Rings of ESCRT III push out buds of membrane (shown in 3D).

Vesicles within vesicles might be made with the help of rings of membrane-bending polymers that are revealed in images from Hanson et al.

The budding of small vesicles into larger endosomes—known as multivesicular bodies (MVBs)—is directed by the sequential action of the ESCRT I, II, and III complexes. ESCRTs I and II bind to ubiquitinated cargo and to endosome-defining phospholipids and probably concentrate the soon-to-be budded cargo on the membrane of the endosome.

The function of ESCRT III has been more elusive, owing to its tendency to form large and insoluble complexes, but mutant phenotypes suggest it acts after the other ESCRTs. As a late player, the authors imagined, ESCRT III might deform the outer membrane to create inward-budding vesicles. Their new striking images support this idea.

Using deep-etch EM, Hanson and colleagues captured images of large ESCRT III polymers. The authors expressed various ESCRT III proteins at high levels, causing them to accumulate on both endosomes and the plasma membrane, where their structures could be more easily viewed.

The images revealed curved 5-nm filaments of ESCRT III polymers tightly associated with the membrane. Upon deletion of the C-terminal half of the protein, rings of the filaments induced curvature in the surrounded membrane, which pushed bud- and tubule-like structures outwards. On endosomes, such budding would result in the creation of internal vesicles. The deletion mutant probably lacks a self-inhibitory domain that normally limits its oligomerization and thus its ability to bend the membrane. The authors suggest that in wild-type cells, accessory proteins probably relieve this inhibition.

Artificially high levels of the ESCRT III proteins were used to obtain the images, but the authors argue that the images represent an exaggerated version of the normal process. The group has yet to figure out, however, how the ESCRT III rings produce buds without themselves getting trapped inside the resultant vesicles. Perhaps either the rings are initially very large and then tighten as the vesicles form (akin to a purse string) or sequential rings form as buds grow, with internal rings disassembling as new outer ones form.


Hanson, P.I., et al.
J. Cell Biol.