impact of UBR5 or MARCH7 knockdown on nsp16 ubiquitination. Knockdown of either UBR5 or MARCHF7 decreased the ubiquitination level of nsp16 compared to that in the negative control (Fig. 3C). Ub contains seven lysine residues (K6, K11, K27, K29, K33, K49, and K63) that form polyubiquitin chains with distinct functions, ultimately dictating the fate of the attached target protein (Grice & Nathan, 2016). To determine the specific type of polyubiquitin chain modification of nsp16 mediated by UBR5 and MARCH7, we constructed a series of Ub mutants retaining only a single lysine residue. Except for K33, all single-lysine Ub mutants promoted nsp16 ubiquitylation to varying extents, suggesting a complex polyubiquitin chain structure on nsp16 potentially regulated by multiple E3 ligases or E2-E3 ligase pairs (Fig. 3D). Notably, analysis of specific lysine linkages revealed that MARCHF7 knockdown primarily reduced the K27-linked ubiquitination of nsp16, while UBR5 knockdown primarily reduced the K48-linked ubiquitination of nsp16 (Fig. 3E and 3F). These findings indicate that MARCHF7 and UBR5 induce K27- and K48-linked ubiquitination on nsp16, respectively.

Figure 3.

UBR5 and MARCHF7 directly interact and colocalise with nsp16 in the endoplasmic reticulum (ER)

MS analysis suggested interactions between nsp16 and both UBR5 and MARCH7. Co-IP experiments confirmed that both Myc-tagged MARCHF7 and endogenous UBR5 interact with nsp16 (Figure 4 — figure supplement 1A and 1B). To determine whether MARCHF7 or UBR5 directly interact with nsp16, fluorescence resonance energy transfer (FRET) assays were performed. When nsp16-YFP was bleached, the fluorescence signals from CFP-UBR5 or CFP-MARCHF7 fusion proteins increased, indicating a direct interaction between these E3 ligases and nsp16 (Figure 4—figure supplement 1C). Image J software was used to quantify the relative fluorescence intensities (Figure 4 — figure supplement 1D). We next investigated whether UBR5 and MARCHF7 require each other for nps16 binding. We compared the binding ability of UBR5 to nsp16 in the presence or absence of MARCHF7 using Co-IP. MARCHF7 knockdown had no effect on the interaction between UBR5 and nsp16, and vice versa (Fig. 4A and 4B), indicating independent binding of UBR5 and MARCHF7 to nsp16. Previous studies have shown that nsp16 is localised to both the nucleus and cytoplasm (Zhang et al., 2020). UBR5 (containing two nuclear localisation signals) and MARCHF7 are present in both compartments (Muñoz-Escobar et al., 2015; Shearer et al., 2018) (Nathan et al., 2008). Immunofluorescence staining revealed that UBR5 and MARCHF7 colocalised primarily with nsp16 in the cytoplasm, with some colocalisation in the nucleus of Hela cells (Figure 4—figure supplement 1E). Similar results were observed in nsp16-transfected HEK293T cells using antibodies against endogenous UBR5 or MARCHF7 (Figure 4 — figure supplement 1F).

Figure 4.

To identify the specific cellular compartment where UBR5 or MARCH7 interacts with nsp16, we co-transfected cells with UBR5-CFP or MARCH7-CFP along with nsp16-YFP. Immunostaining was performed with organelle-specific antibodies to label the mitochondrial marker COX5A, the ER marker PDI, and the Golgi apparatus marker GM130. Notably, both UBR5 and MARCHF7 interacted with nsp16 and colocalised with PDI, but not with COX5A or GM130 (Fig. 4C and 4D). These findings indicate that UBR5 and MARCHF7 directly interact with nsp16 in the ER.

Functional domains of UBR5 and MARCHF7 are required for nsp16 interaction and ubiquitination

UBR5 is a four-domain E3 ligase containing two nuclear localisation signals. The four domains are UBA, UBR, PABC, and HECT (Muñoz-Escobar et al., 2015) (Figure 4—figure supplement 2A). Notably, the HECT domain is essential for its E3 ligase activity, where UBR5 must be conjugated to Ub before transferring it to the substrate (Kim et al., 2021).

To identify the UBR5 domain responsible for nsp16 ubiquitination, we utilised UBR5 mutants with inactivated individual domains. Only the HECT domain mutant failed to degrade nsp16 (Figure 4—figure supplement 2B), consistent with previous findings (Zhou et al., 2022), and was unable to induce nsp16 ubiquitination, further supporting its essential function (Figure 4—figure supplement 2C).

To identify the MARCHF7 region responsible for nsp16 degradation, we constructed a series of truncation mutants, as previously described (Figure 4 — figure supplement 2D) (Zhao et al., 2018) (Nathan et al., 2008). All mutants lost the ability to degrade nsp16 (Figure 4—figure supplement 2E). Furthermore, we explored the ability of these mutants to interact with nsp16 and their impact on nsp16 ubiquitination. Only the mutant with an intact N-terminal region (aa 1–542) retained strong binding to nsp16, while mutants with the active RING domain region (aa 543–616) did not. Consistently, only the wild-type MARCHF7 induced the formation of K27-linked Ub chains on nsp16 (Figure 4 — figure supplement 2F). These findings suggest that the N-terminal of MARCH7 mediates binding to nsp16, while the RING domain is required for its K27-linked ubiquitination activity.

UBR5 and MARCHF7 mediate antiviral activity against SARS-CoV-2 by targeting nsp16 for degradation

To examine the functional impact of nsp16 degradation by UBR5 and MARCHF7 on viral replication, we used a biosafety level 2 cell culture system to generate SARS-CoV-2 transmissible virus-like particles capable of infecting and replicating in Caco2-N^int cells (Ju et al., 2021). Knockdown of either UBR5 or MARCHF7 in these cells significantly increased SARS-CoV-2 replication compared with that in the control group. Notably, simultaneous knockdown of both UBR5 and MARCHF7 further increased SARS-CoV-2 replication (Figure 5—figure supplement 1A and 1B).

Figure 5.

Given the critical antiviral roles of UBR5 and MARCH7, we investigated the modification of nsp16 following viral infection. HEK293T cells stably expressing the SARS-CoV-2 entry receptor ACE2 (HEK293T-ACE2) were infected with the Wuhan strain for 48 h. Co-IP using an anti-nsp16 antibody revealed that nsp16 interacts with endogenous MARCHF7 and UBR5 and undergoes ubiquitination (Fig. 5A). We further validated the antiviral effects of UBR5 and MARCH7 in a biosafety level 3 facility using the Wuhan strain at a multiplicity of infection (MOI) of 0.01 and the Omicron strain at an MOI of 0.001. Compared to the control group, UBR5 or MARCH7 knockdown in Caco2 cells significantly increased intracellular and secreted viral mRNA levels (M and E genes) of both strains. Additionally, with the Wuhan strain, MARCH7 knockdown resulted in a 0.6-log increase in viral titre in the supernatants, while UBR5 knockdown caused an over 1-log increase. With the Omicron strain, knockdown of the two E3 ligases resulted in an even greater increase in viral titres (Fig. 5C–E for the Wuhan strain and 5F–H for the Omicron strain). Immunoblotting (IB) confirmed these findings, demonstrating elevated intracellular and secreted N protein levels upon UBR5 or MARCH7 knockdown (Fig. 5I). The knockdown efficiency of MARCHF7 and UBR5 was also confirmed (Fig. 5B). Due to low transfection efficiency in Caco2 cells, we overexpressed UBR5-Myc or MARCH7-Myc in HEK293T-ACE2 cells. Overexpression of UBR5 or MARCH7 significantly decreased viral mRNA levels of the M and E genes in both the Wuhan and Omicron strains, as measured by real-time quantitative polymerase chain reaction (RT-qPCR). IB analysis also showed decreased N protein levels in cells and supernatants, accompanied by a decrease in viral titres of more than 0.5-log. However, co-transfection with increasing amounts of nsp16 significantly abrogated the inhibitory effects of UBR5 and MARCH7 on SARS-CoV-2 replication (Fig. 6A–H and Figure 6—figure supplement 1A–H). To eliminate the potential influence of nsp16’s intrinsic function on experimental outcomes, we overexpressed UBR5 with an inactive HECT domain (UBR5-ΔHECT) or MARCHF7 with a RING domain deletion (MARCHF7-1-542), along with nsp16. These E3 ligase mutants failed to inhibit viral replication. Moreover, a gradual enhancement of nsp16 protein expression did not enhance viral replication. While there was a slight increase in the mRNA levels of M, protein levels of E and N remained unchanged (Figure 6 — figure supplement 2A–H). These results suggest that UBR5 and MARCH7 exhibit antiviral activity against SAR-CoV-2 of UBR5 and MARCHF7 through their enzymatic function, significantly suppressing viral replication by targeting nsp16 for degradation.

Figure 6.

UBR5 and MARCHF7 mediate broad-spectrum degradation of nsp16 variants

Given the ongoing emergence of SARS-CoV-2 variants, broad-spectrum antiviral activity remains critical. To investigate whether UBR5 and MARCH7 can mediate broad-spectrum degradation of various nsp16 variants, we aligned the amino acid sequences of nsp16 proteins obtained from the National Center for Biotechnology Information (Figure 6 — figure supplement 3A). Among these, only three variants, XBB.1.9.1, XBB.1.5, and XBB.1.16, exhibited a higher number of mutation sites compared to other variants, which showed only one or two mutations, suggesting a more conserved sequence. Consequently, using mutagenesis techniques, we synthesised the nsp16 sequences for these three variants along with several single-site mutants from other variants. Treatment with the proteasome inhibitor MG132 restored nsp16 protein levels for all variants (Figure 6—figure supplement 3B), confirming that degradation occurs through the UPS. Furthermore, knockdown of either UBR5 or MARCHF7 increased the stability of nsp16 proteins from these variants, albeit to varying degrees, indicating that these nsp16 proteins are sensitive to UBR5- or MARCHF7-mediated degradation (Figure 6 — figure supplement 3C). Taken together, these findings suggest that UBR5 and MARCHF7 may confer broad-spectrum antiviral activity by targeting nsp16 from diverse SARS-CoV-2 for degradation.

SARS-CoV-2 infection reduces UBR5 and MARCHF7 expression

To explore the relationship between SARS-CoV-2 infection and the expression levels of UBR5 and MARCHF7, we analysed changes in their mRNA and protein levels after infection with either the Wuhan or Omicron strains at varying MOIs, using RT-qPCR and IB, respectively. Both mRNA and protein levels of MARCFH7 decreased with increasing viral titres. However, UBR5 expression initially increased at low titres but continuously declined with increasing titres (Figure 6 — figure supplement 4A-4C). Previous studies have reported a role for UBR5 in regulating interferon-γ-mediated pathways, which might explain the observed increase in UBR5 expression at low titres (Wu et al., 2022). To confirm these findings in vivo, we examined UBR5 and MARCHF7 mRNA levels in peripheral blood mononuclear cells from SARS-CoV-2-infected patients with varying disease severity. While UBR5 mRNA levels negatively correlated with disease progression, MARCHF7 levels showed no significant correlation (Figure 6 — figure supplement 4D). These results suggest that SARS-CoV-2 may suppress host antiviral defences mediated by UBR5 and MARCHF7.

UBR5 and MARCHF7 protect mice from SARS-CoV-2 challenge

Given the lack of known activators or inhibitors for E3 ligases, we developed a transient overexpression method for UBR5 and MARCH7 in mice using high-pressure tail vein injection of plasmids (Bonamassa et al., 2011). To assess their in vivo antiviral effects, mice were injected with plasmids encoding UBR5 or MARCH7, followed by SARS-CoV-2 challenge (Fig. 7A). Both MARCHF7 and UBR5 overexpression significantly decreased viral E gene copy numbers in mouse lungs compared to those in the control group, as well as markedly decreased viral titres in lung tissue homogenates (Fig. 7C-D). Additionally, treated mice experienced less weight loss (Fig. 7E). Histopathological analysis of major organs 5 d post-infection revealed that lungs from mice treated with UBR5 or MARCH7 displayed attenuated lesions, including alveolar contraction and pulmonary oedema, compared to those typically observed in control mice (Fig. 7F). These treatments also reduced the abundance of N proteins (Fig. 7G). Collectively, these results suggest that UBR5 or MARCH7 overexpression inhibits SARS-CoV-2 virulence in vivo by promoting nsp16 protein degradation.

Figure 7.

Figure 8.

SARS-CoV-2 infection can trigger a severe cytokine storm, thought to be responsible for its high mortality rates (Song et al., 2020) (Hojyo et al., 2020). To examine the effect of UBR5 and MARCH7 on inflammatory responses, we measured levels of key cytokines, including interleukin-6, interleukin-1 receptor antagonist, and interleukin-1β, in the spleens of SARS-CoV-2-infected mice (Makaremi et al., 2022) (Hu et al., 2021). Consistent with their antiviral activity, treatment with UBR5 or MARCH7 significantly reduced the production of these cytokines (Fig. 7H).

Discussion

SARS-CoV-2 nsp16 functions as a 2’O-MTase, catalysing the methylation of the penultimate nucleotide of the viral RNA cap. This modification produces a 5’-RNA cap structure that mimics host RNA, thereby evading host immune detection and responses (Lin et al., 2020) (Balieiro et al., 2022; Russ et al., 2022). High-resolution structures of nsp16 and nsp16-nsp10 heterodimers have been characterised (Rosas-Lemus et al., 2020) (Klima et al., 2022; Lugari et al., 2010), paving the way for antiviral drug development targeting these heterodimers (Balieiro et al., 2022; Melo-Filho et al., 2022). However, the host factors that interact with and regulate nsp16 remain largely unknown.

Previously, our group and others have demonstrated the importance of the UPS in modulating SARS-CoV-2 infection (Xu et al., 2022). For example, host E3 ligases, such as RNF5, Cullin4-DDB1-PRPF19, ZNF598, and TRIM7, ubiquitinate and degrade key SARS-CoV-2 proteins, thereby suppressing viral replication (Li et al., 2023; Liang et al., 2022; Maimaitiyiming et al., 2022; Zhang et al., 2024). Conversely, SARS-CoV-2 employs DUBs to counteract these host defences and enhance its replication (Chen et al., 2024; Gao et al., 2024; Gao et al., 2022; Guo et al., 2021). In this study, we identify nsp16 as a short-lived protein with a half-life of only 2–3 h that is targeted for ubiquitination and degradation by the UPS. Using MS and bioinformatics screening, we identified two E3 ligases, UBR5 and MARCHF7, that directly interact with and target nsp16 for ubiquitination and degradation.

UBR5 is a member of the UBR box protein family and contains a HECT domain essential for its E3 ligase activity (Kim et al., 2021). It is known to regulate tumour growth, metastasis (Xiang et al., 2022) (Qiao et al., 2020), and viral infections. For example, URB5 targets Middle East respiratory syndrome coronavirus ORF4b protein for degradation, dampening the host immune response (Zhou et al., 2022). Conversely, UBR5 acts as a cofactor for Zika virus replication by helping TER94/VCP degrade the capsid protein to expose the viral genome to the cytoplasm (Gestuveo et al.). MARCHF7 is a RING E3 Ub ligase that regulates T cell proliferation, neuronal development, inflammasomes, and tumour progression (Zheng, 2021) (Zhang et al., 2016) (Cai et al., 2022) (Zhao et al., 2018). However, its role in viral infections has been largely unexplored, with only one study identifying it as a host protein interacting with the 2C of foot-and-mouth disease virus (Mahajan et al., 2021).

Our findings reveal that UBR5 and MARCHF7 independently interact with nsp16 and facilitate its degradation, primarily within the cytoplasmic side of the ER. This reflects an evolutionary arms race between viruses, which develop mechanisms to resist host antiviral responses, and hosts, which deploy multiple strategies to counteract the virus. One such strategy is seen with multiple E3 ligases that target the same viral protein for degradation. Functional analyses demonstrate that UBR5 or MARCHF7 knockdown enhanced SARS-CoV-2 replication, whereas their overexpression impaired replication. Furthermore, we found that when the enzymatic activity of these E3 ligases is inactivated, their antiviral effect is abolished, indicating that their antiviral effects are dependent on their E3 ligase activity. Additionally, we observed that in the absence of functional UBR5 or MARCH7, overexpression of nsp16 does not increase viral levels. This suggests that although nsp16 plays an important role, it cannot independently promote viral replication without the involvement of E3 ligases. This underscores the critical function of UBR5 and MARCHF7 as antiviral factors that suppress SARS-CoV-2 replication by promoting nsp16 degradation.

Notably, while nsp16 possesses 18 lysine residues, identifying specific ubiquitination sites proved challenging. Based on structural data of nsp16 (Decroly et al., 2011), we constructed truncated nsp16 mutants to assess their degradation profiles. We found that nsp16 could still be degraded to varying degrees (Appendix-figure S2A-B), suggesting that multiple lysine residues serve as recognition sites for E3 ligases. Nonetheless, the results from these truncated mutants regarding the MG132 recovery provided us with valuable insights. For instance, the Δ2-17 mutant, which lacks this segment, showed more than a two-fold increase in nsp16 expression levels. Conversely, the Δ204-212 mutant exhibited significantly reduced recovery after MG132 treatment compared to the wild-type. These observations offer important clues for further investigation into the interaction regions between nsp16 and MARCHF7 or UBR5, as well as potential ubiquitination sites. MS analysis identified K76 of nsp16 as a ubiquitination site (Appendix-figure S2C). However, an nsp16 K76R mutant was still degraded by E3 ligases, although the degree of ubiquitination was reduced compared to that of the wild-type, indicating that K76 might be one of several potential ubiquitination sites (Appendix-figure S2D-E). Therefore, we hypothesise that nsp16 undergoes ubiquitination at multiple sites and may involve non-canonical ubiquitination pathways at non-lysine residues (McClellan et al., 2019). A comprehensive understanding of the ubiquitination landscape of nsp16 requires further investigation. In addition to UBR5 and MARCHF7, other E3 ligases could contribute to nsp16 regulation, which warrants further studies.

Currently, no specific activators or inhibitors exist for UBR5 or MARCH7. To validate their antiviral effects in vivo, we employed high-pressure tail vein injection, a method commonly used for gene delivery to the liver (Kamimura et al., 2014). This method has also been shown to facilitate gene delivery to other tissues, including the lungs (Bonamassa et al., 2011). Gene expression was successfully detected in lung tissues. While this method induced mild immune activation, its effect is relatively minimal compared with that of other techniques (Suda et al., 2023). Moreover, studies suggest that mice and non-human primates exhibit higher resistance to innate immune activation than humans, suggesting that our findings were only minimally influenced by immune activation (Raper et al., 2003). However, this highlights that translating new drug targets into therapeutic treatments will require considerable time and effort. For example, human trials involving the injection of naked DNA (e.g., OTC cDNA) have shown variable outcomes. While no severe immune responses were observed in most participants, one case of fatal immune activation was reported (Raper et al., 2003), highlighting the variability of human immune responses and a key limitation in extrapolating findings from animal models to human trials. This variability might also explain why we did not observe a clear trend for MARCHF7 mRNA levels in peripheral blood samples from patients with varying disease severities. The limited number of patients in our study could further contribute to this observation.

In this study, we identified the host E3 ligases, UBR5 and MARCHF7, as critical regulators of SARS-CoV-2 nsp16 through ubiquitination and subsequent degradation. We demonstrated a negative correlation between the expression levels of these ligases and the degree of infection, underscoring their antiviral roles. Furthermore, both ligases significantly inhibit SARS-CoV-2 replication. Our findings pave the way for the development of novel therapeutic strategies targeting the UPS for the treatment of COVID-19.

Materials and Methods

Reagents and antibodies

The drugs used in this study were as follows: MG132 (catalog no. S2619), Bafilomycin A1 (catalog no. S1413), Bortezomib (catalog no. S1013), Carfilzomib (catalog no. S2853) and Vinblastine (catalog no. S4504) were purchased from Selleck (Houston, TX, USA). Cycloheximide (catalog no. 66-81-9) was purchased from Sigma (Saint Louis, MO, USA). The antibodies used for immunoblotting analysis (IB) were as follows: anti-UBR5 mAb (Proteintech, Rosemont, IL, USA, catalog no. 66937-1-Ig), anti-MARCHF7 mAb (Santa Cruz Biotechnology, Dallas, TX, USA, catalog no. sc-166945), SARS-CoV-2 2’-O-ribose Methytransferase Antibody (nsp16) (Cell signaling, Danvers, Massachusetts, USA, catalog no. #70811), anti-Flag mAb (Sigma, catalog no. F1804), anti-tubulin mAb (Abcam, Cambridge, Cambridgeshire, UK, catalog no. ab11323), anti-hemagglutinin (anti-HA) pAb (Invitrogen, catalog no. 71-5500), anti-Myc pAb (Proteintech, Rosemont, IL, USA, catalog no. 16286-1-AP), and anti-GFP mAb (Abcam, catalog no. ab1218). Primary antibody: anti-COX5A pAb (Sangon Biotec, Shanghai, CHN, catalog no. D261450), anti-PDI-mAb (Proteintech, catalog no. 2E6A11) and Human GM130/GOLGA2 Antibody (R&D Systems, Minneapolis, MN, USA, catalog no. AF8199) were used for immunofluorescence, and SARS-CoV-2 nucleocapsid monoclonal antibody (mAb) (GeneTex, Irvine, CA, USA, catalog no. GTX635679) were used for Immunohistochemistry. Fluorescent secondary antibodies: goat anti-Rabbit IgG (H+L) Highly Cross Adsorbed Secondary Antibody, Alexa Fluor Plus 488 (Invitrogen, catalog no. A11001), and goat anti-Rabbit IgG (H+L) Highly Cross Adsorbed Secondary Antibody, Alexa Fluor Plus 568 (catalog no. A-11011) and immunohistochemical secondary antibodies: Streptavidin-Peroxidase Anti-Rabbit IgG kit (catalog no. KIT-9706) were purchased from Invitrogen and maixin (Fuzhou, CHN) respectively.

Cell lines and viruses

The cell lines used in this study include HEK293T cells (catalog no. CRL-11268), Hela cells (catalog no.CRM-CCL-2), Caco2 cells (catalog no. HTB-37). All cell lines were purchased from American Type Culture Collection (ATCC; Manassas, VA, USA) and tested for mycoplasma. In order to conduct virus related experiments, stable cell lines including Caco2-N^int (stably expressing SARS-CoV-2 N gene) and 293T-ACE2 (stably expressing ACE2) were generated in our laboratory. The cells were cultured in a cell incubator containing 5% CO2 at 37°C using Dulbecco’s modified Eagle’s medium (DMEM; HyClone, Logan, UT, USA) containing 10% heat-inactivated fetal calf serum (FCS, GIBCO BRL, Grand Island, NY, USA), 100 mg/ml penicillin, and 100 ug/ml streptomycin to provide nutrition. We expanded SARS-CoV-2 virus-like-particles (Wuhan-Hu-1, MN908947, Clade:19A) with high replication capacity in Caco2-N^int using a biosafety level-2 (BSL-2) cell culture system. The virus infecting the cells include Omicron BA. 1 strain (human/CHN_CVRI-01/2022) and SARS-CoV-2 Wuhan strain (BetaCoV/Beijing/IME-BJ05-2020, GenBank access no. MT291831.1), and mouse adapted SARS-CoV-2/C57MA14 variant (GenBank: OL913104.1) used for experiments in mice were provided by Key Laboratory of Jilin Province for Zoonosis Prevention and Control, and all experiments for virus infection were performed in Biosafety level 3 (BSL-3) cell culture system.

Plasmids

Eukaryotic expression plasmids encoding SARS-CoV-2 nsp protein were provided by Professor Wang Peihui (Shandong University). The nsp16-HA expression vector was constructed by adding HA tag at the C-terminus. UBR5 (Gnen ID: 51366) and its mutants (UBR5-ΔHECT, UBR5-ΔPABC, UBR5-ΔUBR) expression plasmids were constructed using purchased plasmids from Addgene (Watertown, MA, USA) as templates with no tag or a MYC tag at the N-terminus. The cDNA of 293T cells was used as a template to construct MARCHF7(Gene ID: 64844) and its truncated mutant with a Myc tag at the C-terminus. All the above expression plasmids were inserted into the VR1012 vector. For immunofluorescence (IF) and Fluorescence Resonance Energy Transfer (FRET) experiments, pCDNA3.1-YFP vector (catalog no. 13033) and pECFP-C1 vector (catalog no. 6076-1) were purchased from Addgene and BD (Biosciences Clontech), and MARCHF7 or UBR5 was constructed on pECFP-C1, and nsp16 was constructed on pCDNA3.1-YFP. Single point mutants of nsp16 of different virus subtypes were obtained by point mutagenesis. Multi-site combined mutants were synthesized by Sangon Biotec Company. Human ubiquitin protein and its mutants have been previously described. Primers required for PCR were listed in Appendix Table S1.

RNA extraction and real-time quantitative PCR

We used the way of Trizol (Invitrogen, catalog no. 15596018CN) to extract RNA. The RNA was subsequently reverse transcribed using a High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Carlsbad, CA, USA, catalog no. 4368814). Finally, PrimeScript RT Master Mix (Takara, Shiga, JPN, catalog no. RR036A) and relative real-time primer were used for qPCR on an Mx3005P instrument (Agilent Technologies, Stratagene, La Jolla, CA, USA). Amplification procedure of the target fragment is as follows: predenaturation at 95°C for 2min, denaturation at 95 °C for 30s, annealing at 55 °C for 30s, and extension at 72 °C for 30s, total of 40 cycles. Real-time primer sequences used in this study were shown in appendix table S1.

Immunoblotting analysis (IB)

Cells or supernatant cultured for a certain time were collected, cell precipitates were resuspended with lysis buffer (50 mM Tris–HCl [pH 7.8], 150 mM NaCl, 1.0% NP-40, 5% glycerol and 4 mM EDTA), 4xloading buffer was added (0.08M Tris [pH 6.8], 2.0%SDS, 10% glycerol, 0.1 M dithiothreitol, and 0.2% bromophenol blue), and samples were lysed by heating at 100℃ for 30min. After removal of cell debris by centrifugation, the supernatant was taken to separate proteins of different sizes by SDS-PAGE. The proteins were transferred onto polyvinylidene fluoride (PVDF) membranes, incubated overnight with the indicated primary antibody, and after 1 hour at room temperature with HRP-conjugated secondary antibodies (Jackson Immunoresearch, West Grove, NJ, USA, catalog no. 115-035-062 for anti-mouse and 111-035-045 for anti-rabbit), the proteins were visualized through Ultrasensitive ECL Chemiluminescence Detection Kit (Proteintech, catalog no. B500024).

Stable cell line generation

We purchased the lentiviral vector pLKO.1-puro (catalog no. 8453) from Adggen. ShRNA targeting target genes were designed and synthesized by Sangon Biotech. After annealing, it was inserted into the vector. Lentivirus was generated by transfection with shRNA (cloned in pLKO.1) or control vector, RRE, VSVG and REV, and the cells were screened with 1/1000 puromycin (3 μg/ml, Sigma, catalog no. P8833) after infection 48 hours. The knockdown efficiency was detected by RT-qPCR or IB. The shRNA target sequences used in this study are shown in appendix table S1.

RNAi

Short interfering RNA (siRNA) used in this study were purchased from RiboBio Co. Ltd. (Guangzhou, CHN). The siRNA sequences for knockdown of MARCHF7 or UBR5 are provided in appendix table S1. The siRNA was transfected into the cells by transfection reagent Lipofectamine 2000 (Invitrogen, catalog no. 11668-019), and the corresponding plasmids were transfected 24 hours later using Lipofectamine 3000 Reagent (Invitrogen, Carlsbad, CA, USA, catalog no. L3000-008). The expression of target genes was detected by RT-qPCR or immunoblotting (IB) analysis 48 hours later.

Immunoprecipitation

After 10 hours of MG132 (10 µM) treatment, the cells were harvested, resuspended in 1ml lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% NP-40) containing protease inhibitor (Roche, catalog no. 11697498001), placed on the shaker oscillator for 4 hours to lyse, then centrifuged to remove cell debris, and the supernatant was incubated with the antibody and protein G agarose beads (Roche, Basel, Basel-City, CH, catalog no. 11243233001) at 4°C overnight. Protein G agarose beads was washed 6-8 times with wash buffer (20 mM Tris-HCL [pH 7.5], 100 mM NaCl, 0.1 mM EDTA, and 0.05% Tween 20) at 4°C, centrifuged at 800 g for 1 min. The proteins were eluted by adding loading buffer (0.08 M Tris [pH 6.8], 2.0% SDS, 10% glycerol, 0.1 M dithiothreitol, and 0.2% bromophenol blue), the samples were boiled at 100°C for 10min to elate the proteins. The lysates and immunoprecipitations were detected by IB.

Mass spectrometry (MS)

HEK293T were transfected with SARS-CoV-2-nsp16-Flag for 36 hours, and then harvested after 10 hours of treatment with MG132 or DMSO. Proteins were enriched by co-immunoprecipitation assay. And the elutions were analyzed by MS. MS analysis was performed by the national center for protein science (Beijing, CHN).

Immunofluorescence

At 48 hours after transfection, the solution was discarded and cells were washed twice with preheated PBS for 5min each, followed by fixation with 4% paraformaldehyde at 37°C for 10min. After three washes with PBS, permeabilized with 0.2% Triton X-100 for 5 min at 37°C, and then immediately followed by blocking with 10% fetal bovine serum at room temperature for 1 hour. The blocked cells were incubated overnight with the indicated primary antibodies. The next day, after three washes with PBS, the cells were incubated with corresponding secondary antibodies for 1 hour at room temperature in the dark. The nuclei were stained with DAPI (49,6-diamidino-2-phenylindole, Sigma, catalog no. 9542) and then stored in 90% glycerol. The fluorescence was detected by a laser scanning confocal microscope (FV 3000, Olympus, Tokyo, Japan).

Viral infectivity assay

The Caco2 cells were transfected with siRNA by Lipofectamine RNAiMAX Reagent (Invitrogen, catalog no. 13778150) to knockdown UBR5 or MARCHF7 and infected with Wuhan strain (MOI=0.01) or Omicron-BA.1 strain (MOI=0.001) at 24 hours after transfection. After two hours of infection, the Caco2 cells were cultured in fresh medium containing 2% fetal bovine serum for 48 hours. The cells and supernatants were collected.

Viral levels were characterized by RT-qPCR and IB analysis of viral structural proteins. The overexpression plasmid UBR5-Myc, MARCHF7-Myc and nsp16-Flag was transfected in 293T-ACE2 by Lipofectamine 3000 Reagent and infected with virus 24 hours later.

Mouse lines and infections

BALB/C mice, 6 weeks, were purchased from Charles River Laboratories, Beijing. Mice were cultured in groups according to experimental groups. Our study examined male and female animals, and similar findings are reported for both sexes. To reduce animal suffering, all welfare and experimental procedures were carried out in strict accordance with the Guide for the Care and Use of Laboratory Animals and relevant ethical regulations during the experiment. The mice were randomly divided into 6 groups and injected with 40 µg/500 µl MARCHF7 or UBR5 plasmids via the tail vein at high pressure. Three groups of mice were infected 50ul with SARS-CoV-2 isolates by nasal challenge at a dose of 10^5.5 TCID₅₀/mL, three groups were not infected with virus as control, and the empty vector injection group was also used as control.

Immunohistochemical analysis

Three mice in each group, a total of 18 mice, were sacrificed after anesthesia. The lungs of mice were collected and fixed in 4% paraformaldehyde fixative. After 2 days, the lungs were progressively dehydrated through different concentrations of ethanol solution, transparently treated, soaked in xylene, and finally embedded in paraffin. Some thin sections obtained through a microtome for hematoxylin and eosin (H&E) staining. The other fraction was incubated with 3% hydrogen peroxide for 5-10 min at room temperature to eliminate endogenous peroxidase activity. SARS-CoV-2 nucleocapsid monoclonal antibody (mAb) (GeneTex, catalog no. GTX635679) and a Streptavidin-Peroxidase Anti-Rabbit IgG kit (Maixin, Fuzhou, CHN, catalog no. KIT-9706) was used to quantify the viral level in lung tissue.

Statistical analysis

The statistical analyses used in the figures have been described in detail in the figure legends. All data were expressed as mean ± standard deviations (SDs). Statistical comparisons were performed using student’s t-test, one-way ANOVA, or repeated measurement ANOVA. The differences were statistically significant as follows: *P < 0.05, **P < 0.01, ***P < 0.001; ns is for no meaning

Ethics statement

Collection of inpatient blood was approved by the Ethics Committee of the First Hospital of Jilin University (21K105-001) in accordance with the guidelines and principles of WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report. All study participants signed an informed consent form. PBMCS were obtained from 9 COVID-19 patients with mild disease, 6 patients with severe disease, and 5 patients with critical disease (Appendix table S2). All animal experiments were approved by the ethics committee of Research Unit of Key Technologies for Prevention and Control of Virus Zoonoses, Chinese Academy of Medical Sciences, Changchun Veterinary Research Institute, Chineses Academy of Agricultural Sciences (IACUC of AMMS-11-2020-006).

Data availability

The authors declare that [the/all other] data supporting the findings of this study are available within the paper [and its supplementary information files]. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identified PXD053961.

Acknowledgements

We thank C.Y. Dai for providing critical reagents.

Additional information

Funding information

This work was supported by funding from the National Natural Science Foundation of China (82341072, 81930062 and 82272316 to ZWY, 82341062 to LZL), the National Key R&D Program of China (2021YFC2301900 and 2301904, 2023YFC2306603), the Science and Technology Department of Jilin Province (YDZJ202301ZYTS521 and YDZJ202201ZYTS587), the Key Laboratory of Molecular Virology, Jilin Province (20102209), and Bethune Project, Jilin University (2023B03). The funding sources were involved in study design, data collection and interpretation, and the decision to submit the work for publication.

Authors’ contributions

L.T.: Conceptualization; Data curation; Formal Analysis; Investigation; Methodology; Writing – original draft. Z.Z.: Methodology; Data curation; Writing – review & editing. Z.L.: Conceptualization; Data curation; Formal Analysis; Investigation; Methodology; Writing – review & editing; Funding acquisition. X.L., Z.L. and W.G.: Methodology; Resources. W.Z.: Conceptualization; Formal Analysis; Funding acquisition; Methodology; Project administration; Resources; Writing – review & editing.