Septins and the Lateral Compartmentalization of Eukaryotic Membranes

Developmental Cell

Review Septins and the Lateral Compartmentalization of Eukaryotic Membranes Fabrice Caudron1 and Yves Barral1,* 1Institute of Biochemistry, Department of Biology, ETH Zurich, Schafmattstrasse 18, 8093 Zurich, Switzerland *Correspondence: [email protected] DOI 10.1016/j.devcel.2009.04.003

Eukaryotic cells from neurons and epithelial cells to unicellular fungi frequently rely on cellular appendages such as axons, dendritic spines, cilia, and buds for their biology. The emergence and differentiation of these appendages depend on the formation of lateral diffusion barriers at their bases to insulate their membranes from the rest of the cell. Here, we review recent progress regarding the molecular mechanisms and functions of such barriers. This overview underlines the importance and conservation of septin-dependent diffusion barriers, which coordinately compartmentalize both plasmatic and internal membranes. We discuss their role in memory establishment and the control of cellular aging. Introduction Differentiated cells such as photoreceptor cells, neurons, and epithelial cells are able to adopt specialized and sophisticated architectures highly adapted to their functions. Specialized architectures generally involve compartmentalizing the cell into segments with distinct organizations, compositions, and activities. For example, the integration and conduction of action potentials and the storage of information in neurons depends on the compartmentalization of these cells into spines, dendrites, somas, axons, and presynaptic buttons. Although the beauty of these cells has been long appreciated, the principles governing the establishment of their elaborate structures remain poorly understood. The organization of eukaryotic cells rarely relies on a rigid carpentry of stable macromolecular assemblies, for example, as in viruses, but instead relies on dynamic, continuously remodeled molecular interactions. This allows for the constant regeneration of the cell’s composition, which provides plasticity to cellular architectures. However, this mode of organization makes the building and maintenance of complex structures a continuous and challenging task for the cell. One particular aspect of this is the requirement to continuously channel the diffusion and exchange of molecules in order to maintain the anisotropic organization of the cell and its organelles. Over the years, evidence has accumulated indicating that lateral diffusion barriers in cellular membranes play an important role in shaping and maintaining specialized cellular architectures. Diffusion barriers help to laterally compartmentalize cellular membranes into separate domains. Such barriers are found in continuous membranes such as the plasma and endoplasmic reticulum (ER) membranes and restrict the lateral diffusion of membraneassociated proteins. As a result, these proteins freely diffuse in the membrane domain where they act but are prevented from diffusing into other membrane areas. Here, we review the current knowledge on diffusion barriers. We begin with an overview of barriers in different cell types and discuss their molecular functions. We then focus on various diffusion barriers found in yeast. Finally, we discuss the role of septins, a family of cytoskeletal GTPases conserved from fungi to animals that are involved in the formation of diffusion barriers.

Diffusion Barriers The Tight Junction of Epithelial Cells The concept of a diffusion barrier was first formulated to explain the organization and function of epithelia. An epithelium consists of cells tightly bound one to another and acts as a barrier between body compartments. This barrier allows the organization of information and molecular exchange with respect to the external environment. As such, their primary function is to prevent free diffusion of molecules through the epithelial sheet. In addition, they ensure the controlled and unidirectional transport of specific solutes and signals into and out of the body. In this manner, epithelia control the uptake of nutrients and the excretion of digestive enzymes, milk, waste materials, and other secreted molecules. In accordance with their roles in polarized transport, epithelial cells are themselves polarized in the apico-basal dimension. Their plasma membranes are divided into apical and basolateral domains that have different compositions, which endow them with distinct properties. Their external context is also distinct, as the apical domain faces the lumen or external space whereas the basolateral domain is in contact with the matrix of the organ. In combination with polarized vesicular trafficking, the asymmetry of these membranes provides directionality to the transcytosis of molecules through the epithelium (Folsch, 2008). One striking feature of apical and basolateral domains is their sharp delimitation by electron dense structures, named the tight junction (TJ) in vertebrates, which encircles the cell. The first function of the TJ is to tightly join the apical membranes of adjacent epithelial cells and block the passage between them. Interestingly, while this paracellular barrier prevents the free diffusion of solutes between epithelial cells, it also restricts the lateral diffusion of some lipids and possibly plasma membrane-associated proteins between the apical and basolateral domains (Figure 1). This function, referred to as the fence function, is thought to ensure that the two membrane domains do not mix. The presence of a lateral diffusion barrier for lipids in the plasma membrane of epithelial cells has been demonstrated. Studies where fluorescent lipids were added to the apical plasma membrane show TJs prevent lipids of the outer leaflet from diffusing into the basolateral membrane, while lipids in Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 493

Developmental Cell

Review Figure 1. Diffusion Barriers in the Tight Junction and Primary Cilium (A) The tight junctions (TJ) (red) restrict the paracellular transit of solutes (arrows) from the lumen (light blue) to the basolateral space (light brown). Each epithelial cell has different membrane compartments (orange for primary cilium, blue for the apical membrane, and brown for the basolateral membrane). TJs separate apical and basolateral membranes. A putative diffusion barrier at its base insulates the primary cilium from the rest of the plasma membrane (dashed red line). (B) TJ fence function. The diffusion barrier (red) concerns primarily the outer leaflet of the plasma membrane (left side, blue and brown) where it affects the diffusion of lipids and probably proteins. Inner membrane lipids (dark green) or lipids that can flip-flop (light green) are free to diffuse between the apical and the basolateral compartments (right). Proteins associated with the inner leaflet of the membrane may diffuse freely. (C) Diffusion barrier at the base of the cilium. The cilium membrane (orange) is separated from the apical membrane (blue) by a ring of highly ordered lipids (dashed red line) at the base of the cilium. This ring is thought to form a diffusion barrier for lipids and proteins. Basal body and cilium microtubules are depicted in gray.

the inner leaflet or flip-flopping can exchange between the two domains (Dragsten et al., 1981; Spiegel et al., 1985; van Meer and Simons, 1986). Thus, the TJ forms a lateral diffusion barrier for lipids specific to the outer leaflet of the plasma membrane. Although evidence for this barrier role exists only for lipids (Matter and Balda, 2003), it is tempting to speculate that the TJ also contributes to the confinement of membrane proteins inserted in the outer or both leaflets. While the protein occludin is part of the TJ complex and was originally proposed to be involved in the fence function (Balda et al., 1996), subsequent studies revealed knockdown of occludin did not perturb lipid compartmentalization (Yu et al., 2005). In contrast, the small GTPases RhoA and Rac1 (Jou et al., 1998) are required for fence function. Indeed, interfering with RhoA and Rac1 functions or overexpressing a C-terminally truncated form of occludin allows the diffusion of lipids to the basolateral compartment. Thus, the current data indicate that diffusion barriers separate apical and basolateral membranes. However, it is unclear through what mechanism RhoA and Rac1 participate in barrier formation. In contrast to lipid compartmentalization, the mechanisms responsible for isolating the protein constituents of the apical and basolateral membrane compartments, and membrane polarity, are poorly understood. Membrane markers such as p75NTR (apical membrane) and the (Na+, K+)-dependent ATPase (basolateral membrane) remain properly localized in conditions where the lipid barrier is compromised (Balda et al., 1996; Jou et al., 1998). Since at least 50% of the (Na+, K+)dependent ATPase freely diffuses at the surface of MDCK cells (Jesaitis and Yguerabide, 1986), the maintenance of its localization upon RhoA and Rac1 inactivation suggest the existence of a Rho- and Rac- independent barrier for these proteins. Therefore it is not yet known how important the fence function is for the maintenance of apico-basal polarity. 494 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

A Diffusion Barrier at the Base of Cilia? Nearly all vertebrate cells form a primary cilium, which protrudes from their apical surface. This cilium acts as a sensory and signaling organelle (for a review see Singla and Reiter, 2006). In some cells, cilia are modified as more complex appendages; for example, the photosensitive outer segment of rod and cone cells in the retina. Although the ciliary membrane is continuous with the plasma membrane, it is highly enriched with specific proteins that support its sensory and signaling functions. It is not known whether the ciliary membrane defines a separate compartment of the plasma membrane, but data hint toward the presence of a diffusion barrier at the base of the cilium (Figure 1). Vieira et al. (2006) observed that the GPI-anchor protein FP-GPI, a resident protein of the apical membrane, is excluded from the ciliary membrane. This exclusion correlates with the presence of a ring of membrane with highly ordered lipid content at the base of the cilium (Vieira et al., 2006). Thus, ciliation and its derivations might correspond to a generic case of plasma membrane compartmentalization. The Flagellar Bracelet of Flagellated Algae Studies on the flagellum of Chlamydomonas reinhardtii, a structure highly related to the cilium of metazoans, support the idea that ciliation is associated with plasma membrane compartmentalization. Freeze-fracture images revealed the presence of an electron-dense, ring-like structure called the ‘‘flagellar bracelet’’ at the base of each flagellum (Weiss et al., 1977). Furthermore, the glycoprotein composition of the flagellar membrane is different from the remainder of the plasma membrane, and membrane proteins cannot freely exchange between these two domains. These findings are consistent with the flagellar bracelet acting as a diffusion barrier. It is possible that the emergence and differentiation of flagella and cilia out of the plasma membrane relies on the presence of a lateral diffusion barrier at their base.

Developmental Cell

Review Figure 2. Diffusion Barrier at the Annulus of Spermatozoa Spermatozoa with membrane compartments, head (black), anterior tail (blue), and posterior tail (green). The electron dense annulus (red) marks the boundary between the anterior tail that comprises the mitochondria (brown) and the posterior tail with the fibrous sheath (gray). Proteins such as CE9 (dark green) only localize to the posterior tail membrane.

Diffusion Barriers in Spermatozoa Sperm cells are divided into four compartments: the head, divided into anterior and posterior heads; the anterior tail, separated from the head by the posterior ring; and the posterior tail, separated from the anterior tail by the annulus, an electron-dense structure (Figure 2). Several membrane-associated proteins localize only to one compartment. For example, the PT-1 antibody (Myles et al., 1981) labels a protein found only in the posterior tail. Furthermore, although the PT-1 antigen freely diffuses within the posterior tail, it does not cross the annulus and does not penetrate the anterior tail (Myles et al., 1984). Therefore, the annulus is thought to function as a diffusion barrier for the PT-1 antigen. Beyond providing an architectural framework for the development of complex cellular organizations, lateral diffusion barriers also help keep proteins away from their sites of action until the appropriate times. The integral membrane protein CE9 (Petruszak et al., 1991) localizes to the posterior tail in immature spermatozoa and, upon maturation, to both the posterior and anterior tail (Nehme et al., 1993; Petruszak et al., 1991). Electron microscopy studies indicate that in many species, the early boundary of CE9 localization correlates with the position of the annulus (Cesario and Bartles, 1994). Remarkably, the entry of CE9 into the anterior tail membrane correlates with the endoproteolytic cleavage of its extracellular N-terminal tail (Petruszak et al., 1991). Cross-linking of CE9 with a polyclonal antibody prevents redistribution, indicating that its translocation depends on the lateral diffusion within the membrane (Nehme et al., 1993). How the annulus functions as a barrier is not known, and it is unclear why cleavage of the extracellular tail of CE9 allows it to overcome a barrier dependent on the intracellular annulus. Therefore, it will be important to experimentally interfere with annulus stability in order to determine the role of the annulus in barrier formation and understand how it interferes with protein diffusion in the plasma membrane. The Initial Segment of the Axon The processes through which long-term maintenance of cellular organization occurs are remarkable, and might be best illus-

trated in neuronal organization and the encoding of neuronal memory that allows us to remember events for decades. Neurons process and propagate information using action potentials, which they generate in their dendrites and pass on through their axons. The functional differentiation of the axons and dendrites from each other depends on the presence of different sets of membrane proteins at their respective surfaces. For example, the Thy-1 glycoprotein localizes specifically to the axonal membrane (Dotti et al., 1991). A lateral diffusion barrier in the plasma membrane at the initial segment of the axon contributes to maintaining the asymmetric distribution of proteins to this compartment (Winckler et al., 1999; Figure 3). When delivered to the axonal membrane by liposome fusion, fluorescent phospholipids freely diffuse within the axonal membrane, but do not leak into the somatodendritic membrane (Kobayashi et al., 1992; Nakada et al., 2003). Furthermore, membrane associated proteins such as Thy-1 or the transmembrane protein L1 show reduced lateral mobility in the initial segment (Winckler et al., 1999). Actin, ankyrin-G, voltage-dependent sodium channels, neurofascin, and NrCAM concentrate in this location and directly contribute to barrier formation. Supporting this view, actin depolymerization causes increased mobility of the voltage-dependent sodium channel, which then spreads into the cell body (Nakada et al., 2003). Barrier loss leads further to the loss of Thy-1 compartmentalization. Therefore, a cortical protein complex establishes a diffusion barrier at the initial segment of the axon, separating the somatodendritic and axonal compartments. This elegant barrier mechanism allows for the formation of the complex architecture of neuronal cells and the differentiation of the axon from the rest of the neuron (Hedstrom et al., 2008). The Neck of Dendritic Spines Relative to axons, the organization of dendritic spines plays an even more important role in memory. The spine forms the postsynaptic compartment and therefore determines the strength with which the receiving neuron is stimulated by a given connection. Spines are dynamic, as they retract from the dendritic shaft when they are inactive and reform in response to stimulation. For example, there are fewer spines in the primary visual cortices of mice raised in the dark, and these spines can recover after a few days in normal light (Valverde, 1967, 1971). The stabilization of specific spines leads to the reinforcement of the corresponding synapses during the establishment of short- and long-term memories. Protocols inducing long-term potentiation (LTP) Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 495

Developmental Cell

Review Figure 3. Diffusion Barriers in Neurons

A

(A) Neurons are compartmentalized into an axon (blue) and a somatodendritic compartment (green). Dendritic spines are depicted as orange circles. The axon initial segment’s diffusion barrier (red dashed line) supports the differentiation of the axon. (B) Dendritic spines have a diffusion barrier at their neck (red dashed line) restricting the diffusion of AMPA receptor (violet) across the spine neck. (C) The initial segment of the axon contains a diffusion barrier composed of an actin network (red filaments) and transmembrane proteins (red) such as b-IV spectrin and AnkyrinG (‘‘pickets’’). These proteins prevent the diffusion of lipids and membrane proteins and thereby separate the axon (blue) and somatodendritic (green) domains from each other.

C

B

60 to 90µm

1µ m

Spine

Somato-dendritic specific membrane protein

‘Pickets’

AMPAr

Axon specific membrane protein

F-actin

modulate spine morphology, particularly by enlarging the spine head (Alvarez and Sabatini, 2007; Nimchinsky et al., 2002). Thus, each spine on a dendrite behaves as an autonomous entity, reflecting its individual history. The properties of spines are the result of isolation of both their plasma membrane and cytosol content from that of the dendritic shaft. The strength of the cytosolic compartmentalization varies according to spine morphology and inversely with respect to the width of its neck (Sabatini et al., 2002; Svoboda et al., 1996). Also, the lateral mobility of the AMPA (Alpha-amino-3-hydroxy5-methyl-4-isoxazole propionic acid) receptor and membranetargeted GFP is strongly restricted within the neck membrane of certain categories of spines, indicating the presence of diffusion barriers at these locations (Ashby et al., 2006). Thus, a major function of spine necks appears to be confinement of postsynaptic components to their resident spines, thereby ensuring spine individualization. However, induction of LTP in one spine also reduces the threshold of activation in neighboring spines from the same dendrite (Harvey and Svoboda, 2007). This is correlated with the diffusion of activated Ras from the induced spine to the dendritic shaft (Harvey et al., 2008). Therefore, the compartmentalization of the spines is not absolute but instead selective, perhaps allowing individualization of each spine as one compartment while still allowing interaction between neighboring spines. Summary This list of known or potential areas where lateral diffusion is restricted in cellular membranes, here called diffusion barriers, is not comprehensive. There is also evidence for the existence of diffusion barriers at the paranodes flanking the nodes of Ranvier of myelinated axons (Sherman and Brophy, 2005), at the leading edge of migrating keratocytes (Weisswange et al., 2005), in the 496 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

cleavage furrow of dividing cells (Schmidt and Nichols, 2004), and at the neck of budding yeast cells (see below; Barral et al., 2000; Takizawa et al., 2000; Dobbelaere and Barral, 2004; Luedeke et al., 2005; Shcheprova et al., 2008). The recurring theme emerging from these studies is that many different diffusion barriers can be idenDiffusion barriers tified in a wide range of cell types, where they assume important roles in the control of cellular organization and particularly in the maintenance of long-term cellular architecture. Although we still know little about how diffusion barriers are established or how they work, the answers to several central questions appear to be within our reach. Do all diffusion barriers function according to the same mechanism or are there different types of diffusion barriers? Are different barriers formed using different types of molecules, or are most barriers similarly structured? Did the barrier principle arise several times during evolution? Before exploring these questions, we will focus on the budding yeast Saccharomyces cerevisiae, where several diffusion barriers have been described at the functional and structural levels, and for which mechanistic insights have begun to be revealed. Septin-Dependent Diffusion Barriers in Budding Yeast Yeast cells divide in a highly asymmetric manner by producing de novo a daughter cell in the form of a bud that emerges from the surface of its mother. During bud formation, the cytoskeleton polarizes and targets exocytic vesicles to the bud, which grows while the size of the mother cell remains static (Pruyne et al., 2004). Over the last decade, the bud neck has been shown to control the exchange of material between mother and bud through the formation of diffusion barriers. Such barriers have been described in the plasma membrane, the endoplasmic reticulum, and the nuclear envelope (Barral et al., 2000; Luedeke et al., 2005; Shcheprova et al., 2008; Takizawa et al., 2000; Figure 4). In the plasma membrane, fluorescence recovery after photobleaching (FRAP) experiments established that over-produced Ist2p protein freely diffuses in the bud part of the membrane. The highly compartmentalized localization of Ist2p depends on the septins, as Ist2p localizes to both the mother and bud plasma membranes in septin-defective cells (Takizawa et al., 2000).

Developmental Cell

Review

Figure 4. Three Diffusion Barriers at the Yeast Bud Neck Sectional view of a yeast cell depicting the plasma (black line), the ER (green), and the nuclear envelope (light red) membranes. These membranes are continuous through the bud neck (middle) yet are compartmentalized into mother (bottom) and bud (top) domains. The septin ring (red) governs the formation of the diffusion barriers responsible for this compartmentalization (dashed lines). Ribosomes (black dots) associate with the ER membranes throughout the cell except at the bud neck, where the ER is smooth. Old nuclear pore complexes (brown) and associated extra-chromosomal rDNA circles (blue) are retained in the mother, and new nuclear pore complexes (orange) are inserted in the nuclear envelope of the bud. Spindle pole bodies (gray) together with spindle and cytoplasmic microtubules (gray lines) are also depicted.

These data suggest that the septin ring establishes a barrier in the plasma membrane to prevent the diffusion of the Ist2p protein through the bud neck. The endoplasmic reticulum (ER) of yeast cells is composed of a reticulated network of tubules underlying the cell cortex (cortical ER) of the nuclear envelope and of a small number of cytoplasmic tubules that connect cortical ER and nuclear envelope. The ER network is a continuous, single entity throughout mother and bud. Accordingly, GFP targeted to the ER lumen freely diffuses throughout the entire ER, as determined by fluorescence loss in photobleaching (FLIP) experiments (Luedeke et al., 2005). However, all ER-membrane proteins tested so far, such as the translocon subunit Sec61p, the SNARE Sec22p, and the HMG-CoA reductase Hmg1p, do not freely diffuse through the bud neck. This compartmentalization is lost in septin-defective cells and other bud neck mutants. Thus, an

ER-specific, septin-dependent barrier limits the lateral diffusion of ER-membrane proteins through the bud neck. Budding yeast undergoes a closed mitosis, and the nuclear envelope remains intact through the end of anaphase. The inner and outer nuclear membranes display different properties during anaphase. Nucleoplasmic GFP (NLS-GFP) and the inner nuclear membrane protein Pmr3p exchange freely between the mother and the bud-localized regions of the nucleus during anaphase as determined by FLIP. In contrast, the outer membrane proteins Nsg1p and Sec61p as well as the nucleoporin Nup49p do not freely cross the bud neck. In septin or bud neck mutants, this compartmentalization is lost, indicating that an additional barrier limits the diffusion of proteins in the outer membrane of the nuclear envelope (Shcheprova et al., 2008). Together, these studies have revealed that diffusion barriers compartmentalize not only the plasma membrane, but also internal membranes. These barriers have been shown to be involved in both the maintenance of cell polarity as well as the control of cellular aging. In each case, barrier establishment depends on septin function, demonstrating the central role of this cytoskeletal component. The known functions of these septin-dependent barriers and the potential mechanisms of their formation and action are discussed below. Maintenance of Cell Polarity Yeast cells polarize cell growth toward the bud by targeting, in an actin-dependent manner, the components of the polarizome and the exocyst to the bud cortex (Tarassov et al., 2008). There, exocyst and polarizome complexes diffuse freely within the bud cortex, without leaking into the mother cell (Barral et al., 2000). In cells lacking the septin collar at the bud neck, both the exocyst component Spa2 and the polarizome components Sec3 and Sec5 are no longer confined to the bud. This is particularly evident in cells that have undergone the switch from apical growth, where the bud is highly polarized, to isotropic bud growth (Barral et al., 2000). In this case, growth is no longer restricted to the bud. Therefore, whereas the polarization of actin cables toward the bud tip originally ensures the establishment of cell polarity, maintaining this polarity during normal bud morphogenesis relies on the presence of a septin-dependent diffusion barrier at the bud neck. Diffusion barriers also play a critical role in cell cleavage (Dobbelaere and Barral, 2004). At the onset of cytokinesis, the yeast septin collar splits into two rings, one on each side of the bud neck. Coincident with splitting, two diffusion barriers form and open a new cortical compartment between them; this is the future site of cleavage. As during bud growth, this compartmentalization helps to confine the localization of polarizome and exocyst complexes to the neck area. Photobleaching experiments indicate that both types of complexes freely diffuse throughout the cortex of the bud neck, yet do not escape into mother and bud (Dobbelaere and Barral, 2004). However, if the septin rings are prematurely disrupted, the barriers are lost and the components normally trapped at the cleavage site escape and cytokinesis aborts. Therefore, during yeast cytokinesis, one function of the septin cytoskeleton is to compartmentalize the cortex at the site of cleavage, thereby ensuring that the different machineries involved in cleavage are coordinately localized to the division area. Bud-neck-dependent compartmentalization of the cell cortex has also been observed for molecules that are not thought to be Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 497

Developmental Cell

Review involved in bud growth or membrane remodeling. For example, the above-mentioned Ist2p protein is a multi-spanning protein involved in osmotolerance (Juschke et al., 2004). It is not known why Ist2p localization is restricted to large buds, although one possibility is that it plays a role in the extensive cell wall and cortical remodeling that occurs in buds undergoing isotropic growth. Compartmentalization of the cell cortex also contributes to the spatial coordination of events relative to cell-cycle progression. For example, yeast cells do not exit mitosis or start cytokinesis until the bud has inherited one of the daughter nuclei. Indeed, the mitotic exit program requires that one spindle pole body (SPB) contacts Lte1p, a protein confined to the bud cortex (Pereira and Schiebel, 2001). The restriction of Lte1p to the bud depends on the septin collar. Accordingly, disruption of septin structures leads to Lte1p escaping into the mother cell and premature mitotic exit in cells with nuclear segregation defects (Castillon et al., 2003; Seshan et al., 2002). In this case, compartmentalization serves to both identify the bud and monitor the location of SPBs relative to it. Together, these data indicate that in budding yeast, septins govern the compartmentalization of the cell cortex. This organization is required for several critical processes: first, to maintain cell polarity and promote the restriction of growth to the bud; second, to help the bud differentiate from the mother cell by accumulating different plasma membrane and cortical proteins; and third, to monitor the proper partition of SPBs between mother and bud at the end of mitosis. Confinement of Age in the Mother Cell Ultimately, cell polarization in budding yeast leads to the separation of the daughter cell from its mother. Whereas the mother has divided at least once, the daughter has not yet divided. Yeast mother cells undergo replicative aging, i.e., they can divide and form only a limited number of buds. However, while the mother ages, the age of the buds is reset; they are born with a full lifespan potential (Steinkraus et al., 2008). One process leading to the aging of the mother cell is the accumulation of extrachromosomal ribosomal DNA circles (ERCs) as they progress through successive rounds of DNA replication (Sinclair and Guarente, 1997). ERCs are circular episomes formed by recombination in the rDNA locus and containing an autonomous replicating sequence. Unlike chromosomes, ERCs and other DNA circles lacking a centromere segregate asymmetrically during anaphase (Murray and Szostak, 1983), accumulate in the mother cell, and somehow contribute to the progressive decay of its fitness (Falcon and Aris, 2003; Murray and Szostak, 1983; Sinclair and Guarente, 1997). What retains DNA circles in the mother compartment has remained elusive until recently. Recent data indicate that this retention relies on a diffusion barrier in the envelope of anaphase nuclei (Shcheprova et al., 2008). This barrier, which forms in a septin-dependent manner at the position where the nucleus crosses the bud neck, prevents proteins embedded in the outer nuclear envelope from freely entering the bud. As a result, preexisting nuclear pores are in majority retained in the mother, whereas new pores appear to preferentially insert in the bud part of the envelope (Figure 4). DNA circles are generally found to localize to the nuclear periphery (Scott-Drew et al., 2002), where they associate with pores (Shcheprova et al., 2008). 498 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

Thus, the barrier also causes the retention of the ERCs in the mother nucleus. Correspondingly, mutation of pore basket components causes the detachment of DNA circles from the nuclear periphery and prevents their retention in the mother cell. Likewise, mutations that weaken the diffusion barrier lead to random segregation of DNA circles between mother and bud. This ultimately results in mothers passing some of their age onto their daughters; the daughters of these cells are then born with the approximate replicative age of their mother (Shcheprova et al., 2008). Thus, this diffusion barrier plays an important role in defining the age of the mother cell and resetting age in its daughters. Whether diffusion barriers help to keep other aging factors in the mother cell during its life span is not yet known, but preliminary evidence suggests that this is the case (Shcheprova et al., 2008). Remarkably, age factors such as carbonylated and aggregated proteins are segregated asymmetrically and are thereby retained (and accumulate) in mothers as they age (Aguilaniu et al., 2003; Erjavec et al., 2007). However, these factors are not located on the nuclear envelope, but rather in the cytoplasm and mitochondria. It will therefore be interesting to determine whether the retention of these other factors in the mother cell also depends on the presence of diffusion barriers at the bud neck. For example, the protein chaperone Hsp104p is involved in the refolding of aggregated proteins, colocalizes with them, and contributes to their retention in the mother cell (Erjavec et al., 2007). Hsp104p primarily functions in the cytoplasm but also contributes to the refolding of ER proteins. One possibility is that Hsp104p aggregates are attached to the ER membrane, and their retention in mother cells might involve the diffusion barrier in the cortical ER. Gametogenesis is a process in which age must particularly be discarded. The process of yeast meiotic sporulation is very similar to that of spermatogenesis in animals. Remarkably, during sporulation, preexisting nuclear pores are also excluded from the nuclear envelope of the spore nuclei (Fuchs and Loidl, 2004). This process depends on the stripping of the preexisting spores from the envelope as the prospore membrane, which encircles the nucleus and forms the spore wall, progressively encapsulates each meiotic nucleus (Figure 5). Whether this contributes to rejuvenation of the spores is unknown, but is an attractive prospect that merits investigation. The observation that septins localize to the leading edge of prospore membrane during this stripping process (Fares et al., 1996) could suggest the presence of a diffusion barrier similar to the one retaining ‘‘old’’ pores in the mother compartment during vegetative growth. As we discuss below, this similarity opens interesting possibilities for our understanding of events during spermatogenesis in higher eukaryotes. Roles of the Different Barriers Taken together, three diffusion barriers have been currently described at the bud neck of budding yeast cells: experiments based on FLIP and FRAP have provided strong evidence for the presence of barriers in the plasma membrane, the cortical ER, and the nuclear envelope, each aligned with the plane of the bud neck. Furthermore, additional barriers may also exist in other organelles such as mitochondria. The presence of a diffusion barrier has been most clearly demonstrated when analyzing the diffusion of proteins normally

Developmental Cell

Review

Figure 5. Nuclear Pore Exclusion during Yeast Meiosis and Spermatogenesis (A) During yeast meiosis, the prospore membrane (red) emanating from each spindle pole body (black) encircles each spore nuclei. During this process, the nuclear pores (violet) are stripped from the envelope (green) of the spore nucleus. (B) During spermatogenesis, the perinuclear ring (red) progresses toward the back of the sperm head in a microtubule-dependent manner (black lines). Most nuclear pore complexes (violet) are stripped (green) and eliminated from the sperm nucleus.

present on both sides of the barrier, such as ER-membrane proteins and resident proteins of the nuclear envelope. In these cases, the membrane identification is straightforward and the presence of the barrier is demonstrated by the lack of exchange between the two sides of the boundary. In the case of polarized proteins, conclusions are more difficult to draw for several reasons. First, the asymmetry of the protein might depend on mechanisms other than retention, such as transport and recycling (Valdez-Taubas and Pelham, 2003). Second, although the best demonstration of the barrier is the loss of partition observed upon septin ring disassembly, this approach does not identify which barrier is involved in the retention of a specific factor, since the septins are required for the formation of all bud neck barriers. The relative contribution of the ER and plasma membrane barriers to the maintenance of cell polarity is therefore difficult to determine. In budding yeast, the cortical ER is so close to the plasma membrane that these two membrane systems cannot be resolved by light microscopy. Therefore, it is not always clear whether a protein is associated with the plasma membrane or with the cortical ER. In the case of Ist2p, the plasma membrane localization is clear (Juschke et al., 2004). However, several components of the exocyst interact both genetically and physically with the translocon (Lipschutz et al., 2003; Toikkanen et al., 2003), suggesting that the exocyst might localize to the ER rather than directly to the plasma membrane.

Whether the polarizome and Lte1p associate with the plasma membrane or the cortical ER is also unclear. Thus, the cortical ER barrier, rather than the plasma membrane barrier, might be required for the confinement of a number of polarity factors to the bud. The cortical ER barrier is likely to be required for the confinement of polarized mRNAs to the bud. A number of yeast mRNAs are transported to the bud by type V myosins. These mRNAs encode various proteins such as the daughter-specific transcription factor Ash1p and the membrane protein Ist2p (see above) and are enriched in mRNAs encoding transmembrane proteins (Shepard et al., 2003; Takizawa et al., 2000). Supporting the idea that these mRNAs are anchored to the bud ER, polarized mRNA localization and anchoring to the bud cortex depends on the protein She3p (Jansen et al., 1996). She3p is an ER protein (Estrada et al., 2003) that localizes specifically to the entire bud ER (Jansen et al., 1996). Accordingly, a number of these mRNAs, such as the ASH1 mRNA, were shown to tightly associate with the ER (Schmid et al., 2006). This association is tight enough to allow cofractionation of these mRNAs with ER membranes in biochemical assays (Gerst, 2008). Thus, both the plasma membrane and the cortical ER barriers are likely to contribute to the maintenance of cell polarity. The diffusion barriers at the bud neck could also involve the compartmentalization of smooth and rough ER. For example, a domain of smooth ER forms the ER barrier at the bud neck, whereas the ER membrane in the rest of the cell is primarily rough (Luedeke et al., 2005). This smooth ER domain might have some additional functions besides forming a barrier, for example, such as cytokinetic functions. Specifically, similar to the sarcoplasm of muscle cells, the bud neck ER could potentially be involved in the control of actomyosin ring contraction, perhaps in controlling Ca2+ release or by providing a scaffold for actomyosin ring organization. While we do not yet know much about ER cleavage prior to cell division, it is tempting to hypothesize that a specialized ER domain at the bud neck could be involved in this process. Similarly, the diffusion barrier in the plasma membrane could reflect a mechanism specifically selected to ensure the compartmentalization of that membrane. Alternatively, ER compartmentalization could be a secondary consequence of the presence of the septin collar, which has more functions than as a simple diffusion barrier, or the convergence of both. In brief, a remarkable feature of diffusion barriers is their functional duality; they can both separate membrane domains and serve as a scaffold to assemble specific structures at the boundary between these domains. Septin Biology In order to investigate how septins might influence barrier formation, more knowledge about septin function is required. Septins were originally discovered in budding yeast as cell-cycle-defective mutants impaired in cytokinesis (Hartwell, 1971). Septins are guanosine-nucleotide binding proteins conserved from budding yeast to mammals (Cao et al., 2007; Pan et al., 2007). The genome of Chlamydomonas encodes one septin-related gene; otherwise, septins are absent in plants. In fungi and animal cells, septins assemble into nonpolar hetero-oligomers, which then form filaments, gauzes, or rings (for recent reviews see Barral Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 499

Developmental Cell

Review and Kinoshita, 2008; Weirich et al., 2008). In budding yeast, septins assemble as a ring at the presumptive bud site. Upon bud emergence, they arrange at the bud neck in an hourglass shape, which splits into two rings during cytokinesis. Septins also form ring or ring-like structures at the annulus of spermatozoa and at the base of dendritic spines. In budding yeast, septins function not only in cytokinesis and the establishment of diffusion barriers, but also in spindle positioning, cell polarity, cell-cycle control, and spore-wall formation (Douglas et al., 2005). In higher eukaryotes, septins contribute to cytokinesis as well as interacting with and regulating the microtubule and actin cytoskeletons during interphase, vesicular trafficking during cell morphogenesis, and chromosome congression during mitosis. In nondividing cells, septins are involved in various cellular processes such as platelet activation, vesicular trafficking at neuronal synapses, phagocytosis, pathogen invasion, sperm, and dendritic spine maturation. Not surprisingly, septins are therefore linked to several human diseases such as hereditary neuralgic amyotrophy (Kuhlenbaumer et al., 2005), Parkinsons disease (Ihara et al., 2003; Ihara et al., 2007), and tumorigenesis (Hall and Russell, 2004). Below, we review some well-understood functions of septins to discuss their mode of action and how these functions could relate to the septin contribution to barrier formation. Septins in Signaling In budding yeast, septins recruit and activate septin-dependent kinases (SDKs) such as Hsl1p and Gin4p (Barral et al., 1999; Hanrahan and Snyder, 2003; Keaton and Lew, 2006; Szkotnicki et al., 2008), which promote signaling from the septins to downstream effectors. For example, SDKs are involved in the regulation of microtubule dynamics at the yeast bud neck during spindle positioning (Kusch et al., 2002) and mediate the coordination of anaphase entry with bud growth and septin collar assembly. In this latter process, Hsl1p and probably Gin4p control the activity of Swe1p, the budding yeast ortholog of Wee1 in fission yeast (Barral et al., 1999; Keaton and Lew, 2006; Shulewitz et al., 1999). Thus, SDKs are thought to act as sensors for proper assembly of the septin collar upon bud emergence and to signal the bud neck position to microtubules during spindle positioning. Septins in Cytokinesis In animals and fungi, septins are required for proper cytokinesis. Early in this process, septins localize to the cleavage plane, and they later relocalize to both sides of the cleavage site (Hartwell, 1971; Kinoshita et al., 1997; Maddox et al., 2007; Nguyen et al., 2000; Surka et al., 2002). Unlike in fungi, animal septins also localize to the leading edge of the contracting furrow (Fares et al., 1995; Neufeld and Rubin, 1994). Beyond their role as diffusion barrier late in cytokinesis, budding yeast septins also serve as a scaffold for the recruitment and assembly of the cleavage machinery at the cleavage plane prior to the onset of cytokinesis. Myosin ring assembly at the bud neck depends on septins (Bi et al., 1998), suggesting that septins might directly bind and recruit myosin II. Furthermore, septins independently recruit other cytokinesis factors to the bud neck, such as the PCH protein Hof1p, the abscission protein Cyk3p, and the IQGap homolog Iqg1p (Epp and Chant, 1997; Iwase et al., 2007; Korinek et al., 2000; Lippincott and Li, 1998; Vallen et al., 2000). 500 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

How septins contribute to the assembly of the cytokinetic machinery in other eukaryotes is not as clear; myosin ring assembly, for example, does not strictly require septin function in animal cells. However, septins do contribute to the proper regulation and contraction of the actomyosin ring. In Drosophila (Field et al., 2005), in the C. elegans embryo (Maddox et al., 2007), and in mammalian cells (Oegema et al., 2000), the role of septins in actomyosin ring contraction is largely mediated through interactions with the protein anillin (Field and Alberts, 1995). Data from Drosophila, mammals and C. elegans suggest that anillin serves as a linker between septins and myosin II and that septins and anillin help stabilize the actomyosin ring (Kinoshita et al., 2002). The function of septins during later steps in cytokinesis is still poorly defined in most eukaryotes. However, the idea that septins in animal cells may function similarly to fungal septins by forming diffusion barriers around the site of cleavage is supported by two observations. First, septins relocalize from the cleavage plane to either side of the cleavage site late in cytokinesis. Second, in late cytokinesis, a diffusion barrier is observed in the plasma membrane around the cleavage area (Schmidt and Nichols, 2004). Thus, septins appear to have dual roles during cytokinesis. In early cytokinesis, they function as a scaffold for the recruitment and/or stabilization of the contractile machinery. In late cytokinesis, they help confine membrane-associated factors to the cleavage site through the formation of diffusion barriers. Septins and Membrane Traffic Septins have also been implicated in the regulation of both membrane internalization and secretion. During growth of the yeast bud, septins control the secretion of chitin in the cell wall, which stabilizes the bud neck. Specifically, the septin collar is required to properly localize the chitin synthase Chs3p to the plasma membrane at the bud neck, and to recruit the Chs3p activator Chs4p and the regulatory protein Bni4p (DeMarini et al., 1997). How septins control Chs3p trafficking is not understood, and evidence for a direct role of septins in the control of vesicular trafficking in yeast is lacking. However, this function has been found for septins in other systems. For example, in mouse macrophages, SEPT2 and SEPT11 localize to the phagocytic cup during FcgR-mediated phagocytosis of IgG-coated beads (Huang et al., 2008), and depletion of Sept2 or Sept11 impairs bead internalization (Huang et al., 2008). In addition, the septin SEPT9_v1–5 is enriched on LnIB phagosomes during the internalization of Listeria monocytogenes (Pizarro-Cerda et al., 2002). In platelets and in the presynaptic compartment of neurons, the septin Sept5 negatively regulates exocytosis (Dent et al., 2002; Martinez et al., 2006). Sept5 coprecipitates with presynaptic vesicles (Beites et al., 1999) and interacts with the t-SNARE syntaxin-1; phosphorylation of Sept5 by Cdk5 stimulates this interaction (Amin et al., 2008; Beites et al., 1999; Taniguchi et al., 2007). In PC12 cells, both Sept5 phosphorylation and interaction with syntaxin leads to the inhibition of secretion (Amin et al., 2008; Beites et al., 1999). Together, these data suggest a direct role for septins in regulating vesicle traffic. How specific septin complexes are recruited to and function on membranes remains an important question. Mammalian Sept4 binds phosphatidyl-inositol-4,5-bis-phosphate (Zhang

Developmental Cell

Review et al., 1999) and budding yeast septins bind phosphatidylinositol-4-phosphate and phosphatidyl-inositol-5-phosphate (Casamayor and Snyder, 2003). Disrupting an N-terminal motif implicated in lipid binding in Sept4 or Cdc11p leads to filament assembly defects. However, the precise mechanism of septinlipid interaction still needs to be clarified and it is key to understanding septin function. How do septin-lipid interactions contribute to diffusion barrier formation? We can speculate that they might contribute directly, for example, through immobilization of underlying lipids in the membrane. These interactions could also create a steric obstacle in the diffusion path of proteins inserted in and extruding from the inner leaflet of the plasma membrane. In addition, septins could help recruit specific lipid or protein membrane components through their involvement in regulating membrane traffic. This could contribute to modulating the lipid or protein composition of the yeast bud neck. Determining the relative impacts of these different barrier formation mechanisms will therefore require clarifying the exact role of septins in vesicle delivery. Septins in Cell Polarity So far, a role for septins in cell polarity is only truly evident in budding yeast, where they not only set up diffusion barriers between mother and bud, but also contribute to early polarization events. A yeast cell selects the site of its next bud in a nonrandom process, depending on cell ploidy (Casamayor and Snyder, 2002; Park and Bi, 2007). In both haploid and diploid cells, bud site selection relies on a spatial cue provided by the septin collar to promote bud formation in the vicinity of a previous site of cytokinesis. Therefore, yeast septins act in a dual manner in cell polarity, first as a scaffold for the recruitment of cell polarity factors early in polarization, and second, through their role in the establishment of diffusion barriers during bud growth (see above). Mechanism of Septin Barrier Formation The most robust model to explain the different roles of septins is that these molecules function as scaffolding platforms for the recruitment and activation of specific molecules, particularly SDKs. One remarkable aspect of this scaffold is that, in addition to organizing proteins, it might also organize lipids through specific lipid-binding motifs. A second aspect of septin scaffolds is that they are large entities that can also serve as landmarks within the cell, keeping the memory of a specific cellular location over longer periods of times. Both features may play important roles for the mechanisms of septin function in barrier formation. Yeast diffusion barriers share three striking features: first, they all colocalize to the bud neck; second, all depend on septin function; and third, they affect both the plasma membrane and internal membranes, a feature so far described only in yeast. What insights does our knowledge of septin function provide on the mechanisms of barrier formation? The yeast septin collar is tightly apposed to the plasma membrane at the bud neck, and this might be sufficient to prevent the diffusion of plasma membrane proteins through the bud neck, as previously proposed. Supporting this model, septins are the only proteins known to be required for the diffusion barrier in the plasma membrane (Barral et al., 2000; Luedeke et al., 2005; Takizawa et al., 2000). However, septins alone cannot be responsible for the formation of the diffusion barrier in the cortical ER. Indeed, septins may directly interact with the

ER membrane facing the plasma membrane, but cannot interact with the membrane on the cytosolic side of the ER (Figure 4). Thus, septins must act indirectly in the formation of this barrier as well as in the formation of the barrier in the outer membrane of the nuclear envelope. One approach to understanding how barriers assemble and how septins function in this process is to determine the molecular composition of these barriers and how they restrict the diffusion of molecules. The ER barrier is perhaps the most amenable for these studies: it can be eliminated without compromising the survival of the cell, opening avenues for genetic studies. Two nonexclusive mechanisms could account for barrier function. First, the immobilization of transmembrane proteins could lead to the formation of a fence, as has been proposed for the barrier at the axon initial segment (Nakada et al., 2003). However, no transmembrane ER protein has yet been found to accumulate at the bud neck. Instead, this diffusion barrier correlates with the presence of smooth ER at the bud neck, from which many ER membrane proteins, such as the translocon subunit Sec61p, the HMG-CoA reductase Hmg1p, and the SNARE Sec22p, are excluded. Moreover, smooth ER domains have been suggested to correspond to areas of lipid biosynthesis. This and the exclusion of ER proteins suggest that the ER membrane may be different at the bud neck; the barrier could be formed by a domain of specialized lipids. In this model, the septin collar would act as a cortical landmark that locally promotes the synthesis of specialized lipids in the ER membrane and, hence, the formation of a lipid-based diffusion barrier in this organelle. Some septin-dependent diffusion barriers could be established through the formation of specialized lipid domains. How could septins control lipid biosynthesis in the ER membrane? The most likely explanation is that they act through their signaling function. Supporting this model, formation of the diffusion barrier in the ER membrane requires the presence of the SDKs Hsl1p and Gin4p (Luedeke et al., 2005). One target of these kinases could be the protein Bud6p, which is required for barrier formation. Bud6p is a peripheral membrane and actin-binding protein (Amberg et al., 1997) that also binds to the formins Bni1p and Bnr1p and positively regulates the actinnucleating activity of Bni1p (Moseley et al., 2004). These interactions are mediated by the C-terminal part of Bud6p. The N-terminal part of Bud6p is important for the correct localization of Bud6p to the cell cortex and is proposed to directly mediate Bud6p interaction with membranes (Jin and Amberg, 2000). However, how Bud6p contributes to barrier formation is not known. One possibility is that Bud6p acts through actin. However, early studies of Bud6p interaction partners suggest that this might not be the case (Luedeke et al., 2005). Instead, Bud6p interaction with membranes might be directly involved in barrier formation, perhaps via interaction with specific lipids, or by formation of a network of protein-protein interaction on the surface of the membrane. Septins may regulate Bud6p function directly or indirectly through the SDKs. Signaling events might also be involved in the assembly of the diffusion barrier in the nuclear envelope. In summary, yeast septins may directly form the diffusion barrier in the plasma membrane and coordinate the localization of other barriers, ensuring that they all form in the same plane. Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 501

Developmental Cell

Review

Figure 6. Summary of the Different Diffusion Barriers In each cell type, the nucleus-containing compartment (blue), the ciliary or flagellar-like compartment (orange), or other compartments (green) are depicted. Epithelial cells have a basolateral (blue), an apical (green), and a primary cilium (orange) domain. Chlamydomonas cells have a cell body (blue) and two flagellar (orange) domains. Spermatozoa have a head (blue), an anterior tail (green), and a posterior tail (orange) domain. Neurons have an axon (green), a somatodendritic (blue), a primary cilium (orange), and dendritic spine (dark blue) domains. Budding yeast have a mother (blue) and a bud (green) domain. Similarly, cytokinetic cells have two daughter (blue) domains and a cytokinetic domain (brown) at the cleavage plane. Photoreceptor cells have an outer segment (orange) and an inner segment (blue) including an axon (green). The different diffusion barriers (dashed lines) can be septin dependent (red), potentially septin dependent (violet), or septin independent (yellow).

Perspectives As diffusion barriers have been described in different organisms and cell types, we can compare their mechanisms and ask whether they all rely on a common, evolutionarily conserved mechanism or whether different mechanisms emerged independently from each other at different times in evolution (Figure 6). Do All Barriers Rely on the Same Molecular Mechanisms? The observation that the different barriers known thus far show different sizes and shapes suggests independent mechanisms. The axon initial segment is approximately 60–90 mm in length (Winckler and Mellman, 1999). In contrast, budding yeast diffusion barriers are restricted to the bud neck and are therefore 502 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

limited in thickness (z0.4 mm) and diameter (z1 mm). Unlike these barriers, the TJ seems to affect mainly the outer leaflet of the plasma membrane. This selectivity is not very well understood mechanistically, but establishes a clear difference with the yeast and neuronal barriers. Therefore, it is unlikely that these three types of barriers share the same mechanism. These distinctions are reflected by the differences observed at the molecular level. Septins have been observed neither in the initial segment of the neuron nor in the TJ complex. Instead, a network of cortical actin that anchors transmembrane proteins forms the diffusion barrier of the initial segment in the axon. These proteins then act as pickets, where their increasing concentration impairs the lateral diffusion of lipids and proteins. Likewise, no occludin-like proteins have been identified in the initial segment or in the yeast bud neck, where the formation of diffusion barriers does not correlate with cell-cell interactions. Thus, the different barriers described so far rely on at least three different types of molecular mechanisms, which must have emerged independently in evolution. Remarkably, however, evolutionary convergence seems to not be limited to barrier formation, but is also evident in the recruitment of regulatory factors to the compartmentalization boundary. For example, reminiscent of the role of septins in the recruitment and regulation of cell-cycle control proteins such as Hsl1p and Swe1p at the yeast bud neck, the TJ probably plays an important role in the control of cell proliferation for example through recruitment of the transcription factor ZONAB (Balda et al., 2003). This might reflect the general importance of coordinating cell compartmentalization and differentiation with cell division. Are Septins Involved in Barrier Formation Apart from Yeast? As septins are conserved from yeast to higher eukaryotes, their roles in the formation of diffusion barriers might also be conserved. The observation that septins are present at the bases of spines (Tada et al., 2007; Xie et al., 2007), in the annuli of spermatozoa (Ihara et al., 2005; Kissel et al., 2005; Steels et al., 2007), and at cytokinetic barriers (Schmidt and Nichols, 2004) strongly suggests that septins are also involved in the formation of diffusion barriers in higher eukaryotes. Interestingly, there are many morphological and size similarities between the annulus, the late cytokinetic bridge, the spine neck, and the yeast bud neck (Barral and Mansuy, 2007), suggesting more profound connections. For example, similar to yeast, septin depletion or overexpression interferes with spine morphogenesis and complexity (Tada et al., 2007; Xie et al., 2007). Furthermore, the presence of septins correlates with the specialization of the ER in the spine to form the spine apparatus, a smooth ER structure. In spermatozoa, septins are thought to be the major constituent of the annulus at the boundary of the anterior and posterior tail compartments (Ihara et al., 2005; Kissel et al., 2005; Steels et al., 2007). Accordingly, depletion of Sept4 affects sperm maturation and leads to male infertility in mice. The annulus also affects the distribution of internal organelles such as mitochondria. During cytokinesis in HeLa cells, the localization of septins in a ring at the cleavage plane correlates with the existence of a diffusion barrier for transmembrane proteins (LYFPGT46) and proteins associated with the inner leaflet of the plasma membrane (Lyn-GFP) (Schmidt and Nichols, 2004). In these three mammalian cases, the barriers localize to the

Developmental Cell

Review ‘‘neck’’ of the cell, the size of which is similar to that of the yeast bud neck. Together, a hallmark of septin-dependent barriers might be that they organize both the plasma membrane and intracellular membranes. Interestingly, in algae, the presence of septin genes correlates with that of centrioles and flagella. Furthermore, the presence of a diffusion barrier at the base of flagella correlates with the presence of a high-ordered lipid phase. Finally, in photoreceptor cells, for example, the connecting cilium, which links the inner segment and the soma to the outer segment where photoreception takes place, contributes to the global compartmentalization of the cell, and not only of the plasma membrane. Therefore, it will be interesting to test whether septins localize to the base of cilia and flagella and the neck of outer segments and whether they are involved in their compartmentalization. Similarities in septin function might also be extended to the process of spermatogenesis. There, as seen during yeast sporulation, nuclear pores are stripped away from the nuclear envelope as the perinuclear ring progresses and combs the nucleus of the spermatocyte (Figure 5). In many species, such as X. laevis, sperm nuclei are therefore devoid of pores. In other species, a few pores remain or are inserted de novo and cluster in the back of the nucleus (Fawcett and Chemes, 1979; Mortimer and Thompson, 1976). Again, it will be interesting to determine whether septins are part of the perinuclear ring and mediate pore stripping. More generally, it will also be interesting to determine whether stripping contributes to the resetting of age and epigenetic information in the male gamete, and whether septins are involved in these events. Whereas septin-dependent barriers appear to be widespread in eukaryotes, the TJ type of barriers are also likely to be widespread in multicellular organisms, where they are not confined to epithelia, but also exist in other cell types, such as endothelial cells. Indeed, these cells also form TJs. In summary, recent data on diffusion barriers and particularly on septin-dependent diffusion barriers promote the model that such structures evolved early in evolution as a mechanism to distinguish and individualize cellular appendages and allow their specialization. More remarkably, these diffusion barriers appear to provide a generic solution for the cell to deal with two very fundamental long-term processes: cellular aging and the generation of memory. In fact, these two processes, the protection of the cell from the age of its mother, and the ability to maintain traces of one’s past over long period of times, appear to represent the heart of what diffusion barriers are there to achieve. Therefore, we suggest that these two processes have been the main driving forces in the emergence of diffusion barriers during evolution.

Alvarez, V.A., and Sabatini, B.L. (2007). Anatomical and physiological plasticity of dendritic spines. Annu. Rev. Neurosci. 30, 79–97. Amberg, D.C., Zahner, J.E., Mulholland, J.W., Pringle, J.R., and Botstein, D. (1997). Aip3p/Bud6p, a yeast actin-interacting protein that is involved in morphogenesis and the selection of bipolar budding sites. Mol. Biol. Cell 8, 729–753. Amin, N.D., Zheng, Y.L., Kesavapany, S., Kanungo, J., Guszczynski, T., Sihag, R.K., Rudrabhatla, P., Albers, W., Grant, P., and Pant, H.C. (2008). Cyclindependent kinase 5 phosphorylation of human septin SEPT5 (hCDCrel-1) modulates exocytosis. J. Neurosci. 28, 3631–3643. Ashby, M.C., Maier, S.R., Nishimune, A., and Henley, J.M. (2006). Lateral diffusion drives constitutive exchange of AMPA receptors at dendritic spines and is regulated by spine morphology. J. Neurosci. 26, 7046–7055. Balda, M.S., Garrett, M.D., and Matter, K. (2003). The ZO-1-associated Y-box factor ZONAB regulates epithelial cell proliferation and cell density. J. Cell Biol. 160, 423–432. Balda, M.S., Whitney, J.A., Flores, C., Gonzalez, S., Cereijido, M., and Matter, K. (1996). Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical-basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J. Cell Biol. 134, 1031–1049. Barral, Y., and Mansuy, I.M. (2007). Septins: cellular and functional barriers of neuronal activity. Curr. Biol. 17, R961–R963. Barral, Y., and Kinoshita, M. (2008). Structural insights shed light onto septin assemblies and function. Curr. Opin. Cell Biol. 20, 12–18. Barral, Y., Parra, M., Bidlingmaier, S., and Snyder, M. (1999). Nim1-related kinases coordinate cell cycle progression with the organization of the peripheral cytoskeleton in yeast. Genes Dev. 13, 176–187. Barral, Y., Mermall, V., Mooseker, M.S., and Snyder, M. (2000). Compartmentalization of the cell cortex by septins is required for maintenance of cell polarity in yeast. Mol. Cell 5, 841–851. Beites, C.L., Xie, H., Bowser, R., and Trimble, W.S. (1999). The septin CDCrel-1 binds syntaxin and inhibits exocytosis. Nat. Neurosci. 2, 434–439. Bi, E., Maddox, P., Lew, D.J., Salmon, E.D., McMillan, J.N., Yeh, E., and Pringle, J.R. (1998). Involvement of an actomyosin contractile ring in Saccharomyces cerevisiae cytokinesis. J. Cell Biol. 142, 1301–1312. Cao, L., Ding, X., Yu, W., Yang, X., Shen, S., and Yu, L. (2007). Phylogenetic and evolutionary analysis of the septin protein family in metazoan. FEBS Lett. 581, 5526–5532. Casamayor, A., and Snyder, M. (2002). Bud-site selection and cell polarity in budding yeast. Curr. Opin. Microbiol. 5, 179–186. Casamayor, A., and Snyder, M. (2003). Molecular dissection of a yeast septin: distinct domains are required for septin interaction, localization, and function. Mol. Cell. Biol. 23, 2762–2777. Castillon, G.A., Adames, N.R., Rosello, C.H., Seidel, H.S., Longtine, M.S., Cooper, J.A., and Heil-Chapdelaine, R.A. (2003). Septins have a dual role in controlling mitotic exit in budding yeast. Curr. Biol. 13, 654–658. Cesario, M.M., and Bartles, J.R. (1994). Compartmentalization, processing and redistribution of the plasma membrane protein CE9 on rodent spermatozoa. Relationship of the annulus to domain boundaries in the plasma membrane of the tail. J. Cell Sci. 107, 561–570.

ACKNOWLEDGMENTS We would like to thank R. Kroschewski, K. Matter, M. The´ry, Z. Shcheprova, and the whole Barral lab for interesting discussions and L. Clay, S. BuvelotFrei, and C. Weirich for critical reading of the manuscript. We also thank J. Vogel (McGill University) for discussions and extensive comments on the manuscript while visiting the ETH. F.C. and Y.B. are supported by the ETH Zu¨rich. REFERENCES Aguilaniu, H., Gustafsson, L., Rigoulet, M., and Nystrom, T. (2003). Asymmetric inheritance of oxidatively damaged proteins during cytokinesis. Science 299, 1751–1753.

DeMarini, D.J., Adams, A.E., Fares, H., De Virgilio, C., Valle, G., Chuang, J.S., and Pringle, J.R. (1997). A septin-based hierarchy of proteins required for localized deposition of chitin in the Saccharomyces cerevisiae cell wall. J. Cell Biol. 139, 75–93. Dent, J., Kato, K., Peng, X.R., Martinez, C., Cattaneo, M., Poujol, C., Nurden, P., Nurden, A., Trimble, W.S., and Ware, J. (2002). A prototypic platelet septin and its participation in secretion. Proc. Natl. Acad. Sci. USA 99, 3064–3069. Dobbelaere, J., and Barral, Y. (2004). Spatial coordination of cytokinetic events by compartmentalization of the cell cortex. Science 305, 393–396. Dotti, C.G., Parton, R.G., and Simons, K. (1991). Polarized sorting of glypiated proteins in hippocampal neurons. Nature 349, 158–161.

Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 503

Developmental Cell

Review Douglas, L.M., Alvarez, F.J., McCreary, C., and Konopka, J.B. (2005). Septin function in yeast model systems and pathogenic fungi. Eukaryot. Cell 4, 1503–1512. Dragsten, P.R., Blumenthal, R., and Handler, J.S. (1981). Membrane asymmetry in epithelia: is the tight junction a barrier to diffusion in the plasma membrane? Nature 294, 718–722.

by the septin cytoskeleton is essential for structural and mechanical integrity of mammalian spermatozoa. Dev. Cell 8, 343–352. Ihara, M., Yamasaki, N., Hagiwara, A., Tanigaki, A., Kitano, A., Hikawa, R., Tomimoto, H., Noda, M., Takanashi, M., Mori, H., et al. (2007). Sept4, a component of presynaptic scaffold and Lewy bodies, is required for the suppression of alpha-synuclein neurotoxicity. Neuron 53, 519–533.

Epp, J.A., and Chant, J. (1997). An IQGAP-related protein controls actin-ring formation and cytokinesis in yeast. Curr. Biol. 7, 921–929.

Iwase, M., Luo, J., Bi, E., and Toh-e, A. (2007). Shs1 plays separable roles in septin organization and cytokinesis in Saccharomyces cerevisiae. Genetics 177, 215–229.

Erjavec, N., Larsson, L., Grantham, J., and Nystrom, T. (2007). Accelerated aging and failure to segregate damaged proteins in Sir2 mutants can be suppressed by overproducing the protein aggregation-remodeling factor Hsp104p. Genes Dev. 21, 2410–2421.

Jansen, R.P., Dowzer, C., Michaelis, C., Galova, M., and Nasmyth, K. (1996). Mother cell-specific HO expression in budding yeast depends on the unconventional myosin myo4p and other cytoplasmic proteins. Cell 84, 687–697.

Estrada, P., Kim, J., Coleman, J., Walker, L., Dunn, B., Takizawa, P., Novick, P., and Ferro-Novick, S. (2003). Myo4p and She3p are required for cortical ER inheritance in Saccharomyces cerevisiae. J. Cell Biol. 163, 1255–1266.

Jesaitis, A.J., and Yguerabide, J. (1986). The lateral mobility of the (Na+,K+)dependent ATPase in Madin-Darby canine kidney cells. J. Cell Biol. 102, 1256–1263.

Falcon, A.A., and Aris, J.P. (2003). Plasmid accumulation reduces life span in Saccharomyces cerevisiae. J. Biol. Chem. 278, 41607–41617.

Jin, H., and Amberg, D.C. (2000). The secretory pathway mediates localization of the cell polarity regulator Aip3p/Bud6p. Mol. Biol. Cell 11, 647–661.

Fares, H., Peifer, M., and Pringle, J.R. (1995). Localization and possible functions of Drosophila septins. Mol. Biol. Cell 6, 1843–1859.

Jou, T.S., Schneeberger, E.E., and Nelson, W.J. (1998). Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases. J. Cell Biol. 142, 101–115.

Fares, H., Goetsch, L., and Pringle, J.R. (1996). Identification of a developmentally regulated septin and involvement of the septins in spore formation in Saccharomyces cerevisiae. J. Cell Biol. 132, 399–411. Fawcett, D.W., and Chemes, H.E. (1979). Changes in distribution of nuclear pores during differentiation of the male germ cells. Tissue Cell 11, 147–162. Field, C.M., and Alberts, B.M. (1995). Anillin, a contractile ring protein that cycles from the nucleus to the cell cortex. J. Cell Biol. 131, 165–178. Field, C.M., Coughlin, M., Doberstein, S., Marty, T., and Sullivan, W. (2005). Characterization of anillin mutants reveals essential roles in septin localization and plasma membrane integrity. Development 132, 2849–2860. Folsch, H. (2008). Regulation of membrane trafficking in polarized epithelial cells. Curr. Opin. Cell Biol. 20, 208–213. Fuchs, J., and Loidl, J. (2004). Behaviour of nucleolus organizing regions (NORs) and nucleoli during mitotic and meiotic divisions in budding yeast. Chromosome Res. 12, 427–438.

Juschke, C., Ferring, D., Jansen, R.P., and Seedorf, M. (2004). A novel transport pathway for a yeast plasma membrane protein encoded by a localized mRNA. Curr. Biol. 14, 406–411. Keaton, M.A., and Lew, D.J. (2006). Eavesdropping on the cytoskeleton: progress and controversy in the yeast morphogenesis checkpoint. Curr. Opin. Microbiol. 9, 540–546. Kinoshita, M., Kumar, S., Mizoguchi, A., Ide, C., Kinoshita, A., Haraguchi, T., Hiraoka, Y., and Noda, M. (1997). Nedd5, a mammalian septin, is a novel cytoskeletal component interacting with actin-based structures. Genes Dev. 11, 1535–1547. Kinoshita, M., Field, C.M., Coughlin, M.L., Straight, A.F., and Mitchison, T.J. (2002). Self- and actin-templated assembly of Mammalian septins. Dev. Cell 3, 791–802. Kissel, H., Georgescu, M.M., Larisch, S., Manova, K., Hunnicutt, G.R., and Steller, H. (2005). The Sept4 septin locus is required for sperm terminal differentiation in mice. Dev. Cell 8, 353–364.

Gerst, J.E. (2008). Message on the web: mRNA and ER co-trafficking. Trends Cell Biol. 18, 68–76.

Kobayashi, T., Storrie, B., Simons, K., and Dotti, C.G. (1992). A functional barrier to movement of lipids in polarized neurons. Nature 359, 647–650.

Hall, P.A., and Russell, S.E. (2004). The pathobiology of the septin gene family. J. Pathol. 204, 489–505.

Korinek, W.S., Bi, E., Epp, J.A., Wang, L., Ho, J., and Chant, J. (2000). Cyk3, a novel SH3-domain protein, affects cytokinesis in yeast. Curr. Biol. 10, 947–950.

Hanrahan, J., and Snyder, M. (2003). Cytoskeletal activation of a checkpoint kinase. Mol. Cell 12, 663–673. Hartwell, L.H. (1971). Genetic control of the cell division cycle in yeast. IV. Genes controlling bud emergence and cytokinesis. Exp. Cell Res. 69, 265–276.

Kuhlenbaumer, G., Hannibal, M.C., Nelis, E., Schirmacher, A., Verpoorten, N., Meuleman, J., Watts, G.D., De Vriendt, E., Young, P., Stogbauer, F., et al. (2005). Mutations in SEPT9 cause hereditary neuralgic amyotrophy. Nat. Genet. 37, 1044–1046.

Harvey, C.D., and Svoboda, K. (2007). Locally dynamic synaptic learning rules in pyramidal neuron dendrites. Nature 450, 1195–1200.

Kusch, J., Meyer, A., Snyder, M.P., and Barral, Y. (2002). Microtubule capture by the cleavage apparatus is required for proper spindle positioning in yeast. Genes Dev. 16, 1627–1639.

Harvey, C.D., Yasuda, R., Zhong, H., and Svoboda, K. (2008). The spread of Ras activity triggered by activation of a single dendritic spine. Science 321, 136–140.

Lippincott, J., and Li, R. (1998). Sequential assembly of myosin II, an IQGAPlike protein, and filamentous actin to a ring structure involved in budding yeast cytokinesis. J. Cell Biol. 140, 355–366.

Hedstrom, K.L., Ogawa, Y., and Rasband, M.N. (2008). AnkyrinG is required for maintenance of the axon initial segment and neuronal polarity. J. Cell Biol. 183, 635–640.

Lipschutz, J.H., Lingappa, V.R., and Mostov, K.E. (2003). The exocyst affects protein synthesis by acting on the translocation machinery of the endoplasmic reticulum. J. Biol. Chem. 278, 20954–20960.

Huang, Y.W., Yan, M., Collins, R.F., Diciccio, J.E., Grinstein, S., and Trimble, W.S. (2008). Mammalian septins are required for phagosome formation. Mol. Biol. Cell 19, 1717–1726.

Luedeke, C., Frei, S.B., Sbalzarini, I., Schwarz, H., Spang, A., and Barral, Y. (2005). Septin-dependent compartmentalization of the endoplasmic reticulum during yeast polarized growth. J. Cell Biol. 169, 897–908.

Ihara, M., Tomimoto, H., Kitayama, H., Morioka, Y., Akiguchi, I., Shibasaki, H., Noda, M., and Kinoshita, M. (2003). Association of the cytoskeletal GTPbinding protein Sept4/H5 with cytoplasmic inclusions found in Parkinson’s disease and other synucleinopathies. J. Biol. Chem. 278, 24095–24102.

Maddox, A.S., Lewellyn, L., Desai, A., and Oegema, K. (2007). Anillin and the septins promote asymmetric ingression of the cytokinetic furrow. Dev. Cell 12, 827–835.

Ihara, M., Kinoshita, A., Yamada, S., Tanaka, H., Tanigaki, A., Kitano, A., Goto, M., Okubo, K., Nishiyama, H., Ogawa, O., et al. (2005). Cortical organization

504 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

Martinez, C., Corral, J., Dent, J.A., Sesma, L., Vicente, V., and Ware, J. (2006). Platelet septin complexes form rings and associate with the microtubular network. J. Thromb. Haemost. 4, 1388–1395.

Developmental Cell

Review Matter, K., and Balda, M.S. (2003). Functional analysis of tight junctions. Methods 30, 228–234.

Seshan, A., Bardin, A.J., and Amon, A. (2002). Control of Lte1 localization by cell polarity determinants and Cdc14. Curr. Biol. 12, 2098–2110.

Mortimer, D., and Thompson, T.E. (1976). Nuclear pores in the spermatozoon of the rat. Experientia 32, 104–106.

Shcheprova, Z., Baldi, S., Frei, S.B., Gonnet, G., and Barral, Y. (2008). A mechanism for asymmetric segregation of age during yeast budding. Nature 454, 728–734.

Moseley, J.B., Sagot, I., Manning, A.L., Xu, Y., Eck, M.J., Pellman, D., and Goode, B.L. (2004). A conserved mechanism for Bni1- and mDia1-induced actin assembly and dual regulation of Bni1 by Bud6 and profilin. Mol. Biol. Cell 15, 896–907. Murray, A.W., and Szostak, J.W. (1983). Pedigree analysis of plasmid segregation in yeast. Cell 34, 961–970. Myles, D.G., Primakoff, P., and Bellve, A.R. (1981). Surface domains of the guinea pig sperm defined with monoclonal antibodies. Cell 23, 433–439. Myles, D.G., Primakoff, P., and Koppel, D.E. (1984). A localized surface protein of guinea pig sperm exhibits free diffusion in its domain. J. Cell Biol. 98, 1905– 1909. Nakada, C., Ritchie, K., Oba, Y., Nakamura, M., Hotta, Y., Iino, R., Kasai, R.S., Yamaguchi, K., Fujiwara, T., and Kusumi, A. (2003). Accumulation of anchored proteins forms membrane diffusion barriers during neuronal polarization. Nat. Cell Biol. 5, 626–632. Nehme, C.L., Cesario, M.M., Myles, D.G., Koppel, D.E., and Bartles, J.R. (1993). Breaching the diffusion barrier that compartmentalizes the transmembrane glycoprotein CE9 to the posterior-tail plasma membrane domain of the rat spermatozoon. J. Cell Biol. 120, 687–694. Neufeld, T.P., and Rubin, G.M. (1994). The Drosophila peanut gene is required for cytokinesis and encodes a protein similar to yeast putative bud neck filament proteins. Cell 77, 371–379. Nguyen, T.Q., Sawa, H., Okano, H., and White, J.G. (2000). The C. elegans septin genes, unc-59 and unc-61, are required for normal postembryonic cytokineses and morphogenesis but have no essential function in embryogenesis. J. Cell Sci. 113, 3825–3837. Nimchinsky, E.A., Sabatini, B.L., and Svoboda, K. (2002). Structure and function of dendritic spines. Annu. Rev. Physiol. 64, 313–353. Oegema, K., Savoian, M.S., Mitchison, T.J., and Field, C.M. (2000). Functional analysis of a human homologue of the Drosophila actin binding protein anillin suggests a role in cytokinesis. J. Cell Biol. 150, 539–552. Pan, F., Malmberg, R.L., and Momany, M. (2007). Analysis of septins across kingdoms reveals orthology and new motifs. BMC Evol. Biol. 7, 103. Park, H.O., and Bi, E. (2007). Central roles of small GTPases in the development of cell polarity in yeast and beyond. Microbiol. Mol. Biol. Rev. 71, 48–96. Pereira, G., and Schiebel, E. (2001). The role of the yeast spindle pole body and the mammalian centrosome in regulating late mitotic events. Curr. Opin. Cell Biol. 13, 762–769. Petruszak, J.A., Nehme, C.L., and Bartles, J.R. (1991). Endoproteolytic cleavage in the extracellular domain of the integral plasma membrane protein CE9 precedes its redistribution from the posterior to the anterior tail of the rat spermatozoon during epididymal maturation. J. Cell Biol. 114, 917–927. Pizarro-Cerda, J., Jonquieres, R., Gouin, E., Vandekerckhove, J., Garin, J., and Cossart, P. (2002). Distinct protein patterns associated with Listeria monocytogenes InlA- or InlB-phagosomes. Cell. Microbiol. 4, 101–115. Pruyne, D., Legesse-Miller, A., Gao, L., Dong, Y., and Bretscher, A. (2004). Mechanisms of polarized growth and organelle segregation in yeast. Annu. Rev. Cell Dev. Biol. 20, 559–591. Sabatini, B.L., Oertner, T.G., and Svoboda, K. (2002). The life cycle of Ca(2+) ions in dendritic spines. Neuron 33, 439–452. Schmidt, K., and Nichols, B.J. (2004). A barrier to lateral diffusion in the cleavage furrow of dividing mammalian cells. Curr. Biol. 14, 1002–1006. Schmid, M., Jaedicke, A., Du, T.G., and Jansen, R.P. (2006). Coordination of endoplasmic reticulum and mRNA localization to the yeast bud. Curr. Biol. 16, 1538–1543. Scott-Drew, S., Wong, C.M., and Murray, J.A. (2002). DNA plasmid transmission in yeast is associated with specific sub-nuclear localisation during cell division. Cell Biol. Int. 26, 393–405.

Shepard, K.A., Gerber, A.P., Jambhekar, A., Takizawa, P.A., Brown, P.O., Herschlag, D., DeRisi, J.L., and Vale, R.D. (2003). Widespread cytoplasmic mRNA transport in yeast: identification of 22 bud-localized transcripts using DNA microarray analysis. Proc. Natl. Acad. Sci. USA 100, 11429–11434. Sherman, D.L., and Brophy, P.J. (2005). Mechanisms of axon ensheathment and myelin growth. Nat. Rev. Neurosci. 6, 683–690. Shulewitz, M.J., Inouye, C.J., and Thorner, J. (1999). Hsl7 localizes to a septin ring and serves as an adapter in a regulatory pathway that relieves tyrosine phosphorylation of Cdc28 protein kinase in Saccharomyces cerevisiae. Mol. Cell. Biol. 19, 7123–7137. Sinclair, D.A., and Guarente, L. (1997). Extrachromosomal rDNA circles– a cause of aging in yeast. Cell 91, 1033–1042. Singla, V., and Reiter, J.F. (2006). The primary cilium as the cell’s antenna: signaling at a sensory organelle. Science 313, 629–633. Spiegel, S., Blumenthal, R., Fishman, P.H., and Handler, J.S. (1985). Gangliosides do not move from apical to basolateral plasma membrane in cultured epithelial cells. Biochim. Biophys. Acta 821, 310–318. Steels, J.D., Estey, M.P., Froese, C.D., Reynaud, D., Pace-Asciak, C., and Trimble, W.S. (2007). Sept12 is a component of the mammalian sperm tail annulus. Cell Motil. Cytoskeleton 64, 794–807. Steinkraus, K.A., Kaeberlein, M., and Kennedy, B. (2008). Replicative aging in yeast: The means to the end. Annu. Rev. Cell Dev. Biol. 24, 29–54. Surka, M.C., Tsang, C.W., and Trimble, W.S. (2002). The mammalian septin MSF localizes with microtubules and is required for completion of cytokinesis. Mol. Biol. Cell 13, 3532–3545. Svoboda, K., Tank, D.W., and Denk, W. (1996). Direct measurement of coupling between dendritic spines and shafts. Science 272, 716–719. Szkotnicki, L., Crutchley, J.M., Zyla, T.R., Bardes, E.S., and Lew, D.J. (2008). The checkpoint kinase Hsl1p is activated by Elm1p-dependent phosphorylation. Mol. Biol. Cell 19, 4675–4686. Tada, T., Simonetta, A., Batterton, M., Kinoshita, M., Edbauer, D., and Sheng, M. (2007). Role of Septin cytoskeleton in spine morphogenesis and dendrite development in neurons. Curr. Biol. 17, 1752–1758. Takizawa, P.A., DeRisi, J.L., Wilhelm, J.E., and Vale, R.D. (2000). Plasma membrane compartmentalization in yeast by messenger RNA transport and a septin diffusion barrier. Science 290, 341–344. Taniguchi, M., Taoka, M., Itakura, M., Asada, A., Saito, T., Kinoshita, M., Takahashi, M., Isobe, T., and Hisanaga, S. (2007). Phosphorylation of adult type Sept5 (CDCrel-1) by cyclin-dependent kinase 5 inhibits interaction with syntaxin-1. J. Biol. Chem. 282, 7869–7876. Tarassov, K., Messier, V., Landry, C.R., Radinovic, S., Serna Molina, M.M., Shames, I., Malitskaya, Y., Vogel, J., Bussey, H., and Michnick, S.W. (2008). An in vivo map of the yeast protein interactome. Science 320, 1465–1470. Toikkanen, J.H., Miller, K.J., Soderlund, H., Jantti, J., and Keranen, S. (2003). The beta subunit of the Sec61p endoplasmic reticulum translocon interacts with the exocyst complex in Saccharomyces cerevisiae. J. Biol. Chem. 278, 20946–20953. Valdez-Taubas, J., and Pelham, H.R. (2003). Slow diffusion of proteins in the yeast plasma membrane allows polarity to be maintained by endocytic cycling. Curr. Biol. 13, 1636–1640. Vallen, E.A., Caviston, J., and Bi, E. (2000). Roles of Hof1p, Bni1p, Bnr1p, and myo1p in cytokinesis in Saccharomyces cerevisiae. Mol. Biol. Cell 11, 593–611. Valverde, F. (1967). Apical dendritic spines of the visual cortex and light deprivation in the mouse. Exp. Brain Res. 3, 337–352. Valverde, F. (1971). Rate and extent of recovery from dark rearing in the visual cortex of the mouse. Brain Res. 33, 1–11.

Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc. 505

Developmental Cell

Review van Meer, G., and Simons, K. (1986). The function of tight junctions in maintaining differences in lipid composition between the apical and the basolateral cell surface domains of MDCK cells. EMBO J. 5, 1455–1464. Vieira, O.V., Gaus, K., Verkade, P., Fullekrug, J., Vaz, W.L., and Simons, K. (2006). FAPP2, cilium formation, and compartmentalization of the apical membrane in polarized Madin-Darby canine kidney (MDCK) cells. Proc. Natl. Acad. Sci. USA 103, 18556–18561. Weirich, C.S., Erzberger, J.P., and Barral, Y. (2008). The septin family of GTPases: architecture and dynamics. Nat. Rev. Mol. Cell Biol. 9, 478–489. Weiss, R.L., Goodenough, D.A., and Goodenough, U.W. (1977). Membrane particle arrays associated with the basal body and with contractile vacuole secretion in Chlamydomonas. J. Cell Biol. 72, 133–143. Weisswange, I., Bretschneider, T., and Anderson, K.I. (2005). The leading edge is a lipid diffusion barrier. J. Cell Sci. 118, 4375–4380.

506 Developmental Cell 16, April 21, 2009 ª2009 Elsevier Inc.

Winckler, B., and Mellman, I. (1999). Neuronal polarity: controlling the sorting and diffusion of membrane components. Neuron 23, 637–640. Winckler, B., Forscher, P., and Mellman, I. (1999). A diffusion barrier maintains distribution of membrane proteins in polarized neurons. Nature 397, 698–701. Xie, Y., Vessey, J.P., Konecna, A., Dahm, R., Macchi, P., and Kiebler, M.A. (2007). The GTP-binding protein Septin 7 is critical for dendrite branching and dendritic-spine morphology. Curr. Biol. 17, 1746–1751. Yu, A.S., McCarthy, K.M., Francis, S.A., McCormack, J.M., Lai, J., Rogers, R.A., Lynch, R.D., and Schneeberger, E.E. (2005). Knockdown of occludin expression leads to diverse phenotypic alterations in epithelial cells. Am. J. Physiol. Cell Physiol. 288, C1231–C1241. Zhang, J., Kong, C., Xie, H., McPherson, P.S., Grinstein, S., and Trimble, W.S. (1999). Phosphatidylinositol polyphosphate binding to the mammalian septin H5 is modulated by GTP. Curr. Biol. 9, 1458–1467.