key: cord-328187-9zd79gai
authors: Zhang, Yali; Wang, Shaojuan; Wu, Yangtao; Hou, Wangheng; Yuan, Lunzhi; Sheng, Chenguang; Wang, Juan; Ye, Jianghui; Zheng, Qingbing; Ma, Jian; Xu, Jingjing; Wei, Min; Li, Zonglin; Nian, Sheng; Xiong, Hualong; Zhang, Liang; Shi, Yang; Fu, Baorong; Cao, Jiali; Yang, Chuanlai; Li, Zhiyong; Yang, Ting; Liu, Lei; Yu, Hai; Hu, Jianda; Ge, Shengxiang; Chen, Yixin; Zhang, Tianying; Zhang, Jun; Cheng, Tong; Yuan, Quan; Xia, Ningshao
title: Virus-free and live-cell visualizing SARS-CoV-2 cell entry for studies of neutralizing antibodies and compound inhibitors
date: 2020-07-22
journal: bioRxiv
DOI: 10.1101/2020.07.22.215236
sha: 
doc_id: 328187
cord_uid: 9zd79gai

The ongoing COVID-19 pandemic, caused by SARS-CoV-2 infection, has resulted in hundreds of thousands of deaths. Cellular entry of SARS-CoV-2, which is mediated by the viral spike protein and host ACE2 receptor, is an essential target for the development of vaccines, therapeutic antibodies, and drugs. Using a mammalian cell expression system, we generated a recombinant fluorescent protein (Gamillus)-fused SARS-CoV-2 spike trimer (STG) to probe the viral entry process. In ACE2-expressing cells, we found that the STG probe has excellent performance in the live-cell visualization of receptor binding, cellular uptake, and intracellular trafficking of SARS-CoV-2 under virus-free conditions. The new system allows quantitative analyses of the inhibition potentials and detailed influence of COVID-19-convalescent human plasmas, neutralizing antibodies and compounds, providing a versatile tool for high-throughput screening and phenotypic characterization of SARS-CoV-2 entry inhibitors. This approach may also be adapted to develop a viral entry visualization system for other viruses.

). The Gamillus (mGam) and mNeonGreen (mNG) were tested 30 as the fused-GFP because mGam is acid-tolerant, which may enable fluorescent 31 tracking when the probe is taken up into acidic cellular organelles 23 , and mNG is the 32 brightest GFP to our knowledge 24 . We designated the RBD-based probes as RBG Figure 1D) , respectively, which 8 were similar to previously reported data for unfused proteins 9, 10 . Together, C-terminal 9 FP-fusion does not influence the structure and ACE2-binding capability of the RBD and 10 S-ectodomain of SARS-CoV-2. 11 Establishment of virus-free assays to visualize SARS-CoV-2 cell entry 12 We established hACE2-overexpressing cell lines using the ACE2hR and 13 ACE2iRb3 constructs (Figure 2A ). Cell-transfection with ACE2hR allowed hACE2-14 overexpression with nucleus visualization (H2B-mRuby3). ACE2iRb3 contains an 15 ACE2-mRuby3 expressing cassette following an IRES-ligated H2B-iRFP670-2A- 16 PuroR. Transfection with ACE2iRb3 simultaneously enabled fluorescent visualization 17 of hACE2 (hACE2-mRuby3) and nucleus (H2B-iRFP670). Using these vectors, we 18 developed three stable cell lines, namely, 293T-ACE2iRb3, 293T-ACE2hR and H1299- 19 ACE2hR. As expected, hACE2 (or ACE2-mRuby3) was expressed at high levels, and 20 the expression of TMPRSS2 (another critical factor for viral entry) 12 did not change in 21 these cells ( Figure 2B ). 22 On 293T-ACE2iRb3 cells, both the SARS-CoV2-RBG and SARS-CoV2-STG 23 probes showed effective-binding to the cells, as membrane-bound and hACE2-24 mRuby3-colocalized mGam signals were observed after a 6-min incubation with the 25 cells ( Figure 2C ). Cytoplasmic mGam signals were detected in cells after incubation 26 for 60-min or longer with the probes, particularly for SARS-CoV2-STG, suggesting that 27 the recombinant probes can not only bind to the cell surface but also be taken up into 28 the cells. The internalization of SARS-CoV2-STG was more evident than that of CoV2-RBG into 293T-ACE2iRb3 ( Figure 2C ). In live-cell dynamic tracking, more 30 internalized mGam signals was observed for SARS-CoV2-STG than SARS-CoV2-31 RBG ( Figure S3A ). For mGam signals, the internalized fluorescence ratio (IFR, Figure   32 S3B) and the internalized vehicle numbers (IVNs, Figure S3D ) of STG-treated cells 33 were both significantly higher than those of RBG-treated cells approximately 30-min 34 after probe-cell incubation. In contrast, no significant difference was noted in hACE2-35 mRuby3 internalization in the presence or absence of probes ( Figure S3C ). 36 Using 293T-ACE2iRb3 cells, we established a cell-based assay mimicking SARS- 37 CoV-2 cell entry based on recombinant probes. It was a one-step wash-free assay 1 ( Figure 2D ). After 1-hour cell-probe incubation, the cells were directly imaged by using 2 a fully automatic high-content screening (HCS) system in confocal mode. For 3 quantitative measurements, the H2B-iRFP670 were used to identify the nucleus, and 4 the ACE2-mRuby3 were used to determine the cell boundary. Based on the detected 5 nucleus and cell outlines, the green fluorescence intensities on the cell membrane and 6 in the cytoplasmic region of each cell could be measured ( Figure 2D ). Generally, we 7 used the mean fluorescence intensity (MFI) in the cytoplasmic region (cMFI) as an 8 index of the amounts of the cell-bound and internalized probes. As the spikes of CoV and HKU1-CoV do not interact with hACE2, the signals of MERS-RBG and HKU1-10 RBG on 293T-ACE2iRb3 were nonspecific background ( Figure 2E ). RaTG13-RBG 11 showed a detectable and dose-dependent cMFI, but the value was significantly lower 12 than those of the probes of SARS-CoV-1 and SARS-CoV-2. Compared to SARS-13 CoV1-RBG, SARS-CoV2-RBG showed slightly stronger signals, possibly due to its 14 higher binding affinity. For SARS-CoV-2, the cMFI of SARS-CoV2-STG, SARS-CoV2-15 STN and SARS-CoV2-ST488 were significantly higher than that of SARS-CoV2-RBG 16 and SARS-CoV2-RBD488, and also stronger than SARS-CoV2-SMG. The dylight488-17 labeled SARS-CoV2-RBD488 probe presented a weaker signal than SARS-CoV2- 18 RBG, suggesting that the NH2-dye modification at some amino acids of the RBD may 19 interfere with its interaction with hACE2. Moreover, mGam-fused probes showed better 20 performance than dylight488-labeled or mNG-fused probes ( Figure 2F ). At a 21 concentration below than 10 nM, the cMFI of SARS-CoV2-STG was approximately 10-22 fold higher than that of SARS-CoV2-RBG. 23 Based on the visualization system, we developed cell-based HCS assays for 24 analyzing the blocking potencies of SARS-CoV-2 entry inhibitors, designated CSBT 25 (using SARS-CoV2-STG) and CRBT (using SARS-CoV2-STG), respectively. The 26 proteins of hACE2-Fc (rACE2), SARS-CoV2-RBD and SARS-CoV2-S1 were 27 employed for inhibition assessments following the procedure described in Figure 2D . 28 As expected, all three proteins exhibited dose-dependent cMFI inhibition in both 29 assays ( Figure 2G ). The Z'-factor coefficients of the CSBT and CRBT were both 30 determined to be over 0.7 ( Figure S3E ), which demonstrated their robustness and 31 reproducibility.

Detecting entry-blocking antibodies in COVID-19-convalescent human plasmas 33 by CSBT and CRBT. 34 Recent studies have suggested that convalescent plasma may be beneficial in 35 COVID-19 treatments 25, 26 . Neutralization antibodies (NAbs) in convalescent plasmas 36 may be essential in suppressing viruses 27 . However, a rapid method for determining 37 the neutralization antibody titer (NAT) of human plasma is still absent. We evaluated 1 the feasibility of the CSBT and CRBT determined entry-blocking antibody titers as NAT 2 surrogates in 32 COVID-19-convalescent human plasmas (Table S1 ). Compared with 3 samples from healthy donors (n=40), all COVID-19-convalescent plasmas showed 4 significant cMFI inhibition on CSBT assay, whereas only 12 samples (37.5%) had 5 detectable CRBT activity ( Figure 3A ). For quantitative analysis, two-fold serial dilution 6 tests were further performed to determine the CSBT and CRBT titers ( Figure 3B ). 7 Moreover, the titers of total antibodies (TAb), IgG, IgM, and lentiviral-pseudotyping-8 particles (LVpp) based NAT (LVppNAT) against SARS-CoV-2 were also measured for 9 comparisons ( Figure 3C ). Among the antibody titers derived from various assays, the 10 CSBT titer showed the best correlation with LVppNAT ( Figure 3D and Table S2,   11 r=0.832, p<0.001), and it also well correlated (r=0.959, p<0.001, Figure 3D ) with the 12 neutralization activity against authentic SARS-CoV-2 virus in 12 representative 13 samples (Table S3) . Together, the CSBT-determined entry-blocking antibody titer is a 14 good NAb surrogate of convalescent plasmas.

Functional phenotyping of mouse anti-spike antibodies by CSBT and CRBT. 16 Serum samples from mice immunized with the SARS-CoV2-RBD, SARS-CoV2-17 S1 and SARS-CoV2-S2 were collected for LVppNAT, CSBT and CRBT measurements. 18 The SARS-CoV2-RBD and SARS-CoV2-S1 immunizations resulted in potent and 19 comparable serum LVppNAT ( Figure S4A ), whereas SARS-CoV2-S2 raised little NAbs. 20 The CSBT ( Figure S4B ) and CRBT ( Figure S4C ) assays also exhibited similar results 21 to LVppNAT measurements. The ID50 correlation coefficient was 0.989 (p<0.001) 22 between CSBT and LVppNAT and 0.925 (p<0.001) between CRBT and LVppNAT. 23 Using RBD-immunized mice, we developed 18 mAbs via RBD-ELISA screening 24 following cell-based functional evaluations (as illustrated Figure 4A ). These mAbs did 25 not display much difference in ELISA-binding to SARS-CoV2-RBD, but 2 of them (8H6 26 and 15A9) showed significantly decreased ELISA-binding activities to SARS-CoV2-ST 27 ( Figure S5A ). Based on epitope-binning assays using a cross-competitive ELISA, the 28 mAbs could be divided into six groups ( Figure S5B ). All mAbs showed detectable but 29 varied surface plasmon resonance (SPR) affinity (0.004-131 nM, Figure S6 ) to SARS-30 CoV2-RBD. Quantitative measurements of CSBT, CRBT and LVppNAT for the mAbs 31 were further performed ( Figure 4B to C, Figure S7A to C). Half of the mAbs exhibited 32 high-to-moderate CSBT blocking potencies (IC50<30 nM), whereas the remaining 33 ones showed low-to-no CSBT activities ( Figure 4B ). In comparisons of the dose-34 dependent cMFI inhibitions against SARS-CoV2-STG, SARS-CoV2-ST488, SARS-35 CoV2-RBG, and SARS-CoV1-RBG ( Figure 4C ), the profiles of most mAbs against 36 SARS-CoV2-STG and SARS-CoV2-ST488 were similar, but the activities of 2B4 and 37 34B4 were dramatically decreased with SARS-CoV2-ST488 compared to SARS-1 CoV2-STG, suggesting that dye-conjugation may modify the epitopes of the two mAbs 2 and hinder their bindings. Notably, seven mAbs exhibited striking enhancement at 3 some dosage in the CRBT assays ( Figure 4C and Figure S7B ), but neither 4 enhancement was noted in the CSBT nor LVppNAT tests. SPR analyses demonstrated 5 that the Fabs of two representative CRBT-enhancing mAbs, 53G2 and 8H6, also 6 showed a dose-dependent promoting effect on the RBD-ACE2 binding, whereas the 7 Fabs of two CRBT-blocking mAbs (36H6 and 2B4) exhibited a dose-dependent 8 reduction in the RBD-ACE2 interaction ( Figure S8 ). Together, the CRBT-enhancing 9 effects of these mAbs may be caused by the antibody-induced RBD conformation 10 changes associated with increases in ACE2-binding capacity. 11 The functional potencies of mAbs determined by various assays are summarized 12 in Figure 4D and Table S4 . The CSBT-IC50 values of the mAbs showed the good 13 correlation with their LVppNAT IC90 (r=0.866, p<0.001, Figure 4E ) or IC50 (r=0.750, 14 p<0.001, Figure S9 ) values, and were also well correlated with their CRBT-IC50 values 15 (r=0.869, p<0.001, Figure 4E ). However, a 53G2 mAb presented CSBT activity but no 16 inhibition in the CRBT assay, suggesting that its CSBT activity is independent of the 17 direct blocking of the RBD-ACE2 interaction ( Figure 4D ). An 83H7 mAb with moderate 18 LVppNAT activity but showed neither CSBT nor CRBT inhibition ( Figure 4D ), 19 suggesting it may act through different mechanisms to achieve neutralization. No 20 significant relationship was noted between the ELISA-or SPR-determined protein-21 binding activities and neutralization potencies of the mAbs ( Figure S9 ). 22 According to the SPR ( Figure S10 ) and CRBT analyses using SARS-CoV1-RBG 23 ( Figure 4C ), the 2B4, 34B4, 5F3, 18C5, and 8H6 mAbs showed cross-reactivity to 24 SARS-CoV-1 and RaTG13-CoV. However, only the 2B4 mAb had neutralization activity 25 in SARS-CoV-1 LVppNAT measurements ( Figure S7D ). Epitope-binning assay ( Figure   26 S5B) suggested that 2B4, 34B4, 5F3 and 14D2 possibly share an overlapping-epitope 27 (cluster C2), and 18C5, 8H6, 83H7 and 65G9 may bind to another similar epitope (mAb 28 cluster C5b). As 2B4 showed comparable LVppNAT potencies against both SARS-29 CoV-1 and SARS-CoV-2, it may recognize a cross-neutralization epitope. The 36H6 30 mAb, which recognizes a unique epitope that differs from other mAbs (mAb cluster C1, 31 Figure S5B ), presented the best performance in LVppNAT, CSBT, and CRBT assays 32 but did not show any cross-reactivity with SARS-CoV-1 or RaTG13. Both 36H6 and The 83H7 mAb showed SARS-CoV-2 neutralization activity against both 2 pseudoparticles (IC50=0.99 nM) and the authentic virus (IC50=13.02 nM) but had 3 neither CSBT nor CRBT activity. We speculated that this mAb may inhibit the SARS-4 CoV-2 via an intracellular neutralization pathway 28 . To validate this, we prepared 5 dylight633-labeled mAbs (Ab633) of 36H6, 53G2, 83H7, and 8H6 and an irrelevant 6 mAb (ctrAb) for dual-visualizing tracking. Among them, 36H6, 53G2 and 8H6 served 7 as controls that had strong, moderate and weak/no activity for both CSBT and 8 neutralization, respectively. In 293T-ACE2iRb3 cells simultaneously incubated with 9 STG and Ab633, we performed time-serial live-cell imaging analyses. To characterize 10 the influence of these mAbs on SARS-CoV2-STG internalization, the dynamic changes 11 of the STG-IVNs, the STG-IVpMFI (the peak MFI of internalized vesicles), the Ab633-12 IVNs, the Ab633-IVpMFI, and the percentage of STG/Ab633-colocalized internalized 13 vesicles, and the STG-IVA were calculated ( Figure 5 ). As expected, 36H6 completely 14 obstructed STG internalization (p<0.001), 53G2 showed significant but incomplete 15 inhibition (p=0.002), and 8H6 presented little/no influence on STG internalization 16 (p=0.70). Compared to ctrAb, the 83H7 showed no significant STG-IVNs reduction 17 (p=0.31, Figure 5A ) but increased the STG-IVpMFI (p<0.001, Figure 5B ). On the other 18 hand, the 83H7 group exhibited higher Ab633-IVNs (p<0.001, Figure 5C ) and Ab633-19 IVpMFI (p<0.001, Figure 5D ). The STG/Ab633 colocalization (p<0.001, Figure 5E ) and 20 the STG-IVA (p<0.001, Figure 5F ) in the 83H7 group were also significantly higher and 21 larger, respectively, than those in the other groups. Representative images at 5-hour 22 post STG/Ab633-cell incubation, as shown in Figure 5G In STG-visualization system, the cMFI measurements following the CSBT 5 procedure showed only a slight reduction in cells treated with a high dose of Amiloride, 6 dynasore, apilimod, and APY0201 ( Figure S11A ). Notably, confocal-images revealed 7 that the STG, colocalizing with internalized ACE2-mRuby3, were trapped on enlarged 8 cytoplasmic vacuoles induced by PIKfyve inhibitors (Figure 6C ), and most of these 9 vacuoles were not stained by pH-dependent LysoView633 dyes, suggesting an 10 abnormal pH-status ( Figure S11B ). In addition, tetrandrine or Baf.A1 also caused Our data provided convincing evidence demonstrating the versatile applicability of 2 the new system. The CSBT-determined entry-blocking potency was a better correlate 3 of NAT against pseudotyping or the authentic SARS-CoV-2 virus than ELISA-binding 4 activity in COVID19-convalescent human plasmas, immunized mouse sera and mAbs. 5 The CSBT may serve as rapid proxy assessment to identify plasma source with 6 therapeutic potential in clinic, and is a useful tool in evaluating vaccine efficacy and 7 neutralizing mAb identification. In this study, 4 of 18 mAbs (36H6, 2B4, 3C8 and 12H8, 8 Figure 4D ) with the strongest CSBT blocking activities (IC50<10 nM) showed the most 9 potent LVppNAT (IC90<3 nM). Notably, the 36H6 mAb presented superior 10 neutralization activity against authentic SARS-CoV-2 (IC50=0.079 nM), which was 11 comparable with recently described potent neutralizing mAbs 16, 31 , and thereby Table   25 S1. live-cell images were acquired at 10-min, 1-hour, 2-hour, 3-hour, 5-hour, 7-hour, 9-hour, 11 11-hour, and 13-hour using a 63x water immersion objective.

To visualize compound-induced influence on viral entry, 293T-ACE2iRb3 ( Figure   13 6C-D, Figure S11A ) or H1299-ACE2hR cells ( Figure S11B ) were pretreated with serial 14 dilutions of compounds for 1-hour. Then the probes were added to the cell cultures for 15 further incubations in the presence of compounds. Cell images shown in Figure 6C 16 and Figure S11B were acquired on TCS SP8 STED confocal microscope using a 100x 17 oil immersion objective. The data of Figure 6D and Figure S11A were derived from 18 images acquired on Opera Phenix using 40x water immersion objective. For pictures 19 of Figure 6C , the cells were gently washed twice with PBS at 5-hour post STG 20 incubation, following a paraformaldehyde fixation before imaging. For experiments as 21 shown in Figure S11B , the cells at 5-hour post STG incubation were stained with 22 Lysoview633 (0.1 μL per well) for 10-min, then the cells were gently washed twice with 23 PBS buffer and fixed with paraformaldehyde treatment before imaging. Cell images 24 involved in Figure 6D and Figure CRBT. Serum samples for mice immunized with recombinant proteins of SARS-CoV2-4 S1 (mouse S1-1, S1-2, S1-3), SARS-CoV2-S2 (mouse S2-1, S2-2, S2-3) and SARS- p01  p02  p03  p04  p05  p06  p07  p08  p09  p10  p11  p12  p13  p14  p15  p16   p18  p19  p20  p21  p22  p23  p24  p25  p26  p27  p28  p29  p30  p31 

(STG vesicles/cell)

Cell entry mechanisms of SARS-CoV-2

Rapid development of an inactivated vaccine candidate for SARS-CoV-15 2

Potent neutralizing antibodies against SARS-CoV-2 identified by high-17 throughput single-cell sequencing of convalescent patients' B cells

Inhibition of SARS-CoV-2 (previously 2019-nCoV) infection by a highly 19 potent pan-coronavirus fusion inhibitor targeting its spike protein that harbors a high 20 capacity to mediate membrane fusion

Neutralization of SARS-CoV-2 spike pseudotyped virus by recombinant 22 ACE2-Ig

A Universal Design of Betacoronavirus Vaccines against COVID-19, MERS, 24 and SARS

Characterization of spike glycoprotein of SARS-CoV-2 on virus entry and 26 its immune cross-reactivity with SARS-CoV

Robust neutralization assay based on SARS-CoV-2 S-bearing 28 vesicular stomatitis virus (VSV) pseudovirus and ACE2-overexpressed BHK21 cells

Establishment and validation of a pseudovirus neutralization assay for 31 SARS-CoV-2

Acid-Tolerant Monomeric GFP from Olindias formosa

A bright monomeric green fluorescent protein derived from 35 Branchiostoma lanceolatum

Treatment of 5 Critically Ill Patients With COVID-19 With Convalescent 37 Plasma

Effectiveness of convalescent plasma therapy in severe COVID-19 39 patients

The convalescent sera option for containing COVID-19

Intracellular neutralization of viral infection in polarized epithelial cells by