COVID-19 patient serum less potently inhibits ACE2-RBD binding for various SARS-CoV-2 RBD mutants
Sample collection for assay validation
16 serum samples consisting of 12 samples from COVID-19 patients (ethical approval #179/2020/BO2, University Hospital Tübingen) and four negative pre-pandemic samples (Central BioHub) were measured by both virus neutralization test and RBDCoV-ACE2 as part of the assay validation.
For technical assay validation, negative pre-pandemic serum samples were purchased from Central BioHub and four previously collected vaccinated samples from healthcare workers vaccinated with the Pfizer BNT-162b2 vaccine30 (222/2020/BO2, University Hospital Tübingen) as well as one sample from a COVID-19 patient (#179/2020/BO2, University Hospital Tübingen) were used.
COVID-19 sample collection
266 serum samples were collected from 168 patients hospitalized at the University Hospital Tübingen, Germany, between April 17, 2020 and May 12, 2021. Longitudinal samples were measured from 35 of the 168 patients ranging from 2 to 12 samples per patient. All individuals were tested positive by SARS-CoV-2 PCR. Key characteristics of the study population are summarized in Table S1.
For serum collection, blood was extracted by venipuncture, with the serum blood collection tube rotated 180° two to three times to extract possible air bubbles in the sample. After a minimum coagulation time of 30 min at room temperature, serum was extracted by centrifugation for 15 min at 2000×g (RT) and then stored at − 80 °C until analysis. Time between blood sampling and centrifugation did not exceed 2 h.
Collection of samples and the execution of this study was approved by the Ethics committee of the Eberhard Karls University Tübingen and the University Hospital Tübingen under the ethical approval numbers 188/2020A and 764/2020/BO2 to Prof. Dr. Michael Bitzer. All participants signed the broad informed written consent of the Medical Faculty Tübingen for sample collection and all methods were performed in accordance with the relevant guidelines and regulations. Samples that were used for assay validation had their collection approved by the Ethics committee of the Eberhard Karls University Tübingen and the University Hospital Tübingen under the ethical approval numbers 222/2020/BO2 to Dr. Karina Althaus and 179/2020/BO2 to Prof. Dr. Juliane Walz. For collection of assay validation samples, informed written consent was obtained and all methods were performed in accordance with the relevant guidelines and regulations.
Expression and purification of SARS-CoV-2 RBD mutants
The expression plasmid pCAGGS, encoding the receptor-binding domain (RBD) of SARS-CoV-2 spike protein (amino acids 319–541), was kindly provided by F. Krammer49. Expression and purification of VOCs alpha, beta and epsilon was carried out as previously described30,50. RBDs of SARS-CoV-2 VOCs gamma, delta, eta, theta, kappa and A.23.1 were generated by PCR amplification of fragments from wild-type or cognate DNA templates and subsequent fusion PCR by overlap extension to introduce described mutations. Based on RBD wild-type sequence, primer pairs RBDfor, E484Krev and E484Kfor, RBDrev for VOC eta and RBDfor, V367Frev and V367Ffor, RBDrev for A.23.1 were used. VOC lambda was generated based on RBD wild-type sequence using primer pairs L452Qfor, L452Qrev and F490Sfor, F490Srev. VOC delta was generated based on VOC epsilon using primer pairs RBDfor, T478Krev and T478Kfor, RBDrev. Based on VOC alpha sequence, VOC theta was generated using primer pairs RBDfor, E484Krev and E484Kfor, RBDrev. VOC kappa was generated based on VOC eta sequence using primer pairs RBDfor, L452Rrev and L452Rfor, RBDrev. VOC gamma was generated based on VOC theta sequence using primer pairs RBDfor, K417Trev and K417Tfor, RBDrev. Amplificates were inserted into the pCDNA3.4 expression vector using XbaI and NotI restriction sites. The integrity of all expression constructs was confirmed by standard sequencing analysis. An overview of the primer sequences is shown in Table S2. Confirmed constructs were expressed in Expi293 cells30,34. Briefly, cells were cultivated (37 °C, 125 rpm, 8% (v/v) CO2) to a density of 5.5 × 106 cells/mL and diluted with Expi293F expression medium. Transfection of the corresponding plasmids (1 µg/mL) with Expifectamine was performed as per the manufacturer’s instructions. Enhancers were added as per the manufacturer’s instructions 20 h post transfection. Cell suspensions were cultivated for 2–5 days (37 °C, 125 rpm, 8% (v/v) CO2) and centrifuged (4 °C, 23,900 × g, 20 min) to clarify the supernatant. Afterwards, supernatants were filtered with a 0.22 µm membrane (Millipore, Darmstadt, Germany) and supplemented with His-A buffer stock solution (20 mM Na2HPO4, 300 mM NaCl, 20 mM imidazole, pH 7.4). The solution was applied to a HisTrap FF crude column on an Äkta pure system (GE Healthcare, Freiburg, Germany), extensively washed with His-A buffer, and eluted with an imidazole gradient (50–400 mM). Amicon 10 K centrifugal filter units (Millipore, Darmstadt, Germany) were used for buffer exchange to PBS and concentration of eluted proteins.
The in-house expressed RBD mutants were immobilized on magnetic MagPlex beads (Luminex) using the AMG Activation Kit for Multiplex Microspheres (# A-LMPAKMM-400, Anteo Technologies). In brief, 1 mL of spectrally distinct MagPlex beads (1.25 *107 beads) were activated in 1 mL of AnteoBind Activation Reagent for 1 h at room temperature. The beads were washed twice with 1 mL of conjugation buffer using a magnetic separator, before being resuspended in 1 mL of antigen solution diluted to 25 µg/mL in conjugation buffer. After 1 h incubation at room temperature the beads were washed twice with 1 mL conjugation buffer and incubated for 1 h in 0.1% (w/v) BSA in conjugation buffer for blocking. Following this, the beads were washed twice with 1 mL storage buffer. Finally, the beads were resuspended in 1 mL storage buffer and stored at 4 °C until further use.
Assay buffer (1:4 Low Cross Buffer (Candor Bioscience GmbH) in CBS (1 × PBS + 1% BSA) + 0.05% Tween20) was supplemented with biotinylated human ACE2 (Sino Biological, # 10108-H08H-B) to a final concentration of 342.9 ng/mL to produce ACE2 buffer. Working inside a sterile laminar flow cabinet, serum samples were thawed and diluted 1:25 in assay buffer before being further diluted 1:8 in ACE2 buffer resulting in a final concentration of 300 ng/mL ACE2 in all 1:200 diluted samples. Spectrally distinct populations of MagPlex beads (Luminex) coupled with RBD proteins of SARS-CoV-2 wild-type and variants alpha, beta, gamma, epsilon, eta, theta, kappa, delta, lambda, Cluster 5 and A.23.1 were pooled in assay buffer to create a bead mix (40 beads/µL per bead population). 25 µL of diluted serum was added to 25 µL of bead mix in each well of a 96-well plate (Corning, #3642). To allow comparison of ACE2 binding inhibition between different RBD mutants on a relative scale, 300 ng/mL ACE2 without added serum was measured in triplicates on every plate as normalization control. Additionally, one quality control sample was analyzed in triplicates on every plate. For blank measurement, 25 µL assay buffer instead of diluted sample was added to two wells per plate. Samples were incubated for 2 h at 21 °C while shaking at 750 rpm on a thermomixer. Following incubation, the beads were washed three times with 100 µL wash buffer (1 × PBS + 0.05 Tween20) using a microplate washer (Biotek 405TS, Biotek Instruments GmbH). For detection of bound biotinylated ACE2, 30 µL of 2 µg/mL RPE-Streptavidin was added to each well and the plate was incubated for 45 min at 21 °C while shaking at 750 rpm on a thermomixer. Afterwards, the beads were washed again three times with 100 µL wash buffer. The 96-well plate was placed for 3 min on the thermomixer at 1000 rpm to resuspend the beads before analysis using a FLEXMAP 3D instrument (Luminex) with the following settings: 80 µL (no timeout), 50 events, Gate: 7500–15,000, Reporter Gain: Standard PMT. MFI values of each sample were divided by the mean of the ACE2 normalization control. The normalized values were converted into percent and subtracted from 100 resulting in the percentage of ACE2 binding inhibition. Negative values were manually set to zero.
MULTICOV-AB34, an in-house produced SARS-CoV-2 antibody assay, was performed with all serum samples to measure RBD/S1/trimeric spike-specific IgG levels. The antigen panel was expanded to include RBD proteins from 11 different SARS-CoV-2 variants from which all, except the Cluster 5 variant from Sino Biological (# 40592-V08H80), were produced in-house. The assay was carried out as previously described30.
All experiments associated with the SARS-CoV-2 virus were conducted in a Biosafety Level 3 laboratory. The recombinant infectious SARS-CoV-2 clone expressing mNeonGreen (icSARS-CoV-2-mNG)51, corresponding to the 2019-nCoV/USA_WA1/2020 isolate, was obtained from the World Reference Center for Emerging Viruses and Arboviruses (WRCEVA) at the UTMB (University of Texas Medical Branch). The mNeon Green reporter gene introduced into ORF7 allows the differentiation between infected and uninfected cells.
The generation of icSARSCoV-2-mNG stocks and the MOI determination was performed as previously described52.
Virus Neutralization Assay (VNT)
VNTs were determined previously53. Briefly, 1 × 104 Caco-2 cells/well were seeded in 96-well plates the day before infection in media containing 5% FCS. Caco-2 cells were co-incubated with the SARS-CoV-2 strain icSARS-CoV-2-mNG51 at a MOI = 1.1 and serum samples in two-fold serial dilutions ranging from 1:40 to 1:5120. 48 h post infection, cells were fixed with 2% PFA and stained with Hoechst33342 (1 µg/mL final concentration) for 10 min at 37 °C. Following this, the staining solution was removed and exchanged for PBS. To quantify infection rates, images were taken with the Cytation3 (Biotek Instruments GmbH) and Hoechst + and mNG + cells were automatically counted by the Gen5 Software (Biotek Instruments GmbH). Infection rate was determined by dividing the number of infected cells through total cell count per condition. Virus-neutralizing titers (VNT50s) were calculated as the half-maximal inhibitory serum dilution.
Assay validation experiments
To determine the intra-assay precision of RBDCoV-ACE2, 12 replicates of four serum samples (Vac1–Vac4) were measured on a 96-well plate (Corning, #3642). Additionally, 15 replicates of the 300 ng/mL ACE2 normalization control and 12 replicates of the blank control containing only assay buffer without sample or ACE2 were measured. For inter-assay precision, five serum samples (Vac1–Vac4 and Inf1) were measured in triplicates in four independent experiments. Additionally, the quality control, the ACE2 normalization control and blank were also processed in triplicates in the same four experiments. Short-term stability was determined by storing ACE2 buffer under six different conditions before proceeding with the assay protocol. The prepared ACE2 buffer was stored 2 h, 4 h and 24 h at both 4 °C and room temperature and compared to ACE2 buffer without storage (fresh). Replicate MFI values of every sample (Vac1–Vac4 (vaccinated), Inf1 (infected) and pre-pandemic) were normalized to the values of the respective ACE2 normalization control. Freeze–thaw stability of the biotinylated ACE2 stocks was determined by analyzing six serum samples (Vac1–Vac4 (vaccinated), Inf1 (infected) and pre-pandemic) in triplicates, with ACE2 stocks undergoing 1 to 5 cycles. In addition to that, every sample was also processed with ACE2 not re-frozen once thawed (fresh, 0 freeze–thaw cycles). The MFI values of every sample were normalized to the values of the respective ACE2 normalization control. To investigate the stability of RBDCoV-ACE2 against variations of the used ACE2 concentration, six samples (Vac1–Vac4 (vaccinated), Inf1 (infected) and pre-pandemic) were analyzed with ACE2 concentrations ranging from 150 ng/mL to 350 ng/mL. Replicate MFI values of every sample were normalized to the values of the respective ACE2 normalization control. For analysis, the mean, standard deviation and coefficient of variation in percent of all replicates were calculated.
To confirm that the multiplex assay format has no undesirable effect on ACE2 binding inhibition values compared to singleplex measurements, 24 samples (pre-pandemic (n = 5) and COVID-19 infected (n = 19)) were analyzed in both singleplex and multiplex (for all VOCs).
One sample from each individual donor (n = 168) was analyzed with the commercially available in-vitro diagnostic test SARS-CoV-2 NeutraLISA (Euroimmun). The assay was performed according to the manufacturer’s instructions. For longitudinal donors with more than one sample available, the sample closest to 20 days after positive PCR diagnosis was picked. Negative values were manually set to zero.
Data collection and assignment to metadata was performed with Microsoft Excel 2016. Data analysis, visualization and curve fitting was performed with Graphpad Prism (version 9.1.2). Virus-neutralizing titers (VNT50s) as the half-maximal inhibitory serum dilution were calculated using 4-parameter nonlinear regression. Longitudinal curves were fitted using a one-site total binding equation. Correlations were analyzed using Spearman’s correlation coefficient. Significances were calculated (where appropriate) using Mann–Whitney U tests. Figures were edited with Inkscape (version 0.92.4). Data generated for this manuscript is available from the authors upon request.