News & Views | Published:

Cell biology

Helices sculpt membrane

Many proteins are carried within cells in bubble-like sacs. These are pinched off from membranes inside the cell, and it seems that the Sar1p protein is key in both starting and finishing this budding process.

The cell contains a network of membrane-bound compartments that exchange proteins with each other and with the cell surface thanks to several haulage systems, each providing a specific link between one station and another. At the departure point, specialized ‘coat’ proteins wrap up a small area of the lipid membrane, shaping it into a bulging ‘bud’ and gathering up proteins due to be transported inside it1. The bud detaches from the membrane — a stage called fission — to form a bubble-like ‘vesicle’ loaded with cargo. Lee et al.2 report in Cell that a coat protein called Sar1p, whose structure contains several α-helices, initiates buds for one type of vesicle by thrusting one of its helices into the membrane, causing it to balloon outwards2.

We knew that Sar1p begins the formation of so-called COPII vesicles, but quite how was unclear. These vesicles transfer proteins from a membrane-bound structure called the endoplasmic reticulum, where they are made, to another such structure, the Golgi apparatus, where they are processed into their final form. A common cellular fuel called guanosine triphosphate (GTP) activates Sar1p. When Sar1p binds to GTP, it exposes a short α-helix at its amino (or N) terminus that anchors the protein to the membrane of the endoplasmic reticulum. There, Sar1p recruits two large COPII protein complexes, Sec23/24p and Sec13/31p, which polymerize into a curved lattice. Studies using artificial lipid vesicles called liposomes show that adding all these components is sufficient to generate coated buds on the liposome and, less efficiently, free coated vesicles3. Now Lee et al.2 report how Sar1p contributes to the initial moulding of the membrane and, less expectedly, to membrane fission.

Using electron microscopy, the authors first show that Sar1p alone can deform liposomes into long, narrow tubules, but only when it is bound to GTP, suggesting an involvement of the N-terminal helix. To demonstrate this, they swap this helix for a peptide that binds to an artificial lipid. As expected, the Sar1p mutant still binds to liposomes containing the artificial lipid but no longer deforms them.

Both normal Sar1p and the domain-swapped mutant can interact with the other COPII complexes, so the next step was to compare incubations conducted with the complete set of COPII proteins. Puzzlingly, though, buds do form in the presence of the swapped Sar1p mutant, yet free vesicles are scarce. In line with this, follow-up experiments conducted on membranes derived from endoplasmic reticulum show that an intact N terminus in Sar1p is key to the efficient release of COPII vesicles. So, if there is no doubt that the spherical shell formed by Sec23/24p and Sec13/31p is central to the sculpting of the membrane, Lee and colleagues' study2 implies that the N terminus of Sar1p is not merely a simple piece of tape that sticks the COPII coat to the membrane, but that it has an active role in membrane deformation and fission.

The N-terminal helix of Sar1p is amphipathic — that is, it has a hydrophobic face and a hydrophilic face. The wide hydrophobic ‘hull’ should insert between the lipid acyl chains of the membrane, while the polar hydrophilic side interacts with the lipid heads and the watery environment of the cytoplasm (Fig. 1a). From model studies, we know that this kind of helix is designed to float on biological membranes, with the axis lying at the interface between the polar and nonpolar lipid regions4. The membrane is a tightly packed bilayer of lipids, so when the N-terminal helices from numerous Sar1p proteins adsorb on its surface, they will expand the outer layer and, because the bilayer has a finite area, compress the inner layer. As a result, the membrane will bend and dome. Indeed, when Lee et al. replaced bulky amino acids in the hydrophobic side of the helix with smaller ones, Sar1p was less able to make tubules from the liposomes and to generate transport vesicles from isolated membranes in vitro.

Figure 1: Sar1p in budding and fission.

a, The N-terminal 18 amino acids of Sar1p form an amphipathic helix, seen here as though looking along the central axis with each amino acid identified by a one-letter code. The highly hydrophobic amino acids (orange) constitute a wide side, whereas polar amino acids (red, negatively charged; blue, positively charged; purple, hydroxylated) make up a smaller one. Other amino acids are in yellow. b, On activation by GTP, Sar1p inserts its N-terminal helix into the membrane. Lee et al.2 show that, once there, it bends the membrane outwards (+) and recruits the Sec23/24p complex, which polymerizes into a coat with Sec13/31p. Eventually, the coated bud is attached to the membrane by a small neck, with a negative curvature (−) in one direction and a positive curvature (+) in the other. Several Sar1p N-terminal helices may orientate parallel to the main neck axis, where they might further constrict the membrane and help fission and release of the bud.

Because Sar1p recruits Sec23/24p, which has a three-dimensional structure that is adapted to a convex surface, it is easy to imagine how the two proteins work in concert to bend the membrane at early stages of coat assembly5 (Fig. 1b). However, the role of Sar1p at the fission step is less intuitive. The curvature of the bud neck resembles that of a horse saddle, being negative in one direction and positive in the other. If the N-terminal helix of Sar1p invades the neck, its most plausible orientation is to align along the neck axis (Fig. 1b). A ring of parallel helices emerging from the coat edge may further constrict the neck and help membrane fission. Notably, COPII-coated buds on liposomes show a wider neck with N-terminal Sar1p mutants than with the unmutated form (Figs 3 and 8 in ref. 2).

The formation of clathrin-coated vesicles, which transport cargoes from the cell surface, follows an analogous process to that of COPII vesicles in that a short N-terminal helix of the protein epsin allows the plasma membrane to deform6. However, the epsin helix is shorter and has a smaller hydrophobic hull. Moreover, its polar side contains several electrically charged residues that bind specifically to PIP2, a negatively charged lipid that is a hallmark of the plasma membrane. So if the insertion of hydrophobic residues from amphipathic helices seems to be a common mechanism for inducing membrane curvature, subtle changes in the sequence may govern the ability to deform specific cellular membranes.

Hydrophobic and polar residues form the two broad classes of the amino-acid alphabet, and their segregation is the basis of the amphipathic helix. Yet hydrophobic amino acids vary in size, and this should influence the helix ‘footprint’ on the membrane. Likewise, the polar amino acids (such as hydroxylated, basic or acidic residues) in the other side do not interact to the same extent with the lipid polar heads7,8. No doubt, the language of membrane-deforming helices at the complex membrane–water interface is very rich and remains to be translated.


  1. 1

    McMahon, H. T. & Mills, I. G. Curr. Opin. Cell Biol. 16, 379–391 (2004).

  2. 2

    Lee, M. C. S. et al. Cell 122, 605–617 (2005).

  3. 3

    Matsuoka, K. et al. Cell 93, 263–275 (1998).

  4. 4

    Hristova, K. et al. J. Mol. Biol. 290, 99–117 (1999).

  5. 5

    Bi, X. et al. Nature 419, 271–277 (2002).

  6. 6

    Ford, M. G. et al. Nature 419, 361–366 (2002).

  7. 7

    Segrest, J. P. et al. J. Lipid Res. 33, 141–166 (1992).

  8. 8

    Bigay, J., Casella, J. F., Drin, G., Mesmin, B. & Antonny, B. EMBO J. 24, 2244–2253 (2005).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.