Budding yeast contain cortical patches, which, in turn, seem to house endocytic adaptors and cytoskeletal proteins. But these patches can vary in composition and in their dynamics, so any relationship between the actin cytoskeleton and endocytosis has, until now, been circumstantial. A recent Cell paper by David Drubin's group, though, outlines a role for actin in the internalization of the budding yeast endocytic complex.

The authors began by studying the localization and dynamics of six yeast proteins involved in endocytosis using green fluorescent protein (GFP)-tagging experiments. The Arc15 subunit of the Arp2/3 complex (which is required for actin nucleation) and Abp1 (an activator of the Arp2/3 complex) both had lifetimes of ∼15 sec. On the other hand, two other Arp2/3 activators, Pan1 and Las17, as well as Sla1 (an endocytic adaptor) and Sla2 (which is thought to function at the actin cytoskeleton–endocytic machinery interface), lasted slightly longer (30–40 sec). All the patches, except those containing Las17 (which typically remained at their site of formation) had an initial, movement-restricted phase of formation followed by a motile phase. The patches also moved from the cortex towards the cell centre.

Next, the authors observed that different proteins were recruited to patches invariantly and sequentially. Sla1 was an early patch component, and was joined by Abp1 (and Arc15), before both proteins disappeared. Similarly, Las17 and Sla2 were later joined by Abp1 and Arc15. All this regularity indicated to the authors that these changes in patch composition would reflect changes in patch behaviour. For example, Sla1–GFP patches started to move slowly towards the centre when Abp1 was recruited to patches, and, as Sla1–GFP disappeared from the patch, the fast phase of motility began. Because filamentous actin and Arc15 colocalize with Abp1 in patches, Drubin's group proposed that actin polymerization might be responsible for 'propelling' endocytic vesicles into the cell, so they treated cells with latrunculin A to sequester actin monomers. Their results indicated that actin polymerization was required for a presumed endocytic vesicle, plus any associated proteins, to move away from the cortex, and for this complex to later disassemble. Furthermore, sla2δ cells also showed inhibited patch motility, consistent with a role for Sla2 in endocytosis.

As patch motility was inhibited in both sla2δ cells and by latrunculin A, Drubin's group took a look at actin in sla2δ cells. Rather than the expected punctate staining that normally occurs, actin 'comet tails' accumulated at the cortex. At the junction with the cell cortex, the comet tails associated with Sla1, and, because Sla1 has recently been reported to link certain types of cargo to the endocytic machinery, the authors studied the localization of Ste2, a receptor cargo protein, in sla2δ cells. Ste2 accumulated in punctae at the junction between comet tails and the cell cortex, indicating that this site probably represents a blocked endocytic site and that, normally, the function of Sla2 is to join actin with the endocytic site to render it productive for internalization. Finally, the authors used fluorescence recovery after photobleaching (FRAP) experiments to show that actin filaments were nucleated at/near the endocytic complexes at the cortex, and were disassembled inside the cell.

So Drubin and colleagues describe a model for early endocytosis in budding yeast that resolves many previously unaddressed issues. The initial step is the assembly, in a non-motile complex at the plasma membrane, of endocytic adaptors (such as Sla1) and Arp2/3 activators (such as Las17 and Pan1), which interact with each other. After ∼20 seconds, actin, the Arp2/3 complex and Abp1 are recruited to the patch; at this point, Sla1, Sla2 and Pan1 move inwards. The early patch components are then disassembled, and patches containing only late components undergo a transition to a second phase of fast movement, during which the late components are disassembled. The early complex is probably propelled into the cytoplasm by forces generated by actin polymerization. These forces might also invaginate the plasma membrane and be involved in the release of the endocytic vesicle. So the pathway is there — how and when each of the steps is regulated is the next challenge.