Human E3 ubiquitin ligases: accelerators and brakes for SARS-CoV-2 infection

Since 2020, the COVID-19 pandemic caused by the severe acute respiratory syndrome coronavirus (SARS-CoV-2) has impacted billions across the globe. This positive-sense, single-stranded RNA virus is closely related to SARS-CoV and more distantly related to MERS-CoV [1–3]. Ubiquitin, a key regulator of cellular homeostasis, contributes to the host antiviral response against SARS-CoV-2 and is up-regulated in patients with COVID-19 [4–6]. Although targeting the ubiquitin-proteasome system (UPS) has been leveraged to develop anti-cancer therapies, the UPS has yet to be exploited for antiviral drug development [7–14]. Therefore, defining the mechanisms and consequences of ubiquitination during SARS-CoV-2 infection may not only deepen our understanding of host cell-virus interactions, but potentially lead to new antiviral strategies, specific for SARS-CoV-2, and possibly more generally for other viruses.

Ubiquitination, the addition of 76-amino acid ubiquitin polypeptides to a protein target, is carried out by the co-ordinated activities of three families of enzymes: E1 ubiquitin-activating enzymes, that use ATP to create a thioester bond between its cysteine residues and the C-terminus of a ubiquitin molecule; E2 ubiquitin-conjugating enzymes that bind activated ubiquitin; and E3 ubiquitin ligases, that position the target and E2 proximally, enabling the transfer of the ubiquitin molecule to specific substrates [15,16]. E3 ligases attach either a single ubiquitin molecule (monoubiquitylation) or chains of ubiquitin molecules (polyubiquitylation). Chains can either be branched and unbranched, depending on the pattern of ubiquitin addition to lysine (K) residues on the growing ubiquitin chain. Less commonly, ubiquitin can be attached to serine, threonine, cysteine or the amino (N)-terminus of protein substrates. Polyubiquitin chains are connected through linkages to the epsilon amino group of various lysine residues within the ubiquitin polypeptide. Although K48- and K11- linkages are typically associated with targeted proteasomal degradation, other linkages, such as K6, K27, K29, K33, or K63, can either trigger degradation or alter protein function, impacting many cellular processes, such as cell cycle progression, initiation of programmed cell death and autophagy, maintenance of genome stability, and activation of signaling pathways [15,17,18].

Many viruses employ strategies to promote infection by hijacking host E3 ubiquitin ligases (Table 1). Through E3 ligase binding, viral proteins can redirect the enzymatic activity towards new substrates. This corrupts the normal E3 function and redirects it to facilitate steps in the viral lifecycle, promoting viral replication and spread. For example, internalization of HSV-1 and Kaposi's sarcoma-associated herpesvirus is dependent on the host E3 ligase Cbl to ubiquitinate the host membrane receptor Nectin-1 [19,20]. Although some host E3 ligases, like Cbl, are successfully exploited by viruses, other E3s have crucial roles in initiating antiviral responses and inhibition of viral replication (Table 2). For example, MARCH8, is known to enhance antiviral responses to HIV-1 and Ebola [20,21].

In this review, we discuss the ‘Janus-faced role’ of ubiquitination during SARS-CoV-2 infection, describing both host E3-mediated lines of defense and mechanisms for viral manipulation of host E3s that suppress immune responses and promote viral replication (Figure 1). Although some host UPS-viral interactions reflect adaptations specific to SARS-CoV-2, other adaptations may have previously evolved for other viruses, but can be deployed against SARS-CoV-2. Additionally, some interactions may not have been under evolutionary selective pressure, and therefore are not necessarily beneficial to either virus or host, but instead might simply be adventitious or pathological. As the UPS impacts almost every step in the SARS-CoV-2 lifecycle, there are many potentially interesting biological problems and therapeutic opportunities. However, the field is at a nascent stage and much remains unclear. In some cases, we have only preliminary findings from artificial experimental systems or cell lines, and the physiological importance of the observations needs to be rigorously established. Below, we survey the field, identify the E3 ligase targets of most importance, and, where feasible, highlight potential therapeutic opportunities.

The first step of SARS-CoV-2 infection is viral particle entry into the body through nasal or oral cavities, which is followed by aspiration and movement of the virus down the respiratory tract into the lungs. In the next step of infection, SARS-CoV-2 must cross the physical barrier created by the respiratory epithelium, the body's initial immune defense. Cilia on respiratory epithelial cells trap the virus in the nasal passages and bronchial tubes, giving immune cells the opportunity to clear viral particles before they can enter host cells.

Angiotensin-converting enzyme 2 (ACE2), which is highly expressed on respiratory epithelial cells, is the host cell surface receptor that the viral spike protein binds to initiate viral entry [22]. The spike protein is a viral glycoprotein embedded in the viral membrane that can be processed into two protein fragments: S1 and S2, both of which contain protease cleavage sites [23]. In the first processing step, S1 undergoes a ‘priming’ cleavage by the host protease furin during viral maturation and budding. This results in the formation of a non-covalent linkage between S1 and S2, enhancing infection rates [24]. Upon spike-ACE2 binding, the S2 domain is cleaved and activated at S2′ by either TMPRSS2, if viral entry occurs via a membrane fusion pathway, or by cathepsin L or cathepsin B if viral entry occurs via an endosomal viral entry pathway [25,26]. Activated S2′ increases the spike protein's binding affinity for ACE2, and, for both viral entry pathways, enhances the release of the SARS-CoV-2 genomic RNA into the host cell's cytoplasm [3,25,27–29].

Given the significance of the ACE2 receptor for entry of SARS-CoV-2, it is intuitive that both the host and the virus have acquired the ability to directly or indirectly regulate ACE2 levels in a manner that serves their opposing needs. In this section, we first highlight altered patterns of ubiquitination that promote viral infection, and second, address how E3 ligase activity may be used to resist SARS-CoV-2 infection.

An increased abundance of ACE2 on respiratory epithelial cells is advantageous for the SARS-CoV-2 virus and can be achieved by viral modulation of host E3 ligases. The E3 ligase UBR4 induces K48-linked polyubiquitination of ACE2 at K26, K112, and K114 and subsequently triggers its proteasomal degradation. UBR4's activity can be blocked via phosphorylation of ACE2 by the host kinase GLK (MAP4K3). Adding increasing amounts of viral spike protein to lung epithelial cells resulted in an increase in GLK mRNA levels in a dose dependent manner, by an unclear mechanism. Chuang et al. show that increasing GLK protein levels stabilize and elevate ACE2 expression. Consistently, in hACE2 knock-in mice infected with SARS-CoV-2 pseudovirus, ACE2 is stabilized, triggering increased expression of GLK. This, in turn, increases the transmissibility of the virus [30]. The E3 ligase Skp2, which is involved in cell cycle progression, also induces the degradation ACE2, potentially via lysosomal or proteasomal degradation [31]. SARS-CoV-2 transmissibility is increased by viral-mediated down-regulation of Skp2 [32]. In lung and liver cancer cell lines arrested in G0/G1 phase with the CDK4/6 inhibitor palbociclib, ACE2 expression is increased, whereas expression of Skp2 is simultaneously decreased [33]. A better understanding of the mechanism by which Skp2 acts on ACE2 is needed to understand its role in SARS-CoV-2 infection.

Viral spike protein cleavage at S2′ and, consequently, virus transmissibility can also be enhanced by viral-mediated modification of the ubiquitin state of TMPRSS2. The overexpression of viral protein nonstructural protein (nsp)16 enhances SARS-CoV-2 entry by facilitating the formation of the STUB1-USP14 de-ubiquitination complex. This complex decreases the ubiquitination of TMPRSS2, thereby increasing its stability and promoting the S′ cleavage required for viral entry [34]. The findings of Temena and Acar support the notion that modulation of spike protein cleavage can alter the rate of viral entry, therefore implying that other viral proteins may exploit ubiquitination to amplify infection. Temena and Acar [35] found that the stability of TMPRSS2-TMPRSS4 is correlated with increased expression of the E3 ligase TRIM31 and a higher rate of SARS-CoV-2 infection, although more research is needed to clarify the significance of this finding. Together, the above data suggest that SARS-CoV-2 can alter ubiquitination states such that the steady-state levels of proteins needed for viral entry (e.g. ACE2 or TMPRSS2) are increased.

Beyond regulation of the ACE2-spike protein interaction, Xu et al. [36] determined that overexpression of the E3 ligases STUB1, TRIM6, TRIM22, and UBE4B promote SARS-CoV-2 infectivity, whereas MARCHF8 diminishes SARS-CoV-2 infection. Xu et al. identified these E3 ligases by screening an overexpression library of more than two hundred E3 ligases in HeLa-ACE2 cells, 24 hours post SARS-CoV-2 infection. STUB1 was also identified in a separate comprehensive analysis by Stukalov et al., which examined global protein levels in A549-ACE2 cells transfected with viral proteins from either SARS-CoV or SARS-CoV-2. Stukalov et al. found that STUB1 ubiquitinates ORF3, which is a viral protein involved in viral entry, induction of the host inflammatory response, and egress; ubiquitination of ORF3 may enhance its activity, ultimately promoting infection [36–38].

Host cell signaling can inhibit SARS-CoV-2 infection. For example, infection by SARS-CoV-2 triggers IFN-γ signaling in A549 cells and primary pulmonary alveolar epithelial cells, upregulating the E3 ligase TRIM28, a negative regulator of ACE2 expression. Accordingly, knockdown of TRIM28 increases both the abundance of ACE2 on the cell surface and the extent of SARS-CoV-2 infection. Dexamethasone, a glucocorticoid with anti-inflammatory effects, can be beneficial for severe cases of COVID-19 [14]. Interestingly, treatment of the TRIM28 knockdown cells with dexamethasone resulted in a partial reversal of increased ACE2 expression, attenuating the infection [39].

Emphasizing the complexity of ACE2 regulation, PIAS4-mediated modification by small ubiquitin-like modifier (SUMO) at K187 on ACE2 can prevent ACE2 degradation. SUMOylation is a PTM similar to ubiquitination that plays roles in numerous cellular processes including protein homeostasis, nuclear-cytoplamic transport, and transcriptional regulation; SUMOylation also has a role in viral replication and the interferon response [18,40–42]. Jin et al. found that ACE2 can either be unmodified, conjugated with K48-linked polyubiquitin by an undefined E3 ligase for TOLLIP-mediated lysosomal degradation, or SUMOylated by PIAS4, which stabilized ACE2 by preventing ubiquitination. The SUMO modification can be reversed by the SUMO protease SENP3. Aiding the host's struggle to eliminate the virus, upon SARS-CoV-2 infection, the ACE2-SENP3 interaction is strengthened, causing ACE2 deSUMOylation and subsequent ubiquitination and degradation. Consistently, a SUMOylation inhibitor destabilizes ACE2 and suppresses SARS-CoV-2 infection [43]. However, given that most SUMOylation events occur in the nucleus further research is needed to determine whether this destabilization represents a direct effect on cytoplasmic ACE2 SUMOylation.

After fusion of the viral envelope with the plasma membrane or after viral endocytosis, SARS-CoV-2 initiates viral replication in the cytoplasm of the host cell. Host ribosomes translate ORF1a and ORF1b from the viral genome into the polyproteins, pp1a and pp1ab, respectively. Two viral proteases, virus-encoded papain-like protease embedded in nsp3 and chymotrypsin-like/main protease embedded in nsp5, cleave pp1a and pp1ab into 16 nonstructural proteins (nsps). Nsp2 through nsp16 comprise the replication-transcriptase complex (RTC). Nsp12 contains the RNA-dependent RNA polymerase that initiates RNA synthesis. Its activity is enhanced by cofactors, nsp7 and nsp8 [44–47]. Additional RTC cofactors aid in viral RNA synthesis and immune-response evasion by remodeling the endoplasmic reticulum (ER) membrane to form double-membrane vesicles that shields viral replication from host cytoplasmic factors [29,48–50]. Within the double-membrane vesicles, the RTC uses the viral positive-sense genomic RNA as a template for the production of a negative-sense strand ‘anti-genome’. This anti-genome then serves as the template to produce numerous (positive-sense) viral RNA genomes [48].

Whereas viral RNA replication is mostly autonomous, with little dependence on host enzymes, SARS-CoV-2 relies heavily on host ribosomes and host proteins for translation, protein maturation, and vesicular transport [51]. Protein synthesis at the ER is monitored for protein misfolding, which triggers extraction of the misfolded proteins from the ER membrane into the cytoplasm where they are ubiquitinated and targeted to the proteasome in a process termed ER-associated degradation (ERAD) [52]. Newly synthesized viral proteins assemble into enveloped virions by budding into the ER-Golgi Intermediate Compartment, leading to encapsulation of the maturing virions into vesicles. These vesicles traffic through the Golgi and ultimately fuse with the plasma membrane, releasing the mature virus via Arl8b-dependent lysosomal exocytosis [29,53]. Many of these processes are modulated by E3 ligases. In fact, the vast majority of SARS-CoV-2 proteins are modified with ubiquitin chains [37]. Below, we discuss how ubiquitination by host E3 ligases can facilitate or limit viral replication.

Ubiquitination can act as a positive regulator of viral genomic RNA transcription. Using functional genomic screens, several host E3 ligases have been identified as regulators of SARS-CoV-2 genomic RNA replication. Zhang et al. [54] used quantitative proteomics in Vero E6 cells infected with SARS-CoV-2 to demonstrate that nsp8 undergoes ubiquitination at K53, which enhances RTC formation and the rate of viral RNA replication. The E3 ligase(s) involved in this process has not been identified.

Following new virion assembly at the Golgi, mature viral particles undergo egress and exit the cell. C2-WW (Tryptophan-Tryptophan)-HECT (NEDD4-like) E3 ligases are part of the ESCRT (endosomal sorting complexes require for transport) machinery and are suggested to promote viral egress. RNA viruses hijack these E3s to promote viral exit and proliferation. The WW domains of these E3 ligases mediate protein-protein interaction by recognizing Proline-rich motifs (PPxY, etc.) on substrates [55–57]. In the context of SARS-CoV-2 infection, the HECT E3 ligases, NEDD4 and WWP1, regulate ESCRT-dependent viral egress through ubiquitination of the viral spike protein, a PPXY-motif-containing substrate. This ubiquitination occurs in the cytosol of infected cells upon assembly of new viral particles. Novelli et al. showed that the abundance of NEDD4 and WWP1 in primary samples derived from patients infected with COVID-19 and mouse models are positively correlated with disease severity, suggesting ubiquitination of the spike protein enhances viral proliferation. Furthermore, the authors were able to significantly diminish viral proliferation by treating Vero E6 cells infected with SARS-CoV-2 with the natural NEDD4 and WWP1 inhibitor, I3C [58], suggesting that selective inhibitors of E3 ligase(s) that promote viral proliferation might be an effective therapeutic strategy.

Essential to the process of viral protein synthesis is the SARS-CoV-2 main protease (known as M^pro or 3CL^pro), which cleaves the polyproteins pp1ab into the nonstructural proteins and is the target of the inhibitors that have been approved for clinical use: nirmatrelvir (Paxlovid), ensitrelvir, azvudine, and deremidevir [59–65]. The SARS-CoV-2 main protease can also suppress innate immune response pathways [59,66,67]. The E3 ubiquitin ligase ZBTB25 negatively regulates viral main protease levels by inducing polyubiquitination of the main protease at K100 for proteasomal degradation, reducing SARS-CoV-2 virulence [68].

A different study found the SARS-CoV-2 main protease to be down-regulated by the E3 ubiquitin ligase Parkin. The activity E3 ubiquitin ligase Parkin is enhanced by PINK1 kinase phosphorylation during mitophagy, which correlates with a decrease in viral replication [69].

In contrast with the findings of Zhang et al. [54] demonstrating ubiquitination of nsp8 at K53 can enhance viral RNA replication, Fan et al. [70] showed that the E3 ligase TRIM22 can induce K48-linked ubiquitination of nsp8 at K97, triggering nsp8 proteasome-mediated degradation in HEK293T cells. TRIM22 mediated ubiquitination and degradation of nsp8 reduced the function of the RTC. Thus, there appears to be conflicting outcomes of ubiquitination at different lysine residues on nsp8 that serve either to enhance or prevent the replication of SARS-CoV-2 genomic RNA.

The E3 ligase TRIM7 is also a negative regulator of the RTC. Liang et al. identified several components of the RTC that contain C-terminal PRY-SPRY domains. These domains are known to be recognized by TRIM7. This finding raises the possibility that TRIM7 promotes RTC destruction via the UPS, and may function as a negative regulator against other viruses. Although an attractive possibility, additional work is needed to assess this hypothesis [70–72].

To prevent assembly of new virions, the E3 ligases MARCHF2 and MARCHF8 promote the degradation of SARS-CoV-2 envelope, spike, and membrane glycoproteins, whereas MARCHF1 only targets the SARS-CoV-2 membrane protein for destruction. This was shown in HEK293T cells co-transfected with MARCHF proteins and various viral proteins to characterize the antiviral functions of this family of E3 ligases [73]. How target specificity is achieved, wherein the vesicular trafficking system ubiquitylation occurs, and how the ubiquitylated proteins are extracted from the membrane all need to be determined. This is especially important given that MARCHF2 localizes to the ER, MARCHF1 localizes to both the lysosome or endosome, and MARCHF8 localizes to the lysosome, endosome, plasma membrane or nucleus [74].

To systematically identify SARS-CoV-2 proteins regulated by the host UPS, Zou et al. fused 21 SARS-CoV-2 ORFs to GFP and overexpressed them in HEK293T cells. The abundance of 10 of the SARS-CoV-2 proteins significantly increased upon chemical inhibition of the host proteasome with MG132. To identify the E3 ligases involved in regulating the stability of SARS-CoV-2 proteins, each reporter cell line was infected with an sgRNA library targeting 700 E3 ligases, E2 conjugating enzymes, and deubiquitinating enzymes [75]. The E3 ligase RNF185, previously proposed to have a role in ERAD [76], was found to promote the degradation of the SARS-CoV-2 viral envelope protein. This finding was supported by microscopy showing that RNF185 and the envelope protein partially co-localize to the ER [35]. Knock-out of RNF185 increased viral titer in cells infected with three SARS-CoV-2 variants (WA, Beta, or Delta) [75].

Similarly, the SARS-CoV-2 envelope protein is polyubiquitinated at K63 for proteasomal degradation by RNF5. Like RNF185, RNF5 normally participates in the targeted degradation of misfolded proteins through ERAD. However, in HEK293T cells overexpression of RNF5 resulted in the degradation of the envelope protein and decreased viral proliferation. Consistently, a chemical activator of RNF5, Analog-1, prevented viral proliferation in SARS-CoV-2 infected mice [77], suggesting that selective activation of RNF5 may warrant further investigation for antiviral therapies.

Separately, it was discovered that RNF5 promotes the formation of the envelope-membrane protein complex necessary for new viral particle synthesis and egression [78]. The envelope-membrane complex is unable to use the autophagosome for exocytosis unless the membrane protein is ubiquitinated at K15 by RNF5 [78]. Both manuscripts suggest that the envelope protein competitively engages with RNF5, causing its proteasomal degradation [77,78], but inhibition of this E3 activity could reduce virus transmissibility.

As discussed above, to initially infect the host, SARS-CoV-2 must bypass the primary barrier of the respiratory epithelium to escape immune recognition. Many epithelial cells in the respiratory tract are lined with cilia which aid in the initial immune defense, as well as the ability to smell, taste, and breath properly; functions that can be disrupted in COVID-19-infected patients [79]. Intraflagellar transport complex B (IFT46) protein is part of the IFT complex required for ciliogenesis, maintenance of functional cilia, and bidirectional movement of particles. Several studies have found that the viral protein ORF10 interacts with the host substrate adaptor ZYG11B of the Cullin-RING E3 ligase complex [8,80,81]. Wang et al. [82] determined that ORF10 hijacks ZYG11B to initiate UPS-mediate degradation of IFT46 and impairs this critical line of respiratory defense. ORF10 is highly conserved across coronavirus strains and is evolving under positive selection [83,84], emphasizing the importance of this viral protein in SARS-CoV-2 proliferation. This highlights a potential cause of cilia dysregulation and related symptoms in patients with COVID-19.

Upon infection, the host immune and epithelial cells detect the SARS-CoV-2 viral pathogen via pattern recognition receptors on the plasma membrane and in the cytosol, resulting in the activation of immune signaling pathways, including the NFκB and MAPK. Immune signaling leads to increased immune-related gene expression and secretion of proinflammatory cytokines [85].

One such proinflammatory pathway is mediated by RIG-I, a cytoplasmic viral RNA receptor, and the E3 ligase TRIM25. TRIM25 can directly conjugate RIG-I with K63-linked polyubiquitin at K172 [86] or conjugate a free polyubiquitin chain that RIG-I will interact with, either of which can activate RIG-I, eliciting an INF-β response [87]. TRIM25 can serve numerous antiviral functions. In the case of influenza transcription, TRIM25 can negatively regulate mRNA synthesis by binding the viral ribonucleoproteins and blocking the movement of the viral RNA polymerase independent of its E3 activity [88]. Savellini et al. found that TRIM25 is upregulated in A549 cells infected with SARS-CoV-2 [89,90]. However, the SARS-CoV-2 attenuates INF-β production by virtue of the fact that the viral nucleoprotein directly binds TRIM25 at its SPRY domain, preventing its ability to ubiquitinate and activate the RIG-I induced interferon response [89].

Looking at the general impact of ubiquitination on the immune system response to SARS-CoV-2, Zhang et al. generated a quantitative proteomic map of the ubiquitin landscape after infection. Bioinformatics analysis indicated that SARS-CoV-2 infection suppressed host immune responses by increasing the ubiquitination states of proteins, such as USP5, IQGAP1, TRIM28, and Hsp90. In Vero E6 cells infected with SARS-CoV-2, ubiquitination of USP5 at K184, K423 K720, and K770 is reduced, while ubiquitination at K558 is increased, leading to elevated protein levels. Increased USP5 negatively regulates IFN-I signaling, thereby suppressing the downstream pro-inflammatory response [54].

The steady-state levels of SARS-CoV-2 protein ORF6 are heavily regulated by UPS system [75]. ORF6 inhibits IFN responses and aids in host immune evasion. In human cells, ORF6 is targeted for ubiquitination by the PRPF19-CRL4B E3 ligase. Overexpression of PRPF19 results in the degradation of ORF6, thereby alleviating ORF6-mediated suppression of IFN responses and inhibiting SARS-CoV-2 replication [91]. The human deubiquitinating enzyme, USP1, was found to prevent the ubiquitin-mediated degradation of ORF6 [92], suggesting that decreased ORF6 expression could be achieved through activation of CRL4 or inhibition or USP1.

IFN responses and human RIG-I-MAVS pathway is also modulated by the SARS-CoV-2 viral protein ORF9b [93,94], which is stabilized by the deubiquitinating enzyme USP29 [95].

Both the innate and adaptive immune responses are critical in the host's fight against SARS-CoV-2 infection. The generation of anti-spike protein and anti-nucleoprotein antibodies by the host are the conventional way that the adaptive immune system blocks SARS-CoV-2. Measuring anti-spike or anti-nucleoprotein antibodies can be used as a diagnostic tool and research into these antibodies has been instrumental in development of spike-based vaccines [96,97]. The process of antibody neutralization is known to require the E3 ligase and cytosolic Fc receptor, TRIM21, that helps the immune system fight off infection. Consistently, TRIM21 is up-regulated in the blood samples of patients with severe COVID-19 infection [98]. These data suggest the nucleoprotein could be an attractive a vaccine target. To quantify the activity of antibodies against the SARS-CoV-2 nucleoprotein, Albecka et al. created an in vitro assay that they termed electroporated-antibody-dependent neutralization. They found that nucleoprotein antibodies are only able to neutralize virions in the presence of TRIM21, likely because TRIM21 promotes the degradation of the nucleoprotein [99]. Additionally, TRIM21 activates innate immunity by appending K27-linked polyubiquitin chains to MAVS, stimulating TANK-binding kinase-1 (TBK1) signaling and subsequent activation of IRF3 and IRF7 [100].

The hyperinflammation or ‘cytokine storm’ associated with COVID-19 can be instigated by the virus, or it could be generated as an unwanted byproduct of the host's attempt to clear the virus. The ‘cytokine storm’ is promoted by NFκB and TBK1 signaling, key drivers of cytokine production. Focusing on the biology of hyperinflammation, Nishitsuji et al. investigated the effect of SARS-CoV-2 proteins on NFκB activity. The authors found that K63-linked polyubiquitination of K61 on Nsp6 by TRIM13 and polyubiquitination ORF7a by RNF121 leads to the recruitment of NEMO with subsequent activation of the NFκB pathway [101]. Nsp9 also triggers a cytokine storm by interacting with TBK1, activating an immune signaling cascade that further increases cytokine release. The E3 ligase MID1 conjugates nsp9 with 48-linked ubiquitin at K59, inducing its degradation [102].

Severe disease in a subset of COVID-19 patients can be a consequence of persistent hyperactivation of the immune system. This process is dynamically regulated by PTMs, including ubiquitination. The dephosphorylation of STAT3, of the JAK/STAT pathway, can be blocked when the viral protein ORF3a recruits the E3 ligase TRIM59 to polyubiquitinate STAT3. Ubiquitination and interaction of STAT3 with ORF3a prevents the TCPTP phosphatase from removing the phosphate groups attached to STAT3, resulting in constitutive activation of this immune signaling pathway. In zebrafish, mice and humans, Cai et al. [103] showed that this hyper- and prolonged activation of STAT3 and the elevation of TRIM59 are associated with the acute renal disease related to COVID-19. Additionally, ORF3a assists SARS-CoV-2 viral egress using the host lysosomal exocytosis pathway and inhibition of autophagy [104].

SARS-CoV-2 can exploit the UPS to promote almost every step of the viral life cycle and dampen the host immune response. These tactics are surprisingly sophisticated and effective, allowing the virus to counteract multiple layers of host defenses. Simultaneously, human respiratory epithelial cells and immune cells engage the UPS to combat the virus. Our review identifies the following critical points of viral and host UPS regulation: the ACE2-spike protein interaction at viral entry, viral protein synthesis at the ER, and cytokine signaling in response to infection. However, these effects balance out and therefore the overall impact of UPS regulation on patient outcomes has yet to be determined. The extent of interactions between SARS-CoV-2 and the host's UPS is truly remarkable, raising the interesting possibility that other viruses may similarly commandeer host E3 ligases.

We anticipate that human E3 ubiquitin ligases may be utilized for targeted protein degradation therapeutics aimed at degrading viral components and modulating the immune system. One therapeutic strategy is to develop proteolysis targeting chimeras (PROTACs), which are bivalent molecules that can recruit an E3 ligase, like VHL or CRBN, to a new target, such as a SARS-CoV-2 protein. For example, the main protease of SARS-CoV-2 has been computationally predicted to interact with E3 ligase CRBN-based PROTACs [105,106]. PROTAC technology is being leveraged to develop new antivirals targeting SARS-CoV-2 M^Pro, with PROTACs showing efficient degradation of M^Pro and resulting in potent antiviral activity [68,107,108]. Additionally, molecular degraders have been developed to target nsp3 for cullin-independent proteasomal degradation [109]. Further research is needed, but the studies highlighted in this review support the notion that molecular glue degraders or PROTAC compounds could be effective anti-SARS-CoV-2 therapeutics.

• Importance: Human E3 ligases conjugate ubiquitin to target proteins for degradation by the UPS. Alternatively, non-degradative ubiquitin chain linkages can activate or otherwise alter protein function. Understanding how E3s regulate the SARS-CoV-2 viral life cycle provides insight into the biology of COVID-19 and may identify new opportunities for antiviral therapeutics.

• Summary of current thinking: Human E3 ubiquitin ligases modulate the SARS-CoV-2 viral life cycle at multiple stages, either promoting viral proliferation and worsening outcomes for COVID-19 patients or suppressing of viral replication and enhancing the host immune response.

• Comment on future directions: Future research should focus on validating preliminary findings in physiological settings and identifying key E3 ligase protein targets to explore as potential therapeutic opportunities for COVID-19.

The authors declare that there are no competing interests associated with the manuscript.

ACE2

angiotensin-converting enzyme 2

ER

endoplasmic reticulum

ERAD

ER-associated degradation

PROTAC

proteolysis targeting chimera

RTC

replication-transcriptase complex

TBK1

TANK-binding kinase-1

UPS

ubiquitin-proteasome system