The formation of clathrin-coated vesicles (CCVs) from clathrin-coated pits (CCPs) during endocytosis has now been visualized using live-cell imaging. It might not quite be the stuff of blockbuster movies, but this observation has enabled Merrifield, Perrais and Zenisek to show that CCP invagination and scission are tightly coupled, and that actin polymerization plays an essential part.

In their study, Merrifield et al. fused the extracellular domain of the transferrin receptor (an endocytic marker) to super-ecliptic phluorin, a pH-sensitive variant of green fluorescent protein (Tfnr–phl). The fluorescence of super-ecliptic phluorin is almost completely quenched when the pH changes from 7.4 to 5.5. They transfected Tfnr–phl into cells that contained fluorescently labelled clathrin, and then observed a high amount of colocalization between clathrin-coated 'structures' (CCSs) and Tfnr–phl. When the external pH was switched from 7.4 to 5.5, most — but not all — fluorescence was quenched. There were some acid-resistant Tfnr–phl spots, which appeared suddenly at CCSs — the first frame in the image series at which these were detected was designated the moment of scission. This protocol gave spectacular movies of CCVs appearing across the membrane of living cells.

Whether or not CCPs can support several rounds of CCV formation has been an unresolved issue, so Merrifield et al. studied CCPs that formed de novo during the imaging session as well as those that were already present from the beginning, and showed that CCVs developed from both newly-formed CCPs and longer-lived CCSs. Moreover, multiple scission events could be detected at several CCSs, with similar kinetics. Each scission, however, gave rise to a Tfnr–phl spot with only a fraction of the fluorescence of the total original Tfnr–phl patch. And some scission events were 'terminal' — that is, no further scission occurred — possibly because the entire CCP was internalized or because the hotspot was undergoing its last round of scission.

Another point of debate has been whether the nanometer-scale movements of CCPs away from the plasma membrane occurs before or after scission. Using a combination of illumination techniques and the cyclical changing of the pH of the buffer surrounding the cells, the authors measured scission and clathrin displacement from the membrane. Movement of CCPs away from the plasma membrane was seen to begin ∼40 sec before membrane scission, with the average movement during invagination being ∼40 nm.

Finally, to address another outstanding question — when actin polymerization occurs relative to scission — Merrifield et al. fluorescently labelled the actin-binding protein cortactin and visualized it, together with Tfnr–phl, at sites of scission. Cortactin recruitment to the prospective membrane scission site began well before scission occurred (coincidentally, ∼40 sec before scission), but peak cortactin recruitment coincided with membrane scission. Preventing actin polymerization using latrunculin B inhibited CCS dynamics and markedly reduced scission events. So, the authors' studies point to a scenario in which CCP movement and scission are normally coordinated and actin polymerization helps to physically separate the budding part of the CCP from its original site of formation.