Construction of pseudo-SARS-CoV-2 virus and screening of specific antibodies directed to Spike glycoproteins
To simulate the functional structure of intact SARS-Cov-2 particles, we constructed a SARS-CoV-2 pseudovirus expressing the Spike glycoproteins outside the envelope and incorporating four viral uncompleted genes of SARS-CoV-2 (ORF1ab, ORFab, N, and M) into the lentivirus encoding green fluorescence protein (GFP) (Figure 1a). A lentivirus-transformed plasmid, pLV-SARS-CoV-2-F1abFabNE-GFP, and an envelope plasmid, pCMV3-2019-nCoV-Spike (S1+S2), were constructed first (Figure S1). After generating SARS-CoV-2 pseudoviruses through co-transfection, we tested their infectivity by transducing HEK-293FT-hACE2 cells (Figure S2). Abundant GFP fluorescence was observed via fluorescence microscopy at 48 and 72 h, suggesting successful construction and transduction of the pseudovirus (Figure 1b). The efficiency of incorporating the SARS-CoV-2 S protein and targeted sequence into the lentiviral cells was evaluated using monoclonal mouse anti-S (S2 domain) via western blotting and qRT-PCR targeted to the F1ab gene, respectively (Figure 1c and 1d). Specific bands at 190 and 80 kDa corresponded to the monomer S protein (S1 +S2) and S2 domains, respectively (Figure 1c, lane 3), and the positive signal for the F1ab gene supported the functional integrity of the SARS-CoV-2 pseudovirus.
Based the SARS-CoV-2 pseudovirus, 11 monoclonal antibodies (CQ2, CQ20, CQ25, CQ8, CQ12, CQ001, CQ100, CQ040, CQ042, CQ023, and M1E1) directed to the Spike glycoproteins of SARS-CoV-2 were prepared and screened according to their affinity using western blotting and particle gel assay. Influenced by its natural shape and charge, the active Spike glycoproteins of SARS-CoV-2 only could be detected with the CQ2, CQ25, and M1E1 antibodies (Figure 1e and 1f). Among these 11 candidates, the CQ25 antibodies showed excellent affinity and specificity for S protein.
Immunomolecular detection platform for SARS-CoV-2/pseudovirus particles
After preparing the SARS-CoV-2 pseudovirus and high-affinity antibodies, an immunomolecular detection platform for intact SARS-CoV-2/pseudovirus was designed. In summary, carboxyl groups were coated on the surface of nanomagnetic beads that were covalently bound to the amino groups of the antibodies to form peptide bonds, resulting in the generation of magnetic beads-antibodies complexes. After co-incubation at 20~25 ℃, anti-S antibodies-magnetic bead complexes specifically captured the SARS-CoV-2/pseudovirus particles. Following separation of the supernatant using a magnetic separator, the SARS-CoV-2/pseudovirus particles were enriched and separated from other “subviral particles” without S glycoprotein and free RNA fragments. Finally, nucleic acid-based detection via qRT-PCR excluded the influence of empty (genome-free) virions to further ensure that the detection platform only targeted intact SARS-CoV-2/pseudovirus particles (Figure 2a).
To establish an immunomolecular detection platform for intact SARS-CoV-2 particles, carboxyl magnetic beads and anti-S antibody (CQ25) were coupled. The effect of conjugation was evaluated by measuring the protein concentration of the complexes following conjugation using a BCA assay (Figure 2b) and particle size analysis (Figure 2c). Compared with untreated carboxyl magnetic beads, the complexes showed much higher absorption intensities (approximately 10-fold) at 562 nm and much larger average effective diameters (approximately 2-fold), indicating that the anti-S antibodies were successfully conjugated with the magnetic beads. The complexes and corresponding supernatant were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Coomassie bright blue staining, which confirmed the successful coupling and optimization of the coupling parameters (Figure 2d). Next, we verified whether the SARS-CoV-2 pseudovirus could be captured by anti-S antibody-magnetic beads complexes (MB-CQ25) at the protein and nucleic acid levels. Because HIV1 P24 is the most abundant marker protein in the lentivirus capsid, we identified the viral capsid proteins of captured virions in a viral particle gel assay with the monoclonal mouse anti-HIV 1 p24 antibody. As shown in Figure 2e, specific bands were observed in the complexes following capture. Importantly, compared with the negative control samples, including the SARS-CoV-2 S pseudovirus, VSV-G pseudovirus, and HBV, only our SARS-CoV-2 pseudovirus generated a positive signal via nucleic acid-based detection following capture (Figure 2f). These results show that the immunomolecular detection platform specifically detected intact SARS-CoV-2/pseudovirus particles.
Immunomolecular assay detects only intact virus particles
To verify the detection target of the immunomolecular assay, sucrose density gradient ultracentrifugation was performed to isolate the pseudovirus particles. As shown in Figure. 3a, the S glycoprotein peaked in the 10–20% sucrose gradient according to western blot analysis, whereas the pseudovirus particles settled at the 40% sucrose gradient in the gel assay, suggesting the presence of abundant free S proteins (large excess over complete virions) in the cell supernatant. Additionally, qRT-PCR and immunomolecular assays were used to determine the RNA levels of total and intact virions in different density gradients. The data indicated that pseudovirus particles were enriched in the 40% sucrose gradient fraction. Nevertheless, the virogene level was significantly higher than that of the intact virion (P < 0.05), suggesting that some subviral particles (without S envelope proteins) were produced during virus packaging; these results are consistent with previous reports on HBV[44, 45]. Using different titers of intact virions to transduce 293FT-HEK-hACE2 cells, we found that the titer of intact virions detected in the immunomolecular assay was positively correlated with infectivity, suggesting that the detection target of the immunomolecular assay is intact virus particles (Figure 3b).
Validation of immunomolecular assay for intact SARS-CoV-2 particles
After optimizing the detection conditions, ten-fold gradient dilutions were prepared and analyzed using fluorescence qRT-PCR in parallel to determine the linear range (Figure S3). When the titer ranged between 6 × 102 and 6 × 107 transduction units (TU)/mL, there was a linear relationship between the Cq value and its titer (log10-transformed) in the immunomolecular assay (R² = 0.99) (Figure 4a). As shown in Figure 4b, there were significant differences between the negative samples with the 24 carboxy magnetic beads coupled with CQ25 antibodies (MB) and positive samples containing 24 instances of captured pseudovirus (P < 0.0001). Using a mean titer of the negative samples of +1.96 standard deviation, the limit of detection was 900 TU/mL. Although the Cq value could be measured when the pseudovirus titer was lower than 900 TU/mL, the results of fluorescence quantitative PCR were negative.
Human serum contains a large amount of albumin and various antibodies, which may affect the assay stability. Additionally, if the patient carries other non-specific viruses, the specificity and stability of the method may be affected. To evaluate these factors, the HBV supernatant collected from AD38 cells and patient serum with different titers of HBV were used to verify the anti-interference ability of this immunomolecular assay. As shown in Figure 4c, for the low-titer group, adding HBV and patient serum with different titers of HBV had a negligible effect on the quantitative qualitative detection rate of intact SARS-CoV-2 pseudovirus particles, with coefficients of variation of 1.5% and 2.0% for intra-assay and inter-assay analyses, respectively; those of the high-titer group were 1.4% and 1.0%. Thus, this method is significantly stable against non-specific viruses and human serum. We further examined the stability of the SARS-CoV-2 pseudovirus over time on various surfaces, including copper, aluminum, paper, and plastic, to mimic different environmental samples (Figure 4d). SARS-CoV-2 was more stable on plastic, aluminum, and paper than on copper, and viable virus was detected up to 96 h after application to these surfaces; however, the virion titers were greatly reduced (from 107.38 to 104.05 TU/mL on aluminum, from 107.38 to 104.58 TU/mL on paper, and from 107.38 to 104.84 TU/mL on plastic). Notably, although the total RNA produced a positive detection signal, live virus was not detected on the copper surface at 72 h. Moreover, the gap in the titer between the total RNA and intact virions increased over time under different conditions. These results suggest that RNA fragments had different stabilities under different conditions after the complete virions were degraded, supporting RNA levels alone should not be used to assess infection risk.