ELISAs were performed as described (Bates et al., 2021b (link)). Plates were coated overnight at 4°C with 1 mg/mL recombinant SARS-CoV-2 spike receptor binding domain (RBD) protein (Bates et al., 2021c ) (BEI Resources NR-52309) or recombinant SARS-CoV-2 nucleocapsid (N) protein (BEI Resources NR-53797). Serum dilutions (6 × 3-fold for RBD, 6 × 4-fold for N) in duplicate were prepared in 5% milk powder, 0.05% Tween-20, in phosphate buffered saline (PBS), starting at 1:1600 (pan-Ig), 1:50 (IgA), 1:200 (IgG). The secondary antibodies used were pan-Ig (1:10,000 anti-human GOXHU IgG/A/M-HRP, A18847 Invitrogen), IgA (1:3,000 anti-human IgA-HRP, 411002 Biolegend), and IgG (1:3,000 anti-human IgG-HRP 555788, BD Biosciences). Plates were developed with o-phenylenediamine (OPD) (ThermoScientific). Absorbance at 492nm was measured on a CLARIOstar plate reader and normalized by subtracting the average of negative control wells and dividing by the highest concentration from a positive control dilution series. ELISA EC50 values were calculated by fitting normalized A492 as described (Bates et al., 2021b (link)). The limit of detection (LOD) was defined by the lowest dilution tested for RBD and half of the lowest dilution for N. Values below the LOD were set to LOD – 1.