Abstract
The emergence of SARS-CoV-2 variants poses greater challenges to the control of COVID-19 pandemic. Here, we parallelly investigated three important characteristics of seven SARS-CoV-2 variants, including two mink-associated variants, the B.1.617.1 variant, and the four WHO-designated variants of concerns (B.1.1.7, B.1.351, P.1, and B.1.617.2). We first investigated the ability of these variants to bind and use animal ACE2 orthologs as entry receptor. We found that, in contrast to a prototype variant, the B.1.1.7, B.1.351, and P.1 variants had significantly enhanced affinities to cattle, pig, and mouse ACE2 proteins, suggesting increased susceptibility of these species to these SARS-CoV-2 variants. We then evaluated in vitro neutralization sensitivity of these variants to four monoclonal antibodies in clinical use. We observed that all the variants were partially or completely resistant against at least one of the four tested antibodies, with B.1.351 and P.1 showing significant resistance to three of them. As ACE2-Ig is a broad-spectrum anti-SARS-CoV-2 drug candidate, we then evaluated in vitro neutralization sensitivity of these variants to eight ACE2-Ig constructs previously described in three different studies. All the SARS-CoV-2 variants were efficiently neutralized by these ACE2-Ig constructs. Interestingly, compared to the prototype variant, most tested variants including the variants of concern B.1.1.7, B.1.351, P.1, and B.1.617.2 showed significantly increased (up to ∼15-fold) neutralization sensitivity to ACE2-Ig constructs that are not heavily mutated in the spike-binding interface of the soluble ACE2 domain, suggesting that SARS-CoV-2 evolves toward better utilizing ACE2, and that ACE2-Ig is an attractive drug candidate for coping with SARS-CoV-2 mutations.
Introduction
As of June 23rd 2021, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the etiological agent of the ongoing coronavirus disease 2019 (COVID-19) pandemic, has already caused about 178 million confirmed infections and over 3.8 million documented deaths over the world, according to World Health Organization (WHO) online updates. The pandemic has triggered unprecedentedly extensive worldwide efforts to develop countermeasures against COVID-19, and a number of encouraging progresses have been achieved in developing prophylactic vaccines and antibody therapeutics1-12. So far, there are more than six prophylactic COVID-19 vaccines that have been authorized by different countries for emergency use. These include two mRNA vaccines (Pfizer-BioNTech, US; Moderna, US)1,2, two inactivated vaccines (Sinopharm, China; Sinovac, China)3-5, and two adenoviral vectored vaccines (Sputnik V, Russia; AstraZeneca-Oxford, UK)6,7. Over 2 billion doses of these vaccines have been administered worldwide. There are also some convalescent patient-derived antibodies that have been authorized for emergency use by the US FDA, such as Regeneron’s antibody cocktail consisting of casirivimab (REGN10933) and imdevimab (REGN10987)8,9, and Eli Lilly’s antibody cocktail consisting of etesevimab (LY-CoV016) and bamlanivimab (LY-CoV555)10-12. So far, all the SARS-CoV-2 vaccines and antibody therapeutics in clinical use were developed on the basis of the prototype SARS-CoV-2 strain.
SARS-CoV-2 is a betacoronavirus that has broad host ranges13-16. Receptor usage is a critical determinant for the host range, as well as an effective neutralization target, of coronaviruses 16,17. SARS-CoV-2 utilizes ACE2 as a key cellular receptor to infect cells13,18,19. Antibodies targeting the interactions between ACE2 and SARS-CoV-2 spike receptor-binding domain (RBD) efficiently neutralize SARS-CoV-2 infection and reduced viral load in animal models and COVID-19 patients8-12. As human ACE2 residues on the spike RBD-binding interface are highly conserved across a number of mammalian-ACE2 orthologs, mutations within the spike RBD region might easily alter cross-species receptor usage by SARS-CoV-2, as well as sensitivity to SARS-CoV-2 neutralization agents. Indeed, we recently found that SARS-CoV-2 can use human ACE2 and a wide range of animal-ACE2 orthologs, but not mouse ACE2, for cell entry 14. But a single amino-acid change within the spike receptor-binding domain (RBD; Q498H, Q498Y, or N501Y) could be sufficient to confer SARS-CoV-2 the ability to utilize mouse ACE2ref.20-23.
SARS-CoV-2 is a single-stranded RNA virus with moderate mutation and recombination frequencies24,25. A number of spontaneous and selection-pressure-driven mutations of SARS-CoV-2 genome have been identified in viral variants that emerged during the course of the pandemic13,26-35. With more and more SARS-CoV-2 variants being identified to carry diverse spike mutations within the RBD region, it’s possible that some of the spike mutations might alter the host range of the virus, or compromise the efficacy of vaccines or neutralizing antibodies developed on the basis of the prototype SARS-CoV-2 strain. In this study, we investigated cross-species receptor usage of multiple circulating SARS-CoV-2 variants that emerged during the pandemic, as well as sensitivity of these variants to four neutralizing antibodies in clinical use and a broad-spectrum anti-SARS-CoV-2 drug candidate.
Results
Spike variants and related mutations investigated in this study
We first cloned twelve spike genes of nine SARS-CoV-2 natural variants. They include a spike gene of an early isolate WHU01ref.36, representing the prototype SARS-CoV-2 variant, a spike gene of the well-studied D614G variant35, two spike genes of a mink-associated variant that carries a Y453F signature mutation in the RBD region31,32,38, a spike gene of the B.1.617.1 variant, and seven spike genes of the four WHO-designated variants of concern (VOCs), B.1.1.7, B.1.351, P.1, and B.1.617.2ref.27-30 (Figure 1A). The VOC B.1.1.7, also called 501Y.V1, was first identified in the UK in late summer to early autumn 2020, then spread rapidly across the UK, and now has been identified in at least 114 countries39-41. B.1.1.7 has an N501Y signature mutation in the RBD region, and carries seven additional mutations in the remaining region of the spike protein27,28. Two spike sequences, one that only carries the N501Y mutation and the other that carries all the spike mutations, were used in the following studies (Figure 1A). The VOC B.1.351, also called 501Y.V2, was first detected in South Africa from samples collected in early August, then rapidly became dominant locally, and now has been identified in at least 68 countries. B.1.351 variant has three signature amino-acid substitutions (K417N-E484K-N501Y) in the RBD region and a number of additional mutations in the remaining regions of the spike protein29. Similarly, two spike sequences, one that only carries the K417N-E484K-N501Y mutations and the other that carries all the spike mutations, were used in the following studies (Figure 1A). The VOC P.1, also called 501Y.V3, emerged in Brazil around mid-November 2020, and now has rapidly spread to at least 37 countries. P.1 variant has three signature substitutions (K417T-E484K-N501Y) in the RBD region, similar to the RBD mutations of the B.1.351 variant. It also carries seven additional lineage-defining amino-acid substitutions in the remaining regions of the spike protein30. Again, two spike sequences, one that only carries the K417T-E484K-N501Y mutations and the other that carries all the spike mutations, were used in the following studies (Figure 1A). The VOC B.1.617.2, emerged in India around September 2020, and now has rapidly spread to over 75 countries. B.1.617.2 variant has two signature substitutions (L452R-T478K) in the RBD region. It also carries six additional lineage-defining amino-acid substitutions in the remaining regions of the spike protein. A spike sequence that carries all the B.1.617.2-signature mutations was used in the following studies (Figure 1A). All the residues associated with the above mentioned RBD mutations are indicated in the structure of human ACE2 in complex with the RBD of a prototype SARS-CoV-2 variant (Figure 1B).
Pseudoviruses of the B.1.1.7, B.1.351, and P.1 variants displayed patterns of animal-ACE2 tropism distinct from the patterns of other variants
To evaluate ACE2 ortholog-mediated viral entry of the above-mentioned SARS-CoV-2 variants, we produced luciferase reporter retroviruses pseudotyped with one of these different spike variants. These reporter pseudoviruses were used to infect 293T cells expressing each of eight ACE2 orthologs (Figure 2 and Table 1), including the human ACE2 and ACE2 orthologs of cattle (Bos taurus), pig (Sus scrofa domesticus), cat (Felis catus), rabbit (Oryctolagus cuniculus), bat (Rhinolophus sinicus isolate Rs-3357), rat (Rattus norvegicus), and mouse (Mus musculus). Parallel infection experiments using 293T cells transfected with an empty vector plasmid were included as controls. Consistent with our previous report14, the early isolate WHU01 efficiently infected 293T cells expressing most of the tested ACE2 orthologs except for that of rat and mouse (Figure 2A). While a similar pattern was also observed with the variants D614G, and Y453F (Figure 2B and C), distinct patterns were observed with the remaining variants (Figure 2D-L). Significant changes were observed with the VOCs B.1.1.7, B.1.351, and P.1, all of which showed ability to use ACE2 orthologs of rat and mouse for entry (Figure 2E-J). In addition, the variants B.1.351 and P.1 showed complete loss of the ability to use bat ACE2 for entry (Figure 2G-J). These data suggest that the mutations acquired by different variants might have changed SARS-CoV-2 RBD affinity to different ACE2 orthologs.
Y453F-RBD and the B.1.617.2-RBD have enhanced affinity to human ACE2
We then sought to quantitatively study the ability of the variants to utilize various animal orthologs of ACE2. The extracellular domains of the eight ACE2 orthologs and the RBD domains of seven spike variants, including WHU01, Y453F, B.1.1.7, B.1.351, P.1, B.1.617.1, and B.1.617.2, were expressed in 293F cells as immunoglobulin Fc fusion proteins. Purified ACE2 and RBD recombinant proteins were then used as immobilized receptors and analytes, respectively, in a bio-layer interferometry (BLI)-based assay to measure the interaction kinetics of fifty-six RBD-ACE2 pairs (Figure 3 and Table 2). When the RBD of early SARS-CoV-2 isolate WHU01 was tested against different ACE2 proteins, strong interactions were observed for human, cattle, pig, cat, and rabbit ACE2 proteins, with human ACE2 showing the highest affinity to this RBD. No any binding signal was detected for mouse ACE2 (Figure 3H). We then compared the kinetics data of each RBD variants to the data of the WHU01 variant. We observed that the Y453F mutation increased RBD affinity to human ACE2 by ∼4.7-fold (Figure 3A and Table 2), consistent with a very recent report42. In addition, the B.1.617.2 RBD but not the B.1.617.1 RBD was also found with >2-fold higher affinity to human ACE2 (Figure 3A and Table 2), consistent with the fast spreading of this variant across the world.
N501Y mutation enhances RBD affinity to most of the tested ACE2 proteins, especially cattle and pig ACE2 proteins
In the case of the B.1.1.7 RBD which has a N501Y signature mutation, we observed 2.9-fold increase in the RBD affinity to human ACE2 (Figure 3A and Table 2). This is consistent with the binding data from other reports and the increased infectivity of the B.1.1.7 variant39-41,43-45. Structural studies show that the Tyr501 residue of the B.1.1.7 RBD inserts into a cavity at the binding interface and forms a perpendicular π–π stacking interaction with the Tyr41residue of ACE2ref.45. Perhaps because of a same mechanism, the B.1.1.7 RBD also displayed significantly increased affinity to all the other tested ACE2 proteins, except for bat ACE2 that has a His41 rather than a Tyr41 (Figure 3B-H and Table 2). More pronounced affinity increase was observed with cattle ACE2 (5.0-fold) and pig ACE2 (6.1-fold), making the affinities of cattle and pig ACE2 proteins to B.1.1.7 RBD slightly higher than that of human ACE2 to WHU01 RBD (Figure 3B and C, and Table 2). Only very weak binding signals were detected for mouse ACE2, though the B.1.1.7 pseudovirus efficiently utilized mouse ACE2 for entry (Figure 2E, F, and 3H), reflecting the sensitivity difference of the two assays for discriminating weak and strong interactions.
The RBDs of B.1.351 and P.1 have increased affinities to cattle, pig, and mouse ACE2 proteins
In the case of the RBDs of B.1.351 and P.1 which share the E484K-N501Y signature mutations, almost identical patterns were observed (Figure 3). Compared to WHU01 RBD, B.1.351 and P.1 RBDs respectively also showed 2.9- and 2.7-fold higher affinities to cattle ACE2, and 3.4- and 2.9-fold higher affinities to pig ACE2 protein, possibly because of the shared N501Y mutation in these two variants (Figure 3B and C, and Table 2). On the other hand, consistent with the pseudovirus infection data (Figure 2G-J), the B.1.351 and P.1 RBDs showed a complete loss of interaction with bat ACE2, and a clear gain of interaction with mouse ACE2 (Figure 3F and H). Additional pseudovirus infection experiments showed that every single RBD mutation (K417N, K417T, E484K, or N501Y) of the B.1.351 and P.1 variants contribute to the utilization of mouse ACE2 as a functional receptor, with the N501Y mutation showed the most prominent effect (Figure 4). A Y41A single mutation in mouse ACE2 significantly impaired infectivity of all the single-spike-mutation pseudoviruses, but not the triple-spike-mutation pseudoviruses, while Y41A-H353A double mutation in mouse ACE2 significantly impaired infectivity of all the pseudoviruses, but much less pronounced effect was observed with the triple-spike-mutation pseudoviruses (Figure 4). These data suggest that the B.1.351 and P.1 RBDs gain binding to mouse ACE2 through the mutated RBD residues forming multiple new interactions with ACE2.
Spike mutations cause escape from potent neutralizing antibodies in clinical use
We then performed pseudovirus neutralization assays to evaluate sensitivity of these diverse SARS-CoV-2 variants to four therapeutic antibodies in clinical use (etesevimab/LY-CoV016, bamlanivimab/LY-CoV555, casirivimab/REGN10933, and imdevimab/REGN10987)8-12. While both the early strain WHU01 and the D614G variant were highly sensitive to all the four antibodies, all the RBD-mutated variants showed partial or complete resistance to at least one antibody (Figure 5). Specifically, both of the variants carrying a Y453F mutation showed strong resistance to REGN10933 (Figure 5C and D). The VOC B1.1.7 showed partial resistance to LY-CoV016 (Figure 5E and F). It is of note that the VOCs B.1.351 and P.1 showed strong resistance to REGN10933, and complete resistance to both LY-CoV016 and LY-CoV555, the two components of an antibody-cocktail therapy authorized for emergency use by the U.S. FDA10-12 (Figure 5G-J). The B.1.617.1 variant showed complete resistance to LY-CoV555 and partial resistance to REGN10933 (Figure 5K). Although the VOC B.1.617.2 also showed strong resistance to LY-CoV555, it’s sensitivity to LY-CoV016 was found significantly increased (Figure 5L). Consistent with these neutralization data, analyzing structural data of these antibodies in complex with the original RBD revealed contact of LY-CoV016 with RBD residues Lys417, LY-CoV555 with RBD residues Glu484 and Phe486, REGN10933 with RBD residues Lys417, Tyr453, Glu484, and Phe486, and REGN10987 with RBD residues Asn439 and Gln498 (Figure 5M-P).
The impact of RBD mutations on antibody affinity
We then performed BLI assays to quantitatively measure the kinetics of antibody-RBD interactions (Figure 6 and Table 3). The binding data are fully consistent with the pseudovirus neutralization data. Specifically, the mink-associated Y453F mutation decreased RBD affinity to REGN10933 (Figure 6C). The B.1.1.7-signature mutation N501Y decreased RBD binding to LY-CoV016 (Figure 6A). The B.1.351 RBD showed significantly decreased binding to REGN10933, and complete loss of interaction with LY-CoV016 and LY-CoV555 (Figure 6A-C). The P.1 RBD showed significantly decreased binding to REGN10933, weakly-detectable interaction with LY-CoV016, and complete loss of interaction with LY-CoV555 (Figure 6A-C). The B.1.617.1 RBD showed complete loss of interaction with LY-CoV555 (Figure 6B). Consistent with the neutralization data, the B.1.617.2 RBD showed significantly decreased affinity to LY-Cov555, but increased affinity to REGN10933 (Figure 5L, 6B, and 6C). These data demonstrate that SARS-CoV-2 variants could easily develop resistance to neutralization antibodies or even antibody cocktails in clinical use, highlighting the necessity of developing broad-spectrum anti-coronavirus agents.
Spike mutations change sensitivity of SARS-CoV-2 variants to different ACE2-Ig constructs
ACE2-Ig is a recombinant Fc fusion protein of soluble human ACE2. Recently we developed a panel of ACE2-Ig constructs that potently neutralize SARS-CoV-2 early isolate and three related but distinct coronaviruses, demonstrating that ACE2-Ig is a promising broadly anti-coronavirus drug candidate 14. Here we evaluated in vitro neutralization sensitivity of the twelve SARS-CoV-2 variants to eight representative ACE2-Ig constructs previously described in three different studies, including three constructs from our recent study (ACE2-Ig-v1, ACE2-Ig-v1.1, ACE2-Ig-v3)14, one construct from Chan et al (ACE2-Ig-Chan-v2.4)46, and four constructs from Glasgow et al (ACE2-Ig-Glasgow-293, ACE2-Ig-Glasgow-310, ACE2-Ig-Glasgow-311, ACE2-Ig-Glasgow-313)47. Because we previously found that the CLD domain of human ACE2 has an ∼20-fold contribution to ACE2-Ig’s neutralization activity against SARS-CoV-2 pseudoviruses14, we used CLD-containing soluble ACE2 domains for all the constructs tested in this study. We got three interesting findings here (Figure 7). First, all the twelve SARS-CoV-2 variants were potently neutralized by all the eight ACE2-Ig constructs, demonstrating that ACE2-Ig is a broad-spectrum anti-SARS-CoV-2 agent. Second, the mink-associated variant Y453F-F486L-N501T showed partial resistance to ACE2-Ig-Glasgow-310 and ACE2-Ig-Glasgow-313, but clearly increased sensitivity to ACE2-Ig-v1, ACE2-Ig-v1.1 and ACE2-Ig-v3, the constructs that are not heavily mutated in the spike-binding interface of the soluble ACE2 domain (Figure 7A-C, F, H). These data suggest that extensively mutating the ACE2 residues near the RBD-binding interface should be avoided. It might result in compromised neutralization breadth, not to mention the risk of eliciting anti-drug antibody (ADA) immune response. Third, compared to the early isolate WHU01, most circulating variants, including the four VOCs B.1.1.7, B.1.351, P.1, and B.1.617.2 showed significantly increased (up to ∼15-fold) neutralization sensitivity to ACE2-Ig-v1, ACE2-Ig-v1.1 and ACE2-Ig-v3, the constructs that carry a wild-type or a D30E-mutated ACE2 domain (Figure 7A-C).
Discussion
In this study, we parallelly investigated multiple circulating SARS-CoV-2 variants, including two mink-associated variants, the B.1.258 and B.1.617 variants, and the VOCs B.1.1.7, B.1.351, and P.1. As receptor usage is a critical determinant for the host range of coronaviruses, we first investigated the ability of these variants to utilize animal orthologs of ACE2 as an entry receptor. We found that, in contrast to an early isolate WHU01, the VOCs B.1.351, and P.1 were able to bind and use mouse ACE2 for entry, suggesting that these variants may have gained ability to infect mouse (Figure 2G-J and 3H). Consistent with our finding, during the preparation of this manuscript, an in vivo study reported that the VOCs B.1.351 and P.1 were able to infect mouse and replicate to high titers in the lungs48. Mice are vaccine-inaccessible rodent species that have large population size and could access the territories of both humans and domestic animals. The fact that the B.1.351 and P.1 variants have gained ability to infect mice raises the possibility of wild rodents becoming a second SARS-CoV-2 reservoir, adding one more concerning factor to these variants.
In addition, we found that multiple variants, especially the VOCs B.1.1.7, B.1.351, and P.1, have significantly increased affinity to cattle and pig ACE2 proteins (Figure 3 and Table 1). Multiple previous studies have shown that cattle and pigs are not or only weakly susceptible to experimental infection with SARS-CoV-215,49-52. Consistently, here we found that both cattle and pig ACE2 proteins have significantly lower affinities than that of human ACE2 to the RBD of an early isolate WHU01. However, in the case of the RBDs of the three VOCs, binding affinities of cattle and pig ACE2 proteins to these RBDs are comparable to that of human ACE2 to WHU01 RBD (Table 2). Cattle and pig ACE2 affinities to B.1.1.7 RBD are even slightly higher than that of human ACE2 to WHU01 RBD. It has been recently reported that the RBDs in the B.1.1.7 and B.1.351 spike proteins are more accessible than that of the early isolate43,44. This could further enable these variants to efficiently use cattle and pig ACE2 proteins. These data altogether suggest that cattle and pigs might now become susceptible to SARS-CoV-2. Cattle and pigs are extremely important livestock animals which serve as two major sources of meat for humans. It might be necessary to perform in vivo studies to re-evaluate the susceptibility of these species to SARS-CoV-2 variants, especially the widely circulating VOCs. On the other hand, the SARS-CoV-2 infected population is already in a huge size, with over 160 million confirmed infections worldwide. As of 29 March 2021, the VOC B.1.1.7 comprises roughly 95% of new SARS-CoV-2 infections in England. It has now been identified in at least 114 countries and exhibits a similar transmission increase (59 to 74%) in Denmark, Switzerland, and the United States39-41. So far, the VOCs B.1.351 and P.1 have also been identified in at least 68 and 37 countries, respectively. The enormous number of SARS-CoV-2 infection cases and high prevalence of these VOCs worldwide make it more and more necessary to closely monitor cattle and pigs for SARS-CoV-2 natural infections.
Our data about neutralization sensitivity of the SARS-CoV-2 variants to four monoclonal antibodies are consistent with multiple other studies53-58. Although, compared to the early isolate WHU01, all the tested variants carrying mutations in the RBD region showed partial or fully escape from at least one and up to three of the four tested antibodies (Figure 5 and 6), none of the variants escaped ACE2-Ig (Figure 7), a broad-spectrum anti-SARS-CoV-2 drug candidate. More importantly, we found that most of the tested SARS-CoV-2 variants had increased (up to ∼15-fold) sensitivity to ACE2-Ig constructs that carry a wild-type or a single-point-mutated ACE2 domain (Figure 7A-B). This might be explained by increased affinity of SARS-CoV-2 variant to human ACE2 or increased accessibility of the RBDs of SARS-CoV-2 variants. Indeed, we observed a 4.7-fold, 2.9-fold, 2.0-fold and 1.6-fold increase in the affinity of human ACE2 to the RBDs of the mink-associated Y453F variant, the B.1.1.7 variant, the B.1.351 variant, and the P.1 variant, respectively (Table 2). On the other hand, recent structural studies have shown that spike trimers of multiple SARS-CoV-2 variants, including a mink-associate variant and the VOCs B.1.1.7, B.1.1.28, and B1.351, have significantly higher propensity to adopt ‘RBD-up’ or open state than the D614G variant does43,44. These data suggest that SARS-CoV-2 was evolving toward better utilization of ACE2 as a receptor, either through increasing RBD affinity to ACE2, or through better exposure of its RBDs, or through both. It further indicates that ACE2 is still likely an essential receptor for SARS-CoV-2. ACE2 also serves as a cellular receptor for a number of other coronaviruses, including SARS-CoV, HCoV-NL63, Pangolin-CoV-2020, Bat-CoV RaTG13, and some other SARS-like CoVs found in bats14,16,59. ACE2-Ig is therefore a promising broadly anti-coronavirus drug candidate that might be used to treat and prevent infection of these diverse coronaviruses and their variants. It might also be a good alternative anti-SARS-CoV-2 agent for the populations who are not responsive to, or don’t have access to any prophylactic vaccines.
There are a couple of limitations of the current study. SARS-CoV-2 spike-pseudotyped reporter viruses, instead of live viruses, were used to study virus-ACE2 interaction and neutralization sensitivity of multiple circulating SARS-CoV-2 variants. It is difficult to obtain or access live SARS-CoV-2 variants, and even more difficult to simultaneously obtain multiple VOC live viruses and the mink-associated variant. Utilization of pseudotyped reporter virus to study virus-receptor interaction and neutralization is a well-established and widely used method in the field for studying various enveloped viruses, such as HIV60,61, influenza virus62-64, and coronaviruses including SARS-CoV-2ref.14,34,47,64-68. Findings about virus-receptor interaction and neutralization based on pseudovirus systems normally correlate well with that based on live viruses. In addition, we applied quantitative binding kinetics assays to validate all the key findings based on pseudovirus infection assays. This largely compensate the limitation of using pseudoviruses. On the other hand, ACE2-Ig is a promising broadly anti-SARS-CoV-2 drug candidate. More studies under in vivo conditions will be necessary to validate its potential for clinical applications. We have recently obtained a preliminary result showing that a single-dose injection or intranasal administration of an ACE2-Ig construct could significantly lower the viral load in the lungs of a COVID-19 mouse model (unpublished data). More mouse studies to test and optimize the protein’s in vivo pharmacokinetic properties and anti-SARS-CoV-2 efficacy are ongoing.
Materials and Methods
Cells
293T cells and HeLa cells were kindly provided by Stem Cell Bank, Chinese Academy of Sciences, confirmed mycoplasma-free by the provider, and maintained in Dulbecco’s Modified Eagle Medium (DMEM, Life Technologies) at 37 °C in a 5% CO2-humidified incubator. Growth medium was supplemented with 2 mM Glutamax-I (Gibco, Cat. No. 35050061), 100 µM non-essential amino acids (Gibco, Cat. No. 11140050), 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco, Cat. No. 15140122), and 10% heat-inactivated FBS (Gibco, Cat. No. 10099141C). HeLa-based stable cells expressing human ACE2 were maintained under the same culture condition as HeLa, except that 3 µg/mL of puromycin was added to the growth medium. 293F cells for recombinant protein production were generously provided by Dr. Yu J. Cao (School of Chemical Biology and Biotechnology, Peking University Shenzhen Graduate School) and maintained in SMM 293-TII serum-free medium (Sino Biological, Cat. No. M293TII) at 37 °C, 8% CO2, in a shaker incubator at 125 rpm.
Plasmids
DNA fragment encoding spike protein of SARS-CoV-2 WHU01 (GenBank: MN988668.1) was synthesized by the Beijing Genomic Institute (BGI, China) and then cloned into pCAGGS plasmid between EcoRI and XhoI restriction sites. Plasmids encoding SARS-CoV-2 spike variants were generated according to the in-fusion cloning protocol. To facilitate SARS-CoV-2 pseudovirus production, spike sequences for WHU01 and all the variants investigated in this study all contain a furin-cleavage site mutation (ΔPRRA) or a GSAS substitution of the PRRA furin-cleavage site. We had shown in our previous study that the ΔPRRA mutation does not affect SARS-CoV-2 cross-species receptor usage or neutralization sensitivity14. The retroviral reporter plasmids encoding a Gaussia luciferase reporter gene were constructed by cloning the reporter genes into the pQCXIP plasmid (Clontech). DNA fragments encoding C-terminally S-tagged ACE2 orthologs were synthesized in pUC57 backbone plasmid by Sangon Biotech (Shanghai, China). These fragments were then cloned into pQCXIP plasmid (Clontech) between SbfI and NotI restriction sites. Plasmids encoding recombinant RBD and soluble ACE2 variants were generated by cloning each of the gene fragments into a pCAGGS-based mouse-IgG2a or human IgG1 Fc fusion protein expression plasmid between NotI and BspEI sites. DNA fragments encoding heavy and light chains of anti-SARS-CoV-2 antibodies were synthesized by Sangon Biotech (Shanghai, China) and then cloned into a pCAGGS plasmid. Four antibodies (LY-CoV016, LY-CoV-555, REGN10933, and REGN10987)ref.8-12 that constitute two antibody cocktails authorized by the U.S. FDA for emergency use were included in this study.
Western Blot to detect S-tagged ACE2 (ACE2-S-tag) expression in 293T cells
293T cells at 30% density in each well of 96 well plates were reverse transfected with 0.60 µg of plasmid in complex with 0.15 µL of lipofectamine 2000 (Life Technologies, Cat. No. 11668019). Twenty-four hours after transfection, cells in each well were lysed with 40 µL lysis buffer and 5 µL of the lysate was used for western blot. ACE2-S-tag expression was detected by using a mouse anti-S-tag monoclonal antibody 6.2 (Invitrogen, Cat. No. MA1-981), and an HRP-conjugated goat anti-mouse IgG Fc secondary antibody (Invitrogen, Cat. No. 31437). Beta-actin was used as an internal control.
Production of reporter retroviruses pseudotyped with SARS-CoV-2 spike variants
MLV retroviral vector-based SARS-CoV-2 spike pseudotypes were produced according to our previous study14, with minor changes. In brief, 293T cells were seeded at 30% density in 150 mm dish at 12-15 hours before transfection. Cells were then transfected with 67.5 µg of polyethylenimine (PEI) Max 40,000 (Polysciences, Inc, Cat. No. 24765-1) in complex with 3.15 µg of plasmid encoding a spike variant, 15.75 µg of plasmid encoding murine leukemia virus (MLV) Gag and Pol proteins, and 15.75 µg of a pQCXIP-based luciferase reporter plasmid. Eight hours after transfection, cell culture medium was refreshed and changed to growth medium containing 2% FBS (Gibco, Cat. No. 10099141C) and 25 mM HEPES (Gibco, Cat. No. 15630080). Cell culture supernatants were collected at 36-48 hours post transfection, spun down at 3000×g for 10 min, and filtered through 0.45 µm filter units to remove cell debris. SARS-CoV-2 spike-pseudotyped viruses were then concentrated 10 times at 2000×g using 100 kDa cut-off Amicon Ultra-15 Centrifugal Filter Units (Millipore. Cat. No. UFC910024).
Pseudovirus Titration
Pseudovirus titer were determined by TCID50 followed a previous protocol69. Viruses were diluted 100 times as a working solution and then serially diluted in a ½ log10 manner. Human ACE2 expressed HeLa cells were infected with those diluted viruses in 96-well plates. Culture supernatants were refreshed every 12 hours and loaded to a Gaussia luciferase assay at 48 hours post infection. TCID50 of each pseudovirus was calculated by the Reed-Muench method.
Western Blot to detect spike protein (C9-tag) and MLV P30 protein in SARS-CoV-2 pseudovirus-containing supernatant
Pseudovirus-containing supernatants (8000 TCID50 of each pseudovirus) were mixed with SDS-PAGE loading buffer and boiled at 95°C for 10min. Spike protein (C9-tag) on the surface of pseudovirus was detected using a mouse anti-C9-tag monoclonal antibody 1D4 (Invitrogen, Cat. No. MA1-722), and an HRP-conjugated goat anti-mouse IgG Fc secondary antibody (Invitrogen, Cat. No. 31437). MLV P30 protein within the pseudovirus capsid was detected using a rabbit anti-MLV-P30 polyclonal antibody (Origene, Cat. No. AP33447PU-N) and an HRP-conjugated goat anti-rabbit IgG Fc secondary antibody (Invitrogen, Cat. No. 31463).
SARS-CoV-2 pseudovirus infection of 293T cells expressing ACE2 orthologs
Pseudovirus infection assay was performed according to our previous study14. In brief, 293T cells at 30% density in each well of gelatin (Millipore, ES-006-B) pre-coated 96-well plates were reverse transfected with 0.15 µL of lipofectamine 2000 (Life Technologies, Cat. No. 11668019) in complex with 60 ng of a vector control plasmid or a plasmid encoding ACE2 orthologs or ACE2 mutants. Twenty-four hours later, cells in each well were infected with 2000 TCID50 SARS-CoV-2 pseudovirus in 100 µL of culture medium containing 2% FBS (Gibco, Cat. No. 10099141C). Culture medium was refreshed every 12 hours. Cell culture supernatants were collected and subjected to a Gaussia luciferase assay at 48 hours post infection.
Gaussia luciferase luminescence flash assay
To measure Gaussia luciferase expression, 20 µL of cell culture supernatant of each sample and 100 µL of assay buffer containing 4 µM coelenterazine native (Biosynth Carbosynth, Cat. No. C-7001) were added to one well of a 96-well black opaque assay plate (Corning, Cat. No. 3915), and measured with Centro LB 960 microplate luminometer (Berthold Technologies) for 0.1 second/well.
Production and Purification of ACE2-Ig protein and SARS-CoV-2 antibodies
293F cells at the density of 6 × 105 cells/mL were seeded into 100 mL SMM 293-TII serum-free medium (Sino Biological, Cat. No. M293TII) one day before transfection. Cells were then transfected with 100 µg plasmid in complex with 250 µg PEI MAX 4000 (Polysciences, Inc, Cat. No. 24765-1). Cell culture supernatants were collected at 48 to 72 hours post transfection. Human IgG1 Fc-containing proteins were purified using Protein A Sepharose CL-4B (GE Healthcare, Cat. No. 17-0780-01), eluted with 0.1 M citric acid at pH 4.5 and neutralized with 1 M Tris-HCl at pH 9.0. Buffers were then exchanged to PBS and proteins were concentrated by 30 kDa cut-off Amicon Ultra-15 Centrifugal Filter Units (Millipore, Cat. No. UFC903096).
SARS-CoV-2 pseudovirus neutralization assay
Pseudovirus neutralization experiments were performed following our previous study14. In brief, SARS-CoV-2 spike variant-pseudotyped luciferase reporter viruses were pre-diluted in DMEM (2% FBS, heat-inactivated) containing titrated amounts of an ACE2-Ig construct or an anti-SARS-CoV-2 antibody. Virus-inhibitor mixtures were incubated at 37 °C for 30min, then added to HeLa-hACE2 cells in 96-well plates and incubated overnight at 37 °C. Virus-inhibitor-containing supernatant was then removed and changed with 150 µL of fresh DMEM (2% FBS) and incubated at 37 °C. Cell culture supernatants were collected for Gaussia luciferase assay at 48 h post infection.
Biolayer interferometry (BLI) assay
The BLI assays were performed on a Fortebio Octet RED384 instrument, with the temperature and shaking speed at 30 °C and 1000 rpm respectively. ACE2-hFc constructs were diluted to 5 μg/mL in 1x assay buffer containing 150 mM NaCl, 0.1% Tween-20, 10mM HEPES and 0.1%BSA (pH 7.4), and used as ligands for the assays. RBD-mFc constructs were serially diluted to 100nM, 50nM, 25 nM, 12.5 nM, and 6.25 nM in the 1x assay buffer. Each experiment group started with a 10 min warm-up for pre-hydration of AHC biosensors, followed by cycles of baseline (60 s), loading (60 s), baseline2 (60s), association (100 s), dissociation (600 s) and regeneration plus neutralization (30 s). A 1:1 Langmuir binding model was applied for data processing. All fitted diagrams (global fit) display the entire association window and the first 200 s (or 100 s only for house mouse and Chinese rufous horseshoe bat ACE2-related assays) of dissociation phase.
Data collection and analysis
All the experiments were independently performed for two or three times by two different people. Image Lab Software (Bio-Rad) was used to collect SDS-PAGE and Western-Blot image data. MikroWin 2000 Software (Berthold Technologies) was used to collect luciferase assay data. The BLI sensorgrams were recorded by Octet Data Acquisition 12.0 software and were fitted using Octet Data Analysis HT 12.0 software. GraphPad Prism 6.0 software was used for figure preparation and statistical analyses.
Data availability
The study did not generate unique datasets or code. Our research resources, including methods, plasmids, and protocols, are available upon reasonable request to qualified academic investigators for noncommercial research purposes. All reagents developed in this study, such as vector plasmids, as well as detailed methods, will be made available upon written request.
Conflict of interests
Shenzhen Bay Laboratory has filed a PCT patent application for multiple ACE2-Ig constructs.
Author contributions
G.Z. conceived and designed this study. W.Y., D.M., H.W., X.T., C.D., H.P., and Y.L. generated experimental materials. W.Y., D.M., and Y.L. performed all experiments. W.Y., D.M., Y.L., and G.Z. analyzed all data. C.L., H.L., M.F., and J.Z. contributed key resources. G.Z. and H.W. wrote the manuscript.
Acknowledgements
We thank the Biochemistry Core of the Shenzhen Bay Laboratory (Shenzhen, China) for providing help on performing the BLI-based binding kinetics assays and data analysis. We thank Dr. Yu J. Cao (School of Chemical Biology and Biotechnology, Peking University Shenzhen Graduate School, Shenzhen, China) for providing the 293F cells used in this study for the production of recombinant proteins. We thank Dr. Michael D. Alpert (Emmune, Inc., USA) for sharing useful comments on this manuscript.
This work was supported by Shenzhen Bay Laboratory Startup Funds (21230041, G.Z.), Major Program of Shenzhen Bay Laboratory (S201101001-2, G.Z.), Key COVID-19 Program of Shenzhen Bay Laboratory (S211410002, G.Z.), Department of Science and Technology of Guangdong province (2020B1111330001, J.Z.), and National Natural Science Foundation of China (82025001, J.Z.).
Footnotes
This version has the following four major revisions. First, it has more SARS-CoV-2 variants included in the study (Figure 1). Second, it contains significantly more infection and neutralization data (Figure 2 and 5). Third, new data on RBD-ACE2 and RBD-antibody interaction kinetics are included (Figure 3 and 6, Table 2 and 3). Fourth, eight ACE2-Ig constructs were parallelly compared for their neutralization activities against diverse SARS-CoV-2 variants (Figure 7).