Inhibition of SARS-CoV-2 infection (previously 2019-nCoV) by a highly potent pan-coronavirus fusion inhibitor targeting its spike protein that harbors a high capacity to mediate membrane fusion

The recent outbreak of coronavirus disease (COVID-19) caused by SARS-CoV-2 infection in Wuhan, China has posed a serious threat to global public health. To develop specific anti-coronavirus therapeutics and prophylactics, the molecular mechanism that underlies viral infection must first be confirmed. Therefore, we herein used a SARS-CoV-2 spike (S) protein-mediated cell-cell fusion assay and found that SARS-CoV-2 showed plasma membrane fusion capacity superior to that of SARS-CoV. We solved the X-ray crystal structure of six-helical bundle (6-HB) core of the HR1 and HR2 domains in SARS-CoV-2 S protein S2 subunit, revealing that several mutated amino acid residues in the HR1 domain may be associated with enhanced interactions with HR2 domain. We previously developed a pan-coronavirus fusion inhibitor, EK1, which targeted HR1 domain and could inhibit infection by divergent human coronaviruses tested, including SARS-CoV and MERS-CoV. We then generated a series of lipopeptides and found that the EK1C4 was the most potent fusion inhibitor against SARS-CoV-2 S protein-mediated membrane fusion and pseudovirus infection with IC50s of 1.3 and 15.8 nM, about 241- and 149-fold more potent than that of EK1 peptide, respectively. EK1C4 was also highly effective against membrane fusion and infection of other human coronavirus pseudoviruses tested, including SARS-CoV and MERS-CoV, as well as SARSr-CoVs, potently inhibiting replication of 4 live human coronaviruses, including SARS-CoV-2. Intranasal application of EK1C4 before or after challenge with HCoV-OC43 protected mice from infection, suggesting that EK1C4 could be used for prevention and treatment of infection by currently circulating SARS-CoV-2 and emerging SARSr-CoVs.


55
In April of 2018, the World Health Organization (WHO) established a priority 56 list of pathogens, including Middle East respiratory syndrome (MERS), severe acute 57 respiratory syndrome (SARS) and Disease X, a disease with an epidemic or pandemic 58 potential caused by an unknown pathogen 1,2 (Fig.1a). 59 In late December 2019, an outbreak of pneumonia with an unknown etiology in 60 Wuhan, China was considered as the first Disease X following the announcement by 61 WHO. Shortly thereafter, a novel coronavirus, 2019-nCoV, as denoted by WHO 3 , 62 was identified as the pathogen causing the coronavirus disease COVID-19 4,5 . 63 2019-nCoV with 79.5% and 96% sequence identity to SARS-CoV and a bat 64 coronavirus, SL-CoV-RaTG13, respectively 6 , was renamed SARS-CoV-2 by the which mediates membrane fusion, has 89.8% sequence identity and 96.9% sequence 83 similarity to those of SARS-CoV, and both of them utilize human 84 angiotensin-converting enzyme 2 (hACE2) as the receptor to infect human cells 6 . 85 Most importantly, the ACE2-binding affinity of the receptor-binding domain (RBD) in 86 S1 subunit of S protein of SARS-CoV-2 is 10-to 20-fold higher than that of 87 SARS-CoV 12 , which may contribute to the higher infectivity and transmissibility of 88 SARS-CoV-2 compared to SARS-CoV. However, it is unclear whether SARS-CoV-2 89 can mediate membrane fusion in a manner that exceeds the capacity of SARS-CoV. 90 After binding of RBD in S1 subunit of S protein on the virion to the ACE2 91 receptor on the target cell, the heptad repeat 1 (HR1) and 2 (HR2) domains in its S2 92 subunit of S protein interact with each other to form a six-helix bundle (6-HB) fusion 93 core, bringing viral and cellular membranes into close proximity for fusion and 94 infection 13 . Therefore, the 6-HB fusion core structure of SARS-CoV-2 and 95 SARS-CoV S proteins should also be compared in order to investigate the structural 96 basis for membrane fusion mediated by their S proteins and thus set the stage for the 97 rational design of coronavirus fusion inhibitors. 98 In our previous studies, we designed a pan-coronavirus fusion inhibitor, EK1, 99 targeting the HR1 domains of HCoV S proteins, which proved to be effective in  In this study, we have shown that SARS-CoV-2 exhibits much higher capacity of 108 membrane fusion than SARS-CoV, suggesting that the fusion machinery of 109 SARS-CoV-2 is an important target for development of coronavirus fusion inhibitors. 110 SARS-CoV-2 infection, we found a typical syncytium phenomenon naturally formed 139 by infected cells, which is rarely reported in SARS-CoV infection (Fig. 1c). To 140 further explore the special characteristic of SARS-CoV-2 infection, we cloned the S 141 gene into PAAV-IRES-GFP vector and established the S-mediated cell-cell fusion 142 system, using 293T cells that express SARS-CoV-2 S protein and EGFP 143 (293T/SARS-CoV-2/EGFP) as the effector cells, and ACE2/293T cells expressing 144 human ACE2 receptor as the target cells ( Fig. 1d and Fig. S1a). After effector cells 145 and target cells were cocultured at 37 °C for 2 h, the fused cells showed at least 2-fold 146 larger size than normal cells and multiple nuclei, and these cells were observed in the 147 SARS-CoV-2 group, but not the SARS-CoV group. After coincubation for 24 h, 148 hundreds of target cells fused together as one big syncytium, containing multiple 149 nuclei (Fig. 1d). Another 24h later, the syncytium grew bigger and could be easily  Previously, we identified that the 6-HB formed by HR1 and HR2 domains of the 166 S2 subunit plays a very important role in the membrane fusion process mediated by 167 MERS-CoV or SARS-CoV S protein 16,17 . Similarly, our recent study suggested that 168 HR1 and HR2 in subunit S2 of SARS-CoV-2 also interacted to form coiled-coil 169 complex to support membrane fusion and viral infection 15 ( Fig. 2a and Fig. S2).

170
However, the specific binding characteristics of SARS-CoV-2 6-HB remained to be 171 explored.

186
The overall 6-HB structure of SARS-CoV-2 is similar to that of other HCoVs 187 with root-mean-square deviation (RMSD) of 0.36 Å to SARS-CoV 6-HB and 0.66 Å 188 to MERS-CoV 6-HB for all the Cα atoms (Fig. 2d). This finding suggested that the 189 overall 6-HB conformation is an important and highly conserved component for these through a weak salt bridge 3.7 Å in distance. In SARS-CoV-2, E918 is mutated to 212 D936 and binds to R1185 in the HR2 domain through a salt bridge 2.7 Å in distance 213 (Fig. 3c). In SARS-CoV, K929 in HR1 binds to E1163 in HR2 through a salt bridge 214 3.2 Å in distance, while T925 is not involved in the interaction. However, when T925 215 was mutated to S943, it could bind to E1182 in the HR2 domain with a hydrogen 216 bond 2.6 Å in distance, and K947 could also bind to E1182 through a salt bridge 3.0 217 Å in distance (Fig. 3d). These results suggested that the multiple replacements in the 218 HR1 domain of emerging SARS-CoV-2 virus could enhance the interactions between 219 HR1 and HR2 domain to further stabilize the 6-HB structure, which may lead to 220 increased infectivity of the virus.

243
Strikingly, EK1C3 peptide with both the 3-amino acid linker "GSG" and the 244 PEG4-based spacer, exhibited 4-fold more potency than EK1C1. It is noteworthy that 245 changing "GSG" in EK1C3 to a longer 5-amino acid linker "GSGSG" significantly 246 increased the inhibitory potency of the hybrid molecule, and EK1C4 had IC 50 value of 247 1.3 nM, which was 43-fold more potent than EK1C1. These findings indicate that the 248 linker length has a significant effect on the overall activity of lipopeptides.

249
Comparison of increasing PEG-based arm lengths in EK1C4 shows that inhibitors 250 potency slightly decreased in the cell-cell fusion assay (Fig. 4e). The data suggest that 251 "GSGSG-PEG4" linker was optimal to bridge both parts of the conjugates. Similarly, 252 EK1C4 showed the most potent inhibitory activity against SARS-CoV-2 PsV 253 infection, with IC 50 value of 15.8 nM, providing 149-fold stronger anti-SARS-CoV-2 254 activity than that of EK1 (IC 50 =2,375 nM) (Fig. 4f).

255
The lipopeptide EK1C4 exhibits the most potent inhibitory activity against in all those coronavirus cell-cell fusion assays (Fig. 5a-h). 280 We also assessed the antiviral activity of EK1C4 on PsV infection by divergent potent than those of EK1, respectively (Fig. 6b-e). We next assessed the cytotoxicity 304 of EK1C4 on various target cells and found that the half cytotoxic concentration 305 (CC 50 ) was beyond 5 μM, which is the highest detection concentration of EK1C4 (Fig.   306 S3). Therefore, the selectivity index (SI=CC 50 /IC 50 ) of EK1C4 is >136, suggesting 307 that EK1C4 is a promising SARS-CoV-2 fusion inhibitor with little, or even no, toxic 308 effect in vitro. Further, we explored the potent antiviral mechanism of EK1C4 and 309 found that the complexes of EK1C4/SARS-2HR1, EK1C4/MERS-HR1, and 310 EK1C4/SARS-2HR1 harbor higher stability and increased Tm values than those of the 311 complexes formed by EK1 and HR1s (Fig. S4). These results suggested that increased 312 antiviral activity of EK1C4 should be related its increased binding affinity with HR1, 313 but their detailed interactions require further studies. significantly along with 100% mortality (Fig. 6f-g). The final survival rates of mice in 325 Pre-0.5, Pre-2, Pre-4, Pre-12 and Pre-24 groups were 100%, 100%, 100%, 83% and 326 0%, respectively ( Fig. 6f-g). In contrast, EK1 with a single dose of 20 mg/kg via nasal 327 administration exhibited very promising prophylactic effect in the Pre-0.5 h and Pre-1 328 h groups, whereas all mice in the EK1-Pre-2 h group eventually died similarly to the 329 mice in the viral control group (Fig. S5). These results suggested that EK1C4 has 330 better stability, antiviral activity, and prolonged half-life in the airway environment 331 when compared with EK1. 332 We then tested the therapeutic effect of EK1C4 0.5 h (Post-0.5 group) and 2 h 333 (Post-2 group) after HCoV-OC43 infection (Fig. 6h-i). The Post-0.5 group and Post-2 334 group mice showed 100% and 16.7% survival rate, respectively, suggesting that 335 EK1C4 harbors good therapeutic effect after a short period of HCoV-OC43 infection, 336 possibly resulting from the establishment of HCoV-OC43 infection in mouse brain 337 where EK1C4 cannot get through the blood brain barrier via nasal administration 14 .

338
As shown in Fig. S6, high viral titer was detected in brains of all 5 mice in Pre-24 339 group and 4 out of 5 mice in Post-2 group, but was not detected in brain tissues of all 340 mice in Pre-0.5, Pre-2, Pre-4, and Post-0.5 groups, while only moderate level of viral 341 titer was detected in brain tissue in one of the 5 mice in Pre-12 group (Fig. S6 a and b).    Inducing the S1/S2 furin-recognition site could significantly increase the capacity of 371 SARS-CoV S protein to mediate cellular membrane surface infection 26 . Interestingly, 372 SARS-CoV-2 harbors the S1/S2 cleavage site in its S protein, but its specific role in S 373 protein-mediated membrane fusion and viral life-cycle remains to be further explored 374 (Fig. S7). A recent report suggested that SARS-CoV-2 mainly used TMPRSS2 for 375 plasma membrane fusion; this means that the TMPRSS2 inhibitor might constitute an 376 option for blocking SARS-CoV-2 fusion with and entry into the host cell 27 .

377
The 6-HB structure formed by HR1 and HR2 regions in the S2 subunit of HCoVs 378 plays a key role during the viral membrane fusion process, which makes it one of the 379 most important targets for drug design. In previous studies, we have found that HR1 380 and HR2 of SARS-CoV-2 could form a stable coiled-coil complex, but the detailed 381 conformations remain unknown. According to the X-ray crystallographic analysis of 382 the complex formed by HR1 and HR2 of SARS-CoV-2 (Fig. 2b), it is a typical 6-HB  However, instead of weakening the interaction between HR1 and HR2, such unilateral 387 difference seems to form new interactions in some regions and enhance the existing 388 ones in other regions (Fig. 3). When K991 in SARS-CoV HR1 was replaced with 389 S929 in SARS-CoV2 HR1, a new, strong hydrogen bond was formed with a distance 390 of 2.4 Å. K933 forms a new interaction with N1192 in SARS-CoV-2 with a distance 391 of 2.7 Å, whereas the corresponding position in SARS-CoV has no such interaction.

392
In the other two regions, E918 binds to R1166 and K929 binds to E1163 in 393 SARS-CoV, both of which were enhanced in SARS-CoV-2. These results suggest that 394 this new HCoV has evolved with improved binding affinity between HR1 and HR2 395 domains, which may accelerate the viral membrane fusion process and enhance viral 396 infectivity or transmissibility. A recent study also found that the binding affinity 397 between ACE2 receptor on the host cell and RBD in S protein of SARS-CoV-2 is 398 more than 10-fold higher than that of SARS-CoV, which may also be associated with 399 the increased infectivity and transmissibility of SARS-CoV-2 12 .

400
The conjugation of cholesterol to viral entry inhibitor has been proved to be an 401 effective strategy to enhance the antiviral activity, such as C34 peptide for HIV-1 28 .

402
However, the mechanism of this enhancement, especially the role of cholesterol group 403 in the C-terminal tail of entry inhibitor, is still unclear. There is a possibility that the 404 cholesterol group could anchor to the target membrane to facilitate the binding of 405 inhibitor to the HR1 targets. However, we noticed that binding affinity between 406 EK1C4 and SARS-CoV-2-HR1P is significantly enhanced than EK1 peptide alone, 407 which suggested that cholesterol group may be involved in binding to HR1P directly 408 (Fig. S4). Therefore, using structural simulation and docking method, we predicted a 409 possible model of EK1C4 in binding with SARS-CoV-2 HR1P (Fig. S8). In this      and by Autodock 4 software 40 for cholesterol moiety (Fig. S8).