Cell Biology International 2000, Vol. 24, No. 11, 819–838 doi:10.1006/cbir.2000.0632, available online at http://www.idealibrary.com on REVIEW DEPLOYMENT OF MEMBRANE FUSION PROTEIN DOMAINS DURING FUSION JOE BENTZ* and ADITYA MITTAL Department of Bioscience and Biotechnology, Drexel University, Philadelphia, PA 19104, U.S.A. Received 30 May 2000; accepted 18 July 2000 It is clear that both viral and intracellular membrane fusion proteins contain a minimal set of domains which must be deployed at the appropriate time during the fusion process. An account of these domains and their functions is given here for the four best-described fusion systems: influenza HA, sendai virus F1, HIV gp120/41 and the neuronal SNARE core composed of synaptobrevin (syn), syntaxin (stx) and the N- and C-termini of SNAP25 (sn25), together with the Ca 2+ binding protein synaptotagmin (syt). Membrane fusion begins with the binding of the virion or vesicle to the target membrane via receptors. The committed step in influenza HA- mediated fusion begins with an aggregate of HAs (at least eight) with some of their HA2 N-termini, a.k.a. fusion peptides, embedded into the viral bilayer (Bentz, 2000a). The hypothesis presented in Bentz (2000b) is that the conformational change of HA to the extended coiled coil extracts the fusion peptides from the viral bilayer. When this extraction occurs from the center of the site of restricted lipid flow, it exposes acyl chains and parts of the HA transmembrane domains to the aqueous media, i.e. a hydrophobic defect is formed. This is the ‘transition state’ of the committed step of fusion. It is stabilized by a ‘dam’ of HAs, which are inhibited from diffusing away by the rest of the HAs in the aggregate and because that would initially expose more acyl chains to water. Recruitment of lipids from the apposed target membrane can heal this hydrophobic defect, initiating lipid mixing and fusion. The HA transmembrane domains are required to be part of the hydrophobic defect, because the HA aggregate must be closely packed enough to restrict lipid flow. This hypothesis provides a simple and direct coupling between the energy released by the formation of the coiled coil to the energy needed to create and stabilize the high energy intermediates of fusion. Several of these essential domains have been described for the viral fusion proteins SV5 F1 and HIV gp120/41, and for the intracellular SNARE fusion system. By comparing these domains, we have constructed a minimal set which appears to be adequate to explain how the conformational changes can produce a successful fusion event, i.e. communication of aqueous compartments.  2000 Academic Press K: enveloped virus; influenza hemagglutinin; Env; HIV; SNARE; hydrophobic; sendai virus. INTRODUCTION Membrane fusion is a crucial event in a multitude of biological processes, and understanding the molecular mechanism is a central goal of biology. The ectodomain of the membrane fusion glyco- protein hemagglutinin (HA) of influenza virus, which requires low pH to initiate fusion, was the first membrane fusion protein whose crystal structure was solved (Wilson et al., 1981) and it remains the prototypical fusion protein (Skehel and Wiley, 1998; Sutton et al., 1998; Baker et al., 1999; Bentz, 2000b). The key differences between the ‘native’ structure of HA and the low pH structure is the formation of an extended coiled coil starting from the N-terminus of the native coiled coil and a helix-turn occurring within the C-terminal end of the native coiled coil, near the *To whom correspondence should be addressed. E-mail: bentzj@drexel.edu 1065–6995/00/110819+20 $35.00/0  2000 Academic Press