key: cord-0742339-xgri8w1w
authors: Seifert, Stephanie N.; Bai, Shuangyi; Fawcett, Stephen; Norton, Elizabeth B.; Zwezdaryk, Kevin J.; Robinson, James; Gunn, Bronwyn; Letko, Michael C.
title: An ACE2-dependent Sarbecovirus in Russian bats is resistant to SARS-CoV-2 vaccines
date: 2022-04-01
journal: bioRxiv
DOI: 10.1101/2021.12.05.471310
sha: fdbe551309a145106ef04d485dcd87daae286b09
doc_id: 742339
cord_uid: xgri8w1w

Spillover of sarbecoviruses from animals to humans has resulted in outbreaks of severe acute respiratory syndrome SARS-CoVs and the ongoing COVID-19 pandemic. Efforts to identify the origins of SARS-CoV-1 and −2 has resulted in the discovery of numerous animal sarbecoviruses – the majority of which are only distantly related to known human pathogens and do not infect human cells. The receptor binding domain (RBD) on sarbecoviruses engages receptor molecules on the host cell and mediates cell invasion. Here, we tested the receptor tropism and serological cross reactivity for RBDs from two sarbecoviruses found in Russian horseshoe bats. While these two viruses are in a viral lineage distinct from SARS-CoV-1 and −2, one virus, Khosta-2, was capable of using human ACE2 to facilitate cell entry. Viral pseudotypes with a recombinant, SARS-CoV-2 spike encoding for the Khosta 2 RBD were resistant to both SARS-CoV-2 monoclonal antibodies and serum from individuals vaccinated for SARS-CoV-2. Our findings further demonstrate that sarbecoviruses circulating in wildlife outside of Asia also pose a threat to global health and ongoing vaccine campaigns against SARS-CoV-2 ONE SENTENCE SUMMARY European bat coronaviruses that are only distantly related to SARS-CoV-2 but use the same cell entry route, escape the immune response against SARS-CoV-2 vaccines, driving the need for broader vaccines.

Zoonotic spillover of sarbecoviruses from animals to humans has led to the 70 emergence of highly pathogenic human viruses, SARS-CoV-1 and -2, with the later 71 Converting Enzyme 2 (ACE2), whereas clade 2, also identified in Asian bats, contains 2 84 deletions and does not use ACE2 and clade 3 viruses, found more widely in Africa and 85

Europe, contain 1 deletion and some can infect using an unknown receptor, while other 86 clade 3 RBDs use ACE2(1-3). Recently, several viruses were identified in China that 87 comprise a fourth clade that also interact with human ACE2(4). 88

In late 2020, two clade 3 sarbecoviruses were identified in Rhinolophus bats in 89

Russia: Khosta-1 was found in Rhinolophus ferrumequinum and Khosta-2 in R. 90 hipposideros(5). Similar to other European and African clade 3 viruses, the Khosta viruses 91 viruses such as SARS-CoV and additionally vary in most of the residues known for clade 115 1 viruses to interact with ACE2 (1, 2, 7, 8) . 116 117 RBD from Khosta viruses mediate entry into human cells 118

Using our scalable Sarbecovirus RBD entry platform, we replaced the RBD from 119 SARS-CoV-1 spike with the Khosta RBDs and generated chimeric spike expression 120 plasmids ( Fig. 1B) (1). For comparison, we also included chimeric RBD spikes for other 121 clade 3 RBDs we have previously tested (BM48-31, Uganda, Rwanda) as well as SARS-122

CoV-2 and related RatG13 viruses. These chimeric spike expression constructs were used 123 to produce BSL2-compatible viral reporter pseudotypes with Vesicular Stomatitis Virus 124 expressing a dual GFP-luciferase reporter(1). All of the chimeric spike proteins expressed 125 to similar levels in mammalian cells and incorporated in Vesicular Stomatitis Virus (VSV). 126

Chimeric spike with the RBD from BM48-31 and RatG13 showed reduced incorporation 127 but this did not correlate with viral entry phenotypes observed in later experiments (Fig.  128 1C, D, E). 129

To test human cell compatibility, we first infected the human liver cell line, Huh-7, 131 with pseudotypes bearing the chimeric Khosta RBD spikes. In the absence of the addition 132 of an exogenous protease, trypsin, the pseudotypes exhibited almost no entry in these 133 cells, which has been observed for other sarbecoviruses and is attributed to low 134 endogenous expression of ACE2. However, when trypsin was included during the 135 infection, entry signal strongly increased for SARS-CoV-1 and -2 RBDs as well as the 136

Khosta RBDs (Fig. 1D ). As we and others have shown previously, trypsin enhancement 137 of sarbecovirus entry is still receptor dependent, suggesting that the Khosta virus RBDs 138 were using a receptor present in human cells to mediate infection (1, 9) . 139

The RBD from Khosta-2 infects cells using human ACE2 141

To characterize potential receptors for the Khosta viruses, we performed a classic 142 receptor tropism test, where we transfected Baby Hamster Kidney (BHK) cells with human 143 orthologues of known coronavirus receptors and then infected with our pseudotype panel. demonstrating these receptors were expressed to functional levels (Fig. 1e) . 156 157

While the RBD from Khosta 2 can use human ACE2 in functional assays ( Fig. 1)  159 and bind ACE2 as a purified protein fragment(6), other domains in spike vary between 160 the Khosta and SARS-CoV spikes. We had the full-length Khosta spike genes 161 synthesized, generated viral pseudotypes and tested their infectivity on human cells (Fig.  162 2). Similar to the chimeric SARS-CoV-based spikes, full-legnth Khsota spikes could also 163 infect Huh-7 cells in the presence of trypsin ( Fig. 2A ) and the Khosta 2 spike was capable 164 of infecting 293T cells expressing human ACE2 even in the absence of trypsin (Fig. 2B) . threat from the Khosta viruses, we generated VSV pseudotyped particles carrying a 176 chimeric SARS-CoV-2-based spike with the RBD from the Khosta viruses (Fig. 3A) . 177

Similar to our earlier SARS-CoV-based spikes, the SARS-CoV-2 chimeric spikes were 178 also infectious in 293T cells expressing human ACE2 (Fig. 3B ). To assess if the ACE2-179 dependent Khosta 2 RBD and SARS-CoV-2 RBD were cross-reactive, we incubated 180 pseudotyped particles with increasing amounts of the SARS-CoV-2 RBD-specific 181 monoclonal antibody, Bamlanivimab. Surprisingly, while SARS-CoV-2 spike was 182 effectively neutralized by the antibody, the SARS-CoV-2 spike with the Khosta 2 RBD 183 was completely resistant, suggesting little cross-reactivity between the two RBDs ( Fig.  184 3C). We repeated the pseudotype experiment using serum from vaccinated individuals 185 and saw a similar trend: the wild-type SARS-CoV-2 spike was easily inhibited by serum 186 from individuals who received either the Moderna or Pfizer vaccine, but the SARS-CoV-187 2-Khosta-2 RBD spike was resistant (Fig. 3D) . At higher dilutions of serum, there was a 188 reduction in the chimeric spike infectivity, but this was significantly less than the wildtype 189 spike at similar serum concentrations (Fig. 3E) . The Khosta 2 RBD shares approximately 190 60% similarity with various SARS-CoV-2 spikes at the amino acid level, which may 191 explain their low cross-reactivity (Fig. 3F) . Taken together, these results demonstrate that In the presence of trypsin, both Khosta-1 and -2 RBDs and spike were capable of 205 infecting human cells, with Khosta-1 performing notably stronger than Khosta-2, however 206 in our receptor-specific assay, only Khosta-2 could infect cells expressing human ACE2 207 without exogenous protease (Fig. 1D, 1E, Fig. 2) . We have previously shown that a small 208 number of clade 2 RBDs, such as As6526, also exhibit protease-mediated, ACE2-209 independent entry, and similar phenotypes have been described for other bat 210 coronaviruses (1, 25) . Because not all of the clade 2 and 3 viruses exhibit this phenotype, 211 these findings collectively suggest that some coronaviruses can infect human cells through 212 a presently unknown receptor. Sarbecoviruses have been shown to co-circulate in bats, 213 so this variation in receptor usage among closely related viruses may even represent an 214 evolutionary strategy for viral persistence within the reservoir host population (2). 215

viruses and one of the clade 2 viruses but do not include any members from clade 3(26, 217

RBDs that use human ACE2 (Fig. 3C ). More concerning was our observation that serum 219 from vaccinated individuals was significantly less effective at neutralizing pseudotypes 220 when just the SARS-CoV-2 RBD was replaced with the Khosta 2 RBD (Fig. 3D-E) . These 221 findings are not too surprising given that the Khosta 2 RBD only shares about 60% 222 sequence identity with SARS-CoV-2, and the serum we tested was from individuals 223 vaccinated only with RBD-specific vaccines from Moderna or Pfizer (Fig. 3E-F) . Curiously, 224

the Khosta 2 RBD is least similar to the currently circulating Omicron variant of SARS-225

CoV-2; with each new variant of concern decreasing in similarity to Khosta 2 (Fig. 3F) . 226

CoV-2 would be neutralized by serum from some individuals. 229

Our findings with chimeric, SARS-CoV-2 spike show that just replacing the RBD is 230 sufficient to reduce efficacy of SARS-CoV-2 spike-directed vaccines (Fig. 3) . However, 231 sarbecovirus recombination in nature typically occurs via template switching resulting in 232 acquisition of regions larger than the NTD (28). Thus, a naturally recombinant virus with 233

Khosta 2 may actually acquire more Khosta 2 spike, which as we show here with full 234 protein, is also infectious against human cells and ACE2 (Fig. 2) . Taken poly-glycine linker and cloned into pcDNA3.1+ as previously described(1). SARS-CoV-2 254 spike (MN997409.1) was codon optimized, modified to including silent cloning sites 255 flanking the RBD, and C-terminal 19 amino acid truncation was introduced to enhance 256 pseudotyping (1, 31) . The SARS-CoV-1 RBD was removed with KpnI and XhoI, and the 257 SARS-CoV-2 RBD was removed with BamHI and PflMI. Codon-optimized, synthesized 258 RBD fragments were cloned into the spike backbones as previously described(1). CoV-2 variants of concern 384 385 386 387  388  389  390  391  392  393  394  395  396  397  398  399  400  401  402  403  404 

Functional assessment of cell entry and receptor usage 406 for SARS-CoV-2 and other lineage B betacoronaviruses

Discovery of a rich gene pool of bat SARS-related coronaviruses provides 20

Transmission of SARS-CoV-2 on mink farms between humans 448 and mink and back to humans

Susceptibility of white-tailed deer (Odocoileus virginianus) to SARS-450 CoV-2

A serological survey of SARS-CoV-2 in cat in Wuhan

Genomic characterization of severe acute respiratory syndrome-454 related coronavirus in European bats and classification of coronaviruses based on partial 455 RNA-dependent RNA polymerase gene sequences

Complete Genome Sequence of a Severe Acute Respiratory Syndrome-457

Trypsin treatment unlocks barrier for zoonotic bat coronaviruses 459 infection

Chimeric spike mRNA vaccines protect against Sarbecovirus 461 challenge in mice

Broad sarbecovirus neutralization by a human monoclonal 463 antibody

Evolutionary origins of the SARS-CoV-2 sarbecovirus lineage responsible 465 for the COVID-19 pandemic

New algorithms and methods to estimate maximum-likelihood 467 phylogenies: assessing the performance of PhyML 3.0

SMS: Smart Model Selection in PhyML

Optimized Pseudotyping Conditions for the SARS-COV-2 Spike 471 Glycoprotein

Adaptive Evolution of MERS-CoV to Species Variation in DPP4