NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 12 NUMBER 2 FEBRUARY 2005 107 and that these interactions are sufficient to form a complex via binding to Sar1p without Sec24p-Bet1p interactions. Hence, the FRET assay was sensitive enough to dissect specific interactions among coat components in the COPII complex. Sato and Nakano 3 then used the FRET assay to examine whether the complex of Sec24/23p bound to cargo (SNARE in this study) remains associated with the membrane after Sar1p-GTP is hydrolyzed (Fig. 1a). When YFP-Sec24/23p was added to CFP-Bet1p-SNARE liposomes preloaded with Sar1p-GTP, an initial increase in FRET signal (comparable to that of GMP- PNP) was observed, indicative of Sar1p- mediated formation of a complex between YFP-Sec24/23p and CFP-Bet1p-SNARE. Soon thereafter, the FRET signal declined owing to disassembly of the complex upon Sar1p-GTP hydrolysis. Because the decline was slower than for Sar1p-GTP hydrolysis (monitored by tryp- tophan fluorescence) the results indicate that the complex of Sec24p and Bet1p-SNARE is stable on membranes for a short period after hydrolysis of Sar1p-GTP. This behavior of COPII coat components is reminiscent of that of COPI components, which remain associated with Golgi membranes after Arf1-GTP hydro- lysis (determined from photobleaching studies in living cells) 8 . Therefore, there seems to be a commonality in the mechanism(s) whereby coat components assemble into metastable coats on the membrane. To determine whether the persistent associa- tion of Sec24/23p with membranes after Sar1p- GTP hydrolysis results from Sec24p interacting specifically with Bet1p-SNARE, the authors repeated the FRET assay using the mutant Bet1p-SNARE, which lacks the Sec24p-binding sequence. In this case, the signal declined at the same rate as Sar1p-GTP hydrolysis. Hence, the delay in membrane dissociation of Sec24/23p after Sar1-GTP hydrolysis requires a specific interaction between SNARE-cargo and Sec24p. This finding raises the possibility that SNAREs serve as scaffolds for Sar1p on the ER, directing Sec24/23p GAP activity toward Sar1p. In the last set of experiments, the authors examined the behavior of YFP-Sec24/23p and CFP-Bet1p-SNARE on membranes under conditions of multiple rounds of Sar1p-GTP hydrolysis (Fig. 1b). Here, the reaction mix was supplemented with Sec12p, which catalyzes efficient GDP-GTP exchange at Sar1p. Notably, the FRET signal persisted through the exchange reaction, indicating that YFP-Sec24/23p asso- ciation with Bet1p-SNARE is maintained, most likely because Sec12p continuously reactivates Sar1p to its GTP-bound form. This finding is important as it explains how kinetically stable COPII complexes can potentially form during Sar1p-GTPase cycles. In this scenario, the con- tinuous interaction of Sec24/23p with cargoes on the membrane—owing to repeated cycles of Sar1p-GTP hydrolysis and reactivation—leads to the formation of ‘coated’ domains or buds that are metastable; that is, they exist as long as Sar1p-GTP is being supplied to the domain. Because Sec12p is strictly localized to ER mem- branes, when the coated domains pinch off of the membrane, their coat is rapidly shed owing to depletion of Sar1p-GTP in the absence of GDP-GTP exchange activity. Not all interactions between different types of cargo and Sec24/23p may be stabilized sufficiently on the membrane during Sar1p-GTPase cycles to produce such coated domains. The authors thus speculate that oligomerization of cargo proteins may create combined signals for high-affinity binding to Sec24/23p. This would lead to the formation of cargo oligomer–Sec24/23p complexes, which in the presence of continuous Sar1p-GTPase cycles would create ‘coated’ sorting domains contain- ing more diverse types of cargo 9 . Many other aspects of COPII coat assembly and its modulation by cargo could potentially be studied using the FRET-based approach of Sato and Nakano 3 . For example, how does Sec13/31p interact with Sec23/24p to crosslink adjacent Sec23/24p complexes into a structural scaffold? Does Sec13/31p play a role in accel- erating Sec23p-mediated GAP activity? As the FRET-based assay can be used to dissect recruitment and cargo-capture events during the assembly of other coat systems on mem- branes, including COPI and clathrin coats, this assay offers a general new approach for clari- fying protein-protein interactions and their kinetics on the membrane. 1. Schekman, R. & Orci, L. Science 271, 1526–1533 (1996). 2. Rothman, J.E. & Wieland, F.T. Science 272, 227–234 (1996). 3. Sato, K. & Nakano, A. Nat. Struct. Mol. Biol. 12, 167– 174 (2005). 4. Nakano, A. & Muramatsu, M. J. Cell Biol. 109, 2677– 2691 (1989). 5. Matsuoka, K. et al. Cell 93, 263–275 (1998). 6. Springer, S., Spang, A. & Schekman. R. Cell 97, 145– 148 (1999). 7. Lee, M.C., Miller, E.A., Goldberg, J., Orci, L. & Schekman, R. Annu. Rev. Cell Dev. Biol. 20, 87–123 (2004). 8. Presley, J.F. et al. Nature 417, 187–193 (2002). 9. Miller, E.A. et al. Cell 114, 497–509 (2003). ACEing GPI release Satyajit Mayor A recent study shows that the testicular angiotensin converting enzyme (ACE) has a new unexpected activity: it is responsible for cleaving GPI off from proteins anchored on the sperm cell membrane. Fertilization is the process by which a sperm and an egg unite to form the first cell in the development of a multicellular organism. In mammals, the membrane properties of a viable and motile sperm undergo drastic alterations during the encounter with the mature oocyte, leading to binding and eventually the fusion with the mature oocyte (for a recent review on the fertilization process, see ref. 1). Many glycosylphosphatidylinositol (GPI)-anchored proteins, such as CD52, PH20 and TESP2, in the sperm are involved in triggering the matu- ration of the sperm and in oocyte recognition events. Most of these proteins are released in the medium surrounding the sperm, some during epididymal maturation, and others after a spe- cific membrane altering process termed ‘sperm capacitation’ (Fig. 1), in which a drastic reorga- nization of the membrane of the sperm takes place, partially due to efflux of β-hydroxysterols (cholesterol in mammals). This culminates in the initiation of the acrosome reaction 2 . Kondoh et al. 3 reported, in a recent issue of Nature Medicine, that testicular ACE specifi- cally cleaves GPI-anchored proteins from the membrane tether. This observation unites three seemingly disparate facets of the fertilization process: (i) the requirement for GPI-anchored sperm membrane proteins, (ii) a specific requirement for ACE in germ cells and (iii) a role for cholesterol removal. Furthermore, their The author is in the Cellular Organization and Signaling Group, National Centre for Biological Sciences, Bellary Road, Bangalore 560 065, India. e-mail: mayor@ncbs.res.in NEWS AND VIEWS © 2005 Nature Publishing Group http://www.nature.com/nsmb