DispatchR811

findings demonstrate that exocytosisregulates BMP secretion andE-cadherin membrane targeting inthe GSC niche (Figure 1A,B).

To test if adherens junctions arerequired for localized activation of BMPsignaling in the GSC niche, Michel et al.[12] knocked down E-cadherin in nichecells but failed to detect any changes inBMP activation in the GSC–nichejunction [12]. Unfortunately,adherens junctions remain intactin the absence of E-cadherin basedon b-catenin/Armadillo expression,probably due to N-cadherin expressionin the GSC niche. For the time being,it thus remains uncertain whetheradherens junctions are requiredfor localized BMP activation.

The study by Michel et al. [12] has,for the first time, directly shown thatBMP activation is restricted toadherens junctions at the stemcell–niche interface, and that BMPsecretion and E-cadherin membranetargeting require exocytosis andrecycling endosomes in the GSC niche[12]. However, three importantquestions remain to be addressed.The first question is whether localizedBMP signaling activation at adherensjunctions is biologically important. Arecent study [15] has shown that stat-depleted GSCs displaced from the hubdue to E-cadherin downregulation canstill maintain BMP signaling andself-renewal, casting some doubt onthe biological importance of adherensjunctions for BMP signaling. Possibly,BMP receptor complexes areco-localized to adherens junctions;without adherens junctions, BMPsignaling activation may be notlocalized but could still proceednormally (Figure 1C). To definitivelyanswer this question, depletion of bothE-cadherin and N-cadherin orb-catenin/Armadillo from the nichecould be used to further test if adherensjunctions are required for activatingBMP signaling in GSCs (Figure 1D).

The second question is whether theTIPF reporter can faithfullycaptureBMPsignaling activation in GSCs. Recently,cyst stem cells have been proposed tobe themajor BMP source for GSCs [15],but TIPF fails to detect BMP signalingactivation in the cyst stem cell–GSCinterface. This raises the concern thatTIPFmay not fully reflect BMP signalingactivation. It will be important togenerate a similar reporter for detectingSAX activation in the GSC niche. Thelast question is whether E-cadherin and

BMP are co-transported or transportedindependently using the same pathway,and whether adherens junctions arerequired for proper BMP targeting(Figure 1A,B). Answers to thesequestions will surely help betterunderstand the role of adherensjunctions in the regulation of BMPsignaling in the GSC niche as well asin other systems.

References1. Affolter, M., and Basler, K. (2007). The

Decapentaplegic morphogen gradient: frompattern formation to growth regulation. Nat.Rev. Genet. 8, 663–674.

2. Xie, T. (2008). Germline stem cell niches. InStemBook (MA: Cambridge).

3. Lopez-Onieva, L., Fernandez-Minan, A., andGonzalez-Reyes, A. (2008). Jak/Stat signallingin niche support cells regulates dpptranscription to control germline stem cellmaintenance in the Drosophila ovary.Development 135, 533–540.

4. Wang, L., Li, Z., and Cai, Y. (2008). The JAK/STAT pathway positively regulates DPPsignaling in the Drosophila germline stem cellniche. J. Cell Biol. 180, 721–728.

5. Hayashi,Y.,Kobayashi,S.,andNakato,H. (2009).Drosophila glypicans regulate the germlinestem cell niche. J. Cell Biol. 187, 473–480.

6. Guo, Z., and Wang, Z. (2009). The glypicanDally is required in the niche for themaintenance of germline stem cells andshort-range BMP signaling in the Drosophilaovary. Development 136, 3627–3635.

7. Liu, M., Lim, T.M., and Cai, Y. (2010). TheDrosophila female germline stem cell lineageacts to spatially restrict DPP function withinthe niche. Sci. Sign. 3, ra57.

8. Wang, X., Harris, R.E., Bayston, L.J., andAshe, H.L. (2008). Type IV collagens regulateBMPsignalling inDrosophila. Nature 455, 72–77.

9. Xia,L., Jia,S.,Huang,S.,Wang,H., Zhu,Y.,Mu,Y.,Kan, L., Zheng,W., Wu, D., Li, X., et al. (2010). TheFused/Smurf complex controls the fate ofDrosophila germline stem cells by generatinga gradient BMP response. Cell 143, 978–990.

10. Harris, R.E., Pargett, M., Sutcliffe, C.,Umulis, D., and Ashe, H.L. (2011). Bratpromotes stem cell differentiation via controlof a bistable switch that restricts BMPsignaling. Dev. Cell 20, 72–83.

11. Wilson, A., and Trumpp, A. (2006).Bone-marrow haematopoietic-stem-cellniches. Nat. Rev. Immunol. 6, 93–106.

12. Michel, M., Raabe, I., Kupinski, A.P.,Perez-Palencia, R., and Bokel, C. (2011).Local BMP receptor activation at adherensjunctions in the Drosophila germline stemcell niche. Nat. Comm. 2, 415.

13. Fuller, M.T., and Spradling, A.C. (2007). Maleand female Drosophila germline stem cells:two versions of immortality. Science 316,402–404.

14. Kawase, E., Wong, M.D., Ding, B.C., and Xie, T.(2004). Gbb/Bmp signaling is essential formaintaining germline stem cells and forrepressing bam transcription in the Drosophilatestis. Development 131, 1365–1375.

15. Leatherman, J.L., and Dinardo, S. (2010).Germline self-renewal requires cyst stem cellsand stat regulates niche adhesion in Drosophilatestes. Nat. Cell Biol. 12, 806–811.

16. Shivdasani, A.A., and Ingham, P.W. (2003).Regulation of stem cell maintenance and transitamplifying cell proliferation by tgf-betasignaling in Drosophila spermatogenesis.Curr. Biol. 13, 2065–2072.

17. Langevin, J., Morgan, M.J., Sibarita, J.B.,Aresta, S., Murthy, M., Schwarz, T.,Camonis, J., and Bellaiche, Y. (2005).Drosophila exocyst components Sec5, Sec6,and Sec15 regulate DE-Cadherin traffickingfrom recycling endosomes to the plasmamembrane. Dev. Cell 9, 365–376.

18. Blankenship, J.T., Fuller, M.T., and Zallen, J.A.(2007). The Drosophila homolog of the Exo84exocyst subunit promotes apical epithelialidentity. J. Cell Sci. 120, 3099–3110.

19. Wang, H., Singh, S.R., Zheng, Z., Oh, S.W.,Chen, X., Edwards, K., and Hou, S.X. (2006).Rap-GEF signaling controls stem cellanchoring to their niche through regulatingDE-cadherin-mediated cell adhesion in theDrosophila testis. Dev. Cell 10, 117–126.

20. Shibasaki, T., Takahashi, H., Miki, T.,Sunaga, Y., Matsumura, K., Yamanaka, M.,Zhang, C., Tamamoto, A., Satoh, T.,Miyazaki, J., et al. (2007). Essential role ofEpac2/Rap1 signaling in regulation of insulingranule dynamics by cAMP. Proc. Natl. Acad.Sci. USA 104, 19333–19338.

Stowers Institute for Medical Research,1000 East 50th Street, Kansas City,MO 64110, USA. Department of Anatomy andCell Biology, University of Kansas School ofMedicine, 3901 Rainbow Blvd, Kansas City,KS 66160, USA.*E-mail: tgx@stowers.org

DOI: 10.1016/j.cub.2011.08.051

Membrane Trafficking: DecodingVesicle Identity with ContrastingChemistries

Proteins involved in membrane traffic must distinguish between differentclasses of vesicles. New work now shows that a-synuclein and ALPS motifsrepresent two extreme types of amphipathic helix that are tuned to detect boththe curvature of transport vesicles as well as their bulk lipid content.

Adam Frost

Eukaryotic life emerged when cellsevolved the ability to isolate

biochemical micro-environmentswithin membranous compartments.Specialized reactions occur moreefficiently within these confined and

Figure 1. Structural features of the ALPS motif of GMAP-210 and the amino terminus ofa-synuclein.

(A) The ALPS of GMAP-210 (amino acids 1–38) is a sensor of curvature for neutral lipidmembranes with mono-unsaturated acyl chains. The ALPS of GMAP-210 is embedded ina packing defect associated with two molecules of palmitoyl-oleoyl-phosphatidyl-cholineand no sterols. (B) The amino terminus of a-synuclein (amino acids 9–41) is a sensor of thecurvature of anionic lipid membranes with sterols and saturated acyl chains. The amphipathicamino terminus is embedded in a leaflet containing sterols and the anionic, saturated lipiddipalmitoyl-phosphatidyl-serine. Yellow: Ala, Val, Leu, Ile, Met, Phe, and Trp. Red: Asp andGlu. Blue: Lys and Arg. Green: all other residues. (Note that the figure is only a schematicrepresentation of the general principles articulated by Pranke et al. [1]; some parameters —such as sterol content — have not yet been studied in depth.)

Current Biology Vol 21 No 19R812

tailored spaces — but the benefits ofcompartmentalization requirecommunication across membraneboundaries. Accordingly, cells evolvedthe ability to mold membranes intospheres and tubules in order to makeconnections between organelles orexchange material with the outsideworld. In every cell, fleets of transportvesicles are fetching and deliveringproteins and lipids with a specificitythat remains largely unexplained. Anew study by Pranke et al. [1] nowprovides insight into the chemistryunderlying vesicle recognition andadvances our understanding ofa protein implicated in Parkinson’sdisease — a-synuclein.

Since the discovery of membranecurvature sensors [2], we havelearned a great deal regarding themechanism(s) by which proteins detectdegrees of curvature in order to identifyvesicles, tubules or cisternae [3].During the same time period, ourknowledge of the diversity of functionsthat appear to be regulated bymembrane curvature-sensing factorshas expanded from traffickingpathways to include cytokinesis,pathogen invasion, phosphoinositidemetabolism and nuclear porebiogenesis [4–7]. However, curvatureis not the only property of membranesthat sensors may be tuned to

recognize. Small GTPases andphosphorylated inositol head-groupsare well-known signposts for differentcompartments [8,9]. Bulk lipid contentalso varies between organelles andhas strong effects on the physicalproperties of membranes. Forexample, the nuclear envelope andendoplasmic reticulum (ER) have lowsterol concentrations and are enrichedfor neutral, mono-unsaturated(‘kinked’) glycerophospholipids [10]. Incontrast, the plasma membrane andendo-lysosomal compartments haveabundant sterols and anionic lipidslike phosphatidylserine [11]. Moreover,the sphingolipids and theglycerophospholipids of the plasmaand endo-lysosomal membranes aremore likely to possess saturated orpolyunsaturated acyl chains [12].

These differences have led to thesuggestion that a basic organizingprinciple for the endomembranesystem may be the absence ofcompartments that combine a highdensity of anionic surface charge withloose lipid packing [3]. But whethercells have sensors to discriminatebetween these bulk bilayer propertieshas not been addressed directly—untilnow. Put another way, Pranke et al. [1]askedwhether the curvature sensors ofthe early secretory pathway haveadapted to the lower surface charge

and looser lipid packing found at theER and early Golgi, while curvaturesensors of the endo-lysosomal systemhave adapted to depend largely onelectrostatic forces. To test this ideathey compared the bilayer-bindingproperties of two amphipathic helices,each having been demonstratedpreviously to have curvature-sensingactivity despite being extremelydifferent in primary sequence [13,14].The first was the amphipathiclipid-packing sensor (ALPS) of thegolgin GMAP-210. The second was theamino terminus of a-synuclein,a synaptic protein that has beenstudied extensively because of its rolein Parkinson’s disease.Multiple studies have reinforced the

concept that ALPSmotifs, including theone found in GMAP-210, bindspecifically to highly curvedmembranes because they have robusthydrophobic faces composed of bulkyamino acids (tryptophan,phenylalanine, leucine), but few or nocationic residues along their polarfaces (Figure 1A) [3]. Without chargedresidues, the membrane partitioncoefficient is determinedpredominantly by hydrophobic forces.Single-liposome imaging assays andatomistic simulations confirm thatpacking defects increase in numberwith increasing degrees of convexity[15,16]. These defects are not justgreasy pockets that can accommodateALPS-like sequences. Simulationssuggest that, even on highly curvedmembranes, packing defects largeenough to accommodate thehydrophobic face of just 2–3 turns of anamphipathic helix rarely formspontaneously [15]. However, largedefects are observed throughcooperative coalescence asamphipathic sequences fold intohelices within the interfacial planebetween the aqueous cytosol andthe hydrocarbon core of themembrane. Specifically, simulationsof convex membranes can actuallydrive amphipathic helix folding byw3 kcal/mol, while flat and concavemembrane surfaces inhibit helix foldingby trapping unfolded peptideskinetically. Importantly, when chargedresidues (e.g. lysines) are substitutedfor neutral ones (e.g. threonine), ALPSsequences lose their specificity for highdegrees of curvature and bind toapproximately flat membranes [1,17].a-Synuclein is at the opposite

extreme in terms of amphipathic

DispatchR813

properties. The polar residues ofa-synuclein are cationic lysines andarginines at the interface between thehydrophilic and hydrophobic faces, andits hydrophobic face consists of smallresidues like valine and alanine [18](Figure 1B). When Pranke et al. [1]considered the lipid content ofendocytosed synaptic vesicles versusthe early secretory pathway, it struckthem that the meager hydrophobicface of a-synuclein might be adaptedto intercalate into membranes enrichedfor saturated acyl chains (with relativelyfewpackingdefects) and an abundanceof anionic phosphatidylserine andphosphoinositides. In accordancewith their hypothesis, in vitro assaysconfirmed that a-synucleindiscriminates between vesicles byboth size and the presence of anionichead-groups. Replacing smallhydrophobic amino acids with bulkierones abolished a-synuclein’s ability todiscriminate smaller from largervesicles.

These extreme contrasts inamphipathic chemistry reveal thatmembrane curvature sensing isheterogeneous and can be driven bya differential weighting of electrostaticversus hydrophobic forces. In the caseof ALPS motifs, curvature sensitivitycan be impaired by enhancingelectrostatic interactions. In the case ofa-synuclein, curvature sensitivity canbe impaired by enhancing hydrophobicinteractions. In both cases, the newdata remind us that cellular sensors areusually weak binders that have tradedaffinity for specificity. It alsostrengthens the notion that factors thatdrive curvature generation in vivo maybe distinguished from curvaturesensors by their ability to oligomerizeand exploit avidity. Sensing versusinducing curvature are the samephenomenon sampled at differentpoints along concentration-dependentbinding curves, but weak bindersthat oligomerize upon the membraneare more likely to drive deformationrather than simply detect curvaturein vivo [19,20].

Going beyond their in vitrocharacterizations of membranebinding, Pranke et al. [1] also report onsome remarkable observations in vivo.Since neither GMAP-210 nora-synuclein exist in budding yeast,the authors rationalized that theywould be less likely to have specificprotein–protein interactions whenexpressed in yeast that could affect

their targeting to vesicles of differentlipid content. When expressed in yeast,fluorescent and electron microscopicvisualization confirmed that thesesensors are targeted to distinct classesof vesicles. GMAP-210 was targeted toperi-ER clusters of w50 nm vesicles.a-Synuclein was targeted to clusters ofmore heterogeneous vesiclesjuxtaposed to the plasma membrane.Co-expression of GMAP-210 (mCherry)and a-synuclein (GFP) indicated thatthese clusters of vesicles were distinctin that they did not overlap or mix. Ina clever control, the authors invertedthe amphipathic sequences in order topreserve membrane partitioningproperties while disrupting anypotential protein–protein interactionsthat could be affecting in vivoobservations. The most parsimoniousinterpretation of their data is that thein vitro binding results for theamphipathic helices formed byGMAP-210 and a-synuclein reflect theirin vivo targeting properties, withGMAP-210 being tuned to bind neutralvesicles of the early secretory pathwayand a-synuclein being tuned to bindanionic vesicles derived fromendocytosis.

In general, our knowledge of thebiological roles of membrane curvaturesensors is still limited and, in particular,our understanding of the roles ofGMAP-210 at the Golgi and ofa-synuclein at the synapse remainsincomplete. The recent report byPranke et al. [1] has deepened ourappreciation of the physical principlesunderlying membrane curvaturesensing while raising new questionsabout their physiological roles. Inparticular, GMAP-210 and a-synucleintarget vesicles that appear by electronmicroscopy to be uncoated and verytightly clustered together, suggestingthat the heterologous presence ofthese proteins interferes with tethering,docking, or fusion steps. Future workinspired by this paper may uncovernovel roles for these contrastingamphiphiles in regulating the fusionof early secretory vesicles with thecis-Golgi or in preparing recycledsynaptic vesicles for another round ofexocytosis.

References1. Pranke, I.M., Morello, V., Bigay, J., Gibson, K.,

Verbavatz, J.-M., Antonny, B., andJackson, C.L. (2011). a-synuclein and ALPSmotifs are membrane curvature sensorswhose contrasting chemistry mediatesselective vesicle binding. J. Cell Biol. 194,89–103.

2. Bigay, J., Gounon, P., Robineau, S., andAntonny, B. (2003). Lipid packing sensed byArfGAP1 couples COPI coat disassembly tomembranebilayercurvature.Nature426, 563–566.

3. Antonny, B. (2011). Mechanisms of membranecurvature sensing. Annu. Rev. Biochem. 80,101–123.

4. Reese, M.L., and Boothroyd, J.C. (2009). Ahelical membrane-binding domain targets theToxoplasma ROP2 family to theparasitophorous vacuole. Traffic 10,1458–1470.

5. Doucet, C.M., Talamas, J.A., and Hetzer, M.W.(2010). Cell cycle-dependent differences innuclear pore complex assembly in metazoa.Cell 141, 1030–1041.

6. Rudner, D.Z., and Losick, R. (2010). Proteinsubcellular localization in bacteria. Cold SpringHarb. Pers. Biol. 2, a000307.

7. Chang-Ileto, B., Frere, S.G., Chan, R.B.,Voronov, S.V., Roux, A., and Di Paolo, G. (2011).Synaptojanin 1-mediated PI(4,5)P2 hydrolysisis modulated by membrane curvature andfacilitates membrane fission. Dev. Cell 20,206–218.

8. Di Paolo, G., and De Camilli, P. (2006).Phosphoinositides in cell regulation andmembrane dynamics. Nature 443, 651–657.

9. Behnia, R., and Munro, S. (2005). Organelleidentity and the signposts for membrane traffic.Nature 438, 597–604.

10. van Meer, G., Voelker, D.R., andFeigenson, G.W. (2008). Membrane lipids:where they are and how they behave. Nat. Rev.Mol. Cell Biol. 9, 112–124.

11. Leventis, P.A., and Grinstein, S. (2010). Thedistribution and function of phosphatidylserinein cellular membranes. Annu. Rev. Biophys. 39,407–427.

12. Takamori, S., Holt, M., Stenius, K., Lemke, E.A.,Grønborg, M., Riedel, D., Urlaub, H.,Schenck, S., Brugger, B., Ringler, P., et al.(2006). Molecular anatomy of a traffickingorganelle. Cell 127, 831–846.

13. Drin, G., Morello, V., Casella, J.-F., Gounon, P.,and Antonny, B. (2008). Asymmetric tethering offlat and curved lipid membranes by a golgin.Science 320, 670–673.

14. Middleton, E.R., andRhoades, E. (2010). Effects ofcurvatureandcompositionona-synucleinbindingto lipid vesicles. Biophys. J. 99, 2279–2288.

15. Cui, H., Lyman, E., and Voth, G.A. (2011).Mechanism of membrane curvature sensing byamphipathic helix containing proteins. Biophys.J. 100, 1271–1279.

16. Hatzakis, N.S., Bhatia, V.K., Larsen, J.,Madsen, K.L., Bolinger, P.-Y., Kunding, A.H.,Castillo, J., Gether, U., Hedegard, P., andStamou, D. (2009). How curved membranesrecruit amphipathic helices and proteinanchoring motifs. Nat. Chem. Biol. 5, 835–841.

17. Drin, G., Casella, J.-F., Gautier, R., Boehmer, T.,Schwartz, T.U., and Antonny, B. (2007). Ageneral amphipathic alpha-helical motif forsensing membrane curvature. Nat. Struct. Mol.Biol. 14, 138–146.

18. Jao, C.C., Hegde, B.G., Chen, J., Haworth, I.S.,and Langen, R. (2008). Structure of membrane-bound alpha-synuclein from site-directed spinlabeling and computational refinement. Proc.Natl. Acad. Sci. USA 105, 19666–19671.

19. Roux, A., Koster, G., Lenz, M., Sorre, B.,Manneville, J.-B., Nassoy, P., andBassereau, P. (2010). Membrane curvaturecontrols dynamin polymerization. Proc. Natl.Acad. Sci. USA 107, 4141–4146.

20. Frost, A., Perera, R., Roux, A., Spasov, K.,Destaing, O., Egelman, E.H., De Camilli, P., andUnger, V.M. (2008). Structural basis of membraneinvaginationbyF-BARdomains.Cell132, 807–817.

Department of Biochemistry and HuntsmanCancer Institute, University of Utah, School ofMedicine, Salt Lake City, UT 84112, USA.E-mail: frost@biochem.utah.edu

DOI: 10.1016/j.cub.2011.08.045