BIOLOGICAL OVERVIEW

The septin family of proteins was first identified in the budding yeast Saccharomyces cerevisiae and subsequently found in other fungi, Drosophila, Xenopus and mammals. The name septin refers to the widespread involvement of these proteins in cytokinesis and septum formation. The septins all contain a P loop nucleotide-binding consensus sequence and additional motifs homologous to the consensus sequences that define the GTPase superfamily. Septin polypeptides contain predicted coiled-coil domains of 36-90 amino acids near their C-termini. In S. cerevisiae, the septins are closely associated with, and apparently constituents of, a ring of filaments, about 10 nm in diameter, that is closely apposed to the cytoplasmic face of the plasma membrane in the mother-bud neck. Mutation in one of the yeast septin genes causes loss of the bud neck filaments, loss of localization of all four of the septins (suggesting that their assembly at the neck is interdependent), a failure to form a chitin ring in the cell wall at the base of the bud (suggesting a defect in the localization of chitin synthase or of an associated regulatory factor), the production of abnormally elongated buds (apparently associated with a hyperpolarization of the cytoskeleton), and a failure of cytokinesis (presumably reflecting a defect in actin/myosin organization and/or in the localization of cell-wall deposition). Septin mutants also show defects in the axial budding pattern, apparently reflecting an inability to localize Bud3p, a putative component of the axial position signal. Thus, the yeast septins appear to play a variety of roles in the organization and development of the cell surface (Fares, 1995 and references).

The first Drosophila septin gene to be discovered was peanut, isolated in a search for factors that interact with seven in absentia (sina), a gene required for the induction of neural fate in the presumptive R7 cells. peanut(pnut) behaves as a dominant enhancer of sina. Homozygous peanut mutants do not survive to adulthood but instead die shortly after pupation. peanut mutants have either severely reduced or else no imaginal discs; these are the epithelial structures that give rise to adult tissues. Such disc-less, pupal-lethal phenotypes often reflect an underlying defect in mitosis. In mitotic mutants, early divisions during embryogenesis are presumably supported by maternally supplied gene products, but later proliferation of imaginal tissues is dependent on the mutant zygotic genome. peanut mutant brains contain a large number of polyploid and multinucleate cells. Such cells have multipolar mitotic spindles and extra centrosomes (Neufeld, 1994).

Drosophila tissue culture cells, in interphase, stain at low intensity with monoclonal antibodies against Peanut; a slight concentration of staining is seen at the plasma membrane. Localization of staining first occurs during cell division, in late anaphase cells, as membrane staining increases at the newly formed cleavage furrow. Such staining is only observed in cells that have initiated cleavage. The intensity of staining at the furrow increases as it progresses inward and appears to occur at the expense of overall membrane staining, which decreases during cytokinesis, suggesting that Pnut protein may be recruited to the furrow from other regions of the membrane. Following mitosis, staining persists well into interphase at the intercellular bridge connecting daughter cells. During cellularization, embryos show intense staining of the advancing membrane front suggesting that Pnut may have a role in cellularization similar to its function during cytokinesis (Neufeld, 1994).

Immunoaffinity studies using antiseptin antibodies reveal that three Drosophia septins, Pnut, Sep1 and Sep2 coimmunoprecipitate and cofractionate as a protein complex. Visualized by electron microcoscopy the septin complex contains filaments of variable lengths (up to 350 nm), meauring 7-9 nm in diameter. The filaments reveal a periodicity of about 26 nm leading to the belief that filaments are constructed from repeats of a 26-nm subunit.The data for complex formation between the three septins best fit a model in which the complex is a linear heterotrimer of septin homodimers, with a final trimer length of 26 nm. Filaments are distributed in lengths between 1 and 14 subunits long, suggesting a simple linear polymerization in which the subunits associate end-to-end (Field, 1996).

What role is played by the septins in promoting furrow formation? In other words, do septins have merely a structural role, as a physical scaffold making up the cytoskeleton underneath the furrow or do the septins have a more active role? In either case, experiments in yeast currently provide the best clues as to septin function. In yeast, bud site selection is defective without septins. Haploid cells normally exhibit an axial pattern in which the new bud is placed immediately adjacent to the bud site used during the preceding cell cycle. Mutations in the septins result in a bipolar budding pattern in haploids but do not affect the budding pattern in diploids. The neck filaments, mainly comprised of septins, form a ring at the presumptive bud site, which persists at the mother-bud neck throughout the cell cycle and is necessary for cytokinesis. Thus in yeast, septins appear to play a very active role in bud site selection (Yang, 1997 and references).

In the axial pattern, cells are constrained to form daughter buds next to the previous division site. Genetic analysis suggests that Bud3, Bud4, and associated factors might provide the cortical marks for axial budding, since mutations in these genes specifically affect the axial pattern. Immunolocalization has revealed that during later stages of the cell cycle, Bud3 encircles the mother-bud neck region as a double ring, at essentially the positions where cells will form axial buds in the next cell cycle. This double ring structure persists until cytokinesis, at which time the double ring is sliced into two single rings, one on each progeny cell. The single rings persist until approximately the time the next axis of polarization forms in anticipation of axial bud formation. The Bud3 rings then dissipate. It appears that Bud3 is initially directed to the neck region by the preexisting septin ring. Conversely, because Bud3 directs the formation of axial buds, it constrains the position of the next septin ring. Thus the axial pattern of budding is potentially produced by a closed cycle of the septin ring acting as a template for accumulation of Bud3 (and associated protein) in the neck, with these factors then directing the positions of the next septal rings. As such, Bud3 behaves as a spatial memory, inherited from one cell cycle to the next. It is quite possible that Bud3 binds the septin ring directly (Chant, 1996).

While certain proteins (including Bud3, Bud4, Axl1 and Axl2 and the septins) are required for proper bud site selection only in haploid cells, other proteins, specifically Bud1, Bud2 and Bud5 are required for bud site selection in all cell types. This second group of proteins are involved in spatially organizing the actin cytoskeleton associated with bud formation, and they are central to directing cell polarity during budding. Bud1 (also known as Rsr1) has strong sequence similarity to the Ras family of proteins and is indeed a GTPase. Bud2 has sequence similarity to GTPase-activating proteins and has been shown to activate GTP hydrolysis by Bud1. Bud5 has sequence similarity to the guanine nucleotide exchange protein Cdc2g and catalyzes GDP-GTP exchange for Bud1. The general bud site selection machinery thus makes up a functional GTPase module: a GTPase and its regulatory proteins. This protein complex has been shown to recruit Cdc24 (an exchange factor for Cdc42 and a Rho-like GTPase) to the bud site. In turn, this complex stimulates the polarization of the actin cytoskeleton toward the presumptive bud site. From this information it becomes clear that the actin based cytoskeleton provides position specific information relevant to septin function (Park, 1997).

What might be the basis for the genetic interaction between peanut and sina that was the basis for the isolation of peanut mutants in Drosophila? A disruption in one of the two copies of pnut increases the severity of a mutant sina allele by about four fold, suggesting that the level of pnut activity is a limiting factor in signaling through sina. Since successful induction of the R7 photoreceptor cells requires a continuous signal from the EGF receptor coded for by sevenless, even a slight delay in the cell cycle due to reduced pnut activity may be enough to decrease further the ability of the presumptive R7 cell in sina mutant flies to respond properly to inductive signals. Alternatively, the effect of pnut mutations on cell signaling may reflect the pleiotropic functions of the septin family of proteins (Neufeld, 1994).

Septins promote F-actin ring formation by crosslinking actin filaments into curved bundles

Animal cell cytokinesis requires a contractile ring of crosslinked actin filaments and myosin motors. How contractile rings form and are stabilized in dividing cells remains unclear. This problem was addressed by focusing on septins, highly conserved proteins in eukaryotes whose precise contribution to cytokinesis remains elusive. The cleavage of the Drosophila melanogaster embryo was used as a model system, where contractile actin rings drive constriction of invaginating membranes to produce an epithelium in a manner akin to cell division. In vivo functional studies show that septins are required for generating curved and tightly packed actin filament networks. In vitro reconstitution assays show that septins alone bundle actin filaments into rings, accounting for the defects in actin ring formation in septin mutants. The bundling and bending activities are conserved for human septins, and highlight unique functions of septins in the organization of contractile actomyosin rings (Mavrakis, 2014).

These findings demonstrate that septins are required during cellularization for generating curved and tightly packed actin filament networks at the tips of cellularization furrows termed furrow canals (FCs). Given that the F-actin bundling and bending activity is conserved for fly and human septins, it is predicted that septins contribute to the assembly, stabilization and contractility of cytokinetic rings. Septin depletion in dividing cultured cells, where anillin is localized to the cleavage furrow, leads to furrow instability, suggesting that the septin F-actin bundling activity could be required for proper cortical tension at the equator or/and the poles, which is in turn critical for the positioning of the cytokinetic ring46. Defective formation or stabilization of curved F-actin could further contribute to furrow shape instability. Decreased cytokinetic ring contractility was also recently confirmed in dividing Drosophila septin mutant epithelia, where anillin localizes to the cleavage furrow (Mavrakis, 2014).

Septins could also potentially contribute to the shape of FCs in actin-independent ways. Recent studies provide compelling evidence that septins regulate the mechanical properties of the cortex in non-dividing mammalian cells. Septins were also proposed to provide the cortical rigidity and membrane curvature necessary for rice blast fungal infection. It will be important to investigate whether septins bind membranes at the FC, and how this might synergize with actin crosslinking for cortex organization (Mavrakis, 2014).

The current findings indicate that the F-actin bending activity of septins requires septin hexamers and not septin filaments, although it cannot be excluded that they both function together in vivo. Septin post-translational modifications might also promote septin filament ring formation, which might in turn organize F-actin in curved bundles. Actin filaments could also potentially act in septin filament nucleation, even under conditions where septins alone do not form filaments. It will be important to compare the actin remodelling activity of septin hexamers and octamers, given that human septins (unlike Drosophila septins form octamers with hSep9 at the ends (Mavrakis, 2014).

A striking feature of septins is that they bend actin into rings and other highly curved geometries. Actin circularization is energetically unfavourable owing to the large bending rigidity of actin filaments. Thus far, stable actin circles have been reported only under strong adhesion mediated by divalent cations or positively charged lipid monolayers. The adhesion energy mediated by septins seems sufficiently high to overcome the large bending energy associated with such a highly curved geometry. In cells, septin-mediated actin curving may act in synergy with myosin-induced actin filament buckling (Mavrakis, 2014).

Septins could conceivably crosslink F-actin into loose contractile networks in processes that do not involve contractile F-actin rings, depending on the local septin and actin filament concentration and turnover, and septin post-translational modifications. Although no direct actin-septin interaction is known in interphase cells, their interplay has been reported in non-dividing cells, despite anillin's nuclear confinement. This study suggests that septins, alone or together with myosin-II, could contribute to these biological processes through their F-actin bundling activity (Mavrakis, 2014).

Mammalian septins are known to interact with exocyst components and SNAREs and to regulate membrane fusion. Membrane growth during cellularization relies on vesicular trafficking, and the exocyst is necessary for this process. As membrane growth is delayed in septin mutants, it will be important to investigate whether septins mediate trafficking events that act together with actin remodelling at the FC to drive membrane growth (Mavrakis, 2014).

GENE STRUCTURE

PROTEIN STRUCTURE

Amino Acids - 539

Structural Domains

The sequence identities between Pnut and yeast CDC3, CDC10, CDC11 and CDC12 and a similar protein in Drosophila, Dmdiff6, range from 35% for Cdc11p to 54% for Dmdiff6, over a region of 300-400 amino acids. Three of the proteins, Pnut, Cdc3p and H5, contain amino termini of about 100 amino acids with limited sequence similarity. These proteins have a predicted coiled-coil domain in the C terminus and an ATP/GTP-binding site motif (P loop) that may be involved in their assembly or regulation (Neufeld, 1994).

date revised:  25 April 2024

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.