A fusion peptide is a short hydrophobic segment of a membrane fusion protein that inserts into a target membrane. Once anchored, the protein refolds and pulls the two membranes together, through a hemifusion stage, until a pore opens. In influenza hemagglutinin it is the first 20 to 25 residues of HA2.
- A fusion peptide is a hydrophobic segment of a fusion protein that inserts into the target membrane.
- Fusion faces a high kinetic barrier from the repulsive hydration force between bilayers.
- The steps run from priming and triggering to an extended intermediate, fold-back, hemifusion and pore opening.
- Class I proteins use an N-terminal fusion peptide; classes II and III use internal fusion loops.
- In influenza hemagglutinin the fusion peptide is the first 20 to 25 residues of HA2, buried until low pH triggers release.
Why do two membranes need help to merge?
Two membranes need help to merge because water bound to their surfaces pushes them apart. Why don't two cells merge when they bump into each other? Their outer membranes are both made of the same oily lipid layers, so you might expect them to flow together like two drops of oil. They don't, and the reason is water. Each membrane surface holds a thin, tightly bound layer of water, and pushing two of those surfaces into contact takes a lot of energy.
A fusion peptide is a short, water-avoiding stretch of a larger protein that gets over that problem. It pokes into the target membrane and anchors the protein there. The rest of the protein then folds up and drags the two membranes together until they merge.
Fusion peptides are best known from enveloped viruses, which have to fuse their own membrane with a cell's membrane to get their contents inside. The same basic physics applies to membrane fusion inside cells, where different proteins do the pulling. This page covers the structural biology at textbook level: what the peptide is, where it sits, and what the protein around it does.
- Priming: a protease cut readies the protein
- Trigger: low pH or receptor binding
- Insertion: the fusion peptide enters the target membrane
- Fold-back: the protein collapses and pulls
- Hemifusion: outer leaflets merge
- Pore: the membranes become one
The energy barrier
In plain terms, the membranes resist being pushed close enough to touch, and something has to do the work of getting them there. Stephen Harrison's 2015 review in Virology describes fusion as facing a high kinetic barrier caused by a repulsive hydration force between the two bilayers. The membranes would be more stable merged than apart. They just cannot get there on their own at any useful speed.
Fusion proteins solve this by storing energy in a folded, spring-loaded shape and releasing it at the right moment. The review puts the central idea as coupling the free energy released when the protein refolds to the job of pulling the membranes into contact. The fusion peptide is the part that lets the protein grab the far membrane, so the refolding has something to pull on.
That arrangement explains a lot about fusion peptides. They must be hydrophobic enough to enter a lipid bilayer. They must also stay hidden until the right moment, because an exposed hydrophobic segment on a protein that has not yet reached its target would either stick to the wrong membrane or waste the stored energy.
What are the steps of viral membrane fusion?
Viral membrane fusion runs from priming and triggering through fold-back and hemifusion to an open pore. The same sequence of steps turns up across very different viral fusion proteins. The review lays them out roughly as follows.
- Priming. The protein is made in an inactive form. For many viruses a protease cuts it, which readies it to act later.
- Triggering. A signal sets off the change in shape. That can be low pH, such as inside an endosome, or binding to a receptor or co-receptor on the cell.
- Extended intermediate. The protein stretches out, and the fusion peptide or fusion loop swings toward the target membrane and inserts into it. For a moment the protein bridges both membranes.
- Fold-back. The protein collapses on itself, drawing the two membranes together.
- Hemifusion. The two facing lipid leaflets merge while the two inner leaflets are still separate.
- Pore opening. A fusion pore opens and widens, and the membranes become one.
Hemifusion is the step that most surprises people new to the topic. For a short time the membranes are joined on the outside and still sealed on the inside, a halfway state that researchers can trap and study.
Fusion peptides and fusion loops in the three classes
Viral fusion proteins fall into three structural classes. They look nothing alike, yet each carries a hydrophobic element that meets the target membrane. Where that element sits is one of the main differences.
| Class | Examples named in the reviews | Membrane-inserting element | Post-fusion shape |
|---|---|---|---|
| Class I | Influenza hemagglutinin (HA), HIV-1 gp41 | Fusion peptide at or near the new N-terminus made by cleavage | Trimer of hairpins built from helices |
| Class II | Flavivirus E protein, alphavirus proteins | Internal fusion loop at the tip of domain II | Trimer that folds back on itself |
| Class III | Rhabdovirus G (such as vesicular stomatitis virus), herpesvirus gB | Two neighboring hydrophobic fusion loops per subunit | Refolded trimer; the change can be reversible |
Strictly, only the class I element is a fusion peptide in the sense of a free end of the chain. In classes II and III the hydrophobic segment is a loop held between two parts of the folded protein, so it is called a fusion loop. Both do the same job at the membrane.
Influenza hemagglutinin as the worked example
Influenza hemagglutinin is the most studied case, and it shows how a fusion peptide is kept safe until it is needed. HA is made as a single chain and later cut into two pieces, HA1 and HA2. Harrison's 2008 review in Nature Structural and Molecular Biology notes that the cut normally happens in the trans-Golgi network. It creates a new N-terminus on HA2, and that new end tucks into a pocket along the protein's three-fold axis. The first 20 to 25 residues of HA2 are the fusion peptide, and in that tucked position they are buried.
When the virus is taken into an endosome and the pH drops, HA goes through a large rearrangement. HA1 separates from HA2, and the fusion peptide is released toward the target membrane.
What the peptide looks like in the membrane was studied directly by Han and colleagues in 2001, using NMR and EPR. They found the HA fusion domain to be a kinked, mostly helical, amphipathic structure, meaning one face is water-loving and the other oily. At pH 5 the kink sharpened and charged residues moved, opening a hydrophobic pocket that let the domain sit deeper in the bilayer. The authors suggested that this deeper insertion could help disturb the lipids enough for fusion to start.
One trimer is probably not enough to do the job alone. The 2008 review cites an estimate that three or four HA trimers may take part in a single fusion event.
Fusion inside cells
Cells fuse membranes all the time: vesicles merging with the plasma membrane, organelles exchanging contents. Much of that work is done by SNARE proteins rather than by anything resembling a viral fusion peptide.
SNAREs on the two membranes zip together into a bundle, and the zipping pulls the membranes into contact. Harrison's 2015 review makes the point that this is much the same generic mechanism as viral fusion, even though the proteins look completely different. In synaptic vesicle fusion the SNAREs pause partway through zipping, and calcium binding releases that pause so fusion can finish on cue.
The comparison is useful because it shows what the fusion peptide adds in viruses. A SNARE complex is anchored in both membranes from the start. A viral protein begins anchored in only one, so it needs a hydrophobic segment it can throw into the other.
A note on the words fusion peptide and fusion protein
The same words mean something else in protein chemistry, and catalog searches mix the two. In recombinant work, a fusion protein is two sequences joined into one chain by design. The 1992 paper by Francis and colleagues that introduced the long IGF-I analogues, for example, called them recombinant fusion protein analogues, because an IGF-I sequence was joined to the first 11 residues of methionyl porcine growth hormone. Nothing about that joined sequence involves membranes. Our page on how IGF-1 LR3 identity is confirmed covers that molecule.
Membrane-fusion peptides also turn up as short synthetic peptides in laboratories studying how lipids behave, since a peptide made on its own can be added to model membranes and measured. For background on what counts as a peptide in the first place, see what peptides are, and for the step up in size to folded chains, polypeptides and proteins.
FOR LABORATORY AND IN-VITRO RESEARCH USE ONLY. NOT FOR HUMAN OR ANIMAL CONSUMPTION.
Common questions
What is the difference between a fusion peptide and a fusion loop?
Why is the fusion peptide hidden before fusion?
What is hemifusion?
Do cells use fusion peptides too?
More science guides
Sources
- Harrison, Viral membrane fusion, Virology, 2015. Priming, triggering, the extended intermediate, fusion peptide or loop insertion, fold-back, hemifusion and pore opening; the three structural classes; the hydration-force barrier; comparison with SNARE-driven fusion.
- Harrison, Viral membrane fusion, Nature Structural and Molecular Biology, 2008. Hemagglutinin cleavage into HA1 and HA2, the fusion peptide as the first 20 to 25 residues of HA2, the low-pH rearrangement, the gp41 trimer of hairpins, and the estimate of three or four HA trimers per fusion event.
- Han, Bushweller, Cafiso and Tamm, Membrane structure and fusion-triggering conformational change of the fusion domain from influenza hemagglutinin, Nature Structural Biology, 2001. The HA fusion domain as a kinked, mostly helical amphipathic structure that bends more sharply at pH 5 and inserts deeper into the bilayer; PMID 11473264.
- Francis et al., Novel recombinant fusion protein analogues of insulin-like growth factor (IGF)-I, Journal of Molecular Endocrinology, 1992. Describes the Long IGF-I analogues built from the first 11 residues of methionyl porcine growth hormone joined to IGF-I, including Long [Arg3]-IGF-I; PMID 1378742.
