Architecture, remodeling, and functions of the septin cytoskeleton

Abstract

The septin family of proteins have fascinated cell biologists for decades due to the elaborate architecture they adopt in different eukaryotic cells. Whether they exist as rings, collars, or gauzes in different cell types and at different times in the cell cycle illustrates a complex series of regulation in structure. While the organization of different septin structures at the cortex of different cell types during the cell cycle has been described to various degrees, the exact structure and regulation at the filament level is still largely unknown. Recent advances in fluorescent and electron microscopy, as well as work in septin biochemistry has allowed new insights into the aspects of septin architecture, remodeling, and function in many cell types. This mini review highlights many of the recent findings with an emphasis on the budding yeast model.

Introduction

The septin genes (CDC3, CDC10, CDC11, and CDC12) were originally discovered in the classic Hartwell screen for mutants that affect the cell division cycle (CDC) in the budding yeast Saccharomyces cerevisiae (Hartwell 1971). The products of these genes were subsequently found to be GTP-binding proteins arranged into filaments, akin to a cytoskeleton, at the bud neck (Byers and Goetsch 1976). Another septin gene was later characterized, SHS1 (also originally termed SEP7), bringing the mitotic septin total to five in budding yeast (Carroll et al. 1998; Mino et al. 1998). The septins are highly conserved throughout eukaryotes with the exception of land plants (Pan et al. 2007) and are involved in numerous cell processes including cytokinesis, ciliogenesis, morphogenesis, mitosis, and exocytosis (Oh and Bi 2011; Mostowy and Cossart 2012; Dolat et al. 2014). Septins form heterooligomeric complexes that polymerize end-to-end into filaments. These filaments can be further organized into higher-order structures such as rings, hourglasses, and gauzes depending on cell types or cell cycle stages. These structures can serve as scaffolds at discrete cellular locations or membrane-bound diffusion barriers to spatially segregate cell compartments, or often a combination of both (Chant et al. 1995; Longtine et al. 1996; DeMarini et al. 1997; Barral et al. 2000; Takizawa et al. 2000). By remodeling the architecture of septin higher-order structures, cells are able to use the same set of proteins to achieve elaborate control of diverse cellular processes. While septin architecture has become increasingly clear due to recent developments in light and electron microscopy, the mechanisms controlling the assembly and remodeling of septin high-order structures still remain elusive. Recent work has begun to elucidate some of these key mechanisms and will be discussed here.

The septin blueprint: assembling the complexes, filaments, hourglasses, and rings

Septin protofilaments are composed of septin hetero-octamers, the “building blocks”, with the core subunits Cdc12-Cdc3-Cdc10-Cdc10-Cdc3-Cdc12 and the use of differential terminal subunits Cdc11 or Shs1, giving rise to two distinct septin hetero-oligomers (Bertin et al. 2008; Garcia et al. 2011). Expression of all five septins in E. coli gives rise to symmetric octamers capped with Cdc11 or Shs1 on both ends and never a combination of Cdc11 and Shs1 (Khan et al. 2018). Octamers capped with Cdc11 polymerize end-to-end to form linear filaments that are often paired like a railway track in vitro (Frazier et al. 1998; Bertin et al. 2008) (Fig. 1, upper left). In contrast, the octamers capped with Shs1 appear to polymerize end-to-end into single filaments and/or laterally associate with each other into curved bundles or rings in vitro (Garcia et al. 2011) (Fig. 1, upper left). This distinct property between the octamers might account for the double and single filaments observed in septin higher-order structures in vivo (Bertin et al. 2012; Ong et al. 2014). Much of the knowledge about septin complex formation and filament assembly has been inferred from the crystal structure of human septin complex SEPT2/6/7 and the proposed interactions between septin subunits during filament assembly (Sirajuddin et al. 2007; John et al. 2007; Bertin et al. 2008; Sellin et al. 2011). The crystal structure of the terminal yeast subunit Cdc11 was recently solved and illustrated that some of the interfaces once thought to mediate septin-septin interaction may be different (Brausemann et al. 2016). In fact, the conserved lysine in the GTP-binding P-loop of septins is an arginine in yeast Cdc11, calling into question the necessity of GTP binding and subsequently GTP hydrolysis of yeast septins in filament assembly. To address this, a recent study used Bimolecular Fluorescence Complementation (BiFC) in combination with the drug mycophenolic acid (MPA) to lower the cytosolic GDP and GTP levels to assess the step-wise association of septin subunits (Weems and McMurray 2017). It was discovered that under normal conditions, Cdc10 forms a homodimer before interacting with two Cdc11-Cdc12-Cdc3 trimers in high GTP conditions and Shs1-Cdc12-Cdc3 trimers in low GTP conditions (Weems and McMurray 2017). Cdc10 and Cdc12 were the only yeast septins previously shown to display GTP hydrolysis activity (Versele and Thorner 2004). Analysis of mutations leading to a loss of the guanine nucleotide binding in Cdc10 and Cdc12 by BiFC provided novel insights into septin complex formation. Without a bound nucleotide, Cdc10 loses the high propensity for homodimerization prior to interaction with Cdc3 (Weems and McMurray 2017). In this case, mutant Cdc10 interacts with Cdc3-Cdc12-Cdc11 (or Cdc3-Cdc12-Shs1) trimer to form a tetramer, and then two tetramers interact to form an octamer. This indicates that when the supposed optimal pathway for octamer formation is no longer available, the cells have developed secondary pathways for octamer formation to avoid detrimental effects. This explains the lack of a strong phenotype at low temperatures for the cdc10 mutation (Weems and McMurray 2017). This also defines an adaptive mechanism for the organism’s survival under different cellular GTP levels. For Cdc12, the nucleotide bound state was found to directly control the ability of Cdc11 (in high GTP) vs. Shs1 (in low GTP) to serve as the terminal subunit (Weems and McMurray 2017). Additionally, Shs1 was found to play a critical role in septin architecture formation in low GTP conditions as shs1Δ cells had relatively normal morphology and septin organization in normal glucose conditions, but were elongated and showed improper septin organization when cultured in glycerol (Weems and McMurray 2017). This, when coupled with the knowledge that altering the Cdc11:Shs1 ratios can change the septin ring diameter (Garcia et al. 2011; Weems and McMurray 2017), illustrates a complex mechanism for cells to regulate octamer formation to alter septin higher-order architecture.

Figure 1. The septin cycle in yeast: architecture, remodeling, and function.

Starting from the upper left and going clockwise: the nascent septin ring forms in late G1 phase, after the launch of a new cell cycle, through Cdc42-controlled septin recruitment (light green boxes) and polarized exocytosis (not depicted here), and the cortical ring serves as a scaffold for myosin-II and a diffusion barrier for polarized growth. The exact filament structure is unknown, but most likely is composed of double filaments formed by Cdc11-capped octamers (green rods) and possibly single Shs1-capped octamers (yellow rods). Next, as the bud emerges and grows in S/G2, the septin structure expands to form an hourglass comprised of double filaments (green rods) that are potentially linked laterally by Shs1-capped octamers (yellow rods). This structure is regulated by interaction with Bni5 (blue box), phosphorylation by Cla4 (orange box), and acetylation (green stars), and the hourglass serves as a scaffold for myosin-II and a diffusion barrier for polarized growth. Third, as the cell enters anaphase, the septin hourglass exhibits a transitional structure that consists of double filaments (green rods in the cartoon and in the EM image) interwoven with single filaments (yellow rods in the cartoon and in the EM image) akin to a gauze-like structure. This transition is regulated by phosphorylation of Shs1 by Gin4 (orange box) and sumoylation of Cdc3, Cdc11, and Shs1 (blue triangles). Lastly, the septins adopt a double ring architecture at the onset of cytokinesis comprised of both double and single filaments arranged circumferentially at the bud neck. This organization is accompanied by a loss of sumoylation and phosphorylation of septins. This structure serves mainly as a diffusion barrier for cytokinesis proteins. In all cases, the light microscopy images are of cells expressing Cdc3-eGFP from its native promoter. The cell cycle timing was determined by bud size and/or spindle morphology. Scale bar, 1 μm. PREM images reproduced from Ong et al. 2014 with permission. Scale bars, 200 nm.

The first septin higher-order structure formed during the yeast cell cycle is a nascent septin ring at the presumed bud site (Fig. 1, upper left). The recruitment of septins to the presumed bud site depends on the small GTPase Cdc42, a master regulator of eukaryotic cell polarity (Cid et al. 2001b; Gladfelter et al. 2002; Caviston et al. 2003; Kadota et al. 2004; Iwase et al. 2006), and involves the function of its effectors Gic1 and Gic2 (Iwase et al. 2006; Sadian et al. 2013). Experimental and modeling studies indicate that a combination of feedback loops between and among Cdc42, septins, and exocytosis gives rise to the recruitment of septins to the presumed bud site, septin ring opening, and daughter cell identity (Okada et al. 2013). Despite these findings, the organization of septin filaments in the nascent septin ring remains unknown.

The second septin higher-order structure formed during the yeast cell cycle is an hourglass at the bud neck, which spans from bud emergence to anaphase (Fig. 1, upper right). The hourglass at the early budding stage, as revealed by platinum-replica electron microscopy (PREM), consists of a layer of membrane-associated, highly packed double filaments of 350 to 400 nm in length that are located at both sides of the bud neck and oriented along the mother-daughter axis (Ong et al. 2014). As Cdc11-capped octamers are rod-shaped and approximately 32-nm long, and have the propensity to form paired filaments in vitro (Frazier et al. 1998; Bertin et al. 2008), the early hourglass is likely mainly made of Cdc11-capped octamers, with 11-to-12 such building blocks constituting each of the paired filament. Shs1 localizes to the presumed bud site and the bud neck as other mitotic septins (Mino et al. 1998; Carroll et al. 1998; Dobbelaere et al. 2003). Shs1-capped octameric rods do not form filaments in vitro (Garcia et al. 2011), but might serve as a linker for the compaction of the double filaments in the early hourglass, a possibility that remains to be tested. The arrangement of double filaments seen by PREM in early budded cells is consistent with previous data obtained by polarized fluorescence microscopy (Vrabioiu and Mitchison 2006; DeMay et al. 2011b) and suggests a rigid structure that is supported by fluorescence-recovery-after-photobleaching (FRAP) analysis of septins in the early hourglass (Caviston et al. 2003; Dobbelaere et al. 2003). However, the mechanism of converting a relatively dynamic nascent septin ring prior to bud emergence to a rigid hourglass after bud emergence remains unknown.

The final septin structure witnessed in the yeast cell cycle is the double ring during cytokinesis. The transition from the hourglass to the double ring may initiate in anaphase and complete at the onset of cytokinesis. As seen by PREM, an array of paired filaments running axially along the mother-daughter axis and interconnecting single filaments running circumferentially to the mother-daughter axis make up a transitional gauze-like structure temporally between the hourglass and double ring (Fig. 1, lower right) (Ong et al. 2014). It is important to note that an earlier study found a gauze-like structure, which consists of circumferentially oriented paired filaments that are interwoven with perpendicularly oriented single filaments along the mother-daughter axis, to be the only hourglass structure in budded cells (Bertin et al. 2012). While this electron-tomography technique preserves the bud-neck shape for visualization (unlike PREM), the use of only asynchronous cell cultures and the fact that a thin layer of membrane-associated septin filaments is surrounded by electron-dense cytoplasm make it challenging to interpret the filament orientation and the timing of the hourglass appearance during the cell cycle. However, it remains possible that, in the PREM study, the circumferential single filaments are actually present in the “early hourglass”, but they are not stable enough to survive the sample-processing conditions. These single filaments might be stabilized by some factors in the “transitional hourglass” (Fig. 1, lower right), which enable their detection by PREM analysis. The double ring consists of both single and double filaments arranged circumferentially (Fig. 1, lower left) (Ong et al. 2014). Live-cell imaging coupled with FRAP, photoactivation, photoconversion, and super resolution microscopy provided a possible mechanism for the architecture change from the gauze-like structure to the double ring (Ong et al. 2014). This dramatic septin remodeling event likely involves rapid disassembly of the parallel double filaments in the transitional hourglass, and re-assembly of the recycled Cdc11-capped octamers into circumferential double filaments coupled with compaction of the circumferential single filaments into a double ring at the external edges of the transitional hourglass (Ong et al. 2014). The disassembly and reassembly of paired septin filaments in different orientations agrees with previous data obtained by both quantitative live imaging (Dobbelaere et al. 2003; Wloka et al. 2011) as well as polarized fluorescent microscopy in which a 90° change in septin positioning was observed (Vrabioiu and Mitchison 2006; DeMay et al. 2011a). Shs1 seems to play a particularly important role in the remodeling process. In shs1Δ cells, very few gauze-like structures and circumferential single filaments were observed, suggesting that Shs1-capped octamers might comprise the single filaments in the transitional structure (Ong et al. 2014). Despite the significant progress in dissecting detailed septin architecture during the cell cycle, the remodeling mechanism for the hourglass-to-double ring transition remains a central fascinating question in the septin field that will surely be addressed with much depth in the future.

Regulating the varieties of septin structures

As alluded to above, assembly and remodeling of the same septin subunits into drastically different higher-ordered structures during the cell cycle must be highly regulated in time and space. The regulation likely occurs through two mechanisms: septin binding proteins and post-translational modifications. Over 100 proteins are known to localize to the bud neck and interact with septins at certain times of the cell cycle (Gladfelter et al. 2001; McMurray and Thorner 2009). A recent proteomics study attempted to identify septin-associated proteins by affinity purification followed by mass spectrometry in yeast cells arrested at different stages of the cell cycle (Renz et al. 2016). This study found 83 septin-associated proteins with some interacting only with the septin complex during a discrete period of the cell cycle. For example, many proteins identified from α-factor-arrested cells represent novel binding partners such as ribosomal RNA processing proteins Nop1, New1, and Rlp7 and nuclear protein Cse1 (Renz et al. 2016). The authors argue that their methods preclude artifacts based on the identification of these interactors only in α-factor-arrested cells only arising in one sample set, confirmation of interactions using a modified and more stringent split-ubiquitin assay (Müller and Johnsson 2008; Moreno et al. 2014; Dünkler et al. 2015), and the use of SILAC labeling prior to mass spectrometry, which should limit artifacts (Ong et al. 2002). While these novel binding partners are potentially very interesting, their interactions with the septins at the bud neck require further investigation. This can be accomplished using a newly developed tripartite split-GFP system that can effectively evaluate the ability of a candidate protein for direct interaction with the septins at the bud neck (Finnigan et al. 2016). The biological significance of such interactions remains to be explored.

Other proteins with known biological functions in regulating septins were also found in the 2016 proteomics study. These include but are not limited to Bni5, Bud3, Bud4, and Gin4 (Renz et al. 2016). Bni5 was found to interact with septins in samples arrested at multiple phases of the cell cycle (Renz et al. 2016). Bni5 is essential for the localization of Myo1, the sole heavy chain of myosin-II in budding yeast, to the bud neck prior to cytokinesis by serving as a linker between the septins and the tail region of Myo1 (Fang et al. 2010; Finnigan et al. 2015). Bni5 dimerizes and regulates the spacing and organization of paired septin filaments by directly interacting with the terminal subunit Cdc11 (Patasi et al. 2015; Booth et al. 2016). Thus, Bni5 is thought to play a key role in stabilizing the septin hourglass. Bud3 and Bud4 were both found to be strong interactors of septins in later stages of the cell cycle (Renz et al. 2016). This finding is consistent with previous studies. Bud3 and Bud4 associate with the septin hourglass in mitosis and then with the double ring during and after cytokinesis (Chant et al. 1995; Sanders and Herskowitz 1996). Both proteins serve as a part of the landmark in directing axial budding (Chant et al. 1995; Sanders and Herskowitz 1996). Both proteins have also been implicated in septin organization (Gladfelter et al. 2005; Wloka et al. 2011; Guo et al. 2011; Eluere et al. 2012; Kang et al. 2013; McQuilken et al. 2017), with Bud4, the anillin-like protein, being involved in the hourglass-to-double ring transition (Wloka et al. 2011; McQuilken et al. 2017).

The kinase Gin4 was also found to strongly interact with the septins in hydroxyurea (S-phase) and cdc15-ts (late anaphase) arrested cells (Renz et al. 2016). Gin4 associates with the septins at the bud neck before cytokinesis and regulates septin organization in mitosis at least in part by phosphorylating Shs1 (Longtine et al. 1998; Mortensen et al. 2002; Asano et al. 2006). Gin4 binds directly to Cdc11 and Shs1, and these interactions may further activate Gin4 kinase activity in mitosis (Mortensen et al. 2002; Asano et al. 2006).

Gin4 is not the only kinase to regulate the septins by phosphorylation. The cyclin dependent kinase Cdc28 as well as the p-21 activated kinase Cla4 phosphorylate Cdc3 and Cdc10, respectively (Tang and Reed 2002; Dobbelaere et al. 2003; Versele and Thorner 2004; Kadota et al. 2004). In total, 59 phosphorylation events have been found to occur on the four septins Cdc3, Cdc10, Cdc11, and Shs1 (Hernández-Rodríguez and Momany 2012). These events are thought to regulate septin higher-order architecture by controlling the disassembly of old septin rings or stabilizing the septin hourglass (Tang and Reed 2002; Versele and Thorner 2004). Due to the plethora of phosphorylation events on septins from unknown kinases and the biological relevance of these phosphorylations being largely unknown (Hernández-Rodríguez and Momany 2012), regulation of septins by kinases remains a highly important topic to study in the field of septin biology.

Besides phosphorylation, septins are also posttranslationally modified by sumoylation and acetylation (Hernández-Rodríguez and Momany 2012; Alonso et al. 2015). Sumoylation of Cdc3, Cdc11, and Shs1 was observed in S. cerevisiae (Johnson and Blobel 1999; Takahashi et al. 2003). Cells carrying sumoylation-deficient alleles of these septin genes displayed no apparent phenotypes, except that ectopic septin rings were associated with the old cell division sites in addition to the bud neck (Johnson and Blobel 1999). This observation suggested that sumoylation of septins might control septin ring disassembly (Johnson and Blobel 1999). However, cells deleted for SIZ1, which encodes the bud neck-localized E3 ligase for sumoylation, failed to produce any phenotypes including formation of the ectopic septin rings, despite the fact that septin sumoylation was nearly abolished in the mutant (Johnson and Gupta 2001). Thus, the biological significance of septin sumoylation in yeast requires further investigation. In mammalian cells, expression of non-sumoylatable septins caused a dramatic defect in cytokinesis in a small, yet significant, population of cells (Ribet et al. 2017). This was attributed to a lack of clearance of large septin bundles from the intracellular bridge during cytokinesis (Ribet et al. 2017), which is strikingly similar to the failed disassembly of the ectopic rings formed by non-sumoylatable septins at the cell division sites in yeast. Together, these observations suggest that sumoylation may play a conserved role in septin filament disassembly, which presumably enables septin architectural remodeling such as the hourglass-to-double ring transition in yeast. A similar hourglass-to-double ring transition might also occur for mammalian septins during cytokinesis, which assume an hourglass-like structure during furrow ingression and then a double ring-like structure during abscission (Estey et al. 2010).

Regulation of septins by acetylation is probably the least studied posttranslational modification. All mitotic septins except Cdc11 are acetylated by the lysine acetyltransferase complex NuA4 (Mitchell et al. 2011). Septins became mislocalized during the nascent ring-to-hourglass transition in NuA4 complex deletion mutant strains, implying that acetylation of septins may regulate septin hourglass formation or its stability (Mitchell et al. 2011). The authors investigated how the loss of acetylation on specific septins could influence septin localization and function using a C-terminal truncation of Shs1 that loses a majority of acetylation sites. They conceded that many of their phenotypes (including elongated bud morphology and septin mislocalization) associated with this truncation allele are complicated by the fact that this region of Shs1 is also heavily phosphorylated and sumoylated (Egelhofer et al. 2008; Garcia et al. 2011; Mitchell et al. 2011; Hernández-Rodríguez and Momany 2012).

The functional significance of the septin architectures

The septins can serve as both scaffolds for proteins involved in cytokinesis and other processes and barriers for diffusion of membrane-associated proteins (Fig. 1). The septin ring at the presumptive bud site as well as the septin hourglass are required for the recruitment and retention of cytokinesis proteins at the bud neck either through direct interactions, for example, the F-BAR protein Hof1 (Oh et al. 2013; Meitinger et al. 2013) or through septin-binding proteins such as Bni5 that links septin filaments to the myosin-II heavy chain Myo1 (Fang et al. 2010; Schneider et al. 2013; Finnigan et al. 2015). Myo1 is required for actin ring formation (Bi et al. 1998). Thus, the septin ring and hourglass scaffold the assembly of the actomyosin ring (AMR) before cytokinesis (Bi et al. 1998; Schneider et al. 2013). The septin hourglass also acts as a scaffold for the morphogenesis checkpoint kinase Hsl1, which, in turn, regulates localization of the downstream component Hsl7 and the effector Wee1-like kinase Swe1 (an inhibitor of CDK1/mitotic cyclins) (Barral et al. 1999; Shulewitz et al. 1999; Longtine et al. 2000; Cid et al. 2001a). This checkpoint module (Hsl1-Hsl7-Swe1) is strategically assembled at the daughter side of the bud neck to monitor bud emergence and morphogenesis and it is activated and delays the entry into mitosis when the morphogenetic process is defective (Keaton and Lew 2006; Howell and Lew 2012). The septin hourglass also serves as a scaffold for many other processes such as bud-site selection and chitin ring synthesis at the mother side of the bud neck during early stage of budding (Longtine et al. 1996; Gladfelter et al. 2001; McMurray and Thorner 2009; Oh and Bi 2011). In animal cells, the septins provide a scaffold for myosin-II and its activating kinases during cytokinesis (Joo et al. 2007). Anillin-dependent organization of septin filaments at the division site also plays a critical role in the maturation of the midbody as well as its anchoring to the plasma membrane (Kechad et al. 2012; El Amine et al. 2013). These septin filaments also promote elongation of the intercellular bridge as well as targeting of the ESCRT-III component CHMP4B to the abscission site (Renshaw et al. 2014). Thus, through their ability to form different higher-order structures, septins are able to function as perfect scaffolds for protein machinery at distinct membranous localities in a variety of situations.

Evidence for the diffusion-barrier model of septin function also comes from work done primarily in budding yeast. During bud growth, the septin hourglass serves to compartmentalize the bud from the mother for membrane or membrane-associated proteins involved in the maintenance of cell polarity or organelle identity (Barral et al. 2000; Takizawa et al. 2000; Luedeke et al. 2005; Shcheprova et al. 2008). Even the nascent septin ring could function as a barrier to restrict Cdc42 activity and exocytosis within the lumen of the ring to control bud emergence and growth (Okada et al. 2013). During cytokinesis, the septin double ring is thought to function as a diffusion barrier that encapsulates diffusible cytokinesis proteins such as the exocyst component Sec3 and chitin synthase II (Chs2) at the division site (Dobbelaere and Barral 2004). The fact that only mild defects in cytokinesis are observed in cells lacking one side of the septin double ring or the entire structure has raised questions as to the necessity of this barrier function in yeast (Wloka et al. 2011). Additionally, Ist2, the marker used for inferring a septin-dependent diffusion barrier at the plasma membrane of the division site (Takizawa et al. 2000), was later found to be an endoplasmic-reticulum (ER) protein (Manford et al. 2012). The compartmentalization seen in the initial study is actually due to the presence of a septin-dependent diffusion barrier at the ER membrane (Manford et al. 2012). This diffusion barrier is critical for preventing the segregation of misfolded proteins from the mother to daughter compartment (Clay et al. 2014; Chao et al. 2014). Septins are also suggested to function as diffusion barriers in different mammalian cell processes such as compartmentalization of the sperm tail (Kissel et al. 2005; Ihara et al. 2005); the forming cilia (Hu et al. 2010), although caution should be taken for this interpretation (Palander et al. 2017); or during dendritic branching in neurons (Xie et al. 2007; Tada et al. 2007). This highlights the conservation of septin-based barriers beyond a specific model system.

Septins appear to play a critical role in modifying cell shape. In animal cells, this function likely occurs through an interaction with the actin cytoskeleton. In mammalian cells, septin filaments co-align with actin filaments in the stress fibers, and these filaments are interdependent for their assembly and organization (Kinoshita et al. 2002; Schmidt and Nichols 2004). HeLa cells show multiple plasma membrane abnormalities upon knockdown of septins in the presence of Latrunculin B, a drug that disrupts actin filaments (Mostowy et al. 2011). Drosophila septin complex crosslinks and bends actin filaments into ring-like structures in vitro and is required for maintaining cell shape during embryogenesis (Mavrakis et al. 2014). In yeast, it is still unclear if septins respond to or influence cell shape. Yeast cells containing a septin mutation exhibit an elongated bud morphology indicative of a role of the septins in regulating cell shape (Hartwell 1971; Adams and Pringle 1984). Yeast septins can also respond to changes in cell curvature as suggested in recent studies in which septin higher-order assembly favored positive curvature at the forming bud neck during bud emergence (Bridges et al. 2016; Kang et al. 2016). It is possible that septins can both detect and modulate cell shape based on their ability to form various higher-order structures that can be dynamically remodeled.

The most striking use of septins that incorporates all the functions discussed above is in the process of fungal pathogenesis. In species such as the opportunistic human pathogen Candida albicans and the rice blast fungus Magnaporthe oryzae, loss of septins or septin architecture results in a significant loss of pathogenicity (Warenda and Konopka 2002; Warenda et al. 2003; Dagdas et al. 2012; Momany and Talbot 2017). In M. oryzae, septins assemble into a ring structure at the base of an infectious peg that is used to penetrate the plant cell, and this septin ring scaffolds the assembly of a toroidal F-actin network that is required for pathogenesis (Dagdas et al. 2012). The septin ring also acts as a lateral diffusion barrier that prevents polarized growth factors from leaving the growing peg (Dagdas et al. 2012). Finally, the septin ring might sensor the membrane curvature and stabilize the neck of the penetration peg, thereby, restricting the protrusive structure to a single location (Dagdas et al. 2012; Momany and Talbot 2017). In summary, septins in different eukaryotic systems employ distinct architecture and dynamic remodeling to regulate diverse cell and developmental processes.

Concluding Remarks

The septins have been investigated since their discovery in the 1970s as critical regulators of cytokinesis, cell morphogenesis, and other processes. Key questions concerning the architecture, remodeling, and functions of the septin cytoskeleton are beginning to be answered largely due to the development of new imaging and gene-editing technologies. The budding yeast S. cerevisiae remains an effective model for defining new concepts and mechanisms for septin structure and function. For example, it is still unclear how various septin higher-order structures such as rings, hourglasses, and gauzes are assembled at the molecular and filament levels and how these structures are remodeled in time and space to perform distinct functions during the cell cycle. Continuing studies of septins in pathogenic fungi will help understand the cause and mechanism of infectious diseases in humans and plants. Septin studies in animal cells not only provide invaluable insights into the conservation of septin structure and function, but also lead to new discoveries regrading processes that are not shared by yeast such as cell migration. Septin studies in animal cells will also help understand the cause of septin-related diseases including male infertility, cancer, and neurodegenerative diseases.