SARS-CoV-2 nsp16 is regulated by host E3 ubiquitin ligases, UBR5, and MARCHF7

Reviewer #1 (Public review):

In this study, Tiang et al. explore the role of ubiquitination of non-structural protein 16 (nsp16) in the SARS-CoV-2 life cycle. nsp16, in conjunction with nsp10, performs the final step of viral mRNA capping through its 2'-O-methylase activity. This modification allows the virus to evade host immune responses and protects its mRNA from degradation. The authors demonstrate that nsp16 undergoes ubiquitination and subsequent degradation by the host E3 ubiquitin ligases UBR5 and MARCHF7 via the ubiquitin-proteasome system (UPS). Specifically, UBR5 and MARCHF7 mediate nsp16 degradation through K48- and K27-linked ubiquitination, respectively. Notably, degradation of nsp16 by either UBR5 or MARCHF7 operates independently, with both mechanisms effectively inhibiting SARS-CoV-2 replication in vitro and in vivo. Furthermore, UBR5 and MARCHF7 exhibit broad-spectrum antiviral activity by targeting nsp16 variants from various SARS-CoV-2 strains. This research advances our understanding of how nsp16 ubiquitination impacts viral replication and highlights potential targets for developing broadly effective antiviral therapies.

Strengths:

The proposed study is of significant interest to the virology community because it aims to elucidate the biological role of ubiquitination in coronavirus proteins and its impact on the viral life cycle. Understanding these mechanisms will address broadly applicable questions about coronavirus biology and enhance our overall knowledge of ubiquitination's diverse functions in cell biology. Employing in vivo studies is a strength.

Weaknesses:

Minor comments:
Figure 5A- The authors should ensure that the figure is properly labeled to clearly distinguish between the IP (Immunoprecipitation) panel and the input panel.

Reviewer #3 (Public review):

Summary:

The manuscript "SARS-CoV-2 nsp16 is regulated by host E3 ubiquitin ligases, UBR5 and MARCHF7" is an interesting work by Tian et al. describing the degradation/ stability of NSP16 of SARS CoV2 via K48 and K27-linked Ubiquitination and proteasomal degradation. The authors have demonstrated that UBR5 and MARCHF7, an E3 ubiquitin ligase bring about the ubiquitination of NSP16. The concept, and experimental approach to prove the hypothesis looks ok. The in vivo data looks ok with the controls. Overall, the manuscript is good.

Strengths:

The study identified important E3 ligases (MARCHF7 and UBR5) that can ubiquitinate NSP16, an important viral factor.

Comments on revisions:

I had gone through the revised form of the manuscript thoroughly. The authors have addressed all of my concerns. To me, the experimental approach looks convincing that the host E3 ubiquitin ligases (UBR5 and MARCHF7) ubiquitinate NSP16 and mark it for proteasomal degradation via K48- and K27- linkage. The authors have represented the final figure (Fig.8) in a convincing manner, opening a new window to explore the mechanism of capping the vRNA bu NSP16.

Author response:

The following is the authors’ response to the original reviews.

Reviewer #1 (Public review):

Major comments:

(1) In Figure 1 the authors could reference and use NSP8 (PMID: 38275298) and Nucleocapsid (PMID: 37185839) in their experiments as positive controls.

Thank you for your suggestion! In Figure 1A, during our screening of SARS-CoV-2 nsp proteins regulated by MG132, we confirmed that nsp8 can also be restored by MG132. This finding indicates that nsp8 is degraded via the proteasome pathway and can therefore serve as a positive control for the experiment. It has been reported that nsp8 undergoes degradation via the ubiquitin-proteasome pathway following its ubiquitination mediated by TRIM22. We have added the description at line 115 in the manuscript.

(2) The data indicating that NSP16 is ubiquitinated come from overexpression systems, and it is possible that NSP16 ubiquitination only occurs in expression contexts, not during coronavirus infection. If NSP16 ubiquitination can't be measured in the context of infection, it is unclear how we can make any conclusions. The authors need to demonstrate the ubiquitination of NSP16 in the context of viral infection.

We greatly appreciate the reviewer's suggestion and have incorporated the corresponding experimental results. As shown in Figure 5A, co-IP experiments using an endogenous nsp16 antibody were conducted following infection with the SARS-CoV-2 Wuhan strain. These experiments confirmed that the nsp16 protein encoded by the virus undergoes ubiquitination in infected cells. This finding highlights the ubiquitination of nsp16 within a biological context, thereby supporting our conclusions in expression contexts.

(3) In Figure 4, adding controls will strengthen the authors' conclusion.

a) Is it possible to observe ubiquitination of NSP16 by transfecting in NSP16-FLAG tagged, immunoprecipitate NSP16, run a western blot, and probe for endogenous ubiquitin?

b) Can the authors please include an empty vector control as well as WT ubiquitin in these panels for comparison?

c) In addition, why are the Ubiquitination patterns different in the IP panels of D and E vs B?? Without an empty vector control, it is challenging to conclude what the background is.

Thank you for your valuable suggestions! We have made the following changes and additions in response to your comments:

a) We have conducted the experiments as per the reviewer's suggestion. Figure 3B shows the result. Co-IP experiments were performed, and endogenous ubiquitination of nsp16 was observed using the endogenous ubiquitin antibody.

b) We apologize for previously focusing solely on presenting multiple ubiquitin mutants on a single panel of nsp16 IP without considering the inclusion of an empty vector control and WT ubiquitin. The experiment has been redesigned and conducted, and the results are now presented in Figures 3E and 3F.

c) The differences in the ubiquitination patterns observed between the IP panels in Figures 3E and 3F compared to 3C may be due to varying plasmids, differences in antibody and depth of exposure. To address this, we have standardized the plasmids in the figure and included an empty vector control as a negative control to clarify the background signal.

(4) Overexpression of the ubiquitin mutants may have an indirect effect on protein homeostasis. The authors can also utilize linkage-specific antibodies in their studies to elucidate the ubiquitin linkage associated with NSP16 ubiquitination. K63-linkage Specific Polyubiquitin (D7A11) Rabbit mAb, 5621S, and K48-linkage Specific Polyubiquitin (D9D5) Rabbit mAb, 8081S from Cell Signaling Technologies?

We greatly appreciate the reviewer's excellent suggestion! Using linkage-specific antibodies to elucidate the ubiquitin linkage associated with nsp16 ubiquitination would indeed provide more direct evidence. However, due to the long lead time for obtaining these antibodies, we plan to conduct further verification in future experiments.

(5) The authors discussed the subcellular localization of overexpressed NSP16- showing the localization of NSP16 in the context of viral infection would strengthen the study. If this is challenging, can the authors express NSP16 along with the co-factor NSP10 and examine its subcellular localization?

Thank you for your suggestion! During viral infection, we observed the ubiquitination of the nsp16 protein through co-IP experiments, indicating that the presence of nsp10 does not influence the regulation of nsp16 ubiquitination by MARCHF7 or UBR5 (Figure 5A). Therefore, we believe that investigating the co-localization of nsp10 and nsp16 would not provide additional value to our results. Additionally, through a literature review, we found studies that have already examined the localization of nsp10 and nsp16 following viral infection. These studies revealed that nsp10 was located in the cytoplasm, while nsp16 can be detected in both the nucleus and cytoplasm (PMID: 33080218; PMID: 34452352). This observation is consistent with the localization of nsp16 that we observed in our overexpression experiments.

(6) a) In Figure 3A, the authors should note that the interaction of NPS16 appears weak with UBR5. The authors should confirm that the interaction of NSP16 and the E3 ligases is relevant in the context of viral infection.

b) In Figure 3B, the scale bars should be labeled in at least one panel, as well as in the legend.

c) The authors discussed nuclear localization of MARCHF7, UBR5, and NSP16, therefore a control with a nuclear stain should be included in this figure to enhance the study.

d) Some panels look overexposed while others are blurry which decreases the robustness of the interaction as the authors stated in line 191. To strengthen the results of Figure 3, consider GST purification and in vitro, cell-free binding assays to confirm a direct interaction between nsp16 and the E3 ligases

Thank you for the reviewer’s thoughtful suggestions! We have made the following changes and adjustments based on your recommendations:

a) On the interaction between nsp16 and UBR5:

The interaction between nsp16 and UBR5 appears to be weak, possibly due to the large size of the UBR5 protein (300 kDa). As a result, there are challenges in presenting the experimental results, including difficulties in both expression and protein level detection. To further confirm the relevance of the interaction between nsp16 and the E3 ligases in the context of viral infection, we have performed experiments, and the results are presented in Figure 5A.

b) On scale bars:

The issue regarding the scale bars in Figure 4 has been addressed, and we have now included them in the figure legend for clarity (Line 885).

c) On nuclear localization control:

For the localization of MARCHF7, UBR5, and nsp16 in Figure 4C, given that both MARCHF7 and UBR5 are tagged with CFP, DAPI staining would result in spectral overlap. However, we conducted co-localization experiments for MARCHF7 or UBR5 with nsp16 in Figure 4—figure supplements 1E and 1F, where DAPI staining was included to illustrate the localization of these three proteins. Our experiments showed that while these proteins are present in both the nucleus and cytoplasm, they are predominantly localized in the cytoplasm.

d) On validation of direct interaction:

We attempted GST purification and in vitro cell-free binding assays to verify the direct interaction between nsp16 and the E3 ligases. However, UBR5 and MARCHF7 are both large proteins, with UBR5 being particularly large, which significantly increased the difficulty of purification. Additionally, we faced challenges in purifying nsp16, as the purified nsp16 protein tended to aggregate. We will continue to optimize purification techniques and conditions in future experiments.

We appreciate your valuable comments, which have greatly contributed to improving our experiments and conclusions.

.

(7) To confirm the knockdown of the E3 ligases by siRNA, the authors should use western blotting to show the presence/absence/decrease of the protein levels in addition to mRNA levels by RT-PCR. The authors have the lysates, and they have shown that the antibodies for MARCHF7 and UBR5 work therefore including this throughout the manuscript to help substantiate the authors' conclusions.

Thank you for the reviewer’s valuable suggestion! We have validated the knockdown efficiency at the protein level for the experiments involving siRNA knockdown. Corresponding Western blot images are now included in the relevant experiments to substantiate our conclusions, in addition to the RT-PCR data, including Figures 2, 4 and 5.

(8) In the overexpression studies of the E3 ligases with viral infection in Figure 5, the authors should include the catalytic mutants for the E3 ligases with the nsp16 gradient experiment. This would strengthen the conclusion of the studies.

Thank you for the reviewer’s suggestion! We have conducted the relevant experiments based on your recommendation, and the corresponding data are presented in the Figure 6—figure supplements 2A-H. These results strengthen the conclusions of our study.

(9) Figure 5: For C and F, for a better comparison of the efficacy against the 2 strains, the authors should use the same scale. This could benefit from a kinetics experiment.

Thank you for the reviewer’s suggestion! We have made revisions in Figures 5E and 5H in responses to your recommendation.

(10) Is there a synergistic effect of double E3 knockdown on viral replication?

Thank you for the reviewer’s question! In Figures 5—figure supplement 1A-B, we conducted experiments by individually and simultaneously knocking down MARCHF7 or UBR5, followed by infection with viral SARS-CoV-2 transmissible virus-like particles. The results revealed that simultaneous knockdown further enhances viral replication, demonstrating a synergistic effect.

(11) In lines 98-100 the authors state "This dual targeting by MARCHF7 and UBR5 impairs the 2'-O-MTase activity of nsp16, blocking the conversion of cap-0 to cap-1 at the 5 'end of viral RNA, ultimately exhibiting potent antiviral activity against SARS-CoV-2". The authors did not examine the 2'-O-MTase activity of nsp16. The authors should rephrase this or provide the data if this experiment was done.

Thank you for the reviewer’s valuable suggestion! Based on your comment, we have revised the ambiguous wording located in lines 100-104.

(12) In the discussion, the authors reported that elucidating a specific lysine residue (s) that is ubiquitinated was challenging and stated that they generated multiple mutants including truncated mutants, and wrote "data not shown". The authors need to include this data as supplementary.

Thank you for the reviewer’s suggestion! Based on your comment, we have included the data regarding the specific lysine residue(s) that is ubiquitinated, along with the truncated mutants, as supplementary data (Appendix-figure S2).

(13) In Figure 7, the authors showed a copy number of SARS CoV-2 E in lung tissue. The authors should show viral titers using either the plaque assay or the TCID50 assay.

Thank you for the reviewer’s suggestion! Based on your comment, we measured the TCID50 of the virus in the lung tissue homogenates, and the results are presented in Figure 7D.

Minor comments:

(1) Line 76: while many E3 ubiquitin ligases directly recognize and bind to their target substrates, cullin-RING ligases directly bind an adaptor, which binds a substrate receptor and/or the substrate directly, while the RING-box protein binds a different surface of the cullin and is also not directly interacting with substrate.

Thank you for the reviewer’s valuable suggestion! Based on your comment, we have revised the ambiguous wording in line 76.

(2) Line 161: having introduced the suggestion that NSP16 is ubiquitinated by these ligases, consider moving Figure 4 to the Figure 3 spot.

Based on your comment, we have rearranged the order of the figures and moved Figure 4 to the Figure 3 spot.

(3) Figure 2: Can the authors please do +/- MG132 for each siRNA? It is possible that the lanes where we don't see NSP16 were because there was no NSP16 expressed, OR it was degraded, MG132 would confirm one or the other.

Thank you for the reviewer’s suggestion! Based on your comment, we have redesigned the experiment and included the MG132 treatment for each siRNA. The results are presented in Figure 2A.

(4) Line 165: The authors write "As confirmed by MS, both Myc-tagged MARCHF7 and endogenous UBR5 interact with nsp16, as seen in the Co-IP experiment" should be the reverse, MS suggests NSP16-E3 interaction, the co-ip confirms this.

Based on your comment, we have revised the wording in line 183 to ensure accuracy. MS suggests the interaction between nsp16 and the E3 ligases, while the Co-IP experiment confirms this interaction.

(5) Line 178: the cited paper doesn't clearly show NSP16 nuclear localization, nor do the authors of said paper claim that they found it there. It is cytoplasmic. Additionally, said paper used overexpression, and it is unclear if NSP16 is nuclear in the context of viral infection.

Thank you for the reviewer’s suggestion! The referenced paper states, "As can be seen in the Supplementary Fig. S2, the viral proteins are either cytoplasmic (NSP2, NSP3C, NSP4, NSP8, Spike, M, N, ORF3a, ORF3b, ORF6, ORF7a, ORF7b, ORF8, ORF9b, and ORF10) or both nuclear and cytoplasmic (NSP1, NSP3N, NSP5, NSP6, NSP7, NSP9, NSP10, NSP12, NSP13, NSP14, NSP15, NSP16, E, and ORF9a)," indicating that nsp16 is localized in both the nucleus and cytoplasm. Upon reviewing the literature, we found that the paper (PMID: 33080218) reports the distribution of nsp16 protein following viral infection. The results indicate that nsp16 is present in both the nucleus and cytoplasm, although the authors of the referenced paper claim that ns16 was located in the nucleus.

(6) Line 197: in addition to the 7 lysine residues, ubiquitin can also form linear N-terminal linkages.

Thank you for the reviewer’s suggestion! Linear N-terminal ubiquitination, with its distinct linkage and substrate recognition mechanism, is typically mediated by a complex consisting of the E3 ubiquitin ligases HOIL-1 and HOIP, and differs from classical ubiquitination. Therefore, this type of ubiquitin chain was not investigated in our experiments.

(7) Line 202: Authors state "Interestingly, all single-lysine Ub mutants promoted nsp16 ubiquitylation to varying degrees, indicating a complex polyubiquitin chain structure on nsp16 potentially regulated by multiple E3 ligases". However, not all the mutants. K33 isn't supported by the blot.

Thank you for pointing that out! Indeed, we made an error in our description. The K33 mutant did not promote nsp16 ubiquitylation, and we have corrected this in the manuscript accordingly in line 173.

(8) Line 204: consider including "E2-E3 ligase pairs" for RING ligases the E2 determines the linkage type see: Cell Research (2016) 26:423-440.

Thank you for your suggestion! We have included the term "E2-E3 ligase pairs" in the article in line 176.

(9) Line 235: The authors used the real virus, the inclusion of the BLS2 virus here is extraneous, it doesn't add anything. The authors can consider removing it.

Thank you for your suggestion! In our experiments, we performed simultaneous knockdown of two E3 ligases, so we believe this data is relevant and should not be removed.

(10) Line 238: Authors state: "led to a significant increase in SARS-CoV-2 levels compared to the control group". What is meant by "levels?"

Thank you for your careful reading. We have updated "levels" to "replication" as suggested to clarify the meaning in line 237.

(11) Line 245: increased titers. This could be improved for specificity by saying, 1-log increase for example.

Thank you for the reviewer's valuable suggestions. We have made the necessary changes and specified "increased titers" as a "1-log increase" in lines 249 and 261.

(12) Line 249: in Figure 5H again, the authors are showing relative mRNA levels. Ideally should show protein levels by western blot.

Thank you for the reviewer's suggestion! We have performed protein-level detection of the knockdown efficiency for the samples, and the bands have been placed in the corresponding positions in Figure 5I.

(13) Line 259: "strongly linked to their ability to modulate..." This appears to be an overextension of the data. The data show nsp16 levels can compensate for E3 overexpression, but not that the E3 ligases are modulating this activity. We can infer this from previous experiments. Perhaps increasing the NSP12 levels would also have the same effect as they don't show that this is specific to NSP16. What about a catalytically dead E3?

Thank you for the reviewer's thoughtful suggestion. We have revised the wording accordingly and designed the viral-related experiments with E3 enzyme activity mutants in Figure 6 supplement 2.

(14) Figure 6: In panel H the MW for UBR5 is incorrect, should be around 300kDa.

Thank you for the reviewer's detailed suggestions. We have made the necessary revisions in Figure 6H.

(15) Line 267: "suggesting a more conserved sequence". What are the authors referring to? More conserved than what? This section would benefit from a discussion of which residues are mutated. Are they potential Ub sites, which could point to differential degradation by the E3s as due to more ubiquitination? Or rather to more efficient interaction with the E3? Is this conserved in related CoVs: original SARS and MERS, for instance?

Thank you for the reviewer’s detailed suggestions. In this context, by “conservation,” we refer to the relative conservation of nsp16 proteins across different subtypes of the Omicron variant. We found that most of the mutation sites contained only 1 to 2 mutations. Additionally, we have constructed and validated multiple-mutant nsp16 proteins, which are still degraded by MARCHF7 or UBR5. Given the ongoing prevalence of the Omicron variant, we aim to explore the broad-spectrum degradation and antiviral effects of these two E3 ligases. While it would be ideal if these experiments could aid in identifying the ubiquitination sites, we have not yet identified any mutant forms that escape degradation. We also compared the nsp16 proteins of several other coronaviruses (such as human coronaviruses 229E, HKU1, MERS-CoV, NL63, OC43, and SARS-CoV-1), and found that these viruses' nsp16 proteins are not highly conserved. As a result, we have not further investigated whether MARCHF7 or UBR5 regulate the nsp16 proteins of these viruses.

(16) Line 347: 2C of what virus?

Thank you for the reviewer’s careful reading. We have made the necessary additions to address this point in line 357.

(17) Line 890: "Scale bars, 25 mm". Should it be 25nm?

Thank you for your feedback! I realized there was an error in the unit labeling, and I have corrected the relevant sections in line 904. I appreciate your careful reading.

Reviewer #2 (Recommendations for the authors):

(1) In Figure 6, the authors found that increasing amounts of nsp16 restored the replication of SARS-CoV-2 in the presence of MARCHF7 or UBR5. The authors better discuss the possibility that nsp16 may stimulate viral replication regardless of these E3 ligases, or provide evidence to further clarify this.

Thank you for your thoughtful suggestion! Given the strong functionality of nsp16 itself, your consideration is very comprehensive. In Figure 6—figure supplement 2A–H, we conducted transfection experiments with E3 activity-deficient proteins and reintroduced nsp16. The results showed that, in the absence of active MARCHF7 or UBR5 antiviral function, overexpression of nsp16 did not promote viral replication, although the RNA levels of the M protein slightly increased. Therefore, in our experiments, excess nsp16 did not significantly stimulate viral replication.

(2) In Figure 7, the in vivo data supports the function of both E3 ligases to reduce viral infectivity. Is it possible that tail vein injection of naked plasmid DNA may stimulate the innate immune system, e.g., induce IFN as a DNA vaccine, which may contribute to the inhibitory effect? The authors are suggested to discuss or address it.

Upon reviewing the relevant literature, we found that the hydrodynamic gene delivery (HGD) method using naked DNA is both highly efficient and associated with a low risk of triggering immune responses or oncogenesis. Studies have shown that HGD only weakly activates host immunity (reference: 37111597), which is less of a concern compared to other gene delivery methods. Although some studies have reported strong immune responses following the injection of naked DNA (e.g., Otc cDNA) in human trials, it is noteworthy that no such responses were observed in 17 other participants. This suggests that the immune reactions observed in some cases may be due to individual variability or limitations in animal models, which may not fully translate to human trials.

Based on these findings, we believe that the antiviral effects observed in our study are primarily attributable to the intrinsic properties and functions of the E3 ligases. Furthermore, it has been reported that mice and non-human primates exhibit significantly greater resistance to innate immune activation compared to humans. This highlights the challenges in translating these findings into effective antiviral therapeutics and underscores the need for further research in this area. We have incorporated the requested discussion into the manuscript in lines 393-410.

(3) The authors shall include some of the key data in supplementary figures in the main text, such as the study on UBR5 and MARCHF7 mediate broad-spectrum degradation of nsp16 variants and SARS-CoV-2 infection decreases UBR5 and MARCHF7 expression, which make it easier for readers to follow.

Thank you for your valuable suggestion regarding the organization of our manuscript. In response to your feedback, we have moved the study on nsp16 variants to the Figure 6—figure supplement 3. Additionally, the data showing changes in UBR5 and MARCHF7 levels following viral infection have been added as supplementary data in Figure 6—figure supplement 4.

(4) The diagrammatic sketches in Figures 1E, S1A and B, 7A, and 8 had low resolutions. Please change them to higher resolutions. Moreover, please state the licensing rights of these diagrammatic sketches.

Thank you for your detailed review! In response to your comment, we have improved the resolution of Figures 1E, S1A and B, 7A, and 8. Additionally, we have specified the drawing tools and source websites in the figure legends (lines 794, 813, 999, and 1013). And we have obtained the necessary licenses for each diagram.

Figure 1E: Created in BioRender. Li, Z. (2025) https://BioRender.com/h43f612

Figure S1B: Created in BioRender. Li, Z. (2025) https://BioRender.com/b98t559

Figure 7A: Created in BioRender. Li, Z. (2025) https://BioRender.com/e76g512

Figure 8: Created in BioRender. Li, Z. (2025) https://BioRender.com/o84p897

(5) The authors suggested that both UBR5 and MARCHF7 had a function in triggering the degradation of NSP16, however, the expression of UBR5 but not MARCHF7 was shown to be associated with the severity of clinical symptoms. Further, why did the host evolve 2 kinds of E3 ligases to adjust only 1 viral target? Please discuss them.

Thank you for your insightful comments. We acknowledge that the limited number of patients with varying degrees of illness in our study could potentially mask some of the observed phenomena. Additionally, individual variability may also play a significant role, which highlights the challenges in translating findings from animal models to human trials.

Regarding the presence of two E3 ligases targeting the same substrate, we view this as part of an evolutionary arms race between the host and the virus. Viruses evolve mechanisms to counteract the host’s antiviral responses, while the host, in turn, develops multiple pathways and strategies to combat viral infection. This dynamic may explain why multiple E3 ligases regulate the levels of the same factor, reflecting the host’s complex and redundant antiviral defense mechanisms. We have incorporated the requested discussion into the manuscript in lines 359-362.

(6) Please standardize the symbol size of the bar charts in the same figure, just like in Figures 1D and 5.

Thank you for your constructive suggestion. We have standardized the symbol sizes of the bar charts in the figure as per your recommendation, ensuring consistency across all panels.

(7) The use of English could be improved.

Thank you for your feedback regarding the language. We have carefully reviewed the manuscript and made revisions to improve the clarity and fluency of the English.

Reviewer #3 (Recommendations for the authors):

Major points:

(1) In Figure 1: The expression level of NSP6, 10, 11, and 12 is weak. Include a higher exposure blot (right next to these blots marking as higher exposure) to show the expression of these plasmids. Here, the NSP12 plasmid has no expression, so it is difficult to conclude the effect of MG132 from this blot. It will be appropriate to show the molecular weight of each gene fragment since some of the plasmids have multiple bands. Verify the densitometric analysis, the NSP4 (+/- MG132) blot, and the densitometric analysis do not correlate. Figure 1B: It is recommended to include appropriate control (media only) for NH4Cl. The DMSO control serves well for the drugs, not for Ammonium Chloride. In Figure 1C, how did the authors arrive at the 15-hour time point? The correlation does not appear as the authors claim. Where is the 15-hour sampling time point for MG132 or CHX chase? The experimental approach to screen the E2/E3 Ub ligase is appreciated.

Thank you for your valuable feedback! Regarding your questions, we have made the following revisions:

On the expression of nsp6, nsp10, nsp11, and nsp12 in Figure 1:

We have replaced the blots for nsp10, nsp11, and nsp12 with higher exposure blots. However, due to the strong expression of NSP14, we were unable to generate a higher exposure blot for nsp6. Based on the current exposure, it is clear that nsp6 is not regulated by the proteasome. Additionally, in the high-exposure blot for nsp12, we were able to observe its expression and found that this protein is weakly regulated by MG132. Following your suggestion, we have labeled the molecular weights of the proteins in the figure.

On the densitometric analysis of nsp4 protein:

We recalculated the densitometric analysis for nsp4 and found no issues. Although the band intensities do not show large changes, the relative fold changes appear more pronounced because we normalized the data using GAPDH as an internal control. We have added detailed description in the figure legend.

On the NH4Cl control:

In this experiment, ammonium chloride was dissolved in DMSO. We reviewed the solubility data and found that ammonium chloride has a solubility of 50 mg/ml in DMSO, which is sufficient to reach the concentrations used in our experiment. While the solubility is higher in water, we believe that DMSO is an appropriate solvent for this compound in our context.

On the 15-hour time point in Figure 1C:

Regarding the 15-hour time point mentioned in Figure 1C, we did not collect samples at that time. We performed semi-quantitative analysis of protein levels at different time points using ImageJ and estimated the half-life time point based on the half-life calculation formula. Thank you for your suggestion; we will clarify this in the figure legend.

Once again, thank you for your thoughtful review and constructive suggestions. We have made the necessary revisions and improvements to the figures based on your feedback.

(2) In Figure 2: I do not find a reason to include DMSO control in the siRNAs for E2/E3 Ub. Please justify why it is necessary. It is requested to include WB for the siRNA-treated samples. It is strongly recommended to show the WB data for siRNA-treated samples because you are showing siRNA treatment of MARCHF7 in shUBR5 cells and vice versa. However, if antibodies for corresponding targets are not available, qPCR can be shown in graphical representation in supplementary data indicating the siRNA target region and qPCR target. Show a graphical representation of domains/ deleted regions of MARCHF7 and UBR5.

Thank you for your valuable feedback! We have addressed your concerns as follows:

On the inclusion of the DMSO control group:

The DMSO group was initially included as a control for the MG132-treated group. By comparing with the MG132 group, we aimed to observe whether nsp16 levels were restored by MG132 treatment. Additionally, in siRNA knockdown experiments, the DMSO group was included to compare nsp16 protein levels after knockdown with those in the NC group, as well as to assess differences in nsp16 restoration between MG132 treatment and factor knockdown. However, we acknowledge some issues in the control design. To address this, we have redesigned and conducted the experiments with improved controls (Figure 2A).

On validating knockdown efficiency:

We have included Western blot data for UBR5 and MARCHF7 knockdown efficiencies. For other factors where specific antibodies were unavailable, we followed your suggestion and provided graphical representations in the Appendix-figure S1, illustrating the siRNA target regions and qPCR target sites to confirm knockdown specificity and efficiency.

(3) In Figure 4 A: Write details on how this IP was done. What was the transfection time of this plasmid? Is the transfection time different from that of NSP16 in Figure 1A which shows a significant degradation of NSP16? Please discuss this in detail. It is recommended that this IP be done in +/- MG132. Since you have used siRNA and performed an IP, It is recommended to repeat the IP (with +/- MG132) using the MARCHF7 and UBR5 plasmids

Thank you for your detailed review and suggestions! We have addressed your concerns as follows:

On the specific protocol for the co-IP in Figure 3A:

The detailed protocol for the immunoprecipitation (IP) experiment is as follows: on day 1, cells were plated, and on day 2, we co-transfected nsp16 and Ub expression plasmids. After 32 hours of transfection, we treated the cells with MG132 for 16 hours, then harvested the cells for IP. We included MG132 treatment in all ubiquitination IP experiments because, without MG132, nsp16 would be degraded, preventing us from observing changes in ubiquitination levels. We apologize for not clearly labeling this in the figure, and we have made the necessary modifications.

On the use of MG132 and NSP16 degradation:

Following your suggestion, we have clarified the use of MG132 in the IP experiments, which differs from the degradation of nsp16 shown in Figure 1A. In Figure 1A, we show the degradation of nsp16 in the absence of MG132 treatment.

On the overexpression of UBR5 and MARCHF7:

The effect of overexpressing UBR5 or MARCHF7 on ubiquitination has been validated in Figure 4 supplement 2. In these experiments, we explored the effect of UBR5 activity domain inactivation on nsp16 ubiquitination, as well as the effect of MARCHF7 truncation on nsp16 ubiquitination modification. In these experiments, overexpression of the wild-type E3 ligases was also included, and the results yielded the same conclusions as those from the E3 knockdown experiments, thereby validating the robustness of our findings.

(4) In Figure 4C: Appropriate controls are missing. The authors claim NSP16 is ubiquitinated and degraded by UBR5 and MARCHF7 via K27 and K48 chains. There is no NSP16 Only control. We cannot compare the NSP16 without an NSP16 transfection. I will suggest the authors repeat these individual controls in both the presence and absence of MG132.

Thank you for your careful review and valuable suggestion! In response to your comment, we have redesigned the experiment and added a control group without nsp16 transfection. We have repeated the validation in the presence of MG132. Without MG132 treatment, nsp16 is degraded, leading to very low protein levels, making it difficult to observe the phenomenon. We have updated the figure accordingly and made the necessary adjustments based on your suggestion (Figure 3E-F).

(5) In my opinion, the Figure 8 needs modification. It is requested to show the levels of strand-specific viral mRNA under UBR5 and MARCHF7 knock-down in +/- of MG312. This figure should also be supported by WB indicating the level of NSP16 (capping activity) and any of the viral proteins. This may validate that if the capping activity is lost, viral translation is affected and hence there is a reduction in virus titre. Alternatively, the figure can be modified by putting a sub-heading box over 7mGppA-RNA section and marking it as a future direction/ hypothesis.

Thank you for your thorough and thoughtful review! Regarding the modification of Figure 8, we completely agree with your suggestion. Currently, examining the impact of viral RNA cap modification is technically challenging for us. Therefore, we have followed your advice and marked the investigation of how nsp16 degradation affects viral RNA cap structures as a future direction/hypothesis in the schematic of Figure 8. This revision helps provide direction for future experiments and enhances the clarity of the figure. Thank you for your thoughtful consideration and valuable suggestion!

Minor points:

(1) Figure 2A: Align NSP16 Blot to actin.

Thank you for your constructive feedback! We have redesigned the experiment and included an MG132 treatment group in Figure 2A. Consequently, the figure has been revised comprehensively, and the nsp16 blot has been aligned with tubulin.

(2) Figure 2C: It is recommended to properly align the lanes where the pLKO and shRNA labelling are overlapping.

Thank you for your thoughtful suggestion! We have revised Figure 2C based on your recommendation to ensure that the pLKO and shRNA labeling no longer overlap. We sincerely apologize for any confusion this may have caused and appreciate your understanding and support.

(3) Just a curious question, what happens if we silence both UBR5 and MARCHF7 and check for virus titre? This is an additional work, but if the authors do not agree, it is ok.

Thank you for your valuable suggestion! Regarding your question about silencing both UBR5 and MARCHF7, we indeed attempted to generate knockout cell lines, but unfortunately, we were not successful at this stage. We plan to explore alternative methods to establish stable knockout cell lines in our future experiments. Meanwhile, as shown in Figure 5 supplement 1, we have performed experiments where both UBR5 and MARCHF7 were knocked down simultaneously, followed by infection with virus-like particles. The results indicate that dual knockdown further enhances viral replication. These findings may partially address your question. Thank you again for your insightful suggestion!