Membrane fusion: a structural perspective on the interplay of lipids and proteins Lukas K Tamm  , Jonathan Crane and Volker Kiessling The fusion of biological membranes is governed by the carefully orchestrated interplay of membrane proteins and lipids. Recently determined structures of fusion proteins, individual domains of fusion proteins and their complexes with regulatory proteins and membrane lipids have yielded much suggestive insight into how viral and intracellular membrane fusion might proceed. These structures may be combined with new knowledge on the fusion of pure lipid bilayer membranes in an attempt to begin to piece together the complex puzzle of how biological membrane fusion machines operate on membranes. Addresses Department of Molecular Physiology and Biological Physics, University of Virginia, PO Box 800736, Charlottesville, VA 22908-0736, USA  e-mail: lkt2e@virginia.edu Current Opinion in Structural Biology 2003, 13:453–466 This review comes from a themed issue on Membranes Edited by Eric Gouaux and Stephen H White 0959-440X/$ – see front matter ß 2003 Elsevier Ltd. All rights reserved. DOI 10.1016/S0959-440X(03)00107-6 Abbreviations CHOL cholesterol DOPC dioleoyl-phosphatidylcholine DOPE dioleoyl-phosphatidylethanolamine EM electron microscopy EPR electron paramagnetic resonance HA hemagglutinin NSF N-ethylmaleimide-sensitive factor PC phosphatidylcholine PE phosphatidylethanolamine PEG polyethylene glycol PS phosphatidylserine PtdIns(4,5)P 2 phosphatidylinositol-4,5-biphosphate SM Sec1/Munc-18 SM sphingomyelin SNAP-25 synaptosome-associated protein of 25 kDa SNARE soluble NSF-attachment protein receptor t-SNARE target membrane SNARE VAMP vesicle-associated membrane protein v-SNARE vesicle membrane SNARE Introduction Membrane fusion is a ubiquitous cell biological process. It occurs intracellularly in membrane trafficking and exo- cytosis, including neurotransmitter release in synaptic transmission, and extracellularly in virus infection, gamete formation in sexual reproduction and myotube formation in organ development. In all these cases, two distinct membranes that separate different cellular com- partments have to merge and thereby connect the two compartments, whose contents are then free to mix and react. Lipid bilayer membranes do not spontaneously fuse. Energy must be invested to overcome hydration repulsion between membranes that approach each other and to disrupt the normal bilayer structure of the fusing membranes. This energy is expended on removing water molecules from the cleft between the two membranes, on bending the membranes that are to be fused and on creating nonbilayer lipid structures that function as fusion intermediates. The energy to drive biological membrane fusion is provided by highly specialized fusion proteins. The structures of several membrane fusion proteins and fragments of such proteins have been solved by X-ray crystallography and NMR. Structural work on membrane fusion proteins up to about 2001 has been covered in several excellent reviews [1,2  ,3  ]. Among the many structures of fusion proteins that have been solved, influenza hemagglutinin (HA) is unique because it is the only membrane fusion protein for which structures of the core fragment before and after fusion are known. It has been known for several years that this core fragment undergoes a dramatic conformational change upon fusion [4]. This advanced knowledge of the perti- nent structural transitions of HA has inspired many experiments on HA-mediated and other fusion systems. Therefore, influenza HA-mediated fusion has often served as the prototype fusion system and has also greatly influenced the interpretation of experimental results obtained with many other fusion systems. Having solved the structures of several soluble membrane fusion protein fragments, the central question in the field now is to determine how these proteins work on their substrates, that is, how they reshape the membranes that they are designed to fuse. In the following, we summarize recent progress in the structural biology of membrane fusion, drawing on influenza HA-mediated fusion as a paradigm. Pathways and possible intermediate lipid structures in membrane fusion Experimental and theoretical studies have suggested the existence of several intermediates on the pathway to membrane fusion. To put biological membrane fusion into perspective, we first discuss the fusion of pure lipid bilayer systems. As fusion of pure lipid bilayers does not occur spontaneously, external forces have to be applied to make them fuse. A conceptually simple and straightfor- ward experiment is to mechanically push two mica- supported lipid bilayers together in the surface forces 453 www.current-opinion.com Current Opinion in Structural Biology 2003, 13:453–466