Dynamic uncapping

A n understanding of de novo organelle construction comes one step closer thanks to He et al. (page 925), who fi nd that Atg11 leads Atg9 to the preautophagosomal structure (PAS). The PAS is intriguing because it is the site where fragments of membrane coalesce to form a new organelle: the autophagosome. In yeast, two fl avors of this process exist. Bulk autophagy is induced by starvation and is essentially a cell nondiscriminately eating itself. The cytoplasm-to-vacuole targeting (Cvt) pathway, however, is constitutive and picks selected cargoes for delivery to the vacuole (the yeast equivalent of the lysosome). Nobody knows what protein gets to the PAS fi rst, but Atg9, as the fi rst characterized transmembrane protein required for both pathways, is a good starting point. It cycles between the PAS and other sites, including mitochondria, probably as a way of collecting membrane fragments to build an autophagosome. It is not yet clear, however, what targets Atg9 to the PAS. He et al. fi nd that Atg11 uses one of its coiled coil regions to bind to Atg9, and this interaction and an intact actin cytoskeleton are required for Atg9's anterograde transport to the PAS. This transport mechanism is only required for the Cvt pathway; during bulk transport, another mechanism somehow ensures Atg9 cycling. Both Atg9 and Atg11 have multiple binding partners, but further efforts will be needed to identify where in the autophagy pathway these interactions occur. An assay using semipermeabilized cells should help determine which complexes are most important for creating an autophagosome. Rac controls nuclear entry R ac proteins are most familiar in the context of cytoskeletal regulation. Now, Kawashima et al. (page 937) fi nd that Rac1 and its regulator help STAT transcriptional activators to get into the nucleus. STAT proteins signal downstream of cyto-kine receptors. These researchers previously reported an association between STAT3 and the GTPase-activating protein MgcRacGAP. Here they report that Rac1 and MgcRacGAP bind STAT5A, and that the association between MgcRacGAP and STAT5A is enhanced by IL-3 signaling. Both Rac1 and MgcRacGAP were necessary for effi cient entry of STAT5A into the nucleus, and in semipermeabilized cells a dominant-negative Rac1 prevented binding of STAT5A to importin-α, and thus nuclear entry. Others have previously shown, using armadillo proteins as import substrates, that Rac1 has a nuclear localization sequence (NLS) that is active when Rac1 is in its active, GTP-bound form. Unpublished data suggest, however, …