What's new

Protein choreography

News and Views

Nature 438, 571-573 (1 December 2005) | doi:10.1038/438571a

Cell biology: Protein choreography

Mara C. Duncan1 and Gregory S. Payne1

Top of page
Just under the cell surface, proteins engage in an intricate ballet to drive a transport process called endocytosis. Much is known about the individual dancers, but now the choreography is revealed.

Endocytosis is the process by which cells gulp up small patches of their outer plasma membrane, sucking them inside to form small membrane-enclosed vesicles. These bubble-like structures provide a transport mechanism for carrying proteins that were embedded in the outer membrane, extracellular molecules associated with some of those proteins and small amounts of extracellular fluid1. The process can affect both normal and disease states of the cell through its roles in nutrient acquisition, cell growth, neural transmission and the entry of viruses into cells.

One well-characterized endocytosis pathway involves the protein clathrin, which coats the vesicles. The individual activities and interactions of many of the protein factors in the clathrin-mediated pathway have been characterized, but how these myriad interactions and activities are integrated to drive endocytosis is not clear. Drubin and colleagues2, writing in Cell, now provide a temporal and spatial framework for the protein dynamics responsible for endocytic vesicle formation in living yeast cells.

More than 40 years ago, electron microscopy provided the first glimpse of clathrin-coated vesicles and half-formed pockets in the plasma membrane3. These types of static image, combined with biochemical and molecular genetic analysis of clathrin and other coat components, engendered a generally accepted model of endocytic vesicle formation1. The clathrin coats assemble as polygonal lattices on the inside of the plasma membrane. Other coat components bind to the intracellular regions of certain plasma membrane proteins, capturing the proteins and ensuring that they will become cargo in the resulting vesicle. The assembly of the coat components is also associated with inward protrusion of the plasma membrane to generate a coated invagination. Finally, the invaginated membrane pinches off to generate a free, coated vesicle, which then sheds its coat as it moves farther into the cell.

Drubin and colleagues2 have examined endocytosis in the yeast Saccharomyces cerevisiae by tagging proteins implicated in the process with green fluorescent protein and following their movements in living cells using fluorescence microscopy4. This team previously demonstrated sequential assembly and disassembly of several proteins at sites of endocytosis5. By monitoring positions of assembled proteins at different times, the group defined three stages in vesicle formation: an initial phase, in which a patch of coat proteins forms but does not move; a second phase, when the endocytic structure moves relatively slowly to a short distance away from the plasma membrane; and a final phase, in which the structure moves rapidly into the cell interior. These phases were proposed to reflect coat assembly, coated membrane invagination and release of the free vesicle.

What emerges from the more comprehensive analysis in the current study is a detailed view of the dynamics of vesicle formation, governed by four sets of proteins, or 'modules' (Fig. 1). Endocytosis begins with assembly at the plasma membrane of a clathrin-coat module containing proteins involved in cargo recruitment and coat formation. In cells lacking clathrin, the other coat components still assemble and their kinetic behaviours are only modestly affected. But in these cells the number of endocytic sites on the plasma membrane is severely reduced, implying that clathrin facilitates the assembly of the endocytic machinery.

Figure 1: The dynamics of endocytic vesicle formation.

Initially, at the outer membrane of the cell, an immobile coat module assembles from clathrin and other cytoplasmic proteins (module 1). The coat captures cargo through interactions with the intracellular regions of the cargo proteins. Next, actin polymerization regulators (module 2) and actin polymerization/stabilizing modules associate with the coat, leading to the formation of actin filaments that drive slow inward movement of the coat (module 3). Finally, a scission module is recruited to separate the budded vesicle from the membrane (module 4). The freed vesicle moves rapidly into the cell and the coat components dissociate from the vesicle for additional rounds of vesicle formation.

High resolution image and legend (35K)

The second module associates with the coat module and consists of a group of proteins involved in regulation of actin polymerization. Actin is a small protein that polymerizes into filaments; filament polymerization can generate the force necessary to move proteins and membranes.

Proteins that promote actin filament formation, accelerate filament assembly and stabilize new filaments make up the third module. Once this module assembles, stable actin filaments become apparent and the endocytic patch on the membrane begins to move, probably indicating membrane invagination (but this remains to be demonstrated). Treatment with a drug that abolishes actin polymerization, or genetic deletion of third-module components, trapped the endocytic patches at the plasma membrane, so it seems that actin polymerization is involved in vesicle invagination.

The fourth module appears transiently and contains proteins that, when purified, can constrict spherical membrane structures into tubules. Remarkably, in cells lacking these proteins, the endocytic patch began to jut into the cell but sometimes snapped back to the cell surface. The authors suggest that this behaviour indicates a defect in vesicle release, consistent with membrane-constricting activity of the proteins. Once the vesicle is released, the coat protein module disassembles and the actin network continues to drive the vesicle deeper into the cell.

How well do the events in yeast correspond to those in mammalian cells? Although many components of the endocytic machinery are evolutionarily conserved between the two, there are differences in their requirements for actin and clathrin. In yeast, endocytosis is strictly dependent on actin assembly, but overall can proceed slowly without clathrin6. In contrast, clathrin-mediated endocytosis in mammalian cells seems to rely more on clathrin than on actin4. The results from yeast indicate that clathrin is an integral component of endocytic coats2, 7, with a significant but non-obligatory role in coat assembly. This confirms a long-standing observation that inactivation of clathrin causes immediate but partial endocytic defects8. Recent live-cell imaging of mammalian fibroblast cells revealed that, as in yeast, actin and actin-polymerizing proteins associate with most, if not all, clathrin coats, and promote invagination and vesicle movement9, 10. However, inhibition of actin polymerization causes only partial defects in endocytosis, indicating that although actin is important it is dispensable for endocytosis in fibroblasts. So, overall, the fundamental features and components of clathrin-mediated endocytosis have been well conserved. The variation between yeast and mammalian cells probably reflects cell-type-specific requirements. For example, the higher internal pressure in yeast might present an energy barrier to vesicle formation that makes actin force-generating mechanisms more significant than in some mammalian cells.

The ability to visualize single vesicles forming in living cells and to perturb the process by gene inactivation provides an unprecedented opportunity to probe the mechanism of endocytosis at a molecular level. Drubin and colleagues have made a good start, but there are still many questions to be answered. How are endocytic sites selected, for example, or are they randomly initiated? What regulatory mechanisms ensure ordered progression from coat assembly to vesicle release and coat disassembly? How is actin polymerization triggered and then harnessed to drive membrane invagination and perhaps scission? And how is endocytosis coordinated with other cellu-lar processes? It is evident that the dance has just begun.