Barnard RJO, Narayan S, Dornadula G, Miller MD, Young JAT. 2004. Low pH is required for avian sarcoma and leukosis virus Env-dependent viral penetration into the cytosol and not for viral uncoating. J Virol 78:10433–10441.
Rein A, Mirro J, Haynes JG, Ernst SM, Nagashima K. 1994. Function of the cytoplasmic domain of a retroviral transmembrane protein: p15E-p2E cleavage activates the membrane fusion capability of the murine leukemia virus Env protein. J Virol 68:1773–1781.
An endosomal fusion receptor. The study of ebolavirus entry into cells has revealed a different fusion trigger: the viral fusion protein binds to a specific fusion receptor in the endosome membrane. Like some other class I viral fusion proteins, the ebolavirus glycoprotein (GP) is cleaved by furin-like proteases in the producer cells into two glycosylated subunits, GP1 and GP2. Following attachment to cells via the viral GP, viral particles are internalized and move to late endosomes. There, the sequential action of cathepsin proteases removes the majority of the glycosylated C terminus of GP1, allowing it to bind to Niemann-Pick C1 protein (NPC1) (Fig. 5.16). NPC1 is a multiple-membrane-spanning protein that resides in the late endosomes and lysosomes and participates in the transport of lysosomal cholesterol to the endoplasmic reticulum and other cellular sites. Individuals with Niemann-Pick type C1 disease lack the protein and consequently have defects in cholesterol transport; fibroblasts from these patients are resistant to Ebolavirus infection.
Figure 5.16 Entry of Ebolavirus into cells. Virus particles bind cells via an unidentified attachment receptor and enter by endocytosis. The mucin and glycan cap on the viral glycoprotein is removed by cellular cysteine proteases, exposing binding sites for NPC1. The latter is required for fusion of the viral and cell membranes, releasing the nucleocapsid into the cytoplasm. Courtesy of Kartik Chandran, Albert Einstein College of Medicine.
The Membrane Fusion Process
Studies with influenza virus envelope glycoproteins indicate that the initial rate of fusion depends on the surface density of HA, suggesting that clustering of several transmembrane protein trimers is required. The number of envelope trimers required to mediate fusion may vary depending on the virus studied, and is frequently debated. One has to bear in mind that assays measuring fusion rely on the use of artificial membrane-forming lipids in vitro; although they are useful tools to probe the mechanism of fusion, they might not reflect accurately the conditions required for fusion between viral and cellular membranes. Indeed, membrane composition is known to affect the fusion rate.
Fusion proceeds by a hemifusion intermediate where the membrane outer, but not inner, leaflets fuse. Such intermediates can be trapped when the HA membrane-spanning region is replaced by a lipid anchor (Box 5.3). The mechanism by which hemifusion progresses to complete fusion is unknown but does not appear very efficient. It was estimated that only 40% of events causing lipid mixing result in mixing of large aqueous molecules. Initially a small aqueous connection between the two membranes, referred to as a fusion pore, opens abruptly. The nascent fusion pore appears unstable and opens and closes repeatedly, flickers, and can ultimately close or remain open and small or open and dilate. Pore widening occurs either by assembly of several small fusion pores or by expansion of individual small pores to allow mixing of large aqueous molecules. Studies with HA proteins bearing amino acid changes in the fusion peptide demonstrated that this peptide plays an important role in pore widening.
A possible model of how fusion is completed could be provided by a comparison of viral fusion proteins with those found in cellular transport vesicles known as SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptors). The pairing of vesicle v-SNAREs to target membrane t-SNAREs drives membrane fusion and delivery of the vesicle cargo to its intracellular or extracellular destination. Each SNARE consists of two domains, a coiled coil and a transmembrane domain. The coiled coils of SNAREs positioned on two different membranes zip up, similar to the hairpin structure identified in viral fusion proteins. This zipping up releases the free energy required to mediate fusion between the two membranes (Fig. 5.17).
Class II Fusion Proteins
The envelope proteins of alphaviruses (E1) and flaviviruses (E) exemplify class II viral fusion proteins. In contrast to type I fusion proteins, E1 and E proteins do not form coiled coils. The proteins share a common fold with a central β-sandwich domain I flanked by domains II and III that tile the surface of the virus particles as dimers (Fig. 5.18A). A helical membrane-proximal domain links this structure to a transmembrane domain that spans the membrane twice. At low pH, the fusion proteins undergo conformational changes that extend domain II toward the endosome membrane, allowing insertion of the fusion loop in the target membrane (Fig. 5.18). During this transition, the dimers dissociate and reassemble into trimers. Refolding of domain III and the membrane-proximal region helix around the central β-sandwich brings the viral membrane close to the target membrane, adopting a hairpin structure as do class I fusion proteins (Fig. 5.13). This same structure is adopted by a eukaryotic protein and supports its function as a fusion protein during sexual reproduction (Box. 5.4).
EXPERIMENTS
Membrane fusion proceeds through a hemifusion intermediate
Fusion is thought to proceed through a hemifusion intermediate in which the outer leaflets of two opposing bilayers fuse, followed by fusion of the inner leaflets and the formation of a fusion pore. Direct evidence for this mechanism has been obtained with influenza virus HA. Mammalian cells in culture producing wild-type HA (left side of figure) are fused with erythrocytes containing two different types of fluorescent dye, one in the cytoplasm (red) and one in the lipid membrane (green). Upon exposure to low pH, HA undergoes conformational changes, the HA1 subunits tilt, and the fusion peptide is inserted into the erythrocyte membrane. The green dye is transferred from the lipid bilayer of the erythrocyte to the bilayer of the HA-producing cell, but the red die is not. Further conformational changes in the HA2 subunits bring the two membranes close together and fusion pores form. As the fusion pores expand, the red dye within the cytoplasm of the erythrocyte is then transferred to the cytoplasm of the HA-producing cell. An altered form of HA (right side of figure) lacking the transmembrane and cytoplasmic domains and with membrane anchoring provided by linkage to a glycosylphosphatidylinositol (GPI) moiety was produced. Upon exposure to low pH, the HA fusion peptide is inserted into the erythrocyte membrane, and green dye is transferred to the membranes of the HA-producing cell, just as in the wild-type protein. However, because no transmembrane domain is present, fusion pores do not form. The diaphragm becomes larger, but there is no mixing of the contents of the cytoplasm, indicating that complete membrane fusion has not occurred. These results prove that hemifusion, or fusion of only the outer leaflet of the bilayer,