This study was performed with the approval of the Institutional Review Boards (IRB) at the Guangzhou Eighth People’s Hospital (20200134). Blood samples were drawn after each participant provided a written informed consent form. Fifty-two plasma samples were collected from convalescent COVID-19 patients at time points from one week to six weeks post-infection. These patients had positive RNA results for SARS-CoV-2 infection and were hospitalized at Guangzhou Eighth People’s Hospital, Guangzhou Medical University, China. In addition, 18 plasma samples from healthy blood donors were collected before the SARS-CoV-2 pandemic and were used for comparative analyses. Plasma samples were clarified, aliquoted, and stored at -80°C until use. Fresh PBMCs from two COVID-19 patients after four months of SARS-CoV-2 infection were used for B cell isolation using fluorescence-activated cell sorting (FACS). All blood samples were collected before or on May of 2020, the early outbreak of COVID-19 in China.

The coding sequence of the full-length N protein of SARS-CoV-2 (NC_045512.2), SARS-CoV (NC_004718.3), MERS-CoV (JX869059.2), OC43 (AY585228.1), or HKU1 (NC_006577.2) was inserted into pcDNA3.1 expression vector (Invitrogen, Waltham, MA) with a D7-tag on the C-terminus of each N protein gene as previously reported (37). The recombinant plasmid was transfected into mammalian 293T cells. After culturing, N protein-containing supernatants were collected and clarified by centrifugation. Aliquots of supernatants were added with a protease inhibitor and then stored at -20°C until use.

Western blot was used for analyzing N protein expression in culture supernatants of transfected 293T cells. Briefly, 20 µl boiled supernatant sample in reducing buffer was loaded into the 10% precast SDS-PAGE gel (Tris-GlyBeyoGel™ Plus PAGE, Beyotime Biotechnology, China) and followed by high-performance electrophoresis. Proteins were blotted onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA). The expression of N protein was probed by the anti-D7 tag Ab (Cliniqa, San Marcos, CA) and the horseradish peroxidase-conjugated secondary Ab (Jackson ImmunoResearch, West Grove, PA). The supernatant collected from the 293T cell culture without transfection or transfected with empty pcDNA3.1 vector was used as a negative control of N protein expression.

Patient-derived mAbs against SARS-CoV-2 N protein were generated from memory B cells (CD19⁺IgD^-IgM^-CD27⁺CD38^low, Supplementary Figure 1 ) in the PBMCs from COVID-19 patients using the antigen-specific B cell approach as previously reported (38). Briefly, memory B cells in fresh PBMCs from convalescent COVID-19 patients were sorted into 96-well culture plates with 100 cells per well. B cells were cultured for 7-10 days in culture medium supplemented with growth factors and stimuli including IL-21, Chk2 inhibitor, and CpG in the presence of feeder cells (irradiated PBMCs from healthy blood donors). The culture supernatants were screened for N protein binding using a capture ELISA. The ELISA microplates precoated with anti-D7 tag Ab were used to capture the N protein with D7-tag in the supernatant of 293T cells transfected with recombinant pcDNA3.1. Cells in the positive wells were subjected to RNA extraction for cDNA synthesis using the SuperScript™ III First-Strand Synthesis System (Invitrogen, Waltham, MA). The cDNA was used as a template of the nested PCR with gene-specific primers or primer mixes to amplify the transcripts of Ab heavy and light variable genes as previously reported (39). The PCR products were purified from agarose gel using QIAEX II Gel Extraction Kit (Qiagen, Hilden, Germany). Purified PCR products were digested with restriction enzymes AgeI/SalI, AgeI/BsiWI, or AgeI/XhoI (ThermoFisher Scientific, Waltham, MA) and followed by ligation into human IgG1, Igκ, or Igλ expression vectors (39). The recombinant Ab-expressing constructs were sequenced and analyzed using the IMGT/V-QUEST (IMGT, France), a sequence alignment software for the immunoglobulin sequences of the variable regions.

The K_D measurement and epitope binning of isolated hmAbs to SARS-CoV-2 N protein were conducted on the Octet K2 System (ForteBio, Fremont, CA) using the Bio-Layer Interferometry (BLI) as previously reported (40). Briefly, the Streptavidin (SA) sensor captured the biotin-labeled (EZ-Link™ NHS-PEG12-Biotin, ThermoFisher Scientific, Waltham, MA) anti-D7 Ab at 80 µg/ml, which subsequently captured the SARS-CoV-2 N protein in the supernatant. Serially diluted Abs against the N protein were added. After washing, the association and dissociation curves were obtained. The K_D values were calculated using Data Analysis SPSS 11.0. For epitope binning, the two Abs were added in tandem. Residual binding was calculated as the percent binding of secondary Ab in the presence of primary Ab relative to the binding of secondary Ab alone.

For epitope mapping of the isolated hmAbs, a panel of series-truncated N proteins including N(1-174), N(1-364), N(46-174), N(245-364), and N(245-419), each with D7-tags at the C-terminus, were expressed in mammalian 293T cells. These N subdomains/fragments covered the NTD (46-174) and CTD (245-364), two main domains of the N protein, as well as the flanking regions of NTD and CTD. The binding of each hmAb to these truncated N proteins was measured using a capture ELISA. The N protein fragments and capture ELISA were also used to test NTD- or CTD-specific Abs in the plasma from COVID-19 patients. The area under the curve (AUC) was calculated using GraphPad Prism 8.

Cross-binding of hmAbs or plasma Abs in COVID-19 patients to other human β-CoV N proteins was determined using the capture ELISA. The N protein of SARS-CoV, MERS-CoV, OC43, or HKU1 was anchored by precoated D7-tag capture Ab. Human mAbs starting at 10 µg/ml were used for preparing a 3-fold serial dilution that was applied to the capture ELISA. To test cross-reactive Abs in plasma samples, each plasma sample at a 1:100 dilution was applied to the capture ELISA. In addition to these COVID-19 patient samples, 18 plasma samples from healthy blood donors collected before the SARS-CoV-2 pandemic were also analyzed. All samples were detected in duplicate. The threshold was calculated by the OD value of negative controls with standard deviation (SD) in each microplate.

Statistical analysis was performed using GraphPad Prism 8.0. Data were expressed as mean ± SD. Differences between 2 groups were compared using unpaired t tests or Mann–Whitney U test. P<0.05 was considered statistically significant.

To isolate mAbs from memory B cells (CD19⁺IgD^-IgM^-CD27⁺CD38^low) in the peripheral blood of COVID-19 patients, we genetically expressed full-length SARS-CoV-2 N protein in mammalian 293 T cells. As shown in Figure 1A , the full-length SARS-CoV-2 N protein was detected in supernatant from 293 T cells transfected with recombinant pcDNA3.1 plasmids with an insert of SARS-CoV-2 N/D7-Tag fusion gene, whereas supernatant from 293 T cells transfected with an empty pcDN3.1 vector did not show any protein band. This genetically expressed full-length SARS-CoV-2 N protein was used to screen Ab-producing memory B cells in B cell culturing plates. B cells in Ab-positive wells were subjected to RNA extraction and subsequently RT-PCR amplification of genes expressing Ab heavy and light chains. As shown in Table 1 , five mAbs against SARS-CoV-2 N protein, namely N3, N5, N31, N83, and 3B7, were successfully obtained from fresh memory B cells in the peripheral blood of two convalescent COVID-19 patients. Three of these hmAbs (N3, N5, and N31) were isolated from one male adult COVID-19 patient. These three hmAbs had the same VH gene, HV1-69, but had different sequences and lengths in their complementarity-determining region 3 (CDR3). These hmAbs were also paired with different light chains, KV4-1, LV1-51, and KV3-15, respectively ( Table 1 ). N3, N5, and N31 hmAbs all exhibited high-affinity binding to SARS-CoV-2 N protein with the K_D values below 1 pM ( Figure 1B ; Table 2 ). The two other hmAbs, N83 and 3B7, were obtained from a female COVID-19 patient. N83 had HV4-61 VH gene that was paired with KV2-28 light chain, while 3B7 had VH1-24 VH gene that was paired with KV3-20 light chain. N83 and 3B7 had K_D values of 1.76 nM and 1.94 nM, respectively ( Figure 1B ; Table 2 ). Thus, N83 and 3B7 also exhibited high-affinity binding to SARS-CoV-2 N, albeit at lower degrees of reactivation strengths when compared to N3, N5 and N31.

To map antigenic epitopes of these five hmAbs to SARS-CoV-2 N protein, a panel of overlapping N protein fragments that covered the entire SARS-CoV-2 N protein sequence were expressed ( Supplementary Figure 2 ). These N protein fragments were used for mapping antigenic epitopes of the five hmAbs using ELISA assays. As shown in Figure 2 , these five hmAbs fell into two categories based on their reactivation strengths. N3, N5, and N31 hmAbs recognized the CTD domain ( Figures 2A, B ), while N83 and 3B7 hmAbs selectively bound to the NTD domain ( Figures 2A, B ). Further analysis of Ab competition by BLI showed that N3, N5, and N31 fully competed with each other, but did not compete with the NTD-specific hmAbs ( Supplementary Figure 3 ). These results suggest that three hmAbs against the CTD may have the same or overlapping binding epitopes, but have different binding epitopes with two hmAbs that were responsive to the NTD.

Since SARS-CoV-2 N protein shares substantial sequence conservation with the protein counterparts of other human β-CoVs, particularly SARS-CoV (20–23), these patient-derived mAbs might cross-react with other human β-CoV N proteins. To clarify this speculation, we genetically expressed N proteins of SARS-CoV, MERS-CoV, OC43, and HKU1 ( Figure 3A ). These N proteins were used for testing cross-activities of the five hmAbs. As shown in Figure 3B , all five hmAbs were cross-reactive to the N protein of SARS-CoV. None of these five hmAbs were reactive to any of the N proteins from MERS-CoV, OC43, or HKU1 ( Figure 3B ; Supplementary Figure 4 ). Thus, the five patient-derived mAbs to SAR-CoV-2 N protein NTD and CTD cross-react with their N protein counterparts of SARS-CoV, but not other human β-CoVs.

We also studied whether Abs against the N protein NTD and CTD were elicited during SARS-CoV-2 natural infection. A total of 52 plasma samples from convalescent COVID-19 patients were used for ELISA analysis of Ab responses to the full-length and truncated fragments of the N proteins. Forty-one (78.8%) plasma samples had detectable levels of N-specific Abs. According to the binding pattern of each sample to NTD and CTD, two different responsive patterns were observed, namely Pattern1 and Pattern2 ( Figure 4 ; Supplementary Table 1 ). In Pattern1, Abs in 26 plasma samples reacted with NTD, CTD, and full-length N proteins at comparable levels, although CTD-Ab level appeared slightly higher than that of total N-Abs (p=0.0525) ( Figure 4 ). While in pattern2, NTD-Abs were dominant in 15 plasma samples (p=0.026, NTD-Abs vs CTD-Abs in pattern2). The levels of N-Abs and NTD-Abs in pattern2 were higher than in pattern1 (p=0.03 and p=0.009, respectively, for their comparisons between pattern1 and pattern2) ( Figure 4 ). The levels of CTD-Abs in these two patterns were comparable (p=0.8375) ( Figure 4 ). These results indicate that Abs against the NTD and CTD of SARS-CoV-2 N protein are elicited during natural infection and that dominant anti-NTD Ab responses occur in some infected individuals.

Recent studies have demonstrated that convalescent COVID-19 sera contain high levels of cross-reactive Abs to SARS-CoV structural proteins indicating robust cross-reactive Abs are elicited during natural SARS-CoV-2 infection (28–30). To confirm these data and also to clarify whether N-specific Abs in COVID-19 patients cross-reacted with N proteins of other human β-CoVs, we first screened Abs in plasma samples that were collected from 18 healthy blood donors to the N proteins of all five human β-CoVs including SARS-CoV-2, SARS-CoV, MERS-CoV, OC43, and HKU1. Abs to OC43 and HKU1 N proteins were detected in 38.9% (7 out of 18) and 22.2% (4 out of 18) of plasma samples, respectively ( Figure 5 ). There was no plasma sample that had detectable Abs to the N proteins of SARS-CoV-2, SARS-CoV, or MERS-CoV( Figure 5 ; Supplementary Figure 5 and Supplementary Table 2 ). Thus, prior to the COVID-19 pandemic, approximately 22 - 39% healthy individuals have detectable Abs against the N proteins of OC43 and HKU1, the causative agents of common cold, while none of them have detectable Abs to the N proteins of SARS-CoV-2, SARS-CoV, or MERS-CoV.

Next, we tested 52 plasma samples collected from COVID-19 patients for Abs against the N proteins of SARS-CoV-2, SARS-CoV, MERS-CoV, OC43, and HKU1. As described above, Abs against SARS-CoV-2 N protein were detected in 78.8% (41 out of 52) of plasma samples from these COVID-19 patients. We also found that these 41 plasma samples had cross-reactive Abs to the N protein of SARS-CoV. The average OD values were 1.35 ± 0.70 and 1.32 ± 0.72 for the N proteins of SARS-CoV-2 and SARS-CoV, respectively ( Figure 5 ), suggesting that the cross-reactive Abs react equally with the N proteins of SARS-CoV-2 and SARS-CoV. No plasma sample had cross-reactive Abs to the N protein of MERS-CoV ( Figure 5 ; Supplementary Figure 5 ). Cross-reactive Abs to the N proteins of OC43 and HKU1 were detected in 36.5% (19/52) and 19.2% (10/52) of samples, respectively ( Figure 5 ). These detection frequencies are similar to those observed in healthy blood donors ( Figure 5 ; Supplementary Table 2 ).