Interactions of HIV-1 antibodies 2F5 and 4E10 with a gp41 epitope prebound to host and viral membrane model systems.

Two HIV-1 recognition domains for the human monoclonal antibodies (MAb) 2F5, which recognises the core sequence ELDKWA, and 4E10, which recognises the core sequence NWFNIT, serve as promising models for immunogens in vaccine development against HIV-1. However, the failure of these recognition domains to generate broadly reactive neutralizing antibodies, and the putative membrane-binding properties of the antibodies raised to these recognition domains, suggest that additional features or recognition motifs are required to form an efficient immunogen, which could possibly include the membrane components. In this study we used an extended peptide epitope sequence derived from the gp41 native sequence (H-NEQELLELDKWASLWNWFNITNWLWYIK-NH), which contains the two recognition domains for 2F5 and 4E10, to examine the role of model cell (POPC) and viral (POPC/cholesterol/sphingomyelin) membranes in the recognition of these two antibodies. By using a surface plasmon resonance biosensor, the binding of 2F5 and 4E10 to membranes was compared and contrasted in the presence and absence of prebound peptide epitope. The recognition of the peptide epitope by each MAb was found to be distinct; 2F5 exhibited strong and almost irreversible binding to both membranes in the presence of the peptide, but bound weakly in the absence of the peptide epitope. In contrast, 4E10 exhibited strong membrane binding in the presence or absence of the peptide epitope, and the binding was essentially irreversible in the presence of the peptide epitope. Overall, these results demonstrate that both 2F5 and 4E10 can bind to membranes prior to epitope recognition, but that high-affinity recognition of gp41-derived epitope sequences by 2F5 and 4E10 occurs in a membrane context. Moreover, 4E10 might utilise the membrane to access and bind to gp41; such membrane properties of 2F5 and 4E10 could be exploited in immunogen design.