EFC Domain Proteins: Some Assembly Required

Developmental Cell

Previews Homem, C.C., and Peifer, M. (2008). Development 135, 1005–1018.

C., Schwartz, M.A., and Yap, A.S. (2014). Curr. Biol. 24, 1689–1699.

Priya, R., and Yap, A.S. (2015). Curr. Top. Dev. Biol. 112, 65–102.

Jodoin, J.N., Coravos, J.S., Chanet, S., Vasquez, C.G., Tworoger, M., Kingston, E.R., Perkins, L.A., Perrimon, N., and Martin, A.C. (2015). Dev. Cell 35, this issue, 685–697.

Martin, A.C., Kaschube, M., and Wieschaus, E.F. (2009). Nature 457, 495–499.

Roh-Johnson, M., Shemer, G., Higgins, C.D., McClellan, J.H., Werts, A.D., Tulu, U.S., Gao, L., Betzig, E., Kiehart, D.P., and Goldstein, B. (2012). Science 335, 1232–1235.

Leerberg, J.M., Gomez, G.A., Verma, S., Moussa, E.J., Wu, S.K., Priya, R., Hoffman, B.D., Grashoff,

Pollard, T.D., and Borisy, G.G. (2003). Cell 112, 453–465.

Sawyer, J.K., Harris, N.J., Slep, K.C., Gaul, U., and Peifer, M. (2009). J. Cell Biol. 186, 57–73.

F-BAR/EFC Domain Proteins: Some Assembly Required Linton M. Traub1,* 1Department of Cell Biology, University of Pittsburgh School of Medicine, 3500 Terrace Street, Pittsburgh, PA 15261, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2015.12.003

Polymeric spirals of crescent-shaped BAR-domain superfamily proteins are touted to girdle eukaryotic phospholipid bilayers into narrow tubules for trafficking and membrane remodeling events. But McDonald et al. (2015) in this issue of Developmental Cell question whether this broadly held view and conceptually appealing mechanism for membrane sculpting is really overhyped.

Most membrane-bounded organelles depend on regular contact with other intracellular compartments for proper long-term functioning. A repulsion-minimizing intermediate in these physical exchanges often involves deformation of a small region of a planar or spherical donor membrane into a perpendicularly oriented cylindrical projection. Because of the universal importance of this membrane restructuring activity, several distinct families of soluble proteins, including the BAR (Bin/Amphiphysin/ Rvs) domain superfamily, can translocate to a membrane site to couple lipid bending with transport. Emblematic is the EFC (extended FCH [Fes/CIP4 homology]) domain anti-parallel dimer, alternatively dubbed the F-BAR because of an overall structural similarity between the EFC and BAR domain a-helicalbundle folds (Shimada et al., 2007) (Figure 1). Two common features in many of these proteins is the presence of a rigid crescent-shaped surface that closely apposes the bilayer, and a propensity to oligomerize into rings or spirals. These properties seem to promote the physical forces necessary to

constrain the bilayer for local restructuring (Frost et al., 2008; Simunovic et al., 2015), with the intrinsic geometry of the concave BAR domain dimer specifying tubule diameter. In the fission yeast Schizosaccharomyces pombe, the F-BAR/EFC domain protein Cdc15p is required late in mitosis for contractile ring biogenesis and ensuing cytokinesis. While plasma membrane binding is essential, McDonald et al. (2015) show convincingly in this issue of Developmental Cell that, unlike FBP17, the Cdc15p F-BAR/EFC domain does not generate tubules in several separate assays. Yet Cdc15p clearly oligomerizes into tip-to-tip-oriented (membraneapposed) filaments, and the bulk of the careful investigation by McDonald and colleagues (2015) dissects how this self-organization occurs and is necessary for proper contractile ring formation and dynamics. The authors find that the F-BAR/EFC domain underpins the ability of Cdc15p to act as a sort of molecular Velcro, assembling into a circumferential string-like armature in the middle of dividing yeast to coordinate actin-based contractile forces by posi-

664 Developmental Cell 35, December 21, 2015 ª2015 Elsevier Inc.

tioning the necessary protein effectors. Critically, yeast Cdc15p is not the only EFC/F-BAR domain protein that seems incapable of dramatic membrane remodeling. The literature is replete with striking EM images of liposome tubules extruded beneath coiled BAR domain superfamily polymers, or with cellular plasma membrane dramatically transformed into a mass of tubules on forced expression of BAR/EFC domains (Frost et al., 2008; Shimada et al., 2007; Takeda et al., 2013). Setting aside the supraphysiological BAR protein concentrations used in many of these experiments, the extensive intermolecular contacts in self-assembled dense helical lattices suggest physical deformation of the underlying membrane. So how can Cdc15p function reliably if not packed into an equivalent membranecovering framework for tubule extension? Both in vitro biophysical and computational investigations reveal that BAR domain-membrane interactions, self-assembly, and collective behavior are complex. Critically, polymerization is not simply a function of protein monomer concentration and binding affinity. In

Developmental Cell

Previews

Figure 1. Multifactorial Regulation of F-BAR/EFC Domain Membrane Binding and SelfOrganization (A) Combined ribbon/surface representations of BAR domain superfamily representatives. Each half of the dimeric functional unit is colored a different shade. Different superfamily members are characterized by varied intrinsic curvature (dashed arc) of the antiparallel dimer assembly; F-BAR/EFC domains have shallow curvature compared with BAR/N-BAR proteins. (B) Factors that impact translocation of EFC/ F-BAR proteins from the reserve cytosolic pool onto the plasma membrane. For clarity, the unstructured polypeptide linker and C-terminal SH3 domain of the intact functional unit is omitted from the depiction of membrane-attached EFC/F-BAR domains.

cells, in addition to membrane-bound density, membrane transformations depend importantly on the lengthened ellipsoidal domain shape and directiondependent anisotropy, as well as on membrane composition, stiffness (tension), and geometry (Simunovic and Voth, 2015; Simunovic et al., 2013, 2015) (Figure 1). And EFC/F-BAR domains do not associate with bilayers in only a single allowable orientation: sidewaysadsorbed F-BAR/EFC domains, with the concave surface oriented roughly parallel to the plane of the membrane, also

occur (Frost et al., 2008; Yu and Schulten, 2013), permitting several self-assembly possibilities. Coarse-grained computational simulations of N-BAR domains (a subset of BAR domains characterized by an additional but integral N-terminal amphipathic a helix with lipid-penetrating properties) depositing on planar membrane at low relative coverage indicate a predisposition to first form linear arrays that, under low tension, dimple the membrane in a corrugated-like pattern (Simunovic et al., 2013). The resemblance of the end-

to-end-arranged Cdc15p F-BAR/EFC domain strands (McDonald et al., 2015) to the computationally predicted linear aggregates (Simunovic et al., 2013; Simunovic and Voth, 2015) is uncanny. Current data thus suggest that on thermally fluctuating membrane expanses, BAR superfamily domain polymers preferentially nucleate as filaments. Satisfyingly, the FBP17 EFC domain crystallizes as tipto-tip-oriented filaments (Shimada et al., 2007). Shifts in the equilibrium between the end-to-end organization of linear F-BAR/EFC domains and the side-toside-associated spiral-wrapped tubules dependent upon membrane tension are known (Frost et al., 2008) and calculated (Simunovic and Voth, 2015) (Figure 1). On stiff membranes, lateral EFC/F-BAR domain-domain contacts are maximized because the diminished bilayer pliability prohibits the strongest electrostatic domain-membrane adherence to optimize shape complementarity. At intermediate membrane tension levels, intermittent side-to-side dimers permit branching in linear arrays (Simunovic and Voth, 2015). But why would F-BAR domain proteins function as linear filaments as opposed to progressing to tightly spiraled cylinders? Branched linear arrays can be arranged into lose meshes, which can define functionally discrete and compositionally distinct microdomains on the inner leaflet of the plasma membrane (Moravcevic et al., 2015; Zhao et al., 2013) or locally manage protein-protein and/or proteinlipid interactions. For Cdc15p, it is the coordination of protein-protein interactions in a spatially restricted manner that is crucial for contractile ring assembly and cell division (McDonald et al., 2015). This underscores another key, but often overlooked, attribute of EFC/F-BAR domain proteins: they typically utilize other linked folded domains (SH3 domains, for instance) to associate with a network of discrete proteins besides membrane phospholipids and themselves. The EFC/F-BAR dimer mandates that pairs of independent domains are displayed, imparting strong avidity effects to assembled polymers. The bound partners can impose physical self-assembly constraints on the fulllength protein, in part by molecular crowding phenomena. Indeed, computational modeling shows that relevant

Developmental Cell 35, December 21, 2015 ª2015 Elsevier Inc. 665

Developmental Cell

Previews morphological membrane effects occur at BAR domain surface densities of only 25% that of the complete spirals seen by EM (Simunovic et al., 2015). Posttranslational phosphorylation of F-BAR domain proteins can introduce additional layers of regulation on membrane attachment (Takeda et al., 2013). McDonald et al. (2015) further show that like yeast Cdc15p, six mammalian F-BAR/EFC proteins also form homotypic assemblages but do not generate rigid tubules. For a first approximation, then, it seems more prudent to generally view the EFC/F-BAR domain as a spatially restricted, conditionally membrane-anchored, polymeric scaffolding and arrangement device that is specialized in some cases, and at high local density upon the appropriate membrane, to deform the surface into cylin-

ders. In most instances, the precise functional significance of EFC/F-BARwrapped plasma-membrane tubules still remains an open question. So, if you have a firm conviction that F-BAR/EFC proteins necessarily tubulate membranes, see whether this elegant and systematic investigation persuades you to re-evaluate. REFERENCES Frost, A., Perera, R., Roux, A., Spasov, K., Destaing, O., Egelman, E.H., De Camilli, P., and Unger, V.M. (2008). Cell 132, 807–817. McDonald, N.A., Vander Kooi, C.W., Ohi, M.D., and Gould, K.L. (2015). Dev. Cell 35, this issue, 725–736. Moravcevic, K., Alvarado, D., Schmitz, K.R., Kenniston, J.A., Mendrola, J.M., Ferguson, K.M., and Lemmon, M.A. (2015). Structure 23, 352–363.

Shimada, A., Niwa, H., Tsujita, K., Suetsugu, S., Nitta, K., Hanawa-Suetsugu, K., Akasaka, R., Nishino, Y., Toyama, M., Chen, L., et al. (2007). Cell 129, 761–772. Simunovic, M., and Voth, G.A. (2015). Nat. Commun. 6, 7219. Simunovic, M., Srivastava, A., and Voth, G.A. (2013). Proc. Natl. Acad. Sci. USA 110, 20396– 20401. Simunovic, M., Voth, G.A., Callan-Jones, A., and Bassereau, P. (2015). Trends Cell Biol. 25, 780–792. Takeda, T., Robinson, I.M., Savoian, M.M., Griffiths, J.R., Whetton, A.D., McMahon, H.T., and Glover, D.M. (2013). Open Biol. 3, 130081. Yu, H., and Schulten, K. (2013). PLoS Comput. Biol. 9, e1002892. Zhao, H., Michelot, A., Koskela, E.V., Tkach, V., Stamou, D., Drubin, D.G., and Lappalainen, P. (2013). Cell Rep. 4, 1213–1223.

Warts Opens Up for Activation Samuel A. Manning1,2,3 and Kieran F. Harvey1,2,3,* 1Peter

MacCallum Cancer Centre, 7 St Andrews Place, East Melbourne, VIC 3002, Australia Peter MacCallum Department of Oncology, University of Melbourne, Parkville, VIC 3010, Australia 3Department of Pathology, University of Melbourne, Parkville, VIC 3010, Australia *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2015.12.004 2Sir

Warts is the central effector kinase of the Hippo growth-control pathway. In this issue of Developmental Cell, by assessing Warts conformation in vivo, Vrabioiu and Struhl (2015) report that the Mob family protein Mats regulates Warts activity allosterically, independent of phosphorylation by Hippo. Intense research over the past decade has elucidated a complex signaling network known as the Hippo pathway. Founding pathway members were identified in Drosophila melanogaster genetic screens as regulators of tissue growth. Subsequently, genetic, proteomic, and cell-based screens in flies and mammals have identified more than 40 pathway proteins (Harvey et al., 2013; Pan, 2010). The Hippo pathway is evolutionarily ancient, with key elements predating metazoan evolution (Sebe´-Pedro´s et al., 2012). Furthermore, pathway deregulation has been linked to many human diseases such as cancer (Harvey et al., 2013).

The Hippo pathway can be sub-classified into three main groups: upstream regulators, the core kinase cassette, and downstream transcriptional regulators. The best-characterized transcriptional regulators of the Hippo pathway are Yorkie and Scalloped. The core kinase cassette limits tissue growth by stimulating Warts-dependent phosphorylation of Yorkie. Upstream proteins influence activity of the core kinase cassette or, in some cases, act directly on Yorkie (Harvey et al., 2013; Pan, 2010). It is likely that the majority of important Hippo pathway proteins have now been discovered. As such, research effort is refocusing to other questions such as key steps

666 Developmental Cell 35, December 21, 2015 ª2015 Elsevier Inc.

of pathway regulation, points of crosstalk with other growth-control networks, and the role of the pathway in specific cell types and in disease. In the present study, Vrabioiu and Struhl (2015) interrogate the mechanism of activation of the Hippo pathway core kinase cassette (Vrabioiu and Struhl, 2015), which consists of the Ser/Thr kinases Hippo and Warts and the non-catalytic proteins Mats and Salvador. Warts (an NDR family kinase), Hippo (a Sterile 20 family kinase), and Mats (a Mob family protein) form a signaling module that is conserved from yeast to humans (Hergovich and Hemmings, 2009). Hippo activates Warts by phosphorylating the