Antigen-specific membrane fusion mediated by the haemagglutinin protein of influenza A virus: separation of attachment and fusion functions on different molecules.

Using genetic engineering techniques, two strategies for changing the receptor-binding specificity of the influenza A virus haemagglutinin (HA) protein whilst preserving its membrane fusion function, have been explored. The aim was to investigate whether the HA protein could be developed as an attachment/entry protein for targeting enveloped virus gene therapy vectors to specific cell populations. In the first strategy, a single chain antibody Fv region (scFv) specific for the hapten NIP was inserted between HA1 codons 139 and 145, to create a scFv-HA chimaeric protein. This protein was shown to possess anti-NIP binding activity, but membrane fusion activity could not be demonstrated. The possibility that linking the scFv domain directly to HA may have inhibited the HA fusion function led to the development of the second strategy. This involved separating the receptor-binding and membrane fusion functions of HA on to two different molecules. The feasibility of this strategy was tested by looking for fusion between NIP-conjugated red blood cells which lacked sialic acid (the HA protein's natural receptor) and Chinese hamster ovary cells that expressed both the above anti-NIP scFv-HA chimaeric protein (as a non-fusigenic, receptor-binding molecule) and wild-type HA protein (as a fusigenic, non-binding molecule) on their surface. Cell-to-cell fusion was observed in this system, indicating that the receptor-binding function of HA can be transferred to an adjacent molecule, and also changed in its specificity, without compromising its membrane fusion activity. This finding strongly suggests that the development of a two-molecule attachment and entry system for retargeting enveloped virus gene therapy vectors, based on HA, is a viable proposition.