Affinity ranking of antibodies using flow cytometry: application in antibody phage display-based target discovery.

Application of antibody phage display to the identification of cell surface antigens with restricted expression patterns is often complicated by the inability to demonstrate specific binding to a certain cell type. The specificity of an antibody can only be properly assessed when the antibody is of sufficient high affinity to detect low-density antigens on cell surfaces. Therefore, a robust and simple assay for the prediction of relative antibody affinities was developed and compared to data obtained using surface plasmon resonance (SPR) technology. A panel of eight anti-CD46 antibody fragments with different affinities was selected from phage display libraries and reformatted into complete human IgG1 molecules. SPR was used to determine K(D) values for these antibodies. The association and dissociation of the antibodies for binding to CD46 expressed on cell surfaces were analysed using FACS-based assays. We show that ranking of the antibodies based on FACS data correlates well with ranking based on K(D) values as measured by SPR and can therefore be used to discriminate between high- and low-affinity antibodies. Finally, we show that a low-affinity antibody may only detect high expression levels of a surface marker while failing to detect lower expression levels of this molecule, which may lead to a false interpretation of antibody specificity.

Molecular and biological characterization of human monoclonal antibodies binding to the spike and nucleocapsid proteins of severe acute respiratory syndrome coronavirus.

Human monoclonal antibodies (MAbs) were selected from semisynthetic antibody phage display libraries by using whole irradiated severe acute respiratory syndrome (SARS) coronavirus (CoV) virions as target. We identified eight human MAbs binding to virus and infected cells, six of which could be mapped to two SARS-CoV structural proteins: the nucleocapsid (N) and spike (S) proteins. Two MAbs reacted with N protein. One of the N protein MAbs recognized a linear epitope conserved between all published human and animal SARS-CoV isolates, and the other bound to a nonlinear N epitope. These two N MAbs did not compete for binding to SARS-CoV. Four MAbs reacted with the S glycoprotein, and three of these MAbs neutralized SARS-CoV in vitro. All three neutralizing anti-S MAbs bound a recombinant S1 fragment comprising residues 318 to 510, a region previously identified as the SARS-CoV S receptor binding domain; the nonneutralizing MAb did not. Two strongly neutralizing anti-S1 MAbs blocked the binding of a recombinant S fragment (residues 1 to 565) to SARS-CoV-susceptible Vero cells completely, whereas a poorly neutralizing S1 MAb blocked binding only partially. The MAb ability to block S1-receptor binding and the level of neutralization of the two strongly neutralizing S1 MAbs correlated with the binding affinity to the S1 domain. Finally, epitope mapping, using recombinant S fragments (residues 318 to 510) containing naturally occurring mutations, revealed the importance of residue N479 for the binding of the most potent neutralizing MAb, CR3014. The complete set of SARS-CoV MAbs described here may be useful for diagnosis, chemoprophylaxis, and therapy of SARS-CoV infection and disease.