ACE2-independent sarbecovirus cell entry is supported by TMPRSS2-related enzymes and reduces sensitivity to antibody-mediated neutralization

  1. Lu Zhang
  2. Hsiu-Hsin Cheng
  3. Nadine Krüger
  4. Bojan Hörnich
  5. Luise Graichen
  6. Alexander S. Hahn
  7. Sebastian R. Schulz
  8. Hans-Martin Jäck
  9. Metodi V. Stankov
  10. Georg M.N. Behrens
  11. Marcel A. Müller
  12. Christian Drosten
  13. Onnen Mörer
  14. Martin Sebastian Winkler
  15. ZhaoHui Qian
  16. Stefan Pöhlmann
  17. Markus Hoffmann

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The COVID-19 pandemic, caused by SARS-CoV-2, demonstrated that zoonotic transmission of animal sarbecoviruses threatens human health but the determinants of transmission are incompletely understood. Here, we show that most spike (S) proteins of horseshoe bat and Malayan pangolin sarbecoviruses employ ACE2 for entry, with human and raccoon dog ACE2 exhibiting broad receptor activity. The insertion of a multibasic cleavage site into the S proteins increased entry into human lung cells driven by most S proteins tested, suggesting that acquisition of a multibasic cleavage site might increase infectivity of diverse animal sarbecoviruses for the human respiratory tract. In contrast, two bat sarbecovirus S proteins drove cell entry in an ACE2-independent, trypsin-dependent fashion and several ACE2-dependent S proteins could switch to the ACE2-independent entry pathway when exposed to trypsin. Several TMPRSS2-related cellular proteases but not the insertion of a multibasic cleavage site into the S protein allowed for ACE2-independent entry in the absence of trypsin and may support viral spread in the respiratory tract. Finally, the pan-sarbecovirus antibody S2H97 enhanced cell entry driven by two S proteins and this effect was reversed by trypsin. Similarly, plasma from quadruple vaccinated individuals neutralized entry driven by all S proteins studied, and use of the ACE2-independent, trypsin-dependent pathway reduced neutralization sensitivity. In sum, our study reports a pathway for entry into human cells that is ACE2-independent, supported by TMPRSS2-related proteases and associated with antibody evasion.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    Response to the reviewer's questions

    __Reviewer #1 (Evidence, reproducibility and clarity (Required)):

    Using the S protein from 14 different sarbecoviruses isolated from bats or pangolin, Zhang et al. makes in this manuscript several points on sabecovirus entry. These points include ACE2 independent entry, trypsin-driven entry, RBD-dependence of trypsin-mediated entry, use of soluble proteases and TRMPRSS-family transmembrane proteases in trypsin-mediated and trypsin-independent entry, and neutralizing antibody evasion in trypsin-mediated entry. Some of these points are supported by the data presented; although there are some discrepancies, they are largely within the range of experimental error. However, some of the statements in the Title, Abstract, and main text, appear to be more than what the data support. Nonetheless, the data authors presented are informative and will help understanding sarbecovirus entry processes. __

    Thank you very much for the positive assessment of our study and for the suggestions for improvement.__

    Major points:

    Below are only a few examples of inaccurate sentences. The authors should rewrite similar statements throughout the manuscript.

    Q1: The title: "ACE2-independent sarbecovirus cell entry is supported by TMPRSS2-related enzymes and reduces sensitivity to antibody-mediated neutralization" does not correctly reflect the presented data (1) because the contribution by TMPRSS2-like enzymes was shown only when they were co-transfected during PV production, but not when they are expressed on the target cell surface, and (2) because "reduces sensitivity to antibody-mediated neutralization" was observed only for one S protein but was not observed for the other two trypsin-dependent S proteins. In addition, this point was made using one monoclonal Ab for trypsin-dependent entry, but not for the entry mediated by TMPRSS2-related enzymes as the title implies. The title sounds like the three points are interconnected and represent general phenomena. Perhaps a more accurate title could be "ACE2-independent sarbecovirus cell entry is supported by trypsin and may reduce sensitivity to a neutralizing antibody". __

    __A1: We appreciate the critique. From our perspective, the statement that ACE2-independent entry is supported by TMPRSS2-related enzymes is correct irrespective of whether these enzymes cleave the viral S protein during entry into uninfected cells or during S protein biogenesis in infected cells (in order to allow for subsequent ACE2-independent entry into uninfected cells). The reviewer is correct that rescue from antibody-mediated neutralization was only observed for one monoclonal antibody. However, we also obtained evidence that ACE2-independent entry allowed for evasion of neutralizing antibodies induced upon infection or vaccination. In order to avoid generalization, we phrased the title in a more careful fashion: "ACE2-independent sarbecovirus cell entry can be supported by TMPRSS2-related enzymes and can reduce sensitivity to antibody-mediated neutralization".

    __

    __Q2: In the Abstract, the authors state "Several TMPRSS2-related cellular proteases but not the insertion of a multibasic cleavage site into the S protein allowed for ACE2-independent entry in the absence of trypsin and may support viral spread in the respiratory tract" (lines 38-41) and "In sum, our study reports a pathway for entry into human cells that is ACE2-independent, supported by TMPRSS2-related proteases...." (lines 44-46). These sentences should be rewritten for the same reason described above for the Title. __

    __A2: __We feel that the statement that TMPRSS2-related enzymes can support ACE2-independent entry is correct. Thus, either trypsin pretreatment of particles or expression of TMPRSS2-related enzymes in particle producing cells allows for ACE2-independent entry. We rephrased our concluding sentence in a more careful fashion and now state: "In sum, our study reports a pathway for entry into human cells that is ACE2-independent, can be supported by TMPRSS2-related proteases and may be associated with antibody evasion."

    __Q3: The lines 102-105 say "...ACE2-independent, trypsin-dependent entry can modulate neutralization by the pan sarbecovirus antibody S2H97..." and the lines 427-9 say "...trypsin-dependent usage of an ACE2-independent entry pathway may result in slightly reduced susceptibility to neutralization by antibodies induced upon infection or vaccination." Because Fig 8 (S2H97 Ab) and Fig 9 (immune plasma) use Vero-ACE2-TMPRSS2 and A549-ACE2-TMPRSS2, respectively, "ACE2-independent," is incorrect here. __

    __A3: __We respectfully disagree. We have shown that certain spikes can facilitate entry into ACE2-expressing cell lines in an ACE2-dependent manner but switch to an ACE2-independent entry route upon pre-treatment of particles with trypsin and blockade of ACE2 by an antibody (Supplementary Figure 4C). In figure 8 and 9, we show that when the ACE2-dependent entry route is blocked by neutralizing antibodies, opening the ACE2-independent route reduces antibody-mediated neutralization. As a consequence, it is fair to conclude that our data indicate that usage of the ACE2-independent entry route may reduce neutralization sensitivity. We feel that this argument is further supported by our most recent data, shown as new figure 3C, which demonstrate that trypsin treatment not only allows for entry into ACE2+ cells pretreated with anti-ACE2 antibody but, more importantly, also permits entry into ACE2 KO cells.

    __

    Q4: The line 46 says "...and associated with antibody evasion", the lines 104-5 says "...and allows for partial antibody evasion in the context of plasma from COVID-19 vaccinees." and the lines 427-9 say "...may result in slightly reduced susceptibility to neutralization by antibodies..." The authors should rewrite them because the resistance to S2H97 Ab was observed with one S protein but all other trypsin-mediated entry was sensitive to S2H97 or immune plasma. __

    __A4: __We have phrased the sentences in question in a more careful fashion and now state:

    "Finally, the pan-sarbecovirus antibody S2H97 enhanced cell entry driven by two S proteins and this effect was reversed by trypsin while trypsin protected entry driven by a third S protein from neutralization by S2H97. Similarly, plasma from quadruple vaccinated individuals neutralized entry driven by all S proteins studied, and availability of the ACE2-independent, trypsin-dependent pathway reduced neutralization sensitivity. In sum, our study reports a pathway for entry into human cells that is ACE2-independent, can be supported by TMPRSS2-related proteases and may be associated with antibody evasion." (Abstract)

    "Finally, we obtained evidence that ACE2-independent, trypsin-dependent entry can modulate neutralization by the pan sarbecovirus antibody S2H97 in a spike-dependent fashion and allows for partial antibody evasion in the context of plasma from COVID-19 vaccinees." (end of introduction).

    "In sum, these results suggest that availability of the trypsin-dependent, ACE2-independent entry pathway may result in slightly reduced susceptibility to neutralization by antibodies induced upon infection or vaccination." (end of results section).

    __

    Q5: If trypsin- independent entry is still controlled by RBD, why LYRa11 and Rs7327 entry is enhanced by and RsSHC014 entry is resistant to S2H97 Ab? The authors may want to discuss possible explanations. __

    __A5: __It is at present unclear why trypsin-treatment increased S2H97-mediated inhibition of LYRa11- and Rs7327-S protein driven entry while it conferred S2H97-resistance to RsSHC014-S. One could speculate that slight differences in the S2H97 epitope of the three spike proteins alter antibody affinity and thus determine whether the antibody enhances or blocks entry.

    __

    Q6: Fig. 2B. The entry supported by ACE2 orthologs was normalized to that utilizing hACE2 after hACE2-supported entry was normalized to background entry (no-S PV). First, it is unclear why background entry is used for normalization instead of being subtracted. Second, two times of such normalization likely created huge experimental errors and might have skewed the outcomes. Thus, 14 PVs should be quantified by RT-qPCR and same genome copy number should be used to directly assess their usage of ACE2 orthologs. This way, normalization by hACE2 entry is not necessary. Background entry should be subtracted, not used for normalization. __

    __A6: __We respectfully disagree. It is fair to ask how much more efficient single cycle particles bearing a viral envelope protein enter target cells as compared to identical particles bearing no viral glycoprotein. Normalization of the data presented as a heat map (Figure 2C) was performed based on the raw data (not the "Fold over Background"-normalized data). Thus, data were only normalized once. Regarding the possibility that different particle numbers were used for the respective pseudoviruses, we would like to state that particle production efficiency was analyzed by immunoblot (based on VSV matrix protein levels) and no major differences for the different pseudoviruses were observed (please see new Supplementary figure 4A). Thus, we are confident that our results are not skewed by gross differences in pseudovirus particle numbers.

    __

    Q7: Because VSV PVs were harvested in culture media, there were serum and divalent cations. Were PVs purified before trypsin digestion? Digestion by trypsin or other proteases should be described in detail. __

    __A7: __Medium without serum was used for PV production to avoid inhibition of trypsin activity by serum components. For immunoblot samples, VSV PVs were further harvested from the culture medium and concentrated using 20% sucrose. The concentrated VSV PVs were aliquoted into separate tubes, each containing an equal volume, and treated with the specified concentrations of proteases at 37{degree sign}C, as detailed in the Materials and Methods section. Subsequently, the treated VSV PVs were mixed with an equal volume of 2x SDS loading buffer and heated at 96{degree sign}C for 10 minutes.

    __

    Q8: How was S2' fragment on the blot determined? Should be described. __

    __A8: The S2' fragment was determined based on the molecular size of the corresponding bands. This information has been added to the respective figure legends.

    Minor points.

    Q9: The line 129 says "...14 S proteins, representing all clades, were selected for detailed analyses". Correct the sentence because the S protein representing clade 5 is not included in the study. __

    __A9: We now state ""...14 S proteins, representing all clades except clade 5, were selected for detailed analyses"

    __

    __Q10: Fig 2. Because 14 S proteins and several TFR1 orthologs were used, a table describing which S isolate is derived from which animal species will help. Organizing Fig 2A and B in the same order will help reading the result. Also, indicate which clades those S proteins belong to. __

    __A10: __We have added a table providing detailed information on the spike proteins under study.

    Supplemental table 1: Information on the spike proteins under study.

    Spike

    Virus

    Identifier

    RBD clade

    Host

    Region

    SARS-2-S

    Human SARS-CoV-2 hCoV-19/Wuhan/Hu-1/2019

    GISAID: EPI_ISL_402125

    1b

    Human (Homo sapiens)

    Asia (China)

    RaTG13-S

    Bat SARSr-CoV hCoV-19/bat/Yunnan/RaTG13/2013

    GISAID: EPI_ISL_402131

    1b

    Bat (Rhinolophus affinis)

    Asia (China)

    P5L-S

    Pangolin SARSr-CoV hCoV-19/pangolin/Guangxi/P5L/2017

    GISAID: EPI_ISL_410540

    1b

    Malayan pangolin (Manis javanica)

    Asia (China)

    cDNA8-S

    Pangolin SARSr-CoV hCoV-19/pangolin/Guangdong/cDNA8-S/2019

    GISAID: EPI_ISL_471461

    1b

    Malayan pangolin (Manis javanica)

    Asia (China)

    Rs4081-S

    Bat SARSr-CoV Rs4081

    GenBank: KY417143.1

    2

    Bat (Rhinolophus sinicus)

    Asia (China)

    Rs4237-S

    Bat SARSr-CoV RS4237

    GenBank: KY417147.1

    2

    Bat (Rhinolophus sinicus)

    Asia (China)

    SARS-1-S

    Human SARS-CoV-1/Frankfurt-1

    GenBank: AY291315.1

    1a

    Human (Homo sapiens)

    Europe (Germany)

    WIV1-S

    Bat SARSr-CoV WIV1

    GenBank: KF367457.1

    1a

    Bat (Rhinolophus sinicus)

    Asia (China)

    LYRa11-S

    Bat SARSr-CoV LYRa11

    GenBank: KF569996.1

    1a

    Bat (Rhinolophus affinis)

    Asia (China)

    RsSHC014-S

    Bat SARSr-CoV RsSHC014

    GenBank: KC881005.1

    1a

    Bat (Rhinolophus sinicus)

    Asia (China)

    Rs4231-S

    Bat SARSr-CoV Rs4231

    GenBank: KY417146.1

    1a

    Bat (Rhinolophus sinicus)

    Asia (China)

    Rs4874-S

    Bat SARSr-CoV Rs4874

    GenBank: KY417150.1

    1a

    Bat (Rhinolophus sinicus)

    Asia (China)

    Rs7327-S

    Bat SARSr-CoV Rs7327

    GenBank: KY417151.1

    1a

    Bat (Rhinolophus sinicus)

    Asia (China)

    BM48-31-S

    Bat SARSr-CoV BM48-31/BGR/2008

    GenBank: GU190215.1

    3

    Rhinolophus blasii

    Europe (Bulgaria)

    __

    Q11: Fig S5. Describe cell lines used. __

    __A11: __We have added a table providing information on the cell lines used.

    Supplemental table 2: Information on the cell lines used.

    Cell line

    Species

    Organ

    Modification

    Culture medium

    Vero

    African green monkey (Cercopithecus aethiops)

    Kidney

    n.a.

    DMEM + 10% FCS + Pen/Strep

    Vero-ACE2+TMPRSS2

    African green monkey (Cercopithecus aethiops)

    Kidney

    Stable expression of human ACE2 and human TMPRSS2

    DMEM + 10% FCS + Pen/Strep + Blasticidin (2 µg/ml) + Puromycin (1 µg/ml)

    Vero-TMPRSS2

    African green monkey (Cercopithecus aethiops)

    Kidney

    Stable expression of human TMPRSS2

    DMEM + 10% FCS + Pen/Strep + Blasticidin (2 µg/ml)

    MyDauLu/47

    Bat (Myotis daubentonii)

    Lung

    n.a.

    DMEM + 10% FCS + Pen/Strep

    PipNi/3

    Bat (Pipistrellus pipistrellus)

    Kidney

    n.a.

    DMEM + 10% FCS + Pen/Strep

    Caco-2

    Human (Homo sapiens)

    Intestine

    n.a.

    MEM + 10% FCS 1% NEA + 10 mM sodium pyruvate + Pen/Strep + Puromycin (1 µg/ml)

    293T

    Human (Homo sapiens)

    Kidney

    n.a.

    DMEM + 10% FCS + Pen/Strep

    293T-ACE2

    Human (Homo sapiens)

    Kidney

    Stable expression of human ACE2

    DMEM + 10% FCS + Pen/Strep + Puromycin (1 µg/ml)

    Huh-7

    Human (Homo sapiens)

    Liver

    n.a.

    DMEM + 10% FCS + Pen/Strep

    Li7

    Human (Homo sapiens)

    Liver

    n.a.

    DMEM + 10% FCS + Pen/Strep

    A549-ACE2

    Human (Homo sapiens)

    Lung

    Stable expression of human ACE2

    DMEM/F-12 + 10% FCS + Pen/Strep + Puromycin (1 µg/ml)

    A549-ACE2+TMPRSS2

    Human (Homo sapiens)

    Lung

    Stable expression of human ACE2 and human TMPRSS2

    DMEM/F-12 + 10% FCS + Pen/Strep + Blasticidin (2 µg/ml) + Puromycin (1 µg/ml)

    Calu-3

    Human (Homo sapiens)

    Lung

    n.a.

    DMEM/F-12 + 10% FCS 1% NEA + 10 mM sodium pyruvate + Pen/Strep

    Calu-3-ACE2

    Human (Homo sapiens)

    Lung

    Stable expression of human ACE2

    DMEM/F-12 + 10% FCS 1% NEA + 10 mM sodium pyruvate + Pen/Strep + Puromycin (1 µg/ml)

    NCI-H522

    Human (Homo sapiens)

    Lung

    n.a.

    RPMI + 10% FCS 1% NEA + 10 mM sodium pyruvate + Pen/Strep

    BHK-21

    Syrian golden hamster (Mesocricetus auratus)

    Kidney

    n.a.

    DMEM + 10% FCS + Pen/Strep

    __ Q12: Fig 3 legend should indicate trypsin digestion condition (concentration and length). __

    __A12: __We have added the requested information to the respective figure legends.

    __

    Reviewer #1 (Significance (Required)):

    Because overwhelming amount of data bear large experimental errors, there are some discrepancies among the data presented. Nonetheless, most of each point the authors claim is largely supported by the data. The problem happened when the authors tried to connect the dots too much and thus overstated some conclusions. If the overstated conclusions are amended throughout the manuscript, presented data provide sufficiently useful information on sarbecovirus entry.

    __

    Thank you. We have rephrased our conclusions in a more careful fashion.

    __Reviewer #2 (Evidence, reproducibility and clarity (Required)):____

    SUMMARY: Recent work from several groups has shown that the majority of bat sarbecoviruses infect cells independent of ACE2, the receptor primarily used by sarbecoviruses that infect humans, and instead infect cells in the presence of exogenous protease including trypsin. In this study, Zhang and colleagues build on these earlier findings by demonstrating that ACE2-independent sarbecovirus entry can be mediated by other exogenous proteases and several different TMPRSS11 enzymes. Using in vitro based methods and viral pseudotypes, the authors reproduce previous findings with trypsin, demonstrate similar effects with alternative proteases and provide lines of evidence suggesting (1) trypsin treatment can impart ACE2-independence and that (2) ACE2-independence provides resistance to neutralizing antibodies. __

    Many thanks for evaluation our manuscript and for the constructive critique.__

    MAJOR COMMENTS:

    Q1: Defining sarbecovirus RBDs into clades by in del features has already been established by other groups and many studies across different disciplines now use these previously-established clades. The authors use slightly different nomenclature without any acknowledgment of the previously defined sarbecovirus RBD clades, which will lead to confusion between studies. For example, SARS-CoV-2 is generally regarded as a clade 1 RBD (with ACE2 use and both loops in tact), clade 3 includes BM48-31 and Khosta-2, clade 4 includes RatG15. __

    __A1: __We have changed the nomenclature of the different groups to "clusters" to avoid confusion. Further, we added for each cluster information on the RBD clade. Please see revised Figure 1.

    __

    Q2: Why did the authors select BM48-31 as the representative of its clade when other members of the clade have known receptors and clear phenotypes in lab assays? BM48-31 has largely failed in every lab assay by every group that has studied it. On the other hand, Khosta2 uses human ACE2, BtKY72 and other African sarbecoviruses can also use ACE2 from their host species and have low but detectable human ACE2 compatibility. It would be interesting to see how the antibody-resistance results compare with other ACE2-dependent sarbecoviruses. __

    __A2: __We have selected BM48-31 at a time when the information stated above was not available. We agree that testing additional spikes for neutralization sensitivity should be considered within future studies but also feel that solid conclusions can be drawn from the 13 spikes tested within this study.

    __

    Q3: What is the aurthors' proposed mechanism for how protease is functioning for ACE2-independent entry? For ACE2-dependent entry, TMPRSS2 cleaves spike after RBD engagement. However, in this study, TMPRSS11 enzymes only function when included in producer cells- prior to RBD engagement. Is TMPRSS11 cleaving spike during spike biogenesis (similar to furin for SARS-CoV-2) or is an alternative mechanism at play? Is TMPRSS11 secreted? If this is the case, then the enzyme may be functioning similar to the other exogenous proteases in this study. __

    __A3: __It is possible that pre-cleavage by a TMPRSS2-like enzymes (or trypsin) is needed for subsequent S protein activation by another protease, likely cathepsin B/L, for ACE2-independent entry. This would be similar to SARS-CoV-2 entry into lung cells, which depends on spike pre-cleavage by furin and spike cleavage-activation by TMPRSS2. Alternatively, the TMPRSS2-like enzymes may cleave spike at the RBD, with the cleavage eluding detection by the methods applied here, and this cleavage might be needed for engagement of the so far unknown receptor responsible for ACE2-independent entry. TMPRSS2-like enzymes can be shed into the extracellular space. However, we feel that extracellular TMPRSS-activity was not responsible for ACE2-independent entry since expression of TMPRSS2-like enzymes in target cells should have also resulted in protease shedding but failed to allow for ACE2-independent entry.

    __

    Q4: Related to comment 3: the authors study trypsin as a pre-treatment, but other studies have shown trypsin exerts activity during entry. How do the authors propose trypsin is functioning prior to RBD engagement? Is it possible that trypsin is not fully inactivated and remains partially active during entry? __

    __A4: For most experiments, trypsin was present/active during the whole entry process. Only for Figures 3B, 8 and 9 trypsin inhibitor was added prior to inoculation of target cells in order to discriminate effects of trypsin on virus particles and cells and to exclude that trypsin compromised the integrity of the antibodies under study. We speculate that trypsin cleavage even before receptor engagement can allow for ACE2-independent entry.

    __

    __Q5: I am not convinced that trypsin is driving ACE2-independent entry for ACE2-dependent viruses. The experiment performed in figure 3C is performed in African green monkey cells using an antibody directed toward human ACE2. The difference in species between antibody and antigen may influence how well the antibody binds ACE2 on the Vero cells, which may only block some ACE2-dependent viruses but not all. Curiously, the only ACE2-dependent spikes that gain "ACE2-independence" are also activated by trypsin. These blocking assay results would be more convincing in a human cell line, or a non-permissive cell line like BHKs that express the human receptor. Alternatively, knocking out ACE2 in the Vero cells may be another way to assess ACE2-independent entry. __

    __A5: __We have now examined entry into 293T WT and 293T ACE2 KO cells. Importantly, the same spikes that allow for trypsin-dependent entry into Vero-TMPRSS2 cells treated with anti-ACE2 antibody also allow for robust entry into 293T ACE2 KO cells when pretreated with trypsin, please see new figure 3C. These results confirm our previous data and validate our conclusion that some spikes facilitate ACE2-dependent entry but can switch to the ACE2-independent entry route upon pre-treatment with trypsin.

    __

    MINOR COMMENTS:

    Q6: line 148: Rs4237 is missing a clade designation __

    __A6: __Rs4237 belongs to the Asian bat cluster (RBD clade 2). This information has been added to the revised figure 1 and is further provided in the new supplemental table 1.

    __ Q7: Figure 3. The figure's main message could be improved by visually grouping the viruses according to clade. __

    __A7: __We modified all figures and now indicate for each spike to which RBD clade they belong.

    __

    Q8: Some details are missing for reproducibility, including the accession numbers of the TMPRSS enzymes used in this study __

    __A8: __We added the requested information to the Materials and methods section.

    __

    Q9: Contrary to claims in the text, this study includes a fairly small panel of spike proteins. Prior studies by Letko 2020, Starr 2022 and Roelle 2022 (cited by the authors) all measured entry for between 20-40 spikes - more twice the number in this study. __

    __A9: We apologize for the mistake and removed the statement that "...these analyses were confined to small numbers of S proteins and.."

    __

    __Q10: Line 472-473: the data presented in figure 2B shows SARS-CoV-2 has slightly better entry with pangolin ACE2 than raccoon dog. I am not sure the authors should cite this data in support of raccoon dogs as an intermediate for SARS-CoV-2. __

    A10: We feel that our statement that - based on ACE2 usage - raccoon dogs should be considered as intermediate hosts is valid since it refers to the finding that diverse sarbecoviruses used this ACE2 orthologue with highest efficiency.

    __

    Reviewer #2 (Significance (Required)):

    SIGNIFICANCE: This study provides some novel insights into proteases and sarbecovirus cell entry and highlights previously unappreciated entry factors that are key for some viruses. A major limitation of this study is its lack of mechanistic exploration. The authors data do not really elucidate how TMPRSS11 proteins mediate ACE2-independent entry, nor do the results explain how ACE2-independence is shielding viruses from neutralizing antibodies. Another limitation is in the choice of using a non-human cell line to study the blocking effect of an antibody directed toward a human protein. __

    We feel that our findings that TMPRSS2-related enzymes can support ACE2-independent entry and that ACE2-independnet entry might allow for some level of antibody evasion are novel and important. We would also like to point out that we employed a human ACE2 KO cell line to address the reviewer's reservations regarding use of a non-human primate cell line. The data obtained with the human KO cell line confirmed those obtained with anti-ACE2 antibody treated non-human primate cell line, validating our conclusions.

    __

    ADVANCE: This study nicely reproduces a number of previous findings, including:

    1. sarbecovirus RBDs can be categorized into clades based on deletions in surface exposed loops
    2. ACE2-independent, trypsin-dependent sarbecovirus entry - notably for Rs4081
    3. the RBD in ACE2-independent sarbecoviruses controls entry
    4. anti-ACE2 antibodies do not block entry for ACE2-independent sarbecoviruses as well as some ACE2-dependent sarbecoviruses
    5. trypsin does not increase S proteins binding to cells
    6. protease expression in target cells does not increase S-driven entry
    7. a multi-basic cleavage site in spike does not compensate for exogenous protease in ACE2-independent entry

    This study has many novel advancements as well:

    1. identification of other exogenous proteases that mediate ACE2-independent entry (elastase, thermolysin)
    2. identification of TMPRSS11 family members that mediate trypsin-free entry for ACE2-independent viruses when produced in cells producing spike proteins but not target cells
    3. ACE2-independent entry may reduce spike susceptibility to antibody neutralization __

    Thank you.__

    AUDIENCE: This study will appeal to the coronavirus research community.

    __

    __Reviewer #3 (Evidence, reproducibility and clarity (Required)):____

    Zhang et al. analyzed the infection mechanisms of various Sarbecovirus primarily using VSV pseudoviruses with individual Sarbecovirus S proteins. The study demonstrated that many Sarbecoviruses, similar to two Sarbecoviruses that do not exhibit infectivity without trypsin, gain infectivity in human cells after processing virus particles with trypsin. This trypsin treatment is closely associated with the cleavage of the S1/S2 site of the S protein. This study demonstrated that the infection of the two viruses is not dependent on ACE2 expression, suggesting infection through receptors other than ACE2. Indeed, this study indicates that the receptor-binding domain of the S protein determines these properties. Furthermore, this study shows that some ACE2-using Sarbecoviruses also acquire ACE2-independent infectivity after trypsin treatment of virus particles. Although similar phenomena have already been reported in some Sarbecoviruses, the data in this study are more extensive, systematically conducted, and thoroughly analyzed, providing sufficient and additional evidence for the points mentioned above. The weaknesses, if pointed out, are that little progress has been made in elucidating the detailed molecular mechanism of this ACE2-independent and trypsin-dependent infection. __

    Thank you very much for reviewing our manuscript and for the positive comments.__

    Q1: To improve the study, the authors may consider the following points: • The Immunoblot data showing the expression level of ACE2-expressing cells used in the analysis of Figure 2 should be presented rather than indicated as "data not shown." __

    __A1: __The immunoblot data are now shown as new supplemental figure 3, panel B, and reveal robust expression of all ACE2 orthologues analyzed.

    __ Q2: In the explanation of Figure 2, it is stated, "all S proteins studied efficiently employed human ACE2 (lines 165-166)," but since there are significant differences in utilization levels, this description needs modification. Is it appropriate to normalize the utilization ability of human ACE2 as "1" in Figure 2B? Supplementary Figure 4 may be more relevant, and it should be considered to use it as a regular figure. __

    __A2: __We modified the text to indicate that although most spike proteins readily interacted with human ACE2, interaction efficacies greatly varied among the spike proteins ("*Thus, all S proteins studied employed human ACE2 for entry with the exception of the aforementioned S proteins of BM48-31, Rs4081 and Rs4237, which had also failed to bind to ACE2 (Figure 2B). However, although most sarbecovirus S proteins were able to readily utilize human ACE2 as an entry receptor, notable differences were observed. For instance, while *

    Particles bearing SARS-2-S, P5L-S, SARS-1-S, WIV-1-S, or Rs4874-S robustly entered BHK-21 cells expressing human ACE2, entry of particles carrying RaTG13-S, cDNA8-S, LYRa11-S, RsSHC014-S, Rs4231-S, or Rs7327-S was roughly 10- to 500-fold less efficient (Figure 2B).", see pages 7-8, lines 182-197). Further, we agree that Figure S4 contains important information for the reader and thus moved the data to main Figure 2 (as new panel B).

    __ Q3: It is concluded that Raccoon dog ACE2 is the most functional ACE2, but is it possible to quantitatively evaluate the level of difference in expression, which is challenging to adjust experimentally? It may be necessary to present data on expression levels or to pay attention to the interpretation of the data. __

    __A3: __The immunoblot data on ACE2 expression are now shown as new supplemental figure 3, panel B-C, and reveal roughly comparable expression of all ACE2 orthologues analyzed.

    __ Q4: No data are presented indicating the functionality of the BM48-31 S protein. While it is assumed that this S protein cannot function as a receptor, it cannot be denied that it may not be adequately expressed. __

    __A4: __Expression of all S proteins studied was readily detectable including BM48-31 S protein, although expression of P5L-S, cDNA8-S and BM4831-S was decreased. Please see new supplementary figure 4, panel A. Consequently, lack of cell entry by pseudoviruses bearing BM48-31-S may in fact be due to inefficient S protein incorporation into particles. This is now stated on page 8, lines 201-202.

    __ Q5: What is meant by "little impact" compared to what is mentioned? (line 306) __

    __A5: __We modified the text for clarity. The paragraph now states: "Expression of TMPRSS11A, TMPRSS11E and furin in cells producing SARS-1-S bearing particles as well as trypsin-treatment slightly improved generation of the S2 fragment (which results from cleavage at the S1/S2 site) (Figure 5E, left panel). Further, TMPRSS11D expression strongly increased production of the S2 fragment and the S2' fragment (which results from cleavage at the S2' site) while TMPRSS2 and TMPRSS13 expression and trypsin treatment only augmented production of the S2' fragment and decreased production of the S2 fragment (Figure 5E)." (please see page 14, lines 346-354).

    __

    __

    __Q6: Although VSV pseudoviruses are used to evaluate infectivity, in experiments using different conditions (e.g., Figure 5F), how is the amount of VSV pseudovirus for infection adjusted to a similar level? __

    __A6: __For infection of target cells, VSV pseudoviruses were normalized for volume. Immunoblot analysis revealed the particle preparations contained comparable amounts of VSV M protein, please see new supplemental figure 4, panel A.

    __ Q7: Citation of the paper. (lines 474-476) __

    __A7: __The requested citations have been inserted.

    __ Q8: What does "(-)" in Supplementary Figure 4 indicate? __

    __A8: __"(-) in former figure S4 (now Figure 2B) indicates empty vector. For clarity (and conformity with the other figures), we have changed the label to "No Spike".

    __ Q9: Is it appropriate to indicate the value of 'Pseudovirus Entry' with background fold ratio ('Fold over Background') in Figure 4B, etc (for example)? __

    __A9: __We feel that adding numerical values indicating the fold change ratios to our graphs would "overload" the figures and reduce clarity of the presented data.

    __ Reviewer #3 (Significance (Required)):

    This study is a comprehensive investigation into the function of the S protein of various Sarbecoviruses within the Coronaviridae family. The S protein is one of the most crucial proteins determining the infectivity of coronaviruses, and understanding the receptors and host cell proteases involved in cleaving the S protein is essential. The importance of furin and TMPRSS2 as proteases, and ACE2 as a receptor, has been clearly demonstrated in the infection of SARS-CoV-2, making them the foremost molecules to understand about SARS-CoV-2. However, in this study, the authors have clearly shown the existence of other significant modes of infection (independent of ACE2 and reliant on other proteases), thereby providing clear significance in this regard. Nevertheless, the current weakness, if point out, lies in the need for more depth of understanding of the specific molecular mechanisms underlying this novel mode of infection. __

    Thank you.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Zhang et al. analyzed the infection mechanisms of various Sarbecovirus primarily using VSV pseudoviruses with individual Sarbecovirus S proteins. The study demonstrated that many Sarbecoviruses, similar to two Sarbecoviruses that do not exhibit infectivity without trypsin, gain infectivity in human cells after processing virus particles with trypsin. This trypsin treatment is closely associated with the cleavage of the S1/S2 site of the S protein. This study demonstrated that the infection of the two viruses is not dependent on ACE2 expression, suggesting infection through receptors other than ACE2. Indeed, this study indicates that the receptor-binding domain of the S protein determines these properties. Furthermore, this study shows that some ACE2-using Sarbecoviruses also acquire ACE2-independent infectivity after trypsin treatment of virus particles. Although similar phenomena have already been reported in some Sarbecoviruses, the data in this study are more extensive, systematically conducted, and thoroughly analyzed, providing sufficient and additional evidence for the points mentioned above. The weaknesses, if pointed out, are that little progress has been made in elucidating the detailed molecular mechanism of this ACE2-independent and trypsin-dependent infection.

    To improve the study, the authors may consider the following points:

    • The Immunoblot data showing the expression level of ACE2-expressing cells used in the analysis of Figure 2 should be presented rather than indicated as "data not shown."
    • In the explanation of Figure 2, it is stated, "all S proteins studied efficiently employed human ACE2 (lines 165-166)," but since there are significant differences in utilization levels, this description needs modification. Is it appropriate to normalize the utilization ability of human ACE2 as "1" in Figure 2B? Supplementary Figure 4 may be more relevant, and it should be considered to use it as a regular figure.
    • It is concluded that Raccoon dog ACE2 is the most functional ACE2, but is it possible to quantitatively evaluate the level of difference in expression, which is challenging to adjust experimentally? It may be necessary to present data on expression levels or to pay attention to the interpretation of the data.
    • No data are presented indicating the functionality of the BM48-31 S protein. While it is assumed that this S protein cannot function as a receptor, it cannot be denied that it may not be adequately expressed.
    • What is meant by "little impact" compared to what is mentioned? (line 306)
    • Although VSV pseudoviruses are used to evaluate infectivity, in experiments using different conditions (e.g., Figure 5F), how is the amount of VSV pseudovirus for infection adjusted to a similar level?
    • Citation of the paper. (lines 474-476)
    • What does "(-)" in Supplementary Figure 4 indicate?
    • Is it appropriate to indicate the value of 'Pseudovirus Entry' with background fold ratio ('Fold over Background') in Figure 4B, etc (for example)?

    Significance

    This study is a comprehensive investigation into the function of the S protein of various Sarbecoviruses within the Coronaviridae family. The S protein is one of the most crucial proteins determining the infectivity of coronaviruses, and understanding the receptors and host cell proteases involved in cleaving the S protein is essential. The importance of furin and TMPRSS2 as proteases, and ACE2 as a receptor, has been clearly demonstrated in the infection of SARS-CoV-2, making them the foremost molecules to understand about SARS-CoV-2. However, in this study, the authors have clearly shown the existence of other significant modes of infection (independent of ACE2 and reliant on other proteases), thereby providing clear significance in this regard. Nevertheless, the current weakness, if point out, lies in the need for more depth of understanding of the specific molecular mechanisms underlying this novel mode of infection.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    Recent work from several groups has shown that the majority of bat sarbecoviruses infect cells independent of ACE2, the receptor primarily used by sarbecoviruses that infect humans, and instead infect cells in the presence of exogenous protease including trypsin. In this study, Zhang and colleagues build on these earlier findings by demonstrating that ACE2-independent sarbecovirus entry can be mediated by other exogenous proteases and several different TMPRSS11 enzymes. Using in vitro based methods and viral pseudotypes, the authors reproduce previous findings with trypsin, demonstrate similar effects with alternative proteases and provide lines of evidence suggesting (1) trypsin treatment can impart ACE2-independence and that (2) ACE2-independence provides resistance to neutralizing antibodies.

    Major comments:

    1. Defining sarbecovirus RBDs into clades by in del features has already been established by other groups and many studies across different disciplines now use these previously-established clades. The authors use slightly different nomenclature without any acknowledgment of the previously defined sarbecovirus RBD clades, which will lead to confusion between studies. For example, SARS-CoV-2 is generally regarded as a clade 1 RBD (with ACE2 use and both loops in tact), clade 3 includes BM48-31 and Khosta-2, clade 4 includes RatG15.
    2. Why did the authors select BM48-31 as the representative of its clade when other members of the clade have known receptors and clear phenotypes in lab assays? BM48-31 has largely failed in every lab assay by every group that has studied it. On the other hand, Khosta2 uses human ACE2, BtKY72 and other African sarbecoviruses can also use ACE2 from their host species and have low but detectable human ACE2 compatibility. It would be interesting to see how the antibody-resistance results compare with other ACE2-dependent sarbecoviruses.
    3. What is the aurthors' proposed mechanism for how protease is functioning for ACE2-independent entry? For ACE2-dependent entry, TMPRSS2 cleaves spike after RBD engagement. However, in this study, TMPRSS11 enzymes only function when included in producer cells- prior to RBD engagement. Is TMPRSS11 cleaving spike during spike biogenesis (similar to furin for SARS-CoV-2) or is an alternative mechanism at play? Is TMPRSS11 secreted? If this is the case, then the enzyme may be functioning similar to the other exogenous proteases in this study.
    4. Related to comment 3: the authors study trypsin as a pre-treatment, but other studies have shown trypsin exerts activity during entry. How do the authors propose trypsin is functioning prior to RBD engagement? Is it possible that trypsin is not fully inactivated and remains partially active during entry?
    5. I am not convinced that trypsin is driving ACE2-independent entry for ACE2-dependent viruses. The experiment performed in figure 3C is performed in African green monkey cells using an antibody directed toward human ACE2. The difference in species between antibody and antigen may influence how well the antibody binds ACE2 on the Vero cells, which may only block some ACE2-dependent viruses but not all. Curiously, the only ACE2-dependent spikes that gain "ACE2-independence" are also activated by trypsin. These blocking assay results would be more convincing in a human cell line, or a non-permissive cell line like BHKs that express the human receptor. Alternatively, knocking out ACE2 in the Vero cells may be another way to assess ACE2-independent entry.

    Minor comments:

    1. line 148: Rs4237 is missing a clade designation
    2. Figure 3. The figure's main message could be improved by visually grouping the viruses according to clade.
    3. Some details are missing for reproducibility, including the accession numbers of the TMPRSS enzymes used in this study
    4. Contrary to claims in the text, this study includes a fairly small panel of spike proteins. Prior studies by Letko 2020, Starr 2022 and Roelle 2022 (cited by the authors) all measured entry for between 20-40 spikes - more twice the number in this study.
    5. Line 472-473: the data presented in figure 2B shows SARS-CoV-2 has slightly better entry with pangolin ACE2 than raccoon dog. I am not sure the authors should cite this data in support of raccoon dogs as an intermediate for SARS-CoV-2.

    Significance

    This study provides some novel insights into proteases and sarbecovirus cell entry and highlights previously unappreciated entry factors that are key for some viruses. A major limitation of this study is its lack of mechanistic exploration. The authors data do not really elucidate how TMPRSS11 proteins mediate ACE2-independent entry, nor do the results explain how ACE2-independence is shielding viruses from neutralizing antibodies. Another limitation is in the choice of using a non-human cell line to study the blocking effect of an antibody directed toward a human protein.

    Advance

    This study nicely reproduces a number of previous findings, including:

    1. sarbecovirus RBDs can be categorized into clades based on deletions in surface exposed loops
    2. ACE2-independent, trypsin-dependent sarbecovirus entry - notably for Rs4081
    3. the RBD in ACE2-independent sarbecoviruses controls entry
    4. anti-ACE2 antibodies do not block entry for ACE2-independent sarbecoviruses as well as some ACE2-dependent sarbecoviruses
    5. trypsin does not increase S proteins binding to cells
    6. protease expression in target cells does not increase S-driven entry
    7. a multi-basic cleavage site in spike does not compensate for exogenous protease in ACE2-independent entry

    This study has many novel advancements as well:

    1. identification of other exogenous proteases that mediate ACE2-independent entry (elastase, thermolysin)
    2. identification of TMPRSS11 family members that mediate trypsin-free entry for ACE2-independent viruses when produced in cells producing spike proteins but not target cells
    3. ACE2-independent entry may reduce spike susceptibility to antibody neutralization

    Audience:

    This study will appeal to the coronavirus research community.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Using the S protein from 14 different sarbecoviruses isolated from bats or pangolin, Zhang et al. makes in this manuscript several points on sabecovirus entry. These points include ACE2 independent entry, trypsin-driven entry, RBD-dependence of trypsin-mediated entry, use of soluble proteases and TRMPRSS-family transmembrane proteases in trypsin-mediated and trypsin-independent entry, and neutralizing antibody evasion in trypsin-mediated entry. Some of these points are supported by the data presented; although there are some discrepancies, they are largely within the range of experimental error. However, some of the statements in the Title, Abstract, and main text, appear to be more than what the data support. Nonetheless, the data authors presented are informative and will help understanding sarbecovirus entry processes.

    Major points:

    Below are only a few examples of inaccurate sentences. The authors should rewrite similar statements throughout the manuscript.

    1. The title: "ACE2-independent sarbecovirus cell entry is supported by TMPRSS2-related enzymes and reduces sensitivity to antibody-mediated neutralization" does not correctly reflect the presented data (1) because the contribution by TMPRSS2-like enzymes was shown only when they were co-transfected during PV production, but not when they are expressed on the target cell surface, and (2) because "reduces sensitivity to antibody-mediated neutralization" was observed only for one S protein but was not observed for the other two trypsin-dependent S proteins. In addition, this point was made using one monoclonal Ab for trypsin-dependent entry, but not for the entry mediated by TMPRSS2-related enzymes as the title implies. The title sounds like the three points are interconnected and represent general phenomena. Perhaps a more accurate title could be "ACE2-independent sarbecovirus cell entry is supported by trypsin and may reduce sensitivity to a neutralizing antibody".

    In the Abstract, the authors state "Several TMPRSS2-related cellular proteases but not the insertion of a multibasic cleavage site into the S protein allowed for ACE2-independent entry in the absence of trypsin and may support viral spread in the respiratory tract" (lines 38-41) and "In sum, our study reports a pathway for entry into human cells that is ACE2-independent, supported by TMPRSS2-related proteases...." (lines 44-46). These sentences should be rewritten for the same reason described above for the Title.

    1. The lines 102-105 say "...ACE2-independent, trypsin-dependent entry can modulate neutralization by the pan sarbecovirus antibody S2H97..." and the lines 427-9 say "...trypsin-dependent usage of an ACE2-independent entry pathway may result in slightly reduced susceptibility to neutralization by antibodies induced upon infection or vaccination." Because Fig 8 (S2H97 Ab) and Fig 9 (immune plasma) use Vero-ACE2-TMPRSS2 and A549-ACE2-TMPRSS2, respectively, "ACE2-independent," is incorrect here.

    The line 46 says "...and associated with antibody evasion", the lines 104-5 says "...and allows for partial antibody evasion in the context of plasma from COVID-19 vaccinees." and the lines 427-9 say "...may result in slightly reduced susceptibility to neutralization by antibodies..." The authors should rewrite them because the resistance to S2H97 Ab was observed with one S protein but all other trypsin-mediated entry was sensitive to S2H97 or immune plasma.

    1. If trypsin- independent entry is still controlled by RBD, why LYRa11 and Rs7327 entry is enhanced by and RsSHC014 entry is resistant to S2H97 Ab? The authors may want to discuss possible explanations.
    2. Fig. 2B. The entry supported by ACE2 orthologs was normalized to that utilizing hACE2 after hACE2-supported entry was normalized to background entry (no-S PV). First, it is unclear why background entry is used for normalization instead of being subtracted. Second, two times of such normalization likely created huge experimental errors and might have skewed the outcomes. Thus, 14 PVs should be quantified by RT-qPCR and same genome copy number should be used to directly assess their usage of ACE2 orthologs. This way, normalization by hACE2 entry is not necessary. Background entry should be subtracted, not used for normalization.
    3. Because VSV PVs were harvested in culture media, there were serum and divalent cations. Were PVs purified before trypsin digestion? Digestion by trypsin or other proteases should be described in detail.
    4. How was S2' fragment on the blot determined? Should be described.

    Minor points.

    1. The line 129 says "...14 S proteins, representing all clades, were selected for detailed analyses". Correct the sentence because the S protein representing clade 5 is not included in the study.
    2. Fig 2. Because 14 S proteins and several TFR1 orthologs were used, a table describing which S isolate is derived from which animal species will help. Organizing Fig 2A and B in the same order will help reading the result. Also, indicate which clades those S proteins belong to.
    3. Fig S5. Describe cell lines used.
    4. Fig 3 legend should indicate trypsin digestion condition (concentration and length).

    Significance

    Because overwhelming amount of data bear large experimental errors, there are some discrepancies among the data presented. Nonetheless, most of each point the authors claim is largely supported by the data. The problem happened when the authors tried to connect the dots too much and thus overstated some conclusions. If the overstated conclusions are amended throughout the manuscript, presented data provide sufficiently useful information on sarbecovirus entry.

  5. Version published to 10.1101/2024.04.18.590061v1 on bioRxiv