Opposing activities of IFITM proteins in SARS‐CoV‐2 infection

Abstract Interferon‐induced transmembrane proteins (IFITMs) restrict infections by many viruses, but a subset of IFITMs enhance infections by specific coronaviruses through currently unknown mechanisms. We show that SARS‐CoV‐2 Spike‐pseudotyped virus and genuine SARS‐CoV‐2 infections are generally restricted by human and mouse IFITM1, IFITM2, and IFITM3, using gain‐ and loss‐of‐function approaches. Mechanistically, SARS‐CoV‐2 restriction occurred independently of IFITM3 S‐palmitoylation, indicating a restrictive capacity distinct from reported inhibition of other viruses. In contrast, the IFITM3 amphipathic helix and its amphipathic properties were required for virus restriction. Mutation of residues within the IFITM3 endocytosis‐promoting YxxФ motif converted human IFITM3 into an enhancer of SARS‐CoV‐2 infection, and cell‐to‐cell fusion assays confirmed the ability of endocytic mutants to enhance Spike‐mediated fusion with the plasma membrane. Overexpression of TMPRSS2, which increases plasma membrane fusion versus endosome fusion of SARS‐CoV‐2, attenuated IFITM3 restriction and converted amphipathic helix mutants into infection enhancers. In sum, we uncover new pro‐ and anti‐viral mechanisms of IFITM3, with clear distinctions drawn between enhancement of viral infection at the plasma membrane and amphipathicity‐based mechanisms used for endosomal SARS‐CoV‐2 restriction.

Please remember: Digital image enhancement is acceptable practice, as long as it accurately represents the original data and conforms to community standards. If a figure has been subjected to significant electronic manipulation, this must be noted in the figure legend or in the 'Materials and Methods' section. The editors reserve the right to request original versions of figures and the original images that were used to assemble the figure.
Further information is available in our Guide For Authors: https://www.embopress.org/page/journal/14602075/authorguide The revision must be submitted online within 90 days; please click on the link below to submit the revision online before 23rd Dec 2020. Shi and colleagues have studied restriction of SARS-CoV2 by IFITM proteins. They show that IFITM 1 and 3 expression restricts infection and that specific IFITM3 mutants actually slightly enhance infectivity. They also show that TMPRSS2 expression reduces IFITM1/3 sensitivity but enhances infection increases in the presence of IFITM mutants. Overall, the data are quite preliminary and not very novel since a series of previous studies have described IFITM restriction of CoV2. There's also actually not a lot of data presented. This is an interesting start and its nicely written but I'm not sure what this study tells us and what insight it brings.
1. The effect sizes for CoV2 are pretty weak, particularly e.g. Fig 2D. Note the comparison with flu shows a complete block to infection in Fig2. Why is that? Are these real inhibition effects? I would have liked to see some replication assays showing that IFITM is really a strong inhibitor. Also do replication assays show strong enhancement in the presence of the IFITM mutants. Its all a bit preliminary.
2. Graph labelling could be clearer. I'm not sure of the value of a control bar without errors, in each of the plots. The axis labels should be consistent, eg relative infection with 1 or 100, not a mixture of the 2.
3. Labeling the bar charts with the virus would improve clarity.
4. The facs analysis is a bit shaky in places. In Fig 2C the gate is too far to the left. It is clear that the effect size in Fig 3C with IFITM1 expression is very strong, ie a complete block to infection. But poor placing of the gate reads it as a 3 fold effect, which it clearly is not given the decent shift in the infected population. Here they're just counting the edge of the uninfected population. A proper look at all the Facs data is advised. 5. I don't like the fact that all the infection data are normalised. Have any titrations been done? Is the effect size MOI dependent?
Referee #2: The manuscript submitted by Shi, et al. is an intriguing study that explores the role of the IFITM proteins in SARS-CoV-2 infection. The authors show that IFITM1, IFITM2 and IFITM3 restrict SARS-CoV-2. The experiments are well-designed and the use of both overexpression and deletion of IFITM proteins strengthens the claim of an important role for these innate immune proteins in viral infection. Overall, the data presented are strong and the authors make a compelling case for the function of IFITM3 that is distinct from other viral infections. The studies that show re-localization of IFITM3 results in enhancement, rather than the inhibition of infection are interesting. However, there are some points that can be addressed to strengthen the studies presented here. 3. The authors show that transfecting cells with IFITM3 can modestly inhibit syncytia formation when expressed in target (non-spike expressing cells) but that IFITM3 mutants enhanced syncytia formation in this context. Is this effect observed only when expressed in target cells? What is the result of co-expressing IFITM3 and mutants in SARS-CoV-2 spike-expressing cells? What is the effect of expressing IFITM1 and IFITM2? 4. The results showing that overexpression of TMPRSS2 decreases IFITM3-mediated SARS-CoV-2 inhibition are interesting. What is the result of overexpressing TMPRSS2 with IFITM1? Or with IFITM3 localization mutants? 5. Human IFITM3-Y20A but not mouse IFITM3-Y20A increased infection compared to vector control cells. Do both mutants localize to the plasma membrane? This should be shown.
We thank the reviewers for taking the time to provide feedback on our work identifying the divergent activities of IFITMs on SARS-CoV-2 infection. Please find below in blue font our responses to the specific points raised by each reviewer. In sum, we have clarified several points of significance within the manuscript text and have made the following major changes: 1. We have provided a supplemental spreadsheet containing all non-normalized and normalized infection data that were utilized to generate each of the graphs in our manuscript. 2. We have added confocal imaging of mouse and human IFITM3 Y20A and L23Q mutants, demonstrating their localization at the cell periphery in comparison to intracellular punctate localization of WT IFITM3 (New Figures 3G and 4D). 3. New data have been added to Figure 7 to further address whether TMPRSS2 overexpression allows IFITM1 or IFITM3-Y20A to enhance infection. We conclude that IFITM1 is not able to enhance infection regardless of TMPRSS2 expression, and that TMPRSS2 does not provide statistically significant enhancement of infection beyond the enhancement already provided by IFITM3-Y20A. 4. We have added an entirely new set of experiments utilizing IFITM3 KO and IFITM locusdeleted MEFs to further confirm an overall restriction of SARS-CoV-2 infection by endogenous IFITMs. By stimulating these cells with type I IFN, we also demonstrate that IFITMs play a role in IFN-mediated inhibition of the virus.
We look forward to publishing this timely and important work with EMBO Journal.

Best regards, Jacob Yount & Alex Compton
Referee #1: Shi and colleagues have studied restriction of SARS-CoV2 by IFITM proteins. They show that IFITM 1 and 3 expression restricts infection and that specific IFITM3 mutants actually slightly enhance infectivity. They also show that TMPRSS2 expression reduces IFITM1/3 sensitivity but enhances infection increases in the presence of IFITM mutants. Overall, the data are quite preliminary and not very novel since a series of previous studies have described IFITM restriction of CoV2. There's also actually not a lot of data presented. This is an interesting start and its nicely written but I'm not sure what this study tells us and what insight it brings.
1. The effect sizes for CoV2 are pretty weak, particularly e.g. Fig 2D. Note the comparison with flu shows a complete block to infection in Fig2. Why is that? Are these real inhibition effects? I would have liked to see some replication assays showing that IFITM is really a strong inhibitor. Also do replication assays show strong enhancement in the presence of the IFITM mutants. Its all a bit preliminary.
The reviewer's comment noting the comparison with influenza highlights an aspect of our work that should have been better discussed, and that bolsters our conclusions. IFITM3 is primarily localized to endosomes and is thus able to very effectively inhibit infection by influenza virus, which enters cells entirely via endocytosis. As a contrast, we previously showed that metapneumovirus, which uses dual cell entry pathways (membrane fusion at either the plasma membrane or within endosomes), is restricted by IFITM3 only in its endocytic entry (McMichael, 21st Oct 2020 1st Authors' Response to Reviewers J Infect Dis, 2018). Our results showing that SARS-CoV-2 is similarly partially inhibited by IFITM3 is consistent with the known ability of this virus to similarly utilize dual cell entry pathways. Our finding that IFITM3 at the plasma membrane enhances SARS-CoV-2 infection adds further unique complexity to our results. Taking all of this together, we would not expect full inhibition of SARS-CoV-2 as is seen for influenza virus. Overall, a partial inhibitory effect of WT IFITM3 is consistent with the known entry pathways of the virus and the opposing roles of IFITM3 that we report on here. We now provide a more thorough discussion of these points. Figure 5). These data further support one of our primary conclusions that IFITMs generally repress infection despite the ability of IFITM3 to enhance infection under certain circumstances. 2. Graph labelling could be clearer. I'm not sure of the value of a control bar without errors, in each of the plots. The axis labels should be consistent, eg relative infection with 1 or 100, not a mixture of the 2.

We have additionally added data in which IFITM3 KO MEFs and IFITM-locus deleted MEFs show increased infection with genuine SARS-CoV-2 as compared to WT cells (New
Control bars are based on a normalization to 100 so error bars are not shown for the controls. We have, however, now provided a supplemental data sheet that contains all nonnormalized and normalized percent infection data that were used in generating all graphs.
We have made axis labels consistent throughout the manuscript as requested.

Labeling the bar charts with the virus would improve clarity.
We have added virus names to the y axes of the infection experiment graphs as requested.
4. The facs analysis is a bit shaky in places. In Fig 2C the gate is too far to the left. It is clear that the effect size in Fig 3C with IFITM1 expression is very strong, ie a complete block to infection. But poor placing of the gate reads it as a 3 fold effect, which it clearly is not given the decent shift in the infected population. Here they're just counting the edge of the uninfected population. A proper look at all the Facs data is advised.
We respectfully disagree with these statements. The flow cytometry gates for infected cells were set based on lack of positive cells in non-infected samples. We also point out that altering the gates as suggested by the reviewer, may strengthen results concerning IFITM1, but would not affect our overall conclusions. 5. I don't like the fact that all the infection data are normalised. Have any titrations been done? Is the effect size MOI dependent?
The infection data is normalized because of day to day variation in the maximum percent infection observed in replicate experiments. We provided representative non-normalized flow cytometry plots for every normalized figure to show the general magnitudes of infections that we achieved, i.e., 7 -20% maximal infection in different experiments. Importantly, despite slight variations in infections, the data trends for effects of IFITMs are consistent across experiments as shown by statistical significance observed in comparisons of the normalized data.
Regarding MOIs, effects of IFITMs on virus infections are generally saturable by increasing virus MOI. For our experiments we chose an MOI of 1, which was the highest virus dose allowed by the titer of our virus stock. An MOI of 1 resulted in reasonable, but not saturating, infection levels in HEK293T cells.

Referee #2:
The manuscript submitted by Shi, et al. is an intriguing study that explores the role of the IFITM proteins in SARS-CoV-2 infection. The authors show that IFITM1, IFITM2 and IFITM3 restrict SARS-CoV-2. The experiments are well-designed and the use of both overexpression and deletion of IFITM proteins strengthens the claim of an important role for these innate immune proteins in viral infection. Overall, the data presented are strong and the authors make a compelling case for the function of IFITM3 that is distinct from other viral infections. The studies that show re-localization of IFITM3 results in enhancement, rather than the inhibition of infection are interesting. However, there are some points that can be addressed to strengthen the studies presented here.
1. There is less inhibition of infection with Caco2 cells which are infectable without overexpression of ACE2. Does this indicate an issue with ACE2 overexpression? What are the baseline levels of infection in each cell line?
We have tried extensively over the past several months to achieve robust infections of Calu3 and Caco2 cells, which as the reviewer notes, endogenously express ACE2. Using an MOI of 1, which provides up to 20% infection of HEK293T-ACE2-GFP cells, we detected infection of Calu3 and Caco2 cells at a very low percentage within the cultures (shown below). We note that most published data with these lines do not measure percent infection, but rather show infection via qPCR, which is not informative as to the number of cells infected within a culture.
Using higher virus doses could possibly give higher infections, but this is not possible given our virus stock titer. We note, however, that we provide data in Figure 1 with Caco2 cells in which concentrated Spike-pseudotyped virus was used to achieve a robust infection allowing us to measure effects of endogenous IFITMs in this relevant line. As shown above, the low infection rates of Caco2 and Calu3 cells with authentic SARS-CoV-2 precludes us from confidently examining roles of IFITMs in these lines. Instead, HEK293T-ACE2-GFP cells have provided an ideal model for us to dissect the opposing roles of IFITMs because 1) they are robustly infected by SARS-CoV-2, 2) the virus can utilize both plasma membrane and endocytic entry pathways in this line, and 3) we can manipulate the virus entry pathway in these cells for mechanistic studies by overexpression of TMPRSS2 and IFITMs. The dual effects of IFITMs on the entry of specific coronaviruses has been controversial and confusing in the field, particularly in the context of SARS-CoV-2 infections. Our results offer a clearer mechanistic understanding of how IFITM3 uses an amphipathicity-based mechanism to inhibit virus entry at endosomes while also enhancing plasma membrane entry in an amphipathicity-independent manner.
As an additional test of effects of endogenous IFITM proteins, we have now added data in which WT, IFITM3, and IFITM-locus deficient (IFITMdel) MEFs were transduced with hACE2 and infected with SARS-CoV-2 (New Figure 5). Compared to WT cells, we observed an increased infection in IFITMdel cells. Consistent with our Caco2 cell experiments in Figure  1, IFITM3 KO and broad IFITM deficiency both prevented type I IFN from fully inhibiting SARS- CoV-2 infections, overall indicating that IFITMs are generally restrictive of infection and that they are among the critical IFN effectors that limit SARS-CoV-2 infections.
3. The authors show that transfecting cells with IFITM3 can modestly inhibit syncytia formation when expressed in target (non-spike expressing cells) but that IFITM3 mutants enhanced syncytia formation in this context. Is this effect observed only when expressed in target cells? What is the result of co-expressing IFITM3 and mutants in SARS-CoV-2 spikeexpressing cells? What is the effect of expressing IFITM1 and IFITM2?
While these are interesting questions, we performed the syncytia assays specifically to have additional confirmation via a distinct assay that IFITM3 is able to enhance SARS-CoV-2 Spikemediated fusion at the plasma membrane. Indeed, this assay confirmed this ability of IFITM3 when expressed in target cells. As for roles of other IFITMs and expression of IFITMs in Spikeexpressing effector cells, we note that a full and comprehensive manuscript on these exact topics has been published as a preprint by the group of Dr. Olivier Schwartz (Pasteur Institute), demonstrating that an in-depth investigation of this topic could easily comprise a full manuscript and is outside the scope of our current study. 5. Human IFITM3-Y20A but not mouse IFITM3-Y20A increased infection compared to vector control cells. Do both mutants localize to the plasma membrane? This should be shown.
We note that both mouse and human IFITM3 have a conserved YxxF endocytosis motif involving Y20, and now make this clear in the manuscript text. Indeed, we previously showed that this motif regulates cellular localization of both mouse and human IFITM3 (Chesarino, JBC, 2014). For the current manuscript, we have added confocal imaging which shows plasma membrane localization for Y20A mutants from both species.
10th Nov 2020 1st Revision -Editorial Decision Dear Jacob and Alex, Thanks for sending me the revised manuscript. The study has now been seen by the original referees and their comments are provided below. As you can see from the comments the referees appreciate the introduced changes. I am therefore very happy to let you know that we will accept the manuscript for publication here.
Before sending you the formal acceptance letter there are just a few things to sort out.
-We need 3-5 keywords -We also need a data availability section. This is the place to enter accession numbers etc. As far as I can see no data is generated that needs to be deposited in a database. If this is correct please state: This study includes no data deposited in external repositories. Please place it after the Materials and methods and before Acknowledgements -I have asked our publisher to do their pre-publication checks on the paper. They will send me the file within the next few days. Please wait to upload the revised version until you have received their comments.
-We include a synopsis of the paper (see http://emboj.embopress.org/). Please provide me with a general summary statement and 3-5 bullet points that capture the key findings of the paper.
-We also need a summary figure for the synopsis. The size should be 550 wide by [200-400] high (pixels). You can also use something from the figures if that is easier.
That should be all -you can use the link below to submit the revised version.
Congratulations on a nice study.

With best wishes
Karin Karin Dumstrei, PhD Senior Editor The EMBO Journal Instructions for preparing your revised manuscript: Please check that the title and abstract of the manuscript are brief, yet explicit, even to nonspecialists.
When assembling figures, please refer to our figure preparation guideline in order to ensure proper formatting and readability in print as well as on screen: https://bit.ly/EMBOPressFigurePreparationGuideline IMPORTANT: When you send the revision we will require -a point-by-point response to the referees' comments, with a detailed description of the changes made (as a word file).
-a word file of the manuscript text.
-individual production quality figure files (one file per figure) -a complete author checklist, which you can download from our author guidelines (https://www.embopress.org/page/journal/14602075/authorguide). The reviewers have addressed my comments effectively. I'm right about the facs but its up to them how they present these data.The work is of good impact and will be of interest to a wide audience. The conclusions are justified. I have no further concerns, suggestions or comments.
Referee #2: In general, I believe this is an interesting and important study that is appropriate for the broad readership of this journal. The authors have addressed the majority of my points and have added sufficient new data and clarifications within the text to strengthen the manuscript. There is still an outstanding concern of the primary use of ACE2 over-expression cell lines. However, the data support the overall conclusions.
19th Nov 2020 2nd Revision -Editorial Decision Dear Jacob, Thank you for submitting your revised manuscript to The EMBO Journal. I have now had a chance to take a look at everything and all looks good. I am therefore very pleased to accept the manuscript for publication here. Please note that it is EMBO Journal policy for the transcript of the editorial process (containing referee reports and your response letter) to be published as an online supplement to each paper. If you do NOT want this, you will need to inform the Editorial Office via email immediately. More information is available here: https://emboj.embopress.org/about#Transparent_Process Your manuscript will be processed for publication in the journal by EMBO Press. Manuscripts in the PDF and electronic editions of The EMBO Journal will be copy edited, and you will be provided with page proofs prior to publication. Please note that supplementary information is not included in the proofs.
Should you be planning a Press Release on your article, please get in contact with embojournal@wiley.com as early as possible, in order to coordinate publication and release dates. No pre-specified effect size was chosen. Experiments with three bilogical replicates were performed a minimum of two times to achieve a minimum of six data points for statistical comparison. Most experiments were performed at least 3 times providing nine data points.

B-Statistics and general methods
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.
a statement of how many times the experiment shown was independently replicated in the laboratory.
Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
In the pink boxes below, please ensure that the answers to the following questions are reported in the manuscript itself. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable). We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects. ANOVA tests were used to evaluate signifcant differences in mean values for >2 groups.

Data
the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way. graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified the exact sample size (n) for each experimental group/condition, given as a number, not a range; Each figure caption should contain the following information, for each panel where they are relevant:

Captions
The data shown in figures should satisfy the following conditions: Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.
Please fill out these boxes ê (Do not worry if you cannot see all your text once you press return)