key: cord-312560-onfabcfv
authors: Klingler, J.; Weiss, S.; Itri, V.; Liu, X.; Oguntuyo, K. Y.; Stevens, C.; Ikegame, S.; Hung, C.-T.; Enyindah-Asonye, G.; Amanat, F.; Baine, I.; Arinsburg, S.; Bandres, J. C.; Kojic, E. M.; Stoever, J.; Jurczyszak, D.; Bermudez-Gonzalez, M.; Simon, V.; Liu, S.; Lee, B.; Krammer, F.; Zolla-Pazner, S.; Hioe, C. E.
title: Role of IgM and IgA Antibodies to the Neutralization of SARS-CoV-2
date: 2020-08-21
journal: medRxiv : the preprint server for health sciences
DOI: 10.1101/2020.08.18.20177303
sha: 
doc_id: 312560
cord_uid: onfabcfv

SARS-CoV-2 has infected millions of people and is on a trajectory to kill more than one million globally. Virus entry depends on the receptor-binding domain (RBD) of the spike protein. Although previous studies demonstrated anti-spike and -RBD antibodies as essential for protection and convalescent plasma as a promising therapeutic option, little is known about the immunoglobulin (Ig) isotypes capable of blocking virus entry. Here, we studied spike- and RBD-specific Ig isotypes in plasma/sera from two acutely infected and 29 convalescent individuals. Spike- and RBD-specific IgM, IgG1, and IgA1 antibodies were produced by all or nearly all subjects at varying levels and detected at 7-8 days post-disease onset. IgG2, IgG3, IgG4, and IgA2 were also present but at much lower levels. All samples also displayed neutralizing activity. IgM, IgG, and IgA were capable of mediating neutralization, but neutralization titers correlated better with binding levels of IgM and IgA1 than IgG.

In December 2019, the first patients with coronavirus disease 2019 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) were identified in the city of Wuhan, Hubei Province, China 1 . Since then, the epidemic has rapidly spread to most regions of the world, infecting millions of people 2 . Effective therapeutics and vaccines against SARS-CoV-2 are urgently needed. To this end, more information about the Ig isotypes present in the plasma of COVID-19 convalescent individuals and their antiviral activities are needed, as convalescent plasma transfusion showed promising results in patients with severe to life-threatening COVID-19 3-5 . The data would also inform vaccine development, as more than 100 vaccine candidates are in different stages of preclinical development, and many are now in phase 2 and 3 clinical trials 6 . Although using different strategies 7 , many vaccines are based on one of the three membrane-anchored proteins present on the virus envelope surface: the SARS-CoV-2 spike protein 8, 9 , which contains the receptor-binding domain (RBD) required for binding to and entry into the cells [10] [11] [12] [13] [14] [15] [16] [17] . These vaccines aim to protect by inducing neutralizing antibodies (Abs) capable of blocking the viral infection. However, which of the immunoglobulin (Ig) isotypes are protective is not yet clear.

Monomeric IgG constitutes approximately 75% of the Abs found in serum and exists as four subtypes: IgG1 (~66% of IgG), IgG2 (~23% of IgG), IgG3 (~7% of IgG) and IgG4 (~4% of IgG) 18, 19 .

IgM Abs represent 10% of total serum Abs and are the first to arise in response to new antigens 19, 20 .

Although IgM Abs do not undergone extensive somatic hypermutation to increase their affinity as do IgG and IgA Abs, their higher valency due to the oligomerization of IgM enhances their avidity and potency against pathogens 19, 21, 22 . IgA Abs exist as two subtypes: IgA1 and IgA2, and represent 15% of total serum Abs 19 . They are dimeric in the mucosa, but in the circulation, these two IgA subtypes are monomeric.

SARS-CoV-2 spike-, RBD-and nucleocapsid-specific serum/plasma Abs of IgM, IgG, and IgA . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint isotypes are found in most COVID-19 patients 23-30 , with Ab neutralizing activities reported among the convalescent patients 26,28,31,32 . However, the neutralizing titers appear to vary greatly 26,28,31,32 and they correlate with Ab binding levels against RBD, spike, and/or nucleocapsid, and also with age, duration of symptoms, and symptom severity 26,28,32 . Several RBD-specific monoclonal Abs of IgG isotype with potent antiviral activities has been generated from individuals with high neutralization titers and these confer protection in animal models 26,31,33,34 . Moreover, a monoclonal Ab of IgA isotype capable of recognizing both the SARS-CoV-1 and SARS-CoV-2 spike proteins, and blocking ACE2 receptor binding was recently described 35 . However, no data are available regarding the neutralizing capacity of plasma IgM and IgA Abs from COVID-19 patients.

Studies on other respiratory viruses such as influenza show that, in addition to IgG, IgA could also mediate virus neutralization, and their relative contribution depends on the physiologic compartment in which they are found, with IgA contributing to the protection of mostly the upper respiratory tract while IgG was protecting the lower respiratory tract 36-39 . Of note, an anti-hemagglutinin monoclonal IgA has been demonstrated to mediate more potent antiviral activities against influenza when compared to a monoclonal IgG against the same epitope 40 . Interestingly, an IgM Ab with potent antiviral activities targeting the receptor binding site of influenza B has also been described 41 . In addition, mucosal respiratory syncytial virus (RSV)-specific IgA neutralizing Abs is a better correlate of protection than serum RSV-specific IgG neutralizing Abs 42 . In the case of SARS-CoV-1, high titers of mucosal IgA in the lungs are correlated with reduced pathology upon viral challenge in animal models 43 .

Whether IgA in the blood and the respiratory tract mucosa offers protection against SARS-CoV-2

infection remains an open question. Moreover, few data are available concerning the contribution of IgM to neutralization and protection against viruses, including SARS-CoV-2.

We have recently published a Luminex assay detecting Ig total against spike and RBD 44 . Based on this assay, we studied here the Ig isotype profiles against spike and RBD in the plasma and serum . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint from acutely infected or convalescent individuals using a Luminex assay that detects antigen-specific IgM, IgG1-4, IgA1 and IgA2. Using a pseudovirus assay 45 , we also measured neutralizing activities in plasma and serum samples and in Ig isotype fractions to determine the neutralizing capacity of IgM, IgA, and IgG. The data demonstrate high prevalence of spike-and RBD-specific IgM and IgA, similar to that of IgG1, in plasma/serum from COVID-19 patients and their significant contributions to virusneutralizing activities. This is the first evidence that purified plasma IgG, IgM, and IgA contributes to SARS-CoV-2 neutralization.

A total of 29 serum (P#5-8) and plasma (TF#1-25) specimens from COVID-19-convalescent individuals was tested. Sera from three uninfected individuals banked as part of an ongoing longitudinal study prior to the COVID-19 outbreak (N#1-3), and an additional ten plasma from COVID-negative contemporaneous blood bank donors (N#4-13) were included for comparison. The specimens were initially titrated for total Ig against spike and RBD (Fig. 1) . All 29 COVID-19 positive specimens exhibited titration curves of total Ig Abs against spike, while none of the negative controls did. Similar results were observed with RBD, except that one COVID-19-negative sample had low titrating levels of RBD-specific Ig. Overall the background MFI values were higher for RBD than spike. The areas under curves (AUCs) highly correlated with the 1:200 dilution MFI (p < 0.0001; Supplementary Fig. 1) , consequently all samples were tested for isotyping at this dilution. To assess the reproducibility of the assay, the samples were tested in at least two separate experiments run on different days, and a strong correlation was observed between the MFIs from these independent experiments (Supplementary Fig.   2 ). To evaluate for the presence of spike-specific and RBD-specific total Ig, IgM, IgG1, IgG2, IgG3, IgG4, IgA1 and IgA2, the specificity of the secondary Abs used to detect the different isotypes were first validated with Luminex beads coated with myeloma proteins of known Ig isotypes (IgG1, IgG2, IgG3, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint IgG4, IgA1, IgA2, and IgM) (Supplementary Fig. 3) .

While all 29 convalescent individuals had anti-spike and anti-RBD total Ig (Fig. 2) , the Ig levels were highly variable, with MFI values ranging from 36,083 to 190,150. All 29 convalescent individuals displayed IgM Abs against spike at varying levels, although only 55% (16/29) were positive for anti-RBD IgM, when evaluated using cut-off values calculated as mean + 3 standard deviation (SD) of the 13 COVID-negative samples. The lower percentage of IgM Abs specific for RBD might be due to the high background observed for IgM against RBD with COVID-negative specimens (Fig. 2b,c) . An IgG1 response was also detected against both spike and RBD in 97% of the convalescent subjects, with MFI values that ranged from 1012.5 to 59,880, and the responses against spike and RBD were highly correlated for every isotype (Supplementary Fig. 4) . In contrast IgG2, IgG3, and IgG4 Abs against spike and RBD were detected in only a small fraction of the subjects, and the levels were very low (MFI values < 1300) (Fig. 2) .

Surprisingly, almost all the individuals produced IgA1 Abs against spike (90%) and RBD (86%) while 21% exhibited IgA2 against spike, and 48% exhibited IgA2 against RBD (Fig. 2) . Overall, these data demonstrate that IgM, IgG1, and IgA1 Abs were all strongly induced against spike and RBD in all or almost all COVID-19 convalescent individuals (Fig. 2) . The levels, however, were highly variable among individuals. While not reaching statistical significance, a general trend was observed toward higher levels of total Ig, IgM, IgG1, and IgA1 in women compared to men (Supplementary Fig. 5 ). In Fig. 3 , comparing levels of total Ig with the different Ig isotypes showed a highly significant correlation with IgG1 for both Abs specific for spike and RBD indicating that IgG1 is the major isotype induced by SARS-CoV-2 infection. There was no other isotype which showed a significant correlation with total Ig Abs for anti-spike Abs, although there was a significant correlation between total Ig and IgA1 for anti-RBD Abs. Moreover, none of the IgM, IgG1, or IgA1 isotypes correlated with one another (Fig. 3c) .

IgA1 is induced early after disease onset in COVID-19 patients. Since almost all convalescent . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint 7 COVID-19 patients displayed IgA Ab responses to spike and RBD antigens, we sought to evaluate the kinetics of IgA versus other Ig isotype induction at early time points. We tested longitudinal samples collected from two patients (P#1 and P#2) from days 7 to 15 post-onset of symptoms. The earliest samples from both patients were positive for IgA1 Abs against spike (Fig. 4a) and RBD (Fig. 4b) , and the levels increased over time. Like IgA1 Ab, total Ig, IgM, and IgG1 Abs were also detectable as early as 7-8 days post symptom onset, and their levels increased over time. On the contrary, IgA2 Ab levels were near or below background on days 7-8 and remained unchanged over two weeks post-onset. IgG4

Abs also remained low or near background, whereas IgG2 and IgG3 Abs increased slightly to above background after 10-15 days.

Neutralizing activities is detected in all convalescent COVID-19 individuals. We subsequently tested the ability of samples from convalescent subjects to neutralize a VSV∆G pseudovirus bearing the SARS-CoV-2 spike protein (COV2pp). The results, shown in All specimens from COVID-19-convalescent individuals were able to neutralize the virus at levels superior to 50% (Fig. 5a) . For 26 of 28 specimens, neutralization reached more than 90% (Fig. 5a) .

Interestingly, one sample demonstrated highly potent neutralization with a reciprocal IC50 titer > 40,960, and neutralization was still 75% at the highest dilution tested. In contrast, one sample had a low neutralization titer (reciprocal IC50 titer = 37) and reached a neutralization plateau of only ~60%. In comparison, none of the samples from COVID-19-negative individuals reached 50% neutralization (Fig.   5b) , while the sRBD positive control demonstrated potent neutralization with an IC50 of 0.06 µg/mL ( Fig. 5c) , similar to that recently reported 45 .

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint 8 isotype levels and neutralization titers varied tremendously among convalescent COVID-19 individuals ( Fig. 6) , we investigated if neutralizing activities correlated with a particular Ab isotype. IgM Abs specific for spike and RBD displayed the strongest correlation with neutralizing Abs (p < 0.0001 and p = 0.0002, respectively), while IgG1 did not show a significant correlation (Fig. 7a) . Neutralization reciprocal IC90 titers weakly correlated with RBD-specific total Ig levels but did not correlate with spike-specific total Ig levels (Fig. 7a) . Correlation with IC50 titers yielded similar results (data not shown). IgA1 levels also correlated with reciprocal IC90 titers (p = 0.013 and p = 0.006 for spike and RBD, respectively; Fig. 7a ). Although there were significant correlations between anti-spike and -RBD Abs of the IgG2 and IgG3s isotypes, most of the neutralization values were below the cut-off and for the few IgG3 and IgG2 responders, the levels were near background (Fig. 7b) .

Neutralizing activities are mediated by IgM, IgG, and IgA fractions. The data above show the strongest correlation of neutralizing activity with IgM. To ask directly to what extent Abs of different isotypes mediate neutralization, we evaluated the neutralization activities of IgM, IgG, and IgA fractions purified from plasma from five COVID-19-convalescent individuals (RP#1-5). The enrichment of IgM, IgG1, and IgA1 Abs reactive with spike and RBD was validated using the isotyping method used above (Supplementary Fig. 6 and not shown). These IgM, IgG and IgA fractions were then evaluated for neutralizing activity along with the original plasma (Fig. 8) . The RP#1-5 plasma neutralizing reciprocal IC50 titers ranged from 35 to 690 (Fig. 8a,b) . The purified IgG, IgA and IgM fractions all displayed neutralization of more than 50%, while the negative control Ig fractions (RN#1) did not (Fig. 8c,d) . The IC50 values for IgM and IgG were similar, and both were significantly better than IgA (Fig 8c,d) .

Our study demonstrates the presence of IgG1, IgA1 and IgM within 7-8 days after the onset of symptom, and the correlation of neutralizing activity is strongest with IgM, followed by IgA1, but does not correlate with IgG1 or other IgG isotypes. While correlations are important, direct testing of . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint neutralizing activities of different isotypes was performed, and these experiments showed that neutralizing activity was displayed most potently by IgM and IgG and less strongly by IgA. These data indicate the protective potential of all three major Ig isotypes and suggest that induction of each of these isotypes by vaccination may offer optimal protection against infection. These data also carry important implications for the use of hyperimmune globulin as treatment and prophylactic modalities.

Several SARS-CoV-2 vaccine candidates tested in animal models and humans were shown to induce IgG responses against spike and RBD as well as virus neutralizing activities, but in many of these studies, the induction of other Ig isotypes was not evaluated 46-49 . DNA vaccines expressing full-length and truncated spike proteins were able to curtail virus infection in the respiratory tract by varying degrees. Virus reduction correlated with levels of neutralization and also Fc-mediated effector functions such as antibody-dependent complement deposition (ADCD) 9 . Interestingly, these DNA vaccines elicited spike-and RBD-specific IgG1, IgG2, IgG3, IgA, and IgM Abs, and similar to our findings, neutralization correlated most strongly with IgM. Adenovirus serotype 26 vaccine vectors encoding seven different SARS-CoV-2 spike variants showed high protection, and virus reduction correlated best with neutralizing Ab titers together with IgM binding levels, FcγRII-binding, and ADCD responses 50 .

We noted that while all COVID-19 convalescent individuals exhibited plasma/serum neutralization activities, reaching 50% neutralization, and 26 of 28 specimens attained 90% neutralization, neutralization levels were highly variable with reciprocal IC50 and IC90 titers ranging over three orders of magnitude. Similarly, the levels of spike-and RBD-binding total Ig and Ig isotypes also varied greatly. A trend to higher levels of total Ig and each Ig isotype was seen in female compared to male subjects, as reported in another study 51 . Moreover, except for TF#11 (a male elite neutralizer), the median neutralizing reciprocal IC90 titer was higher in females than males, although the difference did not reach significance (data not shown). Gender differences in Ab induction have been observed following vaccination against influenza in humans and mice and were shown to result from the impact . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint of sex steroids 52,53 . Whether and to what extent this contributes to gender differences seen in clinical outcomes of COVID-19 54 remains to be investigated. Other studies have shown that the Ab levels were associated with multiple factors, including time from disease onset 55 and disease severity 25 . However, clinical data are not available for the subjects studied here, limiting our analysis only to neutralization and Ig isotypes.

One remarkable finding from our study is that neutralization levels correlated with binding levels of IgM and IgA1, but not IgG1. This is consistent with our data showing the neutralization activities mediated by purified IgM and IgA fractions from COVID-19 patients. Nonetheless, purified IgG fractions from convalescent plasma also exhibited potent neutralization, but since the fractions contained all four IgG subtypes, it remains unknown which IgG subtypes contribute to neutralization, although IgG1, the most abundant IgG isotypes in the blood, did not correlate with neutralization titers, and IgG2 and IgG3 present at low levels displayed weak correlations. The lack of correlation between IgG1 binding titers and neutralization activities may be explained by the fact that the dominant IgG1 responses may target sites not critical for blocking virus entry. Indeed, among mAbs isolated from six COVID-19 convalescent patients, > 40% of RBD-specific mAbs did not display neutralizing activities 26 and another study demonstrated that mAbs that bind to non-RBD epitopes on the spike protein had poor neutralization potencies 33 . Nonetheless, the absence of a correlation between IgG1 binding titers and neutralizing activities reported above requires more study given that data remain controversial since other studies have demonstrated correlations between neutralization titers and SARS-CoV-2 specific IgG levels 26,28,32 .

In addition to neutralization, non-neutralizing Ab activities have been implicated in protection from virus infection through potent Fc-mediated functions such as antibody-dependent cellular cytotoxicity (ADCC), antibody-dependent cellular phagocytosis (ADCP), and complement-mediated lysis; this is true for HIV, influenza, Marburg, and Ebola viruses 38,56-59 . These Fc activities were not evaluated in our . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint study, and their contribution to protection against infection and disease progression in humans is yet unclear 9,60,61 . Interestingly, a recent study demonstrated enrichment of spike-specific IgM and IgA1 Abs and spike-specific phagocytic and ADCD activity in plasma of individuals who recovered from infection, while nucleocapsid-specific IgM and IgA2 responses and nucleocapsid-specific ADCD activity were features enriched in deceased patients 62 . Defining the full potential of Abs against SARS-CoV-2 that includes neutralizing, non-neutralizing and enhancing activities is vital for developing the most effective vaccines and determining the optimal convalescent Ab treatment against COVID-19.

When we examined plasma specimens collected within 7-8 days after COVID-19 onset, we detected IgG and IgA against spike and RBD, as well as IgM. This is consistent with published reports showing that 100% of COVID-19-infected individuals developed IgG within 19 days after symptom onset and that seroconversion for IgG and IgM occurred simultaneously or sequentially 25 . IgA was also found early after infection (4-6 days after symptom onset) in another study 23 . These studies suggest that measuring total Ig, rather than IgG, would provide a better outcome for diagnosis of early disease.

Indeed, we found no correlation between the levels of different isotypes examined in our study. This lack of correlation may result from their non-synchronous, sequential induction (IgM first, then IgG, and finally IgA), but the presence of IgA early during acute infection also suggests the potential contribution of natural IgA, which, similar to natural IgM, arise spontaneously from innate B1 cells to provide the initial humoral responses before the induction, maturation, and class-switching of adaptive classical B cells 63,64 .

In summary, this study demonstrates that spike-and RBD-specific IgM, IgG1, and IgA1 Abs were present in serum or plasma of all or almost all analyzed COVID-19 convalescent subjects and were detected at extremely early stages of infection. The plasma of convalescent individuals also displayed neutralization activities that were mediated by IgM, IgG, and IgA1, although neutralization titers correlated more strongly with levels of IgM and IgA1 than IgG. The contribution of IgM and IgA Abs to . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint the neutralizing activities against SARS-CoV-2 demonstrates their importance of protective immunity against this virus.

Recombinant proteins. The recombinant spike and RBD proteins were produced as previously described 65 in Expi293F cells (ThermoFisher) by transfections of purified D.N.A. using an ExpiFectamine Transfection Kit (ThermoFisher). The soluble version of the spike protein included the protein ectodomain (amino acids 1-1213), a C-terminal thrombin cleavage site, a T4 foldon trimerization domain, and a hexahistidine tag. The protein sequence was also modified to remove the polybasic cleavage site (RRAR to A) and two stabilizing mutations (K986P and V987P, wild type numbering).

The RBD (amino acids 319-541) included the signal peptide (amino acids 1-14) and a hexahistidine tag.

Supernatants from transfected cells were harvested on day three post-transfection by centrifugation of the culture at 4000 g for 20 min. The supernatant was then incubated with 6 mL Ni-NTA agarose (Qiagen) for one to two hours at room temperature. Next, gravity-flow columns were used to collect the Ni-NTA agarose, and the protein was eluted. Each protein was concentrated in Amicon centrifugal units is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint 13 obtained from study participants enrolled in IRB-approved protocols at Icahn School of Medicine at Mount Sinai and the James J. Peter VA Medical Center. Study participants provided written consent at enrollment and agreed to sample banking and future research use of their banked biospecimen. Samples from these protocols included sera from seven participants with documented SARS-CoV-2 infection: specimens from P#1 (d8, d11, and d15 after symptom onset), P#2 (d7, and d10 after symptom onset), and RP#1-5. In addition, sera were used, which had been collected from healthy donors prior to the spread of SARS-CoV-2 in the U.S.A.

Ig fractionation. IgA was first isolated from plasma by mixing 1:2 diluted plasma with peptide M agarose beads (600 µL/28 mL plasma, InvivoGen #GEL-PDM) for 1.5 hours at room temperature.

Beads were then collected on a column and washed with PBS until protein reading (280 nm) by Nanodrop reached background. IgA was eluted from beads with a pH 2.8 elution buffer (Thermo Scientific #21004) and neutralized with pH 9 Tris buffer. The pass-through plasma sample was collected for IgG enrichment using protein G agarose beads (InvivoGen #GEL-AGG) using the same protocol as above and subsequently for IgM isolation using a HiTrap IgM column (G.E. Healthcare #17-5110-01) according to the manufacturer's instruction. An additional purification step was performed using Protein A Plus mini-spin columns to separate IgG from IgM. Protein concentrations were determined with Nanodrop prior to use in Luminex and neutralization experiments.

Luminex binding Ab assay. The SARS-CoV-2 antigens used in this assay were a soluble recombinant trimerized form of the spike protein and a recombinant RBD protein 66 . Antigens were coupled as previously described, with minor changes 44 . Each antigen was covalently coupled individually to a uniquely labeled fluorochrome carboxylated xMAP bead set at 2.0 μg protein/million beads using a twostep carbodiimide reaction with the xMAP Ab Coupling (AbC) Kit following to the manufacturers' instructions (Luminex, Austin, TX). The coupled beads were pelleted, resuspended at 5x10 6 beads/mL in storage buffer (PBS, 0.1% bovine serum albumin (B.S.A.), 0.02% Tween-20, and 0.05% sodium azide, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint pH 7.4), and stored at -80°C. Three to five million beads per batch were prepared in a 1.5 mL conical tube.

Before each experiment, the beads needed for a single run (2,500 beads/well x number of wells) were pelleted and resuspended in assay buffer (PBS, 0.1% B.S.A., 0.02% Tween-20) to deliver 2,500 beads in a volume of 50 μL/well. Sera/plasma samples were serially titrated (1:50 to 1:6400 final dilution) or diluted in assay buffer to 1:100 (for a final dilution of 1:200). The samples were then added as 50 μL/well to the wells containing the beads and incubated at room temperature for 1 hour at 600 rpm. After two washes in assay buffer, 100 μL/well of biotinylated antibodies specific for total Ig, IgG1, IgG2, IgG3, IgG4, IgA1, IgA2, or IgM was added and incubated for 30 min at room temperature on a plate shaker; these antibodies were rabbit biotinylated-anti-human total Ig (Abcam, catalog #ab97158) at 2 μg/mL, mouse biotinylated-anti-human IgG1 Fc (Invitrogen #MH1515) at 4 μg/mL, mouse biotinylated-anti-human IgG2 Fc (Southern Biotech #9060-08) at 1 μg/mL, mouse biotinylated-antihuman IgG3 Hinge (Southern Biotech #9210-08) at 3 μg/mL, mouse biotinylated-anti-human IgG4 Fc (Southern Biotech #9200-08) at 4 μg/mL, mouse biotinylated-anti-human IgA1 Fc (Southern Biotech #9130-08) at 4 μg/mL, mouse biotinylated-anti-human IgA2 Fc (Southern Biotech #9140-08) at 4 μg/mL or goat biotinylated-anti-human IgM (Southern Biotech #2020-08) at 3 μg/mL. After two washes, 100 μL/well of Streptavidin-Phycoerythrin (P.E.) at 1 μg/mL was added (BioLegend #405204) followed by a 30 min incubation at room temperature on a plate shaker. After two additional washes, 100 μL of assay buffer/well was added and put on a shaker to resuspend the beads. The plate was read with a Luminex Flexmap 3D instrument. Specimens were tested in duplicate, and the results were recorded as mean fluorescent intensity (MFI).

COV2pp production and titration. The SARS-CoV-2 pseudovirus (COV2pp) was produced as previously described 45 . Briefly, 293T cells were transfected to overexpress SARS-CoV-2 glycoproteins.

For background entry with particles lacking a viral surface glycoprotein, pCAGG empty vector was . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint R.L.U.) / virus control R.L.U.) *100). The inhibitory concentration 50% (IC50) and 90% (IC90) were respectively defined as the reciprocal sample dilution or purified Ig fraction concentration achieving 50% and 90% neutralization.

Statistical analysis. The statistical significance was determined by a two-tailed Mann-Whitney test and correlations analyzed with a Spearman rank-order correlation test using GraphPad Prism 8.

The raw data that support the findings of this study are available from the corresponding author upon request.

Zhu . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. Immunol. 20, 392-394 (2020).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2020. . https://doi.org/10.1101/2020.08.18.20177303 doi: medRxiv preprint of COV2pp by samples from (a) 28 COVID-19-convalescent individuals and (b) 11 COVID-19-negative individuals, compared to (c) a recombinant soluble RBD (sRBD) control. Each plasma or sera specimen was tested at 4-fold dilutions from 1:10 to 1:40,960, and sRBD was tested at 4-fold dilutions from 100 to 0.02 µg/mL. The data are shown as mean percentage of neutralization + SD of triplicate. The extrapolated titration curves were generated using a nonlinear regression model in GraphPad Prism (Inhibitor versus response -variable slope (four parameters), least squares regression). The dotted horizontal lines highlight 50% neutralization. TF#1   TF#2   TF#3   TF#4   TF#5   TF#6   TF#7   TF#8   TF#9   TF#10   TF#11   TF#12   TF#13   TF#14   TF#15   TF#16   TF#17   TF#18   TF#19   TF#20   TF#21   TF#22   TF#23   TF#24   TF#25 COVID-19-positive subjects TF#1   TF#2   TF#3   TF#4   TF#5   TF#6   TF#7   TF#8   TF#9   TF#10   TF#11   TF#12   TF#13   TF#14   TF#15   TF#16   TF#17   TF#18   TF#19   TF#20   TF#21   TF#22   TF#23   TF#24   TF#25 COVID-19-positive subjects TF#1   TF#2   TF#3   TF#4   TF#5   TF#6   TF#7   TF#8   TF#9   TF#10   TF#11   TF#12   TF#13   TF#14   TF#15   TF#16   TF#17   TF#18   TF#19   TF#20   TF#21   TF#22   TF#23   TF#24   TF#25 COVID-19-positive subjects control. Each plasma or sera specimen was tested at 4-fold dilutions from 1:10 to 1:40,960, and sRBD was tested at 4-fold dilutions from 100 to 0.02 µg/mL. The data are shown as mean percentage of neutralization + SD of triplicate. The extrapolated titration curves were generated using a nonlinear regression model in GraphPad Prism (Inhibitor versus responsevariable slope (four parameters), least squares regression). The dotted horizontal lines highlight 50% neutralization. 

Phase 1/2 Study to Describe the Safety and Immunogenicity of a COVID

RNA Vaccine Candidate (BNT162b1) in Adults 18 to 55 Years of Age: Interim Report

DNA vaccine protection against SARS-CoV-2 in rhesus macaques

Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2

The novel coronavirus 2019 (2019-nCoV) uses the SARS-coronavirus receptor ACE2 and the cellular protease TMPRSS2 for entry into target cells

SARS-CoV-2 invades host cells via a novel route: CD147-spike protein

Neuropilin-1 facilitates SARS-CoV-2 cell entry and provides a possible pathway into the central nervous system

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Plasma were tested at 4-fold dilutions from 1:10 to 1:40,960 or 1:20 to 1:81,920. Data are shown as the mean percentage of neutralization. The dotted horizontal lines highlight 50% neutralization. (b) Neutralization reciprocal IC 50 and IC 90 titers of RP#1-5 plasma. (c) Neutralization of COV2pp by purified IgM, IgG, and IgA fractions from 5 COVID-19-infected individual plasma (RP#1-5) compared to a control Ig fraction. The fractions were tested at 4-fold dilutions from 500 to 0.02 µg/mL. Data are shown as the mean percentage of neutralization. The dotted horizontal lines highlight 50% neutralization. (d) IC 50 of purified IgM, IgG, and IgA fractions from RP#1-5 plasma. The statistical significance was determined by a two

We thank all the donors for their contribution to research and Dr. Arthur Nadas (NYU School of Medicine) for statistical consultation.

The authors declare no competing interests.. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Fig. 6 . Summary of relative Ig isotype levels and neutralization titers.