The mechanism of translocation of peroxisomal proteins from the cytoplasm into the matrix is largely unknown. We have been studying this problem in yeast. We show that the peroxisomal targeting sequences SKL or AKL, with or without a spacer of nine glycines (G9), are sufficient to target chloramphenicol acetyltransferase (CAT) to peroxisomes of Saccharomyces cerevisiae in vivo. The mature form of CAT is a homotrimer, and complete trimerization of CAT was found to occur within a few minutes of synthesis. In contrast, import, measured by immunoelectron microscopy and organellar fractionation, occurred over several hours. To confirm that import of preassembled CAT trimers was occurring, we co-expressed CAT-G9-AKL with CAT lacking a peroxisomal targeting sequence but containing a hemagglutinin-derived epitope tag (HA-CAT). We found that HA-CAT was not imported unless it was co-expressed with CAT-G9-AKL. Both proteins were released from the organelles under mild conditions (pH 8.5) that released other matrix proteins, indicating that import had occurred. These results strongly suggested that HA-CAT was imported as a heterotrimer with CAT-G9-AKL. The process of oligomeric import also occurs in animal cells. When HA-CAT was co-expressed with CAT-G9-AKL in CV-1 cells, HA-CAT co-localized with peroxisomes but was cytoplasmic when expressed alone. It is not clear whether the import of globular proteins into peroxisomes occurs through peroxisomal membrane pores or involves membrane internalization. Both possibilities are discussed.

Is membrane fusion an essentially passive or an active process? It could be that fusion proteins simply need to pin two bilayers together long enough, and the bilayers could do the rest spontaneously. Or, it could be that the fusion proteins play an active role after pinning two bilayers, exerting force in the bilayer in one or another way to direct the fusion process. To distinguish these alternatives, we replaced one or both of the peptidic membrane anchors of exocytic vesicle (v)- and target membrane (t)-SNAREs (soluble N-ethylmaleimide-sensitive fusion protein [NSF] attachment protein [SNAP] receptor) with covalently attached lipids. Replacing either anchor with a phospholipid prevented fusion of liposomes by the isolated SNAREs, but still allowed assembly of trans-SNARE complexes docking vesicles. This result implies an active mechanism; if fusion occurred passively, simply holding the bilayers together long enough would have been sufficient. Studies using polyisoprenoid anchors ranging from 15–55 carbons and multiple phospholipid-containing anchors reveal distinct requirements for anchors of v- and t-SNAREs to function: v-SNAREs require anchors capable of spanning both leaflets, whereas t-SNAREs do not, so long as the anchor is sufficiently hydrophobic. These data, together with previous results showing fusion is inhibited as the length of the linker connecting the helical bundle-containing rod of the SNARE complex to the anchors is increased (McNew, J.A., T. Weber, D.M. Engelman, T.H. Sollner, and J.E. Rothman, 1999. Mol. Cell. 4:415–421), suggests a model in which one activity of the SNARE complex promoting fusion is to exert force on the anchors by pulling on the linkers. This motion would lead to the simultaneous inward movement of lipids from both bilayers, and in the case of the v-SNARE, from both leaflets.

Synaptic transmission relies on an exquisitely orchestrated series of protein-protein interactions. Here we show that fusion driven by neuronal SNAREs is inhibited by the regulatory protein complexin. Furthermore, inner-leaflet mixing is strongly impaired relative to total lipid mixing, indicating that inhibition by complexin arrests fusion at hemifusion. When the calcium sensor synaptotagmin is added in the presence of calcium, inhibition by complexin is relieved and full fusion rapidly proceeds.