Septin ring size scaling and dynamics require the coiled-coil region of Shs1p

How the size and dynamics of higher-order septin structures is determined is not well understood in any system. In this paper, we show that the coiled-coil domain of the septin Shs1p limits septin ring size and dynamics in the filamentous fungus Ashbya gossypii, providing a link between protein exchange and the scaling of septin assemblies.

Abstract

Septins are conserved GTP-binding proteins that assemble into heteromeric complexes that form filaments and higher-order structures in cells. What directs filament assembly, determines the size of higher-order septin structures, and governs septin dynamics is still not well understood. We previously identified two kinases essential for septin ring assembly in the filamentous fungus Ashbya gossypii and demonstrate here that the septin Shs1p is multiphosphorylated at the C-terminus of the protein near the predicted coiled-coil domain. Expression of the nonphosphorylatable allele shs1-9A does not mimic the loss of the kinase nor does complete truncation of the Shs1p C-terminus. Surprisingly, however, loss of the C-terminus or the predicted coiled-coil domain of Shs1p generates expanded zones of septin assemblies and ectopic septin fibers, as well as aberrant cell morphology. The expanded structures form coincident with ring assembly and are heteromeric. Interestingly, while septin recruitment to convex membranes is increased, septin localization is diminished at concave membranes in these mutants. Additionally, the loss of the coiled-coil leads to increased mobility of Shs1p. These data indicate the coiled-coil of Shs1p is an important negative regulator of septin ring size and mobility, and its absence may make septin assembly sensitive to local membrane curvature.

INTRODUCTION

The septins are a class of conserved eukaryotic cytoskeletal proteins that assemble into heteromeric filaments and higher-order structures in vivo (Beise and Trimble, 2011; Oh and Bi, 2011; Mostowy and Cossart, 2012). They were first discovered in budding yeast to be required for cytokinesis (Hartwell, 1971) and have since been shown to participate in many cellular functions, including exocytosis, apoptosis, and membrane compartmentalization (Barral et al., 2000; Takizawa et al., 2000; Luedeke et al., 2005; Amin et al., 2008; McMurray et al., 2011b). Septins can also act as scaffolds whose purpose is to recruit signaling molecules to a specific place at a specific time (Gladfelter et al., 2001; Hagiwara et al., 2011). Misexpressed or dysfunctional septins in humans are associated with several neurological disorders, including Alzheimer's and Parkinson's diseases (Kinoshita et al., 1998; Ihara et al., 2007), and with many different types of cancer (Peterson and Petty, 2010).

The formation of septins into heteromeric complexes and higher-order structures is tightly linked to septin function (McMurray et al., 2011a). Recombinant or purified septins can self-assemble in vitro into nonpolar heteromeric complexes that spontaneously polymerize into long, frequently paired, filaments under low-salt conditions (Field et al., 1996; Frazier et al., 1998). Under high-salt conditions that inhibit the polymerization of recombinant septin complexes into filaments in solution, the presence of phosphatidylinositol-4,5-bisphosphate (PIP₂) membranes promotes polymerization and assembly of crossed networks of filaments (Bertin et al., 2010), indicating a role for membranes in septin filament formation. In vivo, septins organize into various higher-order structures, such as rings, bars, and gauzes (Kinoshita et al., 2002; Rodal et al., 2005; DeMay et al., 2011). These higher-order structures are often associated with the cell cortex, and their formation at the correct time and place is required for normal septin function. Many aspects of septin assembly in the cell remain mysterious: How do septin filaments assemble into higher-order structures? What directs the size and shape of the structures? How does the local geometry of the cell cortex influence the formation and persistence of septin assemblies?

It has been shown that the septins interact with one another and membranes via a series of conserved domains. In general, the septins are made up of a variable N-terminal region with a polybasic domain, which may contribute to septin–membrane interactions; a central GTP-binding domain; and a variable C-terminal extension that sometimes includes a predicted coiled-coil domain (Casamayor and Snyder, 2003; Beise and Trimble, 2011). The crystal structure of the mammalian Sept2, 6, 7 complex revealed that the septins interact with one another via an interface encompassing the G domain and an interface comprising the N- and C-termini (Sirajuddin et al., 2007). GTP binding contributes to septin–septin interactions and complex formation (Nagaraj et al., 2008; Bertin et al., 2010). The order of septin subunits in complexes and filaments is determined by which interface a given septin subunit associates with another septin subunit. The ability to oligomerize and form filaments via these interfaces is necessary to form higher-order septin structures and has been shown to be essential for function in budding yeast (McMurray et al., 2011a).

Based on electron microscopy data showing paired septin filaments with a periodic electron density between the pairs, akin to railroad tracks, it has been suggested that coiled-coil domains of the septins may be playing a role in linking filaments (Bertin et al., 2008). However, recombinantly expressed yeast septin complexes lacking their coiled-coil domains were shown to be capable of paired filament formation in vitro on PIP₂-containing membranes but unable to form extended lateral or crossing associations to make septin meshes (Bertin et al., 2010). Because different septin subunits neighbor one another in the yeast septin complex, heteromeric coiled-coil interactions could help stabilize formation of septin protofilaments in the absence of membranes (Bertin et al., 2008, 2010; de Almeida Marques et al., 2012). It is still unknown where the coiled-coils are oriented relative to septin filaments or with respect to membrane surfaces. Thus, the precise role of coiled-coil domains in septin organization and the mechanisms by which the oligomers and filaments assemble into higher-order structures in cells remain unclear.

In this study, we are using Ashbya gossypii as a model for studying septin higher-order structure formation (Dietrich et al., 2004). Ashbya has three distinct septin structures: a diffuse cloud at the tips of growing hyphae, cortically associated rings at the interregion (IR) of the cell, and rings around the cortex of branch sites (DeMay et al., 2009). Unlike septin rings that form at the mother-bud neck of yeast, the septin rings that form at the Ashbya IR and branches are composed of discrete bars of septin filaments, persist over many hours, and are not influenced by the cell cycle (DeMay et al., 2009). Like yeast, Ashbya expresses five septins during vegetative growth, Cdc3p, Cdc10p, Cdc11p, Cdc12p, and Shs1p (Sep7p), and all are essential for ring formation (DeMay et al., 2009). In yeast, however, Shs1p is dispensable for septin ring assembly, whereas the other septins are essential for ring formation, dependent somewhat on strain background (Iwase et al., 2007; Garcia et al., 2011; McMurray et al., 2011a).

Several studies have shown that posttranslational modifications to the septins, including phosphorylation, sumoylation, and acetylation, play a role in normal septin organization (Johnson and Blobel, 1999; Mortensen et al., 2002; Tang and Reed, 2002; Sinha et al., 2007; Egelhofer et al., 2008; Mitchell et al., 2011). We previously identified a pair of kinases, Elm1p and Gin4p, that are required for assembly of septin rings in the filamentous fungus A. gossypii (DeMay et al., 2009). In the absence of either of these kinases, septins localizing as a diffuse cloud to the tips of growing hyphae cannot assemble into cortically attached rings. We show that the septin subunit Shs1p has multiple phosphorylation sites in its C-terminus in proximity to the predicted coiled-coil domain. This makes the Shs1p C-terminus an appealing potential regulatory region for septin filament and higher-order structure formation. Data presented in this study demonstrate that the C-terminal coiled-coil domain of Shs1p, but surprisingly not phosphorylation, is required to limit septin ring size and mobility within higher-order assemblies.

RESULTS

The C-terminus of Shs1p is a candidate regulatory region

We hypothesized that phosphorylation of septins may regulate septin assembly because Ashbya cells lacking the Elm1p or Gin4p kinases have defects building higher-order rings (DeMay et al., 2009). Immunoprecipitation of the septin Shs1p indicated this septin migrates at a higher than predicted molecular weight (∼75–85 kDa, as opposed to the 65 kDa predicted by primary sequence) as monitored by SDS–PAGE (Figure 1A). This mobility shift is eliminated when immunoprecipitated Shs1p is treated with phosphatase (Figure 1A). The high-molecular-weight Shs1–green fluorescent protein (Shs1-GFP; migrating at 100–110 kDa) associates with the other four septins, Cdc3p, Cdc11p, Cdc12p, and Cdc10p (Figure 1B), indicating Ashbya septins form heteromeric complexes composed of five different septin subunits. In vitro, these purified complexes form protofilaments that are on average 32.8 ± 4.3 nm (n = 529), suggesting Ashbya septin protofilaments, like those purified from yeast (Frazier et al., 1998), are generally octameric, given an individual subunit length of approximately 4 nm (Figure 1, C, left, and D). Interestingly, yeast protofilament lengths were measured for complexes composed of only four of the septin subunits (Frazier et al., 1998), yet the purified Ashbya complex represents all five septin subunits. On incubation in a low-salt solution, the septin complexes were able to polymerize into paired filaments (Figure 1C, right).

FIGURE 1:

The Shs1p C-terminus is a candidate regulatory region. (A) Shs1p is phosphorylated. Cdc11-TAP was immunoprecipitated from Ashbya lysates (AG495.1) to pull down the entire septin complex. The immunoprecipitate was treated with λ-phosphatase, phosphatase buffer only, or phosphatase buffer with phosphatase inhibitors. Proteins were separated on SDS–PAGE, transferred to PVDF membrane, and probed with anti-Shs1 antibody. (B) Purification of the septin complex. Shs1-GFP was immunoprecipitated from Ashbya lysates (AG124) using the GFP-Trap-M system. Proteins were separated on SDS–PAGE and stained with colloidal-blue stain. (C) Ashbya septins form filaments in vitro. Cdc11-TAP was immunoprecipitated from Ashbya (AG495.1) to precipitate the entire complex. The complex was separated from protein A–Sepharose beads using TEV protease. Protein complexes were transferred to carbon grids and negatively stained with uranyl acetate. Boxes have been enlarged to show individual complexes in 1M KCl and filaments in 75 mM KCl solutions. Scale bar: 100 nm. (D) Distribution of septin complex lengths in 1 M KCl solutions. Mean complex length is 32.8 ± 4.3 nm, consistent with octameric complexes. n = 529 complexes. (E) Schematics for full-length Shs1p and the C-terminal truncation mutant, Shs1ΔC.

For determination of which sites are modified on Shs1p, both bands corresponding to the size of Shs1-GFP were cut out and submitted for mass spectrometry analysis. We found that Shs1-GFP is phosphorylated at nine sites (Ser-359, Ser-362, Ser-380, Thr-394, Ser-396, Ser-408, Ser-431, Thr-555, Ser-558) in the C-terminus of the protein, but is neither acetylated nor sumoylated. These phosphorylation sites lie in proximity to, but not within, the predicted coiled-coil domain (Figure 1E). Thus, the C-terminus, encompassing the phosphorylation sites and the coiled-coil domain, is an attractive candidate region for examining the regulation of Shs1p and septin ring assembly.

Loss of the Shs1p C-terminus leads to aberrant, temperature-sensitive septin assemblies

For evaluation of the role of the C-terminus of Shs1p in higher-order septin structure and function, SHS1 was truncated at the first phosphorylation site (Ser-359, as indicated in the Shs1ΔC schematic, Figure 1E), tagged with GFP, and expressed as the only copy of the allele at the endogenous locus of SHS1. Surprisingly, loss of the Shs1p C-terminus (residues 359–580), encompassing all of the phosphorylation sites as well as the coiled-coil domain, did not lead to a loss of septin rings. Septin rings are still able to form at branch sites and throughout the IR of the cell (Figure 2, A and B). However, branch rings with Shs1ΔC-GFP are larger and are made of whisker-like, elongated fibers that follow the convex cell cortex. These cortical structures can be seen in cross-section as individual, expanded bars encircling the bases of branches and continuing along the main hypha (Figure 2B, bottom panels). This mutant thus extends the size of septin structures specifically associated with the curved membrane of lateral branches. Ectopic fibers of Shs1ΔC-GFP are also found at the cortex throughout the IR (Figure 2, A and B). Shs1ΔC-GFP is expressed at levels comparable with those of Shs1-GFP, indicating that enhanced levels of Shs1ΔC-GFP are not leading to the increased zones of assembly and additional septin fibers we see in the mutant strain (Figure 2C). Interestingly, these fibers do not form if Shs1ΔC-GFP is expressed in a strain with a wild-type Shs1p, indicating this allele is recessive (Supplemental Figure S1).

FIGURE 2:

Loss of Shs1p C-terminus leads to an increased zone of septin ring assembly. (A) Loss of the Shs1p C-terminus leads to expanded branch rings and ectopic fiber formation throughout the IR. Cells expressing Shs1-GFP (AG124) and Shs1ΔC-GFP (AG539.2) were visualized by wide-field fluorescence microscopy. IR rings are indicated by white arrows and branch rings are indicated by white arrowheads. Ectopic fibers in the IR are indicated by a red arrow. Scale bar: 10 μm. (B) Zoomed-in view of IR and branch ring structures in Shs1-GFP compared with Shs1ΔC-GFP cells. Cell outlines are represented by dotted white lines. Scale bar: 2.5 μm. (C) Expression for Shs1ΔC is comparable with wild-type levels. Ashbya lysates for the indicated strains were separated on SDS–PAGE (10% acrylamide) and probed with anti-GFP antibody. Tubulin was used as a loading control. (D) Expanded rings are temperature sensitive. Shs1ΔC-GFP was grown overnight at 30°C or 37°C. Cells were imaged live by wide-field fluorescence microscopy. Scale bar: 10 μm. (E) Shs1ΔC-GFP is stably expressed at 37°C. Ashbya lysates for the indicated strains grown at 37°C were separated on SDS–PAGE (10% acrylamide) and probed with anti-GFP and anti–α-tubulin antibodies.

We tested the stability of expanded branch rings and ectopic filaments in the Shs1ΔC mutant by growing cells at 37°C overnight. At this temperature, septin rings are normally able to form and persist. However, as shown in Figure 2D, for cells expressing the Shs1p C-terminal truncation, there is a decrease in the frequency of enlarged branch rings (64% at 30°C compared with 42% at 37°C, n > 20 rings each). This indicates that the expanded structures are defective in assembly or are relatively unstable compared with the normal rings at high temperature, although overall expression levels are unchanged for Shs1ΔC-GFP compared with Shs1-GFP (Figure 2E). This is especially interesting, because even though Shs1ΔC-GFP is incorporating into the branch and IR rings in the mutant strains, these “normal”-sized rings do not appear to be affected by the increase in temperature; only the expansion zones and elongated fibers appear to be affected. It is possible that the increased temperature alters membrane composition such that the zone of septin recruitment to the cortex is more limited at high temperature.

Loss of the Shs1p C-terminus impacts cell morphology

To determine whether there were functional consequences of the expanded septin structures, we assessed cell morphology in shs1ΔC mutants. Introduction of Shs1ΔC-GFP led to a slight, but statistically significant, defect in colony growth (Figure 3A; p < 0.05), but did not affect ascospore production, a process that requires septins. Looking at individual cells, it was apparent that the loss of the Shs1p C-terminus had an effect on overall cell morphology (Figure 3B). The hyphae of this mutant appear swollen, possessing average hyphal widths 0.5-μm wider than those seen in wild-type cells (Figure 3C, left; n ≥ 196 hyphae each; p < 0.0001). Additionally, Shs1ΔC cells form branches more frequently, with an average interbranch distance of 11.0 ± 7.2 μm, compared with 15.5 ± 10.9 μm average distance between branches for the wild-type (Figure 3C, right; n > 300 distances each strain; p < 0.0001).

FIGURE 3:

Loss of the Shs1p C-terminus causes cell morphological defects. (A) Shs1ΔC has a slight, but statistically significant, growth defect. Shs1-GFP (AG124) and Shs1ΔC-GFP (AG539.2) cells were grown on AFM G418 agar plates for 10 d (n = 3 each strain). Plates were imaged every 24 h, and colony growth was measured in area. Average slope for Shs1ΔC is 458.9 mm²/d vs. 528.7 mm²/d for Shs1-GFP (p < 0.05, determined with a two-tailed t test). (B) Shs1 C-terminal deletion leads to cell morphological defects. Phase-contrast images of Shs1-GFP and Shs1ΔC-GFP cells grown for 12 h are shown. Scale bar: 10 μm. (C) Hyphal width and interbranch distance are altered in the Shs1ΔC strain. Hyphal widths were measured for each hypha in cells grown for 14 h (left). n ≥196 hyphae each (p < 0.0001, two-tailed t test). The length between branches was measured for cells cultured for 14 h (right). n > 300 interbranch distances for each strain (p < 0.0001, two-tailed t test). (D) Actin polarization is not altered in Shs1ΔC. Shs1-GFP and Shs1ΔC-GFP cells were grown overnight, fixed, and stained for actin. Actin patches polarized to several tips for each strain are shown. Scale bar: 10 μm. (E) Septin localization to the tips is defective in Shs1ΔC cells. Zoomed-in view of septin localization to the tips of cells expressing Shs1-GFP and Shs1ΔC-GFP. Scale bar: 2.5 μm.

These cell morphological defects are not a result of an alteration in the actin cytoskeleton. Actin patches are heavily concentrated at the same percentage of hyphal tips (86% for both strains) with a comparable relative mean fluorescence intensity for a given volume for both the wild-type (57.8 ± 18.0 a.u., n = 99) and the mutant (56.1 ± 17 a.u., n = 76, Figure 3D). This suggests that while there is an increase in branch frequency, there is no deficiency in actin polarity. It is possible that the hyphal width defect arises in the mutant cells, because the concentration of Shs1p at the tips is substantially decreased (Figure 3E, quantified in Figure 6C later in the paper, n > 100 tips each; p < 0.0001). Overall, the C-terminus of Shs1p is required for normal recruitment of septin to hyphal tips, but not to septin rings, and leads to larger rings at branches and aberrant cell morphology.

FIGURE 6:

The coiled-coil of Shs1p, but not phosphorylation, is necessary for normal septin morphology. (A) Schematics for Shs1p mutants. (B) Deletion of the Shs1p coiled-coil produces expanded branch rings. Shs1-GFP (AG124), Shs1-GFP 9A (AG643.1), Shs1ΔC-GFP (AG539.2), and Shs1ΔCC-GFP (AG623) were grown for 14 h at 30°C and imaged live. Scale bar: 5 μm. (C) Quantification of septin fluorescence above cytosolic background levels at hyphal tips. Shs1-GFP, Shs1ΔC-GFP, Shs1ΔCC-GFP, and Shs1-GFP 9A cells were imaged on a wide-field fluorescence microscope, and images were contrasted to the same levels. Fluorescence intensities for a given volume of individual tips were measured using Volocity (n ≥ 98 each). Cytosolic background was subtracted, and values for tips with a relative mean fluorescence above background are shown. Statistical significance was determined using a two-tailed t test, *p < 0.0001. Error bars represent SE. (D) Shs1ΔCC-GFP is expressed at levels comparable with those of Shs1-GFP. Ashbya lysates for the indicated strains grown at 30°C were separated on SDS–PAGE (10% acrylamide) and probed with anti-GFP and anti–α-tubulin antibodies.

Expansion of septin rings occurs at ring assembly in a kinase-independent manner

The size of a septin ring in normal Ashbya cells is determined at the birth of the ring and remains constant throughout the lifetime of the structure until septation is initiated many hours after ring assembly (DeMay et al., 2009). The expanded septin rings and ectopic fibers seen in Shs1ΔC-GFP cells could arise during ring assembly such that newly assembled rings are abnormally large, which would suggest a defect in the initial scaling of the ring. Alternatively, the expanded zone of the septin ring could result from continued septin polymerization after the initial assembly, which would suggest a defect in limiting the window of assembly in both time and space. For determination of when the expanded septin rings form, Shs1ΔC-GFP–expressing cells were captured by time-lapse fluorescence microscopy. Frames from a representative movie (Supplemental Movie) shown in Figure 4A illustrate that the expanded zone of septin is present at the birth of the ring (arrow) and that the expanded rings persist over time but do not continue lengthening after the ring is initially built (arrowhead). Furthermore, additional fibers do not appear to form after establishment of the ring. Thus the C-terminus of Shs1p is required for setting the initial size of septin rings at the time of assembly.

FIGURE 4:

Septin ring assembly timing and FCF-induced fiber formation are unaffected by the loss of the Shs1 C-terminus. (A) Expanded branch ring size is established at assembly. Shs1ΔC-GFP (AG539.2) cells were imaged every 10 min for 1 h. A preexisting expanded branch ring is indicated by arrowheads. An emerging ring is indicated by arrows. Scale bar: 2.5 μm. (B) Polymerization of Shs1p by the small molecule FCF is unaltered in Shs1ΔC strain. Shs1-GFP and Shs1ΔC-GFP cells were grown 14 h at 30°C and were treated with various concentrations of FCF. Cells were sampled every 30 min for 2 h. Representative images after 2 h in the indicated concentration of FCF are shown. Scale bar: 10 μm. (C) Ectopic fiber formation does not require the Elm1p kinase. elm1Δ cells expressing Shs1-GFP (AG120) or Shs1ΔC-GFP (AG632.1) were imaged live. Ectopic fibers at the IR are indicated with an arrow, and an expanded branch ring is indicated with an arrowhead. Scale bar: 5 μm.

It is not known what determines the length of a septin ring during assembly and what restricts its expansion along the cortex after that length has been reached. One possibility is that assembly is limited to a finite time period, and the final size is determined by the rate of septin polymerization in that time. If so, Shs1ΔC rings potentially are larger as there is simply more polymerization per unit time. To investigate whether Shs1ΔC promotes increased septin polymerization, we assessed whether these mutants were more sensitive to the septin-polymerizing small molecule forchlorfenuron (FCF). Treatment of Ashbya cells with FCF leads to extensive polymerization of septin-based fibers throughout the hyphae (DeMay et al., 2010). We hypothesized that if Shs1ΔC-GFP promotes more rapid septin fiber formation, FCF-induced septin fibers would arise more quickly and/or at lower FCF concentrations. Cells were treated with 10, 25, 62.5, and 125 μM FCF for 30, 60, 90, and 120 min. No gross differences for the induction of septin fibers between wild-type and Shs1ΔC were apparent at any time or concentration of FCF (Figure 4B). This result suggests loss of the Shs1p C-terminus does not increase septin polymerization, or at least not in a way that is synergistic with FCF's mode of action.

We previously showed that the Elm1p and Gin4p kinases are required to assemble septin rings throughout the IR, although branch rings still form normally in those mutants (DeMay et al., 2009). Despite the loss of all Shs1p phosphorylation sites in the Shs1ΔC mutant, septin rings are built normally throughout the IR of the cell, but with the addition of ectopic fibers in the IR that are not a component of a ring. We sought to test the role of the kinases in assembling expanded septin rings and ectopic fibers in the Shs1ΔC-GFP mutant by deleting the ELM1 allele in the shs1ΔC background. Without Elm1p, IR rings are unable to form for both wild-type and Shs1ΔC cells. However, the expanded branch rings and ectopic fibers in the IR in Shs1ΔC-GFP still form, despite the loss of the kinase (Figure 4C). This indicates Elm1p is not required for increasing the zone of septin structure formation.

Shs1ΔC-GFP interacts with other septin subunits

We hypothesized that loss of the C-terminus may be promoting homomeric Shs1 complex formation and that the expanded assemblies potentially are not composed of pure heteromeric septin filaments. Notably, however, Shs1ΔC-GFP can coimmunoprecipitate Cdc11p comparable with wild-type Shs1p (Figure 5A) and Shs1ΔC-GFP interacts with all four of the other septin subunits in a stoichiometric manner (Figure 5B). Additionally, we found that Cdc11 localizes to the elongated fibers both at the branch rings and throughout the IR, as detected by immunofluorescence, further indicating Shs1ΔC-GFP is not the only constituent of these expanded rings and aberrant filaments (Figure 5C). Thus Shs1ΔC-GFP is still able to efficiently interact with the other septins, indicating the C-terminus of Shs1 is dispensable for complex formation and excess septin structures seen in the mutant are not aberrant homomeric fibers. Intriguingly, Cdc11p is not concentrated at hyphal tips in the Shs1ΔC-GFP background (Figure 5D), suggesting that when Shs1p is not localizing normally to the tips, the other septins cannot concentrate there either. Therefore Shs1p oligomerizes with the other septins in this mutant, and deletion of the Shs1p C-terminus diminishes the ability of other septins to localize to the hyphal tips.

FIGURE 5:

Shs1ΔC interacts with other septins. (A) Shs1ΔC-GFP interacts with Cdc11p. Shs1-GFP (AG124) and Shs1ΔC-GFP (AG539.2) were immunoprecipitated from Ashbya lysates using anti-GFP antibody. Immunoprecipitates were separated by SDS–PAGE, transferred to PVDF membrane, and probed with anti-GFP or anti-Cdc11 antibodies. (B) Shs1ΔC-GFP complexes with the other septins. Shs1-GFP and Shs1ΔC-GFP were immunoprecipitated from Ashbya lysates using anti-GFP antibody. Immunoprecipitates were separated by SDS–PAGE and silver-stained to assess coprecipitating proteins. (C) Shs1-GFP and Shs1ΔC-GFP cells were fixed with formaldehyde, spheroplasted with zymolyase, and probed with anti-Cdc11 antibody. An ectopic fiber in the IR is indicated with an arrowhead, and an expanded branch ring is indicated with an arrow. Scale bar: 5 μm. (D) Cdc11 localization is diminished at the tips in Shs1ΔC. Scale bar: 5 μm.

The coiled-coil of Shs1p is necessary for normal septin morphology

Because the C-terminus of Shs1p encompasses nine phosphorylation sites and a coiled-coil domain, either the phosphorylation sites or the coiled-coil domain could be responsible for regulating the formation of properly sized septin rings. Mutants expressing a nonphosphorylatable or phosphomimetic form of Shs1p (Shs1-GFP 9A or 9D, respectively) or Shs1p with the coiled-coil deleted (Shs1ΔCC-GFP, residues 459–508 deleted) were assessed for their ability to form septin rings (Figure 6A). The loss of the phosphorylation sites had no effect on the appearance of any of the septin rings (Figure 6B). The phosphomimetic Shs1-GFP 9D was dominant and lethal, so its effect on higher-order septin structure could not be assessed. Intriguingly, the loss of the Shs1p coiled-coil phenocopied the C-terminal truncation (Figure 6B). Shs1ΔCC-GFP localizes to a broader zone at branch rings and to ectopic fibers throughout the IR. We also assessed the mean fluorescence of Shs1p localizing to tips in these strains and found that, similar to Shs1ΔC-GFP, Shs1ΔCC-GFP is diminished at tips; however Shs1-GFP 9A recruitment to the tips is not significantly different from wild-type (Figure 6C, n ≥ 98 for each strain). As for Shs1ΔC-GFP, the expanded branch rings and depletion of septins at tips are not a result of different expression of Shs1ΔCC-GFP compared with wild-type (Figure 6D). These data indicate the coiled-coil of Shs1p is playing an important role in septin ring geometry and scaling in Ashbya.

Shs1ΔCC-GFP alters septin mobility and concentration in rings

Septin rings are assembled at a fixed size in Ashbya but are able to incorporate new protein through time and are dynamic (DeMay et al., 2009, 2010). It is not known whether septin dynamics contribute to scaling of septin rings in any system. Therefore we were interested to evaluate how loss of the Shs1p coiled-coil and/or phosphorylation may impact either the amount or rate of septin incorporation in rings or the stability of septin-based structures. To assess the relative amount of Shs1p incorporating into septin structures at steady state, we measured the fluorescence intensity of a region of fixed size for IR and branch rings, in Shs1ΔC-GFP, Shs1ΔCC-GFP, and Shs1-GFP 9A compared with Shs1-GFP (IR ring: n ≥ 80 each strain; branch ring: n ≥ 48 each strain). Mean fluorescence intensity at IR rings was significantly decreased for Shs1ΔCC-GFP and significantly increased for Shs1-GFP 9A, compared with wild-type (Figure 7A; p < 0.05). This suggests the coiled-coil and phosphorylation of Shs1 are performing opposing roles in concentrating septin at the IR. The mean fluorescence intensities for branch rings in all mutant strains were comparable with wild-type (Figure 7B). Thus, despite the increase in the zone of septin assembled at branches, the concentration of septin in a given area of either ring type is unchanged. This indicates septins are recruited to a larger region at branch sites, but the amount of septin that gets recruited per unit area is the same in wild-type and Shs1 mutant cells.

FIGURE 7:

Shs1ΔC-GFP alters higher-order structure dynamics. (A) Concentration of Shs1ΔCC in IR rings is increased and Shs1 9A is decreased compared with wild-type levels. Shs1-GFP (AG124), Shs1ΔCC-GFP (AG623), Shs1-GFP 9A (AG643.1), and Shs1ΔC-GFP (AG539.2) cells were imaged on a wide-field fluorescence microscope, and images were contrasted to the same levels. Fluorescence intensities for a given volume of individual rings were measured (n ≥ 80 for each strain). Cytoplasmic background was subtracted. Error bars represent SE, *p < 0.05. (B) Concentration of Shs1 is unaltered in branch rings, regardless of mutations. Images and measurements were performed as described for A (n ≥ 48 for each strain). (C) Shs1ΔCC-GFP is more mobile than wild-type Shs1-GFP in the IR rings. Shs1-GFP, Shs1ΔCC-GFP, Shs1-GFP 9A, and Shs1ΔC-GFP were imaged using confocal spinning-disk microscopy. Individual rings in each strain were bleached and then imaged every 30 s for 5 min (n ≥ 7 rings per type). Fluorescence intensity of the bleached region was recorded at each time point. Intensities were corrected for cytoplasmic background and photobleaching. Error bars on the recovery curves represent SE. Statistical significance for mobile fraction and t_1/2 were determined by two-tailed t tests (Table 1). (D) Cdc11-mCherry dynamics are unchanged despite the loss of the Shs1 coiled-coil. FRAP was performed on Cdc11-mCherry in either a wild-type Shs1 (AG208) or Shs1ΔCC (AG664.1) background, as described for C (n = 9 IR rings for each strain). Summary statistics are presented in Table 2.

The coiled-coil and/or phosphorylation of Shs1p could be responsible for regulating septin ring dynamics. It is easy to imagine that the melting and annealing of coiled-coil structures may be rate-limiting for dynamics or that phosphorylation induces structural changes that influence septin turnover. Potentially, altered dynamics could contribute to defects in ring scaling or changes in septin concentration. Entire septin rings in Shs1ΔC-GFP, Shs1ΔCC-GFP, and Shs1-GFP 9A cells were photobleached and fluorescence recovery was followed through time to test this (n ≥ 7 rings for both ring types in all strains). No statistically significant change in mobile fraction or t_1/2 was observed for any of the mutant branch rings compared with wild-type, although there was a trend toward somewhat higher mobility in the coiled-coil mutants at branch rings. At the IR rings, Shs1-GFP 9A showed dynamics and mobile fraction similar to those of wild-type. However, the mobile fraction of Shs1ΔCC-GFP significantly increased (Table 1 and Figure 7C; p < 0.05). The relative t_1/2 of Shs1ΔCC-GFP recovery at IR rings is comparable with the wild-type t_1/2. However, the nearly doubled mobile fraction and this similar t_1/2 indicate that the absence of the coiled-coil increases the amount of Shs1p that can turn over and increases the actual rate of recovery by about twofold. This suggests the Shs1p coiled-coil normally limits septin dynamics, while phosphorylation has no significant effect on Shs1p dynamics in either ring type.

TABLE 1:

Shs1 FRAP summary statistics.

Ring type	n	Mobile fraction ± SD (%)	t1/2 ± SD (s)
Shs1-GFP IR	14	32.2 ± 13.9	43.8 ± 10.4
Shs1ΔCC-GFP IR	8	60.9 ± 10.0a	41.4 ± 6.7
Shs1-GFP 9A IR	10	23.5 ± 6.9	51.6 ± 16.3
Shs1ΔC-GFP IR	7	47.5 ± 18.3b	31.8 ± 10.9
Shs1-GFP branch	20	45.5 ± 19.1	39.4 ± 13.9
Shs1ΔCC-GFP branch	8	57.4 ± 24.4	41.3 ± 10.4
Shs1-GFP 9A branch	10	39.4 ± 16.6	46.8 ± 8.1
Shs1ΔC-GFP branch	7	56.2 ± 27.5	42.1 ± 8.1

^ap < 0.05 (analysis of variance).

^bp < 0.1.

Shs1p may be turning over at the ring as a monomer or in complex with the other members of a heteromeric septin protofilaments. We hypothesized that the loss of the Shs1p coiled-coil may also increase the dynamics of other septins at the IR rings. Cdc11-mCherry was expressed in wild-type and Shs1ΔCC strains, and fluorescence recovery after photobleaching (FRAP) was monitored for Cdc11-mCherry at the IR rings to test this (Table 2 and Figure 7D). Regardless of the whether the Shs1p coiled-coil was present, there was no change in mobile fraction or t_1/2 of Cdc11-mCherry. These results indicate that the increased dynamics of Shs1ΔCC-GFP do not affect those of Cdc11p and provide evidence that the turnover of individual septin subunits may be regulated independently of other septins.

TABLE 2:

Cdc11-mCherry IR ring FRAP summary statistics.

Background	n	Mobile fraction ± SD (%)	t1/2 ± SD (s)
Shs1-GFP	9	44.7 ± 15.6	40.9 ± 12.0
Shs1ΔCC-GFP	9	40.2 ± 23.1	41.9 ± 14.0

No statistically significant differences were observed (two-tailed t test).

Ag Shs1ΔCC alters yeast septin ring and cell morphology

To test whether the Shs1p coiled-coil has a conserved function in regulating the geometry and scale of septin rings, we heterologously expressed Ashbya gossypii Shs1ΔC-GFP (Ag Shs1ΔC-GFP) in the budding yeast Saccharomyces cerevisiae. When expressed on a plasmid as the only copy of Shs1p (in an shs1Δ), Ag Shs1ΔC-GFP altered both cell and septin ring morphology in yeast (Figure 8A). The majority of cells expressing Ag Shs1ΔC-GFP as the sole form of Shs1p are elongated and have asymmetric septin rings. Although these yeast cells do not form ectopic fibers as in Ashbya, it is evident that Shs1ΔC has a striking effect on cell and septin ring shape. As in Ashbya, the mutant phenotype caused by Shs1ΔC is recessive, such that expressing Ag Shs1ΔC-GFP in a strain with the wild-type S. cerevisiae SHS1 (Sc SHS1) allele produces cells with normal shape and septin rings (Figure 8B). The loss of the coiled-coil domain alone is sufficient to produce the aberrant cell and septin phenotypes seen in the Shs1p C-terminal deletion mutant. Expression of Ag Shs1ΔCC-GFP as the sole copy of Shs1p in yeast phenocopied Shs1ΔC and led to elongated cells and asymmetrical septin rings (Figure 8C). Deletion of the Sc Shs1 C-terminus (residues 470–551), including its predicted coiled-coil, displayed no apparent cell or septin defects (Figure S2), indicating the septin and cell size regulatory roles performed by the Ashbya Shs1p coiled-coil may not be shared by S. cerevisiae Shs1p.

FIGURE 8:

Ashbya Shs1ΔCC-GFP expression leads to aberrant cell and septin ring morphology in yeast. (A) Ashbya Shs1ΔC expressed as the only copy of Shs1p in the cell leads to yeast cell and septin defects. S. cerevisiae shs1Δ cells heterologously expressing plasmidic A. gossypii Shs1-GFP or Shs1ΔC-GFP were grown to log phase and imaged live. Arrowheads indicate misshapen septin rings at the mother-bud neck. Scale bar: 5 μm. (B) Shs1ΔC cell and septin morphology defects are recessive in yeast. S. cerevisiae cells with a wild-type copy of SHS1 heterologously expressing plasmidic A. gossypii Shs1-GFP or Shs1ΔC-GFP were grown to log phase and imaged live. Scale bar: 5 μm. (C) Ashbya Shs1ΔCC-GFP expressed as the only copy of Shs1p in yeast produces defective cell and septin morphology in yeast. S. cerevisiae cells expressing A. gossypii Shs1-GFP, Shs1ΔC-GFP, or Shs1ΔCC-GFP on a plasmid were grown to log phase and imaged live. Scale bar: 5 μm.

DISCUSSION

The roles for the Ashbya Shs1p coiled-coil are summarized in Figure 9. In wild-type cells, septins localize to a cloud at the tip of growing hyphae and to rings of a fixed size at branches and in the IR. The deletion of the Shs1p coiled-coil leads to a loss of tip septin recruitment, increases the zone branch rings occupy, and causes ectopic fiber formation in the IR. Additionally, Shs1p lacking its coiled-coil is more mobile. Overall, the data we present in this paper are consistent with a role for the coiled-coil domain of Shs1p in limiting septin ring size and dynamics.

FIGURE 9:

Summary of limitation of septin ring size and mobility by the Shs1p coiled-coil. Loss of the Shs1p coiled-coil leads to wider hyphae, enlarged septin rings at branches, ectopic fibers throughout the IR, and increased Shs1p mobility at the IR rings.

These results were surprising, considering that the coiled-coil is generally thought to enhance septin–septin interactions and facilitate oligomerization and filament formation. It is possible that the role in limiting size and dynamics of septin rings is particular to the coiled-coil of Shs1p. Of the five septins in Ashbya and S. cerevisiae, Shs1p remains the most mysterious. It is nonstoichiometric in both organisms, although it is substoichiometric in yeast and superstoichiometric in Ashbya based on fluorescence intensities at rings (Mortensen et al., 2002; DeMay et al., 2009). It has been proposed that Shs1p is an accessory septin that can substitute for Cdc11p at the terminal position in the hetero-octamer (Garcia et al., 2011). In vitro complexes formed by purified septins from Ashbya are hetero-octamers, suggesting that only four of the five septins are present in any given heteromer. Thus, in Ashbya, the complexes that are immunoprecipitated from lysates are either polar protofilaments with Cdc11p and Shs1p at opposite ends or are nonpolar decamers in which the capping septin has a weaker association with the complex than the underlying octamer and can fall apart after the initial immunoprecipitation step, before visualization with electron microscopy. The salt concentration of the immunoprecipitation buffer is sufficiently high that it is unlikely a single Shs1 molecule is pulling out multiple protofilaments. The existence of polar protofilaments is an intriguing possibility, because it would suggest septin polymers are nonpolar when an even number of different subunits are available, but can potentially form polar filaments when given an odd number of different subunits. This has important ramifications for septin polymerization in mammalian cells, in which many different septins and isoforms are coexpressed.

When recombinant yeast septin complexes containing Shs1p in place of Cdc11p are incubated in a low-salt solution that normally promotes long, straight filament formation, the complexes self-assemble into rings (Garcia et al., 2011). This suggests Shs1p is capable of promoting curvature to septin assemblies. Given that septin higher-order complexes are found in highly curved areas, such as the yeast mother-bud neck, the branch sites of Ashbya, and the bases of dendritic spines in neurons, it makes sense that the septin complex would have intrinsic curvature. In vitro, the C-terminus of yeast Shs1p is required for the septin complexes to self-assemble into rings, but it is not required for oligomerization into complexes (Garcia et al., 2011). These special features of the S. cerevisiae Shs1p and the requirement for the C-terminus, which includes a predicted coiled-coil domain, suggest that Shs1p has the potential to play important roles in septin ring organization in the cell and may react to and/or generate membrane curvature.

Because the loss of the Shs1p coiled-coil domain in Ashbya leads to an expansion of septin branch rings and a loss of septin localization to the hyphal tips, it is tempting to speculate that membrane curvature has a role for directing septin higher-order structure organization. Shs1ΔCC-GFP localizes more readily to branch sites than to tips, which suggests that in the absence of the Shs1p coiled-coil, septin complexes or filaments may be more easily recruited to convex rather than concave membrane surfaces inside the cell. The coiled-coil may be required for capping filaments or influencing septin complex curvature, such that it limits how far septin rings can expand along the membrane. Shs1p coiled-coil interactions with the membrane may depend on membrane composition. We speculate that the branch sites of Ashbya may have a different membrane lipid or protein composition than the membrane at the hyphal tips. Wild-type Shs1p, in this case, would have a higher affinity for specific lipids in the membrane at the tips than the Shs1p coiled-coil deletion. The decrease of ectopic fibers and branch ring expansion in the coiled-coil mutant at high temperatures may be consistent with an alteration in membrane composition or fluidity.

The association of septins with the cell cortex is still a very poorly understood facet of septin organization. It is clear that septins can interact directly with static lipids in vitro and the polybasic domain may contribute to this interaction, but additional surfaces, and in particular Cdc10p, may be critical for membrane bilayer association (Casamayor and Snyder, 2003; Bertin et al., 2010). Additional in vitro data demonstrate that purified septins can self-assemble into filaments and orthogonal arrays when incubated with a PIP₂-1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) lipid monolayer and that arrangement into orthogonal arrays, but not filament formation, under these conditions relies on septin coiled-coil domains (Bertin et al., 2010). This suggests that formation of more elaborate septin assemblies are facilitated by membranes in a way that involves the septin coiled-coils, though it is not clear whether the coiled-coils can directly interact with the membranes. How septin interactions with the membrane are regulated in the cell remains an interesting open question.

The Ashbya Shs1p coiled-coil deletion leads to an increase in hyphal width and branching frequency, consistent with increased branching phenotypes seen in other filamentous fungal cells lacking coiled-coil domain-containing septins (Alvarez-Tabares and Perez-Martin, 2010; Lindsey et al., 2010; Hernandez-Rodriguez et al., 2012). This suggests that the mutant septin complexes themselves may play a role in promoting polarized growth throughout the hyphae, perhaps due to increased septin recruitment to the cortex at future branch sites. It is possible that the expanded area of septin localization can act as a scaffold to concentrate polarity machinery, leading to more hyphal branching.

We were intrigued to find that phosphorylation of Shs1p is not required for septin organization into rings, and in fact nonphosphorylatable Shs1-GFP 9A is significantly more concentrated at IR rings than wild-type Shs1-GFP. Localization to tips and branch rings, however, appears normal, and there is no change in Shs1p dynamics in this mutant. There are likely functional consequences of Shs1p phosphorylation, because mutating all Shs1p phosphorylation sites to the phosphomimetic residue aspartic acid was lethal. It has been established in S. cerevisiae and Candida albicans that the kinase Gin4p directly phosphorylates septins (Mortensen et al., 2002; Sinha et al., 2007), and loss of Gin4p activity leads to defects in septin ring formation and dynamics (Longtine et al., 1998; Wightman et al., 2004; Li et al., 2012). We previously showed that the Elm1p and Gin4p kinases are required for septin ring formation in Ashbya (DeMay et al., 2009). Because the cells with nonphosphorylatable Shs1p did not mimic the kinase mutant phenotype, the kinases must have other targets. It remains possible that Ashbya Shs1p is phosphorylated by Elm1p or Gin4p, but because additional targets—perhaps other septins—also get phosphorylated, simply the loss of Shs1p phosphorylation sites is not sufficient to cause a defect in ring assembly. Interestingly, in Shs1ΔC cells lacking Elm1p, ectopic fiber formation remains intact, although the normal IR rings are not able to assemble. This indicates that not all septin structure formation in the IR is reliant on Elm1p kinase activity. Again, we speculate that increased interaction of septins with the membrane in the mutant may be sufficient to overcome the need for phosphorylation.

The increased recovery after photobleaching of Shs1ΔCC compared with wild-type Shs1p indicates that in addition to limiting septin ring size, the Shs1p coiled-coil also limits the dynamics. The increased dynamics and increased ring size may be two distinct properties of the mutant protein or an indication that dynamics contribute to ring scaling. Approximately twice the amount of Shs1ΔCC can recover to IR rings in the same amount of time as its wild-type counterpart. It is possible, given the decreased steady-state concentration of Shs1ΔCC compared with wild-type, that individual bars comprising IR rings made of bundles of septins are thinner from membrane to cytoplasm, and the percent of septin exposed to the cytoplasm is therefore higher than it would normally be. This could result in the increased mobile fraction we observe for this mutant. Alternatively, the increased turnover could simply be a result of removing a dynamics-limiting property of the Shs1p coiled-coil. In C. albicans, it has been shown that Cdc10p, a septin that does not have a coiled-coil domain, is dynamic in hyphal septin rings, while other septins at the same location are not (Gonzalez-Novo et al., 2008), suggesting septin coiled-coil domains may inhibit dynamics in other systems. Similarly, we demonstrate that the loss of the Shs1p coiled-coil increases the dynamics of Shs1p, but this mutation does not affect the dynamics of Cdc11p, whose coiled-coil remains intact. It is possible that melting and annealing of the coiled-coils is rate-limiting for septin turnover or that the coiled-coil restricts incorporation into the rings by forcing the subunits to be in a certain conformation for entry to or exit from the polymer. Much still remains to be known about how septin dynamics are regulated in any cell type, but we show here that the septin coiled-coils have the potential to contribute to regulation of septin turnover at higher-order assemblies. It will be interesting to test whether this is specific to the Shs1p coiled-coil, or whether this is a conserved function among other septins.

The data we present in this paper demonstrate a novel role for septin coiled-coils in directing septin higher-order structure scaling and dynamics. How the coiled-coil is functioning specifically—whether it is modulating septin–septin interactions, septin–membrane interactions, or interactions of septins with other proteins—remains to be uncovered. This work is an important first step in understanding how septin ring size and dynamics may be controlled by individual septin proteins.

MATERIALS AND METHODS

Growth conditions and strain construction

A. gossypii media, culturing, and transformation protocols have been described previously (Ayad-Durieux et al., 2000; Wendland et al., 2000). Strains generated and used in this study can be found in Table 3. AG442.1 was obtained by transforming ΔlΔt cells with AGB088. To obtain AG495.1, we amplified TAP-GEN3 from AGB275 using AGO649/650, and the PCR product was used to directly tag CDC11A. The strain was verified using primer pairs AGO653/654 and AGO651/652. AG539.2 was constructed using direct integration of GFP-GEN3 (amplified from AGB005 using AGO996/997) immediately upstream of the bases encoding Ser-359 in SHS1. The same PCR product was used to create AGB412.1 via gap repair of AGB088 cut with AatII and BamHI HF. The entire open reading frame (ORF) of the plasmid was sequence-verified using AGO5, 557, and 593. AG647.1 was created by transforming ΔlΔt cells with the AGB412.1. AGB405 was constructed via overlap-extension PCR. Regions neighboring, but not including, the Shs1 coiled-coil were amplified using primer pairs AGO1061/1063 and AGO1062/5. The overlap-extension product was made using the products of the previous reactions and AGO1061/5. AGB88 was cut with AatII and BamHI HF, and this cut plasmid was gap-repaired via yeast cotransformation with the overlap-extension product. The final Shs1 coiled-coil deletion plasmid was confirmed by a test digest with ApaI and sequencing across the entire ORF with AGO5, 521, 557, 593, 684, and 1063. For integration into ΔlΔt, AGB405 was cut with ApaLI and transformed into cells. The strain (AG623) was verified with primer pairs AGO365/520 and AGO428/5. AG632.1 was created by transforming AG539.2 cells with PCR product amplified from AGB009 with AGO452/453. The deletion was verified with primer pairs AGO26/235, AGO234/27, and AGO28/27. A gene encoding SHS1 9A (the nine phosphorylation sites changed to nonphosphorylatable alanine) was synthesized by Genscript (Piscataway, NJ) and contained in the pUC57 vector (AGB358). SHS1 9A was cut from the vector using SacI and subcloned into pRS416, also digested with SacI (AGB404). GFP-GEN was amplified from AGB005 using AGO953/AGO775 and cotransformed in yeast with AGB404 to yield AGB406. AGB406 was verified by test digests with EagI and sequencing across the entire ORF with oligos AGO428, 520, 521, 557, 595, and 684. For integration into ΔlΔt, the plasmid was digested with SapI and transformed into cells. The resulting strain (AG643.1) was verified using primer pairs AGO365/520 and AGO557/684, as well as sequencing using AGO428. A similar cloning strategy was used for SHS1 9D. In short, a gene encoding SHS1 9D (the phosphorylation sites changed to phosphomimetic aspartic acid) was synthesized by Genscript in the pUC57 vector (AGB359). SHS1 9D was cut from pUC57 with SacI and subcloned into pRS416 (AGB373). GFP-GEN was amplified from AGB005 using AGO953/AGO776 and cotransformed in yeast with AGB373 to yield AGB378. The entire ORF was sequenced as for AGB406. Strains AG208 and AG664.1 were obtained by transforming AG124 and AG623, respectively, with pCdc11a-mCherry-Nat (AGB141).

TABLE 3:

Ashbya strains used in this study.

Strain	Relevant genotypea	Source
wild-type	Δleu2Δthr4	Altmann-Johl and Philippsen, 1996
AG120	Agelm1Δ::GEN3/AgSHS1-GFP-NAT1	DeMay et al., 2009
AG124	AgSHS1-GFP-GEN3	Helfer and Gladfelter, 2006
AG208	pAGB141 [pAg CDC11a-mCHERRY-NAT1]/AgSEP7-GFP-GEN3	This study
AG442.1	pAGB088 [pAgSHS1-GFP-Gen]	This study
AG495.1	AgCDC11a-TAP-GEN3	This study
AG539.2	AgSHS1ΔC-GFP-GEN3	This study
AG623	AgSHS1ΔCC-GFP-GEN3	This study
AG632.1	Agelm1Δnat/AgSHS1ΔC-GFP-GEN3	This study
AG643.1	AgSHS1-GFP-GEN3 9A	This study
AG647.1	pAGB412.1 [pAgSHS1ΔC-GFP-GEN3]	This study
AG664.1	pAGB141 [pAg CDC11a-mCHERRY-NAT1]/AgSHS1ΔCC-GFP-GEN3	This study

^aWith the exception of plasmidic strains (in brackets), all analyzed mycelia were homokaryotic (all nuclei have the same genotype).

S. cerevisiae was grown overnight in yeast–peptone–dextrose media, plus 200 μg/ml G418 for GEN3 selection, if applicable. Yeast strains generated and used in this study can be found in Table 4. Yeast lacking a copy of SHS1 were transformed with AGB088, AGB412.1, and AGB405 to yield AGY50, 51, and 52, respectively. Wild-type yeast (DHD5) were transformed with AGB088, AGB412.1, and AGB405 to yield AGY53, 54, and 55, respectively. The C-terminus of S. cerevisiae SHS1 was truncated with a GFP-GEN cassette amplified from AGB005 using AGO1129/1130. The strain was confirmed via fluorescence microscopy.

TABLE 4:

Yeast strains used in this study.

Strain	Relevant genotypea	Source
DHD5 (wt)	MATa/MATa ura3-52/ura3-52 leu2-3112/leu2-3112 his3-11,15/his3-11,15	Schmitz et al., 2006
AGY032	MATa shs1Δtrp	Erfei Bi, University of Pennsylvania, Philadelphia, PA
AGY50	pAGB088 [pAg SHS1-GFP-GEN3]//AGY032	This study
AGY51	pAGB412.1 [pAg SHS1ΔC-GFP-GEN3]//AGY032	This study
AGY52	pAGB405 [pAg SHS1ΔCC-GFP-GEN3]//AGY032	This study
AGY53	pAGB088 [pAg SHS1-GFP-Gen]//DHD5	This study
AGY54	pAGB412.1 [pAg SHS1ΔC-GFP-GEN3]//DHD5	This study
AGY55	pAGB405 [pAg SHS1ΔCC-GFP-GEN3]//DHD5	This study
AGY59	Sc SHS1ΔC-GFP-GEN3//DHD5	This study

^aBrackets indicate expression from a plasmid.

Plasmids used in this study are listed in Table 5. PCR was performed using standard methods with polymerases from Roche Diagnostics (Indianapolis, IN) and Invitrogen (Carlsbad, CA). Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA), and all restriction enzymes came from New England Biolabs (Beverly, MA). Oligonucleotide primers are listed in Table 6. Plasmid isolation from yeast and sequencing were performed as described in Schmitz et al. (2006). All sequencing was performed by Dartmouth College Core Facilities (Hanover, NH).

TABLE 5:

Plasmids used in this study.

Number	Name	Vector	Relevant insert	Source
	pRS416			Sikorski and Hieter, 1989
AGB005	pAGT141	pUC19	GFP-GEN3	Kaufmann, 2009
AGB009	pAGT100	pUC19	NAT1	Kaufmann, 2009
AGB088	pAgShs1-GFP-Gen	pRS416	SHS1-GFP-GEN3	Helfer and Gladfelter, 2006
AGB141	pAg Cdc11a-mCherry-Nat	pRS416	CDC11a-mCHERRY-NAT1	DeMay et al., 2009
AGB275	pTAP-Gen	pUC19	TAP-GEN	Peter Philippsen, University of Basel, Switzerland
AGB358	pAg Shs1 9A	pUC57	SHS1 9A	This study
AGB359	pAg Shs1 9D	pUC57	SHS1 9D	This study
AGB373	pAg Shs1 9D pRS416	pRS416	SHS1 9D	This study
AGB378	pAg Shs1-GFP-Gen 9D	pRS416	SHS1-GFP-GEN3 9D	This study
AGB404	pAg Shs1-GFP 9A	pRS416	SHS1-GFP 9A	This study
AGB405	pAg Shs1ΔCC-GFP-Gen	pRS416	SHS1ΔCC-GFP-GEN3	This study
AGB406	pAg Shs1-GFP-Gen 9A	pRS416	SHS1-GFP-GEN3 9A	This study
AGB412.1	pAg Shs1ΔC-GFP-Gen	pRS416	SHS1ΔC-GFP-GEN3	This study

TABLE 6:

Oligonucleotides used in this study.

Number	Name	Sequence
AGO5	Green2.2	TGTAGTTCCCGTCATCTTTG
AGO26	Elm1V1	CCCGTACGCCAACAGAGGAGATG
AGO27	Elm1V2	GTGCAGAACAAAGGGCTCTCGAC
AGO28	Elm1IV1	CAGATCTCGAGTGGCTCTACGTC
AGO37	VG3	ATGTTGGACGAGTCGGAATC
AGO234	V3 NAT1	ACATGAGCATGCCCTGCCCC
AGO235	V2 NAT1	GTGGTGAAGGACCCATCCAG
AGO365	5′ Shs1 tag	CTCCATCTGCCGACTCTAGTC
AGO427	Shs1V2	GCGCCCACAGTAGTGGTTTCAGA
AGO428	Shs1IV1	GCGGCTCTCAGGATAGGACGAA
AGO452	Elm1 Nat F	GACGACAATAGAAAACCAGTTGCATGGTGCTAGGGCATAGGGAGCCGACGGCCAGTGAATTCGAG
AGO453	Elm1 Nat R	CAGACGATTACAAGCGCAGTAGTAACTGCGGTGTGTGTGGGGATGGACCATGATTACGCCAAGCT
AGO520	GFP seqR	CGAGATTCCCGGGTAATAAC
AGO521	Gen seq F	TGGTGTCGCTCCTTCTAGTG
AGO557	Shs1 int seq R	CCCAGTGGAACTGGAAAGAC
AGO593	point mutate PCR V2	GTGTCTCCTCTGCGAGTACC
AGO649	Cdc11-TAP F	GAACGTCTGGAAAAAGAAGCCAAAACCAAGCAGGAAATTGAGGATAAAACGACGGCCAGTGAATTCGAGCTC
AGO650	Cdc11-TAP R	CTAGCGTATGATGCTCACATGTTTCAGCCATTCTTTTCTTGCTTCCAGGCATGCAAGCTTAGATC
AGO651	Cdc11 int F	CACGGGAAGAGCAGATCAGG
AGO652	TAP int R	CATTACGACCGAGATTCCCG
AGO653	TAP ds F	GGACGAGTCGGAATCGCAGAC
AGO654	TAP teft2 R	CAGGCATGCAAGCTTAGATC
AGO684	Shs1 PM seq 1 R	GAGCCATTCCGTTTCTGTTG
AGO775	Shs1-6A-GFP-Gen R	GCAAACCGCCTCTCCCCGCGCGTTGGCCGATTCATTAATGCCGCGTTCTGTTACATACCC
AGO776	Shs1-6D-GFP Gen R	GGCCAGTGAGCGCGCGTAATACGACTCACTATAGGGCGCCGCGTTCTGTTACATACCC
AGO953	Shs1-GFP-Gen 9PM F	CATCTGCCGACTCTAGTCAATTTGCTAGTGATGGATATGCTTCCCGCGCAACAGG
AGO996	Shs1ΔC GFP Gen F	TTCAAAGGGTACTTCAATAGTAAAGACCGAGAAGAATTCCAAAACGACGGCCAGTGAATTCG
AGO997	Shs1ΔC GFP Gen R	GTCTTATTACCATTATGGTGGGATATCTGCGTATAGACCATGATTACGCCAAGCTTGC
AGO1129	Sc Shs1ΔC GFP Gen F	GTTAAGAAATATTTCAGAAACCGTTCCATATGTCTTGAGACATAAAACGACGGCCAGTGAATTCG
AGO1130	Sc Shs1ΔC GFP Gen R	TTATTTATTTGCTCAGCTTTGGATTTTGTACAGATACAACCATGATTACGCCAAGCTTGC
AGO1061	Shs1ΔCC a F	TGCCGGGTAACGGAAATAAC
AGO1062	Shs1ΔCC b R	ATGTCTCAACACATATGGCAGTGTCTCCG
AGO1063	Shs1ΔCC c F	GTTGAGACATACGACGTCAATAGGATCAACTGTGACC

Radial growth assays were performed as previously described (Anker and Gladfelter, 2011).

Microscope setup and image processing

A Zeiss Axioimage-M1 upright light microscope (Carl Zeiss, Jena, Germany) equipped with a Plan-Apochromat 63×/1.4 numerical aperture (NA) oil objective was used for wide-field fluorescence microscopy. GFP signal was visualized with a Zeiss 38HE filter set and Alexa Flour 568 signal was visualized with Chroma filter set 41002B. An Exfo X-Cite 120 lamp was used as the fluorescent light source. Images were acquired on an Orca-AG charge-coupled device (CCD) camera (C4742-80-12AG; Hamamatsu, Bridgewater, NJ) driven by Volocity 5 (Perkin Elmer-Cetus, Waltham, MA). Live and immunofluorescence images were taken at 0.5-μm steps to encompass the depth of the entire hypha and were fast-deconvolved in Volocity 4. For live images used to ascertain the fluorescence intensity of a given area, 0.5-μm steps were taken over a total thickness of 9 μm. Time-lapse movies were acquired at a time interval of 10 min and iteratively deconvolved (45 iterations, Volocity 4).

Confocal time-lapse movies were acquired using a Yokogawa spinning-disk attached to a Nikon (Melville, NY) Eclipse Ti utilizing a Nikon 100×/1.4 NA Plan-Apo VC objective, controlled by MetaMorph 7 (Molecular Devices, Sunnyvale, CA). This system was assembled by Quorum Technologies (Guelph, Ontario, Canada). Photobleaching of entire septin rings was performed using a 405-nm laser at 70% power (1000-ms, 450-mW laser) in conjunction with a Mosaic micromirror array system. Images were captured using a Hamamatsu ImagEM C9100-13 electron multiplying (EM)-CCD camera. GFP was imaged using 30% laser power (491-nm laser) exposed for 200 ms/slice. mCherry was imaged using 30% laser power (561-nm laser) exposed for 250 ms/slice. Seventeen-slice Z-stacks (0.5-μm steps) through the entire hypha were imaged at 30-s intervals.

Live cells (Ashbya and yeast) were imaged on 2% agarose gel pads made with 2× low-fluorescence minimal media, covered with a glass coverslip, and sealed with lanolin or Vaseline–lanolin–paraffin. All images and movies (wide-field and confocal) presented here are maximum projections of three-dimensional volumes.

Fluorescence intensity and FRAP analysis

All fluorescence intensity and FRAP measurements were taken in Volocity 4.

Mean fluorescence intensity measurements for septin rings and septin clouds at hyphal tips were measured using a region of interest of the same area for each structure type. Cytosolic background was measured by subtracting an average of five background mean fluorescence measurements in the same cell as the measured ring from the mean fluorescence intensity of the ring.

To measure the intensity of a given ring over the course of a FRAP movie, we drew a region of interest around the entire ring, and the mean fluorescence value per pixel of the selected area was recorded. An adjacent region of cytosol was measured, and this mean fluorescence value was subtracted from the mean fluorescence of the ring to yield the mean fluorescence of the ring minus the background. Photobleaching was corrected for by using the mean fluorescence intensity of a ring of the same type in the same field. The regions of interest around the rings were shifted as necessary to compensate for cell drift during the movie. The mobile fractions for each ring were calculated by dividing percent recovery by percent bleach, and mobile fractions were averaged to obtain the average mobile fractions and SDs listed in Tables 1 and 2. The t_1/2 for each ring was calculated using the formula t_{1/2 =} t_maxfl × (log(0.5)/log(1/mobile fraction), where t_maxfl is the time point at which recovery plateaus. Average t_1/2 and SD values are listed in Tables 1 and 2. Statistical significance of the mobile fraction and t_1/2 measurements were determined by two-tailed t tests.

FCF treatment

FCF (Sigma-Aldrich, St. Louis, MO) was dissolved in 99.5% ethanol to a stock concentration of 250 mM. The stock was diluted directly in the liquid media of cell cultures grown overnight for 14 h to bring FCF to a final concentration of 25–125 μM. After addition of FCF, cells were incubated at growth conditions for an additional 30–120 min. Treated cells were fixed for 10 min in 2% paraformaldehyde (Sigma-Aldrich) to preserve GFP signal. Cells were washed twice in 1× PBS and resuspended in ProLong Gold antifade reagent (Invitrogen).

Immunofluorescence

Immunofluorescence was performed as described previously (Gladfelter et al., 2006). Rabbit anti-ScCdc11p (Santa Cruz Biotechnology, Santa Cruz, CA) was diluted in 1× PBS + 1 mg/ml bovine serum albumin and used at a 1/50 dilution. Cells were incubated with primary antibody overnight at 4°C. Primary antibody was detected and visualized using Alexa Fluor 568 anti-rabbit (Invitrogen) at a 1/200 dilution.

Visualization of F-actin

For viewing F-actin, cells were fixed with 3.7% formaldehyde (Fisher Scientific, Fair Lawn, NJ) at 30°C for 1 h. Cells were washed twice in 1× phosphate-buffered saline (PBS) and resuspended in 1× PBS. Cells were incubated with Alexa Fluor 568 phalloidin (Invitrogen) at a final concentration of 6.6 μM for 1 h at room temperature in the dark. Cells were washed twice with 1× PBS and resuspended in ProLong Gold antifade reagent (Invitrogen).

Protein extraction and purification

Cells were grown in AFM for 16 h and harvested by vacuum filtration. For lysis, 500 mg of cells were resuspended in 1 ml ice-cold lysis buffer (50 mM HEPES-KOH, pH 7.6, 1 M KCl, 1 mM MgCl₂, 1 mM ethylene glycol tetraacetic acid (EGTA), 5% glycerol, 0.45% Tween-20, 2× protease inhibitor cocktail). For lysates used to analyze the phosphorylation status of Shs1 by mass spectrometry and Western blot, phosphatase inhibitors (50 mM NaF and 100 mM β-glycerophosphate) were added to the lysis buffer. Cells were lysed by bead-beating with 0.5-mm zirconia/silica beads in a MiniBeadbeater-8 (BioSpec Products, Bartlesville, OK) at top speed for five intervals of 90 s with 30-s rests. Cell lysates were clarified by centrifugation for 10 min at 13,200 rpm. Protein concentrations were determined by Bradford assay.

Septin complexes to be analyzed by mass spectrometry were purified from AG124 (Shs1-GFP) lysates using the GFP-Trap-M system (Chromotek, Munich, Germany). Lysates containing 18 mg total protein were incubated with 12 μl GFP-Trap-M beads for 2 h at 4°C. Beads with precipitated proteins were separated from lysates on a magnet and washed seven times with 150 μl IP wash buffer (50 mM HEPES-KOH, pH 7.6, 200 mM KCl, 1 mM MgCl₂, 1 mM EGTA, 5% glycerol, 0.1% Tween-20). Precipitated proteins were eluted in 35 μl 5× reducing sample buffer and separated on a 16 × 18 cm 10% acrylamide gel by SDS–PAGE at 25 mA for 7 h. Proteins were stained using the colloidal-blue staining kit (Invitrogen). Bands representing Shs1p were excised and analyzed for phosphorylation, acetylation, and sumoylation by the Taplin Biological Mass Spectrometry facility (Harvard Medical School, Boston, MA).

Septin complexes to be analyzed for Shs1p phosphorylation by phosphatase treatment were purified from AG495.1 (Cdc11-TAP). Lysates containing 10 mg total protein were incubated with 100 μl protein A–Sepharose beads (Sigma-Aldrich) and 50 μl rabbit immunoglobulin G (IgG; Fisher) at 4°C for 2 h. Immunoprecipitates were washed twice with IP wash buffer, and samples to be treated with phosphatase were washed twice with 1× phosphatase buffer (New England Biolabs). Samples were treated with λ-phosphatase (New England Biolabs), phosphatase buffer only, or phosphatase buffer with phosphatase inhibitors for 1 h at 30°C. Samples were washed twice with IP wash buffer and eluted in 1× reducing sample buffer. Proteins were separated by SDS–PAGE.

For comparison of septin complexes formed in AG539.2 (Shs1ΔC-GFP) with those in AG124 (Shs1-GFP), lysates containing 7.5 mg total protein were incubated with 50 μl Dynabeads (Invitrogen) coupled to anti-GFP (ab290; Abcam, Cambridge, MA) per the manufacturer's instructions for 2 h at 4°C. Beads were washed three times with wash buffer (100 mM HEPES, pH 8.0, 20 mM Mg acetate, 300 mM Na acetate, 10% glycerol, 10 mM EGTA, 0.1 mM EDTA, 0.5% NP-40) on a magnet and eluted in 1× sample buffer at 65°C for 10 min. Samples were reduced by adding 0.35 μl β-mercaptoethanol (Fisher) and boiling for 5 min. Proteins were separated by SDS–PAGE on a 10% acrylamide gel and silver-stained.

Septin complexes purified for electron microscopy were immunoprecipitated from AG495.1 (Cdc11-TAP). Lysates containing 40 mg total protein were precleared with 500 μl protein A–Sepharose beads for 2 h, which was followed by an incubation with 800 μl protein A–Sepharose beads and 100 μl rabbit IgG for 3 h at 4°C. Beads were transferred to a 0.8 × 4 cm poly-prep chromatography column and washed with 30 ml purification wash buffer (100 mM HEPES, pH 8.0, 20 mM Mg acetate, 300 mM Na acetate, 10% glycerol, 10 mM EGTA, 0.1 mM EDTA, 0.5% NP-40); this was followed by 10 ml TEV cleavage buffer (10 mM Tris HCl, pH 8.0, 150 mM NaCl, 0.1% NP-40, 0.5 mM EDTA, 1.0 mM dithiothreitol). Beads were treated overnight with 300 U ProTEV Protease Plus (Promega, Madison, WI) in 1 ml TEV cleavage buffer. Purified proteins were collected by gravity flow-through and ProTEV Protease Plus was removed from the eluate using Magne-His Ni Particles (Promega). The sample was then split in half, and dialyzed into either 1 M KCl or 75 mM KCl buffer (20 mM HEPES, 1M or 75 mM KCl, 0.5 mM EDTA, 0.5 mM EGTA, 8% sucrose).

Western blotting

Whole-cell extracts or immunoprecipitates were loaded onto a 10% polyacrylamide gel and separated by SDS–PAGE for 60 min at a constant voltage of 165 V. Proteins were transferred to polyvinylidene fluoride (PVDF) membrane (Bio-Rad, Hercules, CA). Membranes were probed with rabbit polyclonal anti-GFP antibody (ab290; Abcam) at a 1:1000, rat monoclonal anti–α-tubulin antibody (AbD; Serotec, Raleigh, NC) both at a 1:1000 dilution, and/or an anti-Shs1 antibody (a gift from Doug Kellogg, University of California, Santa Cruz) at a final concentration of 2 μg/ml. All primary antibodies were diluted in PBS-Tween-20 (0.1%) with 3% milk. Alkaline phosphatase–conjugated secondary antibodies (Bio-Rad) were used at a 1:10,000 dilution in PBS–Tween-20 (0.1%). Blots were developed with the enhanced chemifluorescence detection system and visualized using a Storm 860 PhosphorImager (GE Healthcare, Piscataway, NJ).

Electron microscopy

Freshly isolated and dialyzed septin complex was applied to glow-discharged carbon-coated 300-mesh Cu grids (Electron Microscopy Sciences, Hatfield, PA). After 2 min, the grids were rinsed in water and stained with 1% uranyl acetate in 50% methanol for 30 s. The excess uranyl acetate was wicked away with filter paper. All transmission electron microscopy images were taken at 100 kV on a JEOL TEM1010 equipped with a digital camera (XR-41B) and capture engine software (AMTV540; Advanced Microscopy Techniques, Woburn, MA). Micrographs were captured at 25,000× magnification. Septin complex length measurements were taken in ImageJ (National Institutes of Health).

Abbreviations used:

Ag Shs1ΔC-GFP

Ashbya gossypii Shs1ΔC-GFP

CCD

charge-coupled device

EGTA

ethylene glycol tetraacetic acid

FCF

forchlor­fenuron

FRAP

fluorescence recovery after photobleaching

IgG

immunoglobulin G

IR

interregion

NA

numerical aperture

ORF

open reading frame

PIP₂

phosphatidylinositol-4,5-bisphosphate

PVDF

polyvinylidene fluoride

Sc SHS1

S. cerevisiae SHS1

Shs1-GFP

Shs1–green fluorescent protein