Septins coordinate cell wall integrity and lipid metabolism in a sphingolipid-dependent process

The cell wall and plasma membrane (PM) are the primary lines of defense against environmental insults for fungal cells. With the large surface area of hyphal networks and intimate contact with the surrounding media, fungi encounter many stresses, ranging from ion imbalance to predation. The cell wall contains several polysaccharide constituents – chitin provides the rigid framework of the cell wall, β-glucan maintains the shape and mannans form the outermost, protective layer (Bulawa et al., 1995; Munro and Gow, 2001; Orlean, 2012; Schmidt, 2004; Shaw et al., 1991; Sudoh et al., 2000). Precise regulation of cell wall and plasma membrane architecture is attained by tightly coordinated signaling pathways, which control genes responsible for maintaining homeostasis between the intracellular and extracellular environments.

Septins are highly conserved small GTPase cytoskeletal proteins that function as molecular scaffolds for dynamic cell wall and plasma membrane-remodeling, as well as diffusion barriers restricting movement of membrane and cell wall-associated molecules (Caudron and Barral, 2009; Dobbelaere and Barral, 2004; Gladfelter et al., 2001; Glomb and Gronemeyer, 2016; Longtine and Bi, 2003; Oh and Bi, 2011; Takizawa et al., 2000; Trimble and Grinstein, 2015). The Saccharomyces cerevisiae septins Cdc3, Cdc10, Cdc11 and Cdc12 have been termed ‘core septins’ because they are monomers that assemble into non-polar heterooligomers and micrometer-scale higher order structures in the form of bars, rings or gauzes (Bertin et al., 2008; McMurray et al., 2011; McMurray and Thorner, 2019; Sirajuddin et al., 2007). Aspergillusnidulans contains four orthologous septin proteins – AspA^Cdc11, AspB^Cdc3, AspC^Cdc12 and AspD^Cdc10, as well as AspE, which has no S. cerevisiae ortholog (Pan et al., 2007). A. nidulans contains two distinct subpopulations of heterooligomers – an octameric oligomer consisting of all four core septins in the same order as proposed in S. cerevisiae (AspA^Cdc11–AspC^Cdc12–AspB^Cdc3–AspD^Cdc10–AspD^Cdc10–AspB^Cdc3–AspC^Cdc12–AspA^Cdc11), and a second hexameric oligomer, with the proposed order (AspA^Cdc11–AspC^Cdc12–AspB^Cdc3–AspB^Cdc3–AspC^Cdc12–AspA^Cdc11) (Glomb and Gronemeyer, 2016; Hernandez-Rodriguez et al., 2014; Oh and Bi, 2011). For clarity, we will refer to AspA^Cdc11, AspB^Cdc3 and AspC^Cdc12 as ‘core hexamer septins’; AspA^Cdc11, AspB^Cdc3, AspC^Cdc12 and AspD^Cdc10 as ‘core octamer septins’; and AspE as the ‘noncore septin’ because it does not assemble into oligomeric structures, though it is required for higher order structure assembly at the multicellular stage (Hernandez-Rodriguez et al., 2012, 2014).

Previous studies in Candida albicans have shown that septins provide the scaffolding for cell wall proteins via the cell wall integrity (CWI) mitogen-activated protein kinase (MAPK) signaling pathway (Blankenship et al., 2010; Caudron and Barral, 2009; Fukuda et al., 2009; Gladfelter et al., 2001; Mostowy and Cossart, 2012). The CWI pathway, along with the other major MAPK signaling pathways [high osmolarity glycerol (HOG), cAMP-protein kinase A (PKA), target of rapamycin (TOR), calcineurin/Ca²⁺ and mating/pheromone response pathways] are highly conserved across the kingdom Fungi, and have been shown to undergo extensive cross-talk to coordinate virtually all biological functions in the cell, from expansion and division to asexual reproduction (Colabardini et al., 2014; Fujioka et al., 2006; Garcia et al., 2017; Manfiolli et al., 2019; Munro et al., 2007; Rodriguez-Pena et al., 2010; Walker et al., 2008; Wiederhold et al., 2005).

Sphingolipids are long chain base-containing lipids (Dickson, 2008) that are metabolized in a highly conserved pathway in plants, animals and fungi; sphingolipid metabolism shares direct connections to other major metabolic pathways, such as sterol metabolism and fatty acid and phospholipid synthesis (Sims et al., 2004). Sphingolipid pathway intermediates, such as phytosphingosine (PHS), have been shown to be involved in CWI pathway signaling in S. cerevisiae (Dickson, 2008; Roelants et al., 2002). Sterol-rich domains (SRDs), also called ‘lipid rafts’ or ‘lipid microdomains’, are regions of the plasma membrane enriched in specific classes of lipids, including sterols [ergosterol, the major sterol found in most fungi (Weete et al., 2010)], sphingolipids and phosphoinositides (Alvarez et al., 2007; Tanaka and Tani, 2018), which have been shown to be functionally important for maintenance of cell polarity (Bagnat and Simons, 2002). Membrane organization, plasticity and overall integrity have been attributed to sphingolipid, sterol and glycerophospholipid interactions (van Meer et al., 2008). In vitro work has shown that yeast septins bind directly to the phosphoinositides phosphatidylinositol 4,5-bisphosphate [PIP2 or PI(4,5)P2], phosphatidylinositol 5-phosphate [PI(5)P] and phosphatidylinositol 4-phosphate [PI(4)P] (Bertin et al., 2010; Krokowski et al., 2018; Tanaka-Takiguchi et al., 2009; Zhang et al., 1999). Septins have also been shown to genetically interact with Sur2, a sphinganine C4-hydroxylase which catalyzes the conversion of sphinganine to PHS in sphingolipid biosynthesis (Costanzo et al., 2010), and to physically interact with GTPases, Bud1 and Cdc42 to maintain sphingolipid-dependent diffusion barriers at cell membranes in yeast (Clay et al., 2014). Interdependent colocalization of septins and sterol rich microdomains have been described in a number of in vivo and in vitro systems (Beh et al., 2009; Martin and Konopka, 2004). In the dimorphic fungus Ustilago maydis, septins and SRDs depend on each other to localize properly at the hyphal tip for cell signaling and cell polarity establishment (Canovas and Perez-Martin, 2009). In the budding yeast S. cerevisiae, long-chain sphingolipids have been implicated in maintaining asymmetric distribution and mobility of some membrane-spanning molecules, such as multidrug resistance transporters (Singh et al., 2017). In contrast to these findings, a more recent comprehensive study of protein segregation during cell division in budding yeast found that the majority of endoplasmic reticulum (ER) and PM proteins with transmembrane domains were actually symmetrically segregated; however, septins were shown to be responsible for partitioning of the PM-associated ER at the bud neck, thereby restricting diffusion of ER lumen and a particular set of membrane-localized proteins (Sugiyama and Tanaka, 2019).

Recent work has started to unravel the functional connections between the septins, CWI MAPK pathway signaling and lipid metabolism; however, most studies have focused on a small subset of septin monomers and/or were conducted primarily in yeast-type fungi (Badrane et al., 2016; Bridges and Gladfelter, 2015; de Almeida, 2018; Merlini et al., 2015; Wu et al., 2010; Zhang et al., 2017). Here, we show in the filamentous fungus A. nidulans that the core septins AspA^Cdc11, AspB^Cdc3, AspC^Cdc12 and AspD^Cdc10 are required for proper coordination of the CWI pathway, that all five septins are involved in lipid metabolism and that these roles require sphingolipids.

To determine whether A. nidulans septins are important for CWI, we used spore dilution assays to test the growth of septin deletion mutants on medium containing a variety of known cell wall-disturbing agents. Wild-type (WT) and septin-null mutants were tested on Calcofluor White (CFW) and Congo Red (CR), cell wall polymer-intercalating agents that perturb chitin and β-1,3-glucan, respectively, as well as caspofungin (CASP), an inhibitor of β-1,3-glucan synthase, and fludioxonil (FLU), a phenylpyrrol fungicide that antagonizes the group III histidine kinase in the osmosensing pathway and consequently affects CWI pathway signaling (Fig. 1) (Dongo et al., 2009; Hohmann, 2002; Kojima et al., 2006; Kopecka and Gabriel, 1992; Kovacs et al., 2013; Kuranda et al., 2006; Onishi et al., 2000; Pancaldi et al., 1984; Roncero and Duran, 1985; Yun et al., 2014). Septin mutants ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 showed hypersensitivity to CFW, CR, CASP and FLU; ΔaspD^cdc10 displayed no sensitivity to CFW, CR and FLU, but sensitivity to CASP (see Table S1 for list of strains used in this study). The ΔaspE mutant showed no sensitivity to CFW, CASP and FLU, and very mild sensitivity to CR (Fig. 1A).

One hallmark of CWI defects is that they can be rescued by the addition of an osmotic stabilizer, such as sorbitol or sucrose. The addition of exogenous 1.2 M sorbitol partially rescued the hypersensitivity to CASP and fully rescued the hypersensitivity to CFW, CR and FLU for ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 (Fig. 1B). Sorbitol also rescued the sensitivity of ΔaspD^cdc10 to CASP and the mild sensitivity of ΔaspE to CR. The osmotic rescue of growth defects in the mutants indicates that their sensitivity to cell wall-disturbing agents is likely the result of a CWI defect.

Because the core septin mutants showed sensitivity to cell wall-disturbing agents, consistent with action via the CWI pathway, we predicted that there might be differences in cell wall polymer localization in the mutants. To examine cell wall polymer localization, we undertook live-cell imaging of WT and ΔaspB^cdc3 since it is a member of both hexamers and octamers. Conidia were incubated on coverslips in liquid medium, stained with CFW or Aniline Blue to observe chitin and β-1,3-glucan, respectively, and immediately observed by fluorescence microscopy. Z-stack images were analyzed one-by-one, compressed into maximum projection images to visualize the fluorescence signal in the entire hyphal structure, and by line scans of Aniline Blue and CFW staining patterns (Fig. 2; Fig. S1). As previously reported, ΔaspB^cdc3 cells showed more presumptive branch initials per hyphal compartment than WT cells (Hernandez-Rodriguez et al., 2012). The Aniline Blue staining showed a reduction of labeling at all ΔaspB^cdc3 hyphal tips. Smaller patches of less intense Aniline Blue labeling also occurred at presumptive branch initials in ΔaspB^cdc3 than in WT (Fig. 2A; Fig. S1A). The CFW staining of ΔaspB^cdc3 showed a shift of the tip band of staining closer to the hyphal apex and a less well-defined endocytic collar zone than in WT (Fig. 2B; Fig. S1B). These data show that cell wall organization is altered in ΔaspB^cdc3 and raise the possibility that it might be altered in other core hexamer septin-null mutants as well.

Previous studies have shown that perturbations to one cell wall component often trigger compensatory changes to others via the CWI pathway (Fortwendel et al., 2010; Garcia-Rodriguez et al., 2000; Walker et al., 2008). To analyze the cell wall composition of septin mutants, two independent biological replicates of a diagnostic glycosyl linkage analysis (Pettolino et al., 2012) were conducted to quantify the cell wall polysaccharide composition of WT, ΔaspB^cdc3 (a core septin-null mutant which showed hypersensitivity to cell wall-disturbing agents) and ΔaspE (non-core septin mutant) (Fig. S2). All samples contained 3- and 4-linked glucose as the primary cell wall components, as well as 4-linked N-acetylglucosamine (the monomer of chitin; 4-GlcNAc) and a relatively minor amount of mannan. The septin mutants ΔaspB^cdc3 and ΔaspE showed 7% and 2% increases in the average area of the detected linkage peak for 4-GlcNAc compared to WT, respectively; the amount of 4-Glc and of mannan (data not shown) did not show significant differences between samples. These data indicate an increase in chitin content of the ΔaspB^cdc3 mutant and raise the possibility that it might be altered in other core hexamer septin-null mutants as well.

Membrane-spanning cell wall synthases are the ultimate effectors of the CWI pathway. A. nidulans contains six genes for chitin synthases – chsA, chsB, chsC, chsD, csmA and csmB. Chitin synthase B localizes to sites of polarized growth in hyphal tips, as well as developing septa in vegetative hyphae and conidiophores, a pattern very similar to septin localization. Deletion of chitin synthase B causes severe defects in most filamentous fungi analyzed thus far, and repression of the chitin synthase B gene expression in chsA, chsC and chsD double mutants exacerbates growth defects from a number of developmental states observed in each single mutant, suggesting it plays a major role in chitin synthesis at most growth stages (Fukuda et al., 2009). For these reasons, we chose chitin synthase B as a candidate to observe in a septin mutant background for possible defects in localization. We hypothesized that recruitment or maintenance of cell wall synthases, such as chitin synthase could be disrupted in septin-null mutants that showed sensitivity to cell wall-disturbing agents and altered cell wall staining patterns. To determine the localization of synthases, a chitin synthase B–GFP (chsB-GFP) strain was crossed with strains in which core hexamer septins were deleted. After repeated attempts, the only successful cross was with core hexamer deletion strain ΔaspA^cdc11. The ΔaspA^cdc11, chsB-GFP strain showed conspicuous differences in chitin synthase localization patterns compared to WT (Fig. 3; Fig. S3). In WT, the chitin synthase–GFP signal was at the tips of >90% of branches, presumptive branch initials and apical hyphal tips (Fig. S3A). In the septin-null mutant ΔaspA^cdc11, GFP signal was absent in most presumptive branch initials, but present in longer side branches that were at least 10 µm long and in the apical hyphal tip(s) (Fig. 3; Fig. S3A). Intriguingly, the chitin label CFW was more intense in the septin deletion strain, consistent with the increase in 4-GlcNAc levels seen in polysaccharide analysis (Fig. S3B). CFW labeling also showed that the polymer chitin was present throughout the hyphae in both the WT and septin deletion strains (Fig. 2) suggesting that chitin synthases had been active in both, but perhaps were not retained as long in the presumptive branch initials of the septin deletion mutant. These data support the hypothesis that core hexamer septins play a role in recruitment and/or retention of cell wall synthases to the proper locations at the plasma membrane during growth and development.

To determine whether there are genetic interactions between the septins and the CWI pathway kinases, double mutants were generated by means of sexual crosses and analyzed. The first CWI pathway kinase, PkcA^Pkc1 (ANID_00106), is essential in A. nidulans, and so null alleles could not be utilized for double-mutant analysis. A null allele of the final kinase in the CWI MAPK signaling cascade, MpkA^Slt2 (ANID_11786), was crossed with ΔaspB^cdc3 and ΔaspE, and progeny were challenged with cell wall-disturbing treatments. If the septins were directly in the CWI pathway, we would expect double mutants to show the same phenotypes as the parental null mutant that acts earliest. If the septins were part of a parallel pathway or alternate node that also affected CWI, we would expect a novel or synergistic phenotype in the double mutants.

Spore dilution assays were conducted, challenging the double mutants and the parental strains with cell wall-disturbing treatments. The double mutants ΔaspB^cdc3ΔmpkA^slt2 and ΔaspE ΔmpkA^slt2 displayed a colony-level radial growth defect and reduced conidiation, which phenocopied the ΔmpkA^slt2 single mutant. When challenged with low concentrations of CASP and CFW, the ΔaspB^cdc3ΔmpkA^slt2 and ΔaspE ΔmpkA^slt2 mutants were more sensitive than ΔaspB^cdc3 and ΔaspE single mutants, but suppressed the colony growth defects of ΔmpkA^slt2. The novel phenotype of the double mutants shows that septins are involved in CWI and raises the possibility that they act in a bypass or parallel node for rescue of cell wall defects (Fig. 4).

One possible explanation for the observed sensitivity to cell wall-disturbing agents could be that septins are involved in ‘cross-talk’ with other MAPK signaling pathways that have been shown to interact with the CWI pathway signaling such as the Ca²⁺/calcineurin (CAMK) signaling pathway. To test this possibility, calcium chloride, EGTA (a Ca²⁺-chelating agent) and FK-506 (a calcineurin inhibitor) were added to each treatment (Fig. S4) (Sio et al., 2005; Wang et al., 2016; Yoshimoto et al., 2002). The treatments caused no obvious colony growth defects, suggesting that the Ca²⁺/calcineurin signaling pathway crosstalk does not significantly contribute to the observed sensitivity of septin-null mutants to cell wall-disturbing agents.

Another pathway closely associated with CWI, lipid biosynthesis and lipid signaling is the TOR MAPK signaling pathway (Harvald et al., 2015). To test the involvement of septins in this signaling pathway, septin-null mutants were challenged with rapamycin (a potent inhibitor of the TOR pathway), as well as the methylxanthine derivatives caffeine and theophylline (Barbet et al., 1996; Cardenas et al., 1999; Scott and Lawrence, 1998). There were no observable growth defects in the presence of rapamycin (final concentration was 600 ng/ml, which is ∼3-fold higher than the inhibitory concentration for known TOR pathway mutants) in the septin-null mutants compared to WT (Fig. S4) (Fitzgibbon et al., 2005; Soulard et al., 2010). Caffeine and theophylline have been shown to interfere with phosphodiesterase activity in the cAMP-PKA and TOR pathways, and ΔpkaA^tpk1 and Δtor1^torA mutants show hypersensitivity to caffeine treatment, which cannot be rescued by exogenous sorbitol (Binder et al., 2010; Butcher and Potter, 1972; Kuranda et al., 2006; Ni et al., 2005; Scott and Lawrence, 1998). If the septins were involved in crosstalk between the cAMP-PKA or TOR pathways and the CWI pathway, we would predict that septin-null mutants would be hypersensitive to caffeine and theophylline treatments. To our surprise, only ΔaspE showed hypersensitivity to both caffeine and theophylline, and the hypersensitivity was not rescued by an osmotic stabilizer (Fig. S4 and data not shown). These data suggest that the TOR, cAMP-PKA and Ca²⁺/calcineurin signaling pathways do not contribute to the cell wall sensitivity or plasma membrane resistance phenotypes in the core septin-null mutants, but the cAMP-PKA or TOR pathways may contribute to the phenotypes of the septin ΔaspE null mutant.

We hypothesized that septins could indirectly modulate MAPK pathways, particularly the CWI pathway, through interactions with plasma membrane lipids within SRDs. Recent work has shown that the CWI pathway in S. cerevisiae is regulated by sphingolipids and ergosterol, facilitating proper deposition of cell wall polymers at actively growing regions and sites of septation (de Almeida, 2018; Tanaka and Tani, 2018). It is well-established that septins localize to sites of polarized growth and septation, where highly dynamic remodeling of the plasma membrane and cell wall via membrane-bound synthases and polarity-associated proteins takes place (Foderaro et al., 2017). These highly dynamic plasma membrane SRDs in fungi often contain ergosterol and sphingolipids. To determine whether loss of septins might affect one of the major SRD-associated lipids, each septin mutant was challenged by drug treatments disrupting the ergosterol biosynthesis pathway in spore dilution assays (Fig. 5). These assays included an azole treatment (itraconazole), an allylamine treatment (terbinafine), and a polyene treatment (natamycin), which each impact a different step in ergosterol biosynthesis (Fig. 5A). Allylamines (terbinafine) inhibit the conversion of squalene to squalene epoxide, azoles (itraconazole) inhibit the conversion of lanosterol to 4,4-dimethylcholesta-8,14,24-trienol, and polyenes bind directly to ergosterol (Onyewu et al., 2003).

The ΔaspA^cdc11, ΔaspB^cdc3, ΔaspC^cdc12 and ΔaspD^cdc10 mutants were more resistant to itraconazole and terbinafine than WT or ΔaspE. Only ΔaspA^cdc11 and ΔaspE showed strong resistance to natamycin treatment (Fig. 5B). The addition of exogenous 50 μM ergosterol was able to fully rescue the sensitivity of all null mutant strains and WT to itraconazole, terbinafine and natamycin, suggesting ergosterol was indeed the primary lipid component disrupted by these treatments (Fig. 5C). These data suggest that all five septins are involved in monitoring ergosterol metabolism and/or deposition.

Sphingolipids are a class of plasma membrane lipids that has been shown to be associated with SRDs, along with sterols and phosphoinositides, and therefore are possible targets for septin-mediated interactions at the membrane. To determine whether loss of septins impacts sphingolipid metabolism in A. nidulans, the septin-null mutants were challenged with sphingolipid biosynthesis-disrupting agents myriocin and aureobasidin A (AbA) (Fig. 6). Myriocin disrupts the first committed step of the biosynthetic pathway, converting palmitoyl-coA and serine into 3-ketodihydrosphingosine, preventing the accumulation of downstream intermediates, such as ceramides and sphingoid bases like PHS, as well as complex sphingolipids at the final steps of the pathway (Fig. 6A) (Alvarez et al., 2007; Martin and Konopka, 2004). Aureobasidin A inhibits IPC synthase, disrupting the conversion of inositolphosphorylceramide from phytoceramide, and consequently causing the accumulation of intermediates such as PHS, which has been shown to be toxic at high concentrations (Fig. 6A) (Cheng et al., 2001).

The core hexamer null mutants ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 were more sensitive to myriocin than the other septin-null mutant or WT cells (Fig. 6B). ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 were also more resistant to AbA than null mutants of the octamer-exclusive septin ΔaspD^cdc10, the noncore septin ΔaspE or WT (Fig. 6B). Strikingly, the addition of exogenous PHS (15 µM) to the myriocin treatment fully rescued the sensitivity of ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 (Fig. 6C). ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 were also more resistant to higher concentrations (20 µM) of the phytosphingosine intermediate (PHS), which has been shown to be toxic at high concentrations (Fig. 6C) (Li et al., 2007).

To address whether cell wall and plasma membrane defects might be associated with one another in septin-null mutants, a combinatory drug treatment approach was taken. Sublethal concentrations of CFW (2 μg/ml) and myriocin (10 μg/ml), in which all strains grew at every spore concentration, were combined (Fig. 6D). When combined, the two drugs resulted in additive, colony-level growth defects for ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12. This sensitivity to both drugs in combination was rescued by the addition of exogenous PHS. Surprisingly the addition of exogenous sorbitol, which had fully rescued the hypersensitivity of the septin-null mutants to all previously tested cell wall-disturbing agents, resulted in a more dramatic growth defect in medium containing only myriocin, or in combination with CFW (Fig. 6C,D). These data together suggest that the core hexamer septins AspA^Cdc11, AspB^Cdc3 and AspC^Cdc12 may monitor sphingolipid metabolism. They further suggest that the hexamer septins may signal sphingolipid status to the CWI pathway, and that this signal is required for proper CWI function.

We predicted that, since core septin hexamer null mutants ΔaspA^cdc11, ΔaspB^cdc3 and ΔaspC^cdc12 were sensitive to drugs that inhibit sphingolipid biosynthesis, the localization of these septins might be altered when exposed to the same treatments. We examined whether disruption of sphingolipid biosynthesis causes changes in septin localization using live-cell imaging with fluorescence microscopy. A strain carrying AspA^Cdc11–GFP was grown in liquid complete medium overnight, treated with exogenous PHS (15 µM) and imaged for 3 h post-treatment. There was a dynamic shift in septin localization under PHS treatment over the course of the experiment, compared to the vehicle control. The septin–GFP signal shifted from a relatively homogenous cortical localization along the hyphal tips to a more stochastic, punctate localization along the entire length of hyphae (Fig. 7, middle panels).

In contrast to these results, treatments with the ergosterol-binding polyene filipin III (25 μg/ml) and the phosphatidylinositol 3-kinase inhibitor wortmannin (20 μg/ml), did not affect the localization pattern of septins as dramatically as the PHS treatment (Fig. 7, left- and right-most panels). Similar patterns of aberrant localization were observed under PHS treatments for AspB^Cdc3–GFP (Fig. S5). Myriocin treatment also resulted in a similar pattern of mislocalization in AspA^Cdc11–GFP cells when compared to treatment with PHS (Fig. S6). These results suggest sphingolipid content and/or distribution within the plasma membrane contributes to the localization and stability of core hexamer septins at the plasma membrane and that ergosterol and phosphoinositides may not be as vital for this process.

Our data show that A. nidulans septins play roles in both plasma membrane and CWI, and that distinct subgroups of septins carry out these roles. Previous work has shown that the five septins of A. nidulans septins form hexamers (AspA^Cdc11, AspB^Cdc3 and AspC^Cdc12) and octamers (AspA^Cdc11, AspB^Cdc3, AspC^Cdc12 and AspD^Cdc10) and that the noncore septin AspE does not appear to be a stable member of a heteropolymer (Hernandez-Rodriguez et al., 2014). The current work suggests that although all septins are involved in coordinating cell wall and membrane integrity, the roles of hexamers, octamers and the noncore septin are somewhat different. Core hexamer septins appear to be most important for sphingolipid metabolism, all five septins appear to be involved in ergosterol metabolism, and core septins are most important for the CWI pathway with the noncore septin possibly playing a minor role. As summarized in Fig. 8 and discussed in more detail below, our previous and current data are consistent with a model in which (1) all five septins assemble at sites of membrane and cell wall remodeling in a sphingolipid-dependent process; (2) all five septins recruit and/or scaffold ergosterol and the core hexamer septins recruit and/or scaffold sphingolipids and associated sensors at these sites, triggering changes in lipid metabolism; and (3) the core septins recruit and/or scaffold CWI machinery to the proper locations and trigger changes in cell wall synthesis. The noncore septin might play a minor role in this process.

We hypothesize that SRD-associated lipids (ergosterol, sphingolipids and phosphoinositides) recruit or facilitate binding and assembly of septins on the membrane, with sphingolipids contributing more to the stabilization of septins than other SRD-associated lipids in A. nidulans (Fig. 8A). Previous studies showed preferential in vitro binding of yeast core septin orthologs to the phosphoinositides PIP2, PI(4,5)P2, PI(5)P and PI(4)P; however, our treatments with wortmannin and filipin III, known disruptors of phosphoinositides and ergosterol respectively (de Kruijff and Demel, 1974; Santos et al., 1998; Takeo, 1985; Walker et al., 2000; Wymann et al., 1996), did not affect septin localization as dramatically as sphingolipid-disturbing treatments (Fig. 7; Figs S3 and S4). The marked septin mislocalization we observed upon treatment with PHS and myriocin strongly supports the idea that sphingolipids at the membrane contribute to core septin localization and help maintain septin assemblies at the proper locations. Previous studies on composition of lipid microdomains showed that relatively minor differences in sphingolipid structure can have significant effects on sterol and phospholipid interactions, consequently resulting in major changes in membrane properties (Björkbom et al., 2010; Ramstedt and Slotte, 2006). Perhaps sphingolipids (and ergosterol to a lesser extent) help to stabilize SRDs in a way that facilitates assembly of septin filaments and higher order structures via diffusion, collision and annealing, as proposed by Bridges and colleagues (Bertin et al., 2010; Bridges et al., 2014; Krokowski et al., 2018; Tanaka-Takiguchi et al., 2009). Consistent with this idea, A. nidulans AspB filaments have been shown to move along the plasma membrane, break apart and ‘snap’ together in a way that suggests collision and annealing (Hernandez-Rodriguez et al., 2012).

Consistent with a role for septins in modulating membrane composition, all septin-null mutants were resistant to ergosterol-disrupting treatments and core hexamer septin mutants were affected by disruption of sphingolipid biosynthesis (Fig. 5A,B; Fig. 6). Based on proposed mechanisms for septins as diffusion barriers or organizational scaffolds of membrane-associated proteins in yeast and smut fungi, we propose septins monitor lipid microdomain composition and/or organization in filamentous ascomycetes (Canovas and Perez-Martin, 2009; Sugiyama and Tanaka, 2019) (Fig. 8B). Septins do not have a transmembrane domain, a feature that often defines established membrane ‘sensors’ that monitor local lipid environments (Cho et al., 2019; Nyholm et al., 2007); however, septins share a highly conserved polybasic domain proposed to facilitate septin–membrane interactions (Valadares et al., 2017; Zhang et al., 1999). In addition to the polybasic domain, septins have recently been shown to contain an amphipathic helix motif, which has been implicated in septin sensing of membrane curvatures (Cannon et al., 2019). Given that septins have been shown to assemble into non-polar higher-order structures along the plasma membrane (Bridges et al., 2014), septin assembly itself might be the mechanism by which lipid composition and protein organization is monitored at the cytosolic face of membranes. Perhaps septin assemblies that pass specific size or geometric thresholds trigger signaling through MAPK and other pathways.

In addition to monitoring and relaying information about the membrane, septins clearly have a role in normal cell wall growth and remediation of cell wall stress via the CWI pathway. We propose a major role of the core septins is to recruit and/or organize integral proteins to sites of polarized growth or remodeling at the cell cortex to ensure CWI pathway functions are carried out (Fig. 8C). The sensitivity to cell wall-disturbing agents, altered cell wall composition and altered polysaccharide deposition in the core septin-null mutants (Figs 1,2; Figs S1,S2) are consistent with phenotypes of CWI pathway mutants in previous studies (Valiante et al., 2008, 2009, 2015). Our glycosyl linkage analysis showed that cell wall chitin content is increased in the septin mutant cells compared to WT cells (Fig. S2). Hyper-synthesis of chitin has been shown to occur during cell wall stress via the CWI pathway in S. cerevisiae and Candida albicans (Dallies et al., 1998; Kapteyn et al., 2000, 1997; Lagorce et al., 2002; Osmond et al., 1999; Popolo et al., 1997; Ram et al., 1998, 1994). Double mutant analyses between septins and CWI pathway kinases also support a role for core septins in maintaining CWI under stress (Fig. 4). Suppression of cell wall defects under cell wall stress by deletion of septins in an ΔmpkA^slt2 background suggests that a parallel node by which septins negatively regulate CWI pathway sensors or kinases could exist. This interpretation is consistent with studies in yeast showing that Bni4 (ANID_00979), a formin which is phosphorylated by and functions downstream of MAPK Slt2^mpkA, directly interacts with the core septin orthologs in order to recruit chitin synthases to the bud neck (Heinisch and Rodicio, 2018; Larson et al., 2010; Perez et al., 2016). Although the involvement of AspE in the CWI pathway appears to be minor based on the very mild sensitivity of the null mutant to Congo Red (Fig. 1), sensitivity to TOR and cAMP-PKA pathway inhibitors suggests that AspE might participate indirectly through MAPK pathway crosstalk (Fig. S4). It could be that, in the septin-null mutants, septin filaments are not able to properly assemble and monitor lipid composition and protein organization at the cytosolic face of membranes, resulting in defects in plasma membrane and CWI (Fig. 8).

Although we have discussed membrane and CWI separately, it is possible that membrane defects in the septin-null mutants contribute to the observed cell wall changes or that cell wall defects contribute to the observed changes in lipid metabolism. When septin deletion mutants were challenged with both membrane-disturbing and cell wall-disturbing agents in combination, rescue of the lipid defect (via PHS) restored proper growth, but rescue of the cell wall defect (via sorbitol) did not rescue lethality (Fig. 6C,D). This suggests that there is a synergistic effect of disrupting sphingolipid metabolism and cell wall architecture in septin-null mutants and that septin–sphingolipid interactions are required for roles in maintaining CWI.

Because A. nidulans septin deletion mutants are viable, we were able to systematically analyze the roles of all septins in this organism. Our data show that all septins are required for proper coordination of lipid metabolism, with core hexamer septins most important for sphingolipid metabolism and all septins involved in ergosterol metabolism. Our data also show that the core septins are most important for the CWI pathway, but that these roles require sphingolipids. Based on our data, we propose that septins are critical for tight coordination of plasma membrane metabolism and cell wall synthesis during normal development and response to exogenous stress, and that the site of this coordination is likely SRDs. Future work will address how subgroups of septins interact with each other and with sphingolipids and ergosterol in SRDs to coordinate the CWI response.

Strains used in this study are listed in Table S1. Media used were previously described (Harris et al., 1994). Aspergillus nidulans strains were harvested in double-distilled water (ddH₂O), spores normalized to 10⁷ or 10⁶ conidia/ml, and then serial diluted 4-fold into separate Eppendorf tubes. All strains were inoculated in 10 µl droplets in a grid pattern on Petri plates containing 25 ml of solid (1.8% agar) complete medium (CM) (1% glucose, 0.2% peptone, 0.1% casamino acids, 0.1% yeast extract, trace elements, nitrate salts and 0.01% vitamins, pH 6.5; with amino acid supplements as noted) or minimal medium (MM) (1% glucose, trace elements, 1% thiamine, 0.05% biotin, pH 6.5; with amino acid supplements as noted) with or without amended supplements. All incubations were conducted at 30°C as indicated for 3–4 days before images were taken. Sorbitol, NaCl and KCl were added at 2 M, 1 M and 1.5 M, respectively, to media before autoclaving. Stock solutions were prepared as follows: Blankophor BBH/CFW (Bayer Corporation; standard-SV-2460; 25 mg/ml in ddH₂O and adjusted pH with 1 M KOH until solubilized), Congo Red (Fisher Scientific; Lot No.8232-6; 10 mg/ml in ddH₂O), caspofungin acetate (1 mg/ml in ddH₂O), fludioxonil/pestanal (2 mg/ml in DMSO) and caffeine monohydrate (63.66 mg/ml in ddH₂O and gently heated with stirring until solubilized); calcium chloride (2 M in ddH₂O), EGTA (0.5 M in ddH₂O), rapamycin/sirolimus (1 mg/ml in acetonitrile), FK-506 (5 mg/ml in DMSO), natamycin (1.5 mg/ml in MeOH), itraconazole (10 mg/ml in DMSO), terbinafine (10 mg/ml in DMSO). Stock solutions of myriocin (5 mg/ml), phytosphingosine (1 mg/ml) and aureobasidin A (1 mg/ml) were prepared in DMSO, ethanol and methanol, respectively and stored at −20°C in the dark. Images of plates were captured using a cellular device with an 8.0 Megapixel camera, and subsequently processed using Photoshop CS5 Version 12.0 X32.

Parental strains were co-inoculated in CM liquid at 10⁵ conidia/ml supplemented with all parental auxotrophic markers and allowed to incubate at 30°C for up to 1 week or until a thick mycelial mat had formed. Mycelial mats were transferred to solid MM plates containing only shared auxotrophic markers from the genetic backgrounds of each parental strain. Plates containing mycelial mats were covered with parafilm and incubated in the dark at room temperature for up to 2 weeks or until mature cleistothecia formed on the mycelial mats. Multiple cleistothecia from each genetic cross were collected in water, diluted and plated onto solid medium containing all auxotrophic supplements from each parental strain to allow growth of all resulting progeny. Approximately 50 progenies were collected from each dissected cleistothecium, and each colony was transferred to master plates and replica-plated onto MM without any supplements, to isolate prototrophic progeny. Five to ten progenies from each cross were then 3-phased streaked to obtain single colony isolates for PCR verification. All progenies of genetic crosses were verified by diagnostic PCR using KOD XTREME Hot Start DNA Polymerase (71975-3, EMD Millipore) or OneTaq^® Hot Start Quick-Load^® 2X Master Mix with Standard Buffer (M0488L, New England BioLabs Inc.) according to the manufacturer's instructions. Double mutant strains of ΔmpkA^slt2 were verified for deletion of mpkA^slt2 by amplification of entire gene using primers, MpkA-806-F’ and MpkA-3779-R’, followed by SacI HF restriction enzyme digestion of PCR product to better visualize band sizes. All progeny from chitin synthase and septin-null mutant crosses were determined to be virtually identical to each septin mutant parental strain in growth/morphology on a colony level and by microscopy, PCR-verified for the deletion of each septin gene, and screened visually by fluorescence microscopy for the presence of chitin synthase–GFP signal. Chitin synthase–GFP strains in a ΔaspA^cdc11 genetic background were verified for deletion of aspA^cdc11 by amplification of the entire gene using primers, AspA-KO-F’ and AspA-KO-R’, followed by XhoI restriction enzyme digestion for verification of band sizes. Strains and primer sets used in this study can be found in Tables S1 and S2, respectively.

Growth and preparation of cells were as previously reported (Momany et al., 1999). Conidia were inoculated on sterile coverslips in 10 ml liquid complete medium or MM at 10⁵ conidia/ml and incubated at 30°C in a small Petri dish for the specified amount of time. Cell walls were stained for chitin with Blankophor BBH (CFW) (American Cyanamid, Wayne, NJ; 25 mg/ml stock solution in ddH₂O and pH adjusted by 1 M KOH until solubilized; working solution made by diluting stock solution by 100× and 8 μl dissolved in 5 ml ddH₂O prepared fresh for working solution and used immediately), β-(1,3)-glucans were stained with Aniline Blue (stock solution prepared fresh to 10 mg/ml final concentration in ddH₂O; working solution prepared at 135.55 μM in 50 mM phosphate buffer, pH adjusted to 9.5 with 5 M KOH, and used immediately; coverslips stained for 5 min in the dark at room temperature). Live-cell imaging experiments tracking septin localization were conducted using filipin III (stock solution prepared at 5 mg/ml in DMSO; working solution was used at 25 μg/ml final concentration in liquid complete medium), phytosphingosine (working solution was used at 15 µM in liquid complete medium) and wortmannin (stock solution prepared to 2 mg/ml in DMSO; working solution was used at 20 μg/ml in liquid complete medium). Vehicle controls and conducted at ≤1% (w/v) in liquid medium. Imaging was performed in the Biomedical Microscopy Core at the University of Georgia.

Microscopy was carried out using Zeiss Axioplan microscope and Zeiss Axiocam MRc charge-coupled device camera and software, as well as a Deltavision I deconvolution inverted microscope and LSM880 confocal fluorescent microscope with diode laser (405 nm), argon (458, 488, 514 nm) and HeNe (543, 633 nm) laser lines. All micrograph comparisons between treatments imaged with identical microscope settings. Subsequent image analysis and scale bars added to micrographs, using ImageJ software 1.48v, Java 1.6.0_20 (64-bit) or Zen 2.3 imaging software, and final figures compiled in Photoshop CS5 software version 12.0 ×32.

Maximum projection images were made and scale bars added for each micrograph in Zen (Black) Edition. Micrographs were false colored black and white, line scans were drawn by hand through hyphal tips, line scan profiles were generated in Zen 3.1 (Blue) imaging software, and associated data points for each profile were exported to Microsoft Excel version 2004 (Build 12730.20270) to create graphs and conduct statistical analysis. Figures were generated using Adobe Photoshop 2020 Version 21.1.2.

Maximum projection images were made for each micrograph, line scans were drawn by hand through hyphal branch tips in ImageJ software 1.48v, and data points were exported to Microsoft Excel version 2004 (Build 12730.20270) to create graphs and conduct statistical analysis. Figures were assembled using Adobe Photoshop 2020 Version: 21.1.2.

Cell walls were isolated from a protocol based on Bull (1970); cell wall extraction and lyophilization were conducted as previously described in Guest and Momany (2000) with slight modifications listed in full procedure here. A single batch of complete medium (as described above) was autoclaved and supplements added: arginine, methionine, pyridoxine and riboflavin, were added to single flask and then distributed to individual flasks to be inoculated. 10⁴ conidia/ml were inoculated in two flasks each of 100 ml liquid complete medium. Flasks were incubated at 30°C in orbital shaker at 200 rpm for 48 h. Mycelia was gravity filtrated, then vacuum-filtrated through #42 Whatman filter paper, subsequently washed with 50 ml each of chilled ddH₂O to remove residual medium and stored at −80°C until completely frozen. Mycelium from each sample was allowed to thaw on ice and then washed sequentially with 50 ml chilled ddH₂O and 0.5 M NaCl. Fungal hyphal mats were transferred to mortar and pestle, then subsequently flash frozen in liquid nitrogen and ground in chilled Tris/EDTA disruption buffer (DB; 20 mM Tris-HCl pH 8.0, 50 mM EDTA) with a pre-chilled mortar and pestle. Samples were monitored by microscopy under a 60× or 100× objective until hyphal ghosts were evident. Cell walls were separated by centrifugation at 13,800 g for 10 min at 4°C. The cell pellet was placed in a beaker with 40–100 ml of chilled Tris/EDTA buffer and stirred at 4°C for 12 h. The cell pellet was collected by centrifugation as above and stirred again with 100 ml chilled ddH₂O at 4°C for 4 h. Cell wall material was collected by vacuum filtration, frozen at −80°C, lyophilized until dry and stored at room temperature (25°C) for further analysis.

To determine the glycosyl linkages, the samples were acetylated using pyridine and acetic anhydride in order to get better solubility, before two rounds of permethylation using sodium hydroxide (15 min) and methyl iodide (45 min). The permethylated material was hydrolyzed using 2 M trifluoroacetic acid (TFA; 2 h in sealed tube at 121°C), reduced with NaBD₄ and acetylated using acetic anhydride/pyridine. The resulting partially methylated alditol acetates (PMAAs; neutral sugars) were analyzed on an Agilent 7890A GC interfaced to a 5975C MSD (mass selective detector, electron impact ionization mode); separation was performed on a 30 m Supelco SP-2331 bonded phase fused silica capillary column using Supelco SP-2331 fused silica capillary column [30 m×0.25 mm internal diameter (ID)]. The PMAAs of amino sugars were separated on Supelco Equity-1 column (30 m×0.25 mm ID). Furthermore, the relative quantities of respective glycosyl linkages were calculated by integrating the peak area of respective peak. Since the neutral and amino sugars were analyzed on different instruments, the peak area of amino sugars was normalized to the value for the 4-Glc peak, which is prominent in both the instruments and the integrated and normalized peak areas were pooled together to calculate the relative percentage of individual linkages. Two independent, biological replicates were conducted for this analysis and processed in tandem. The average area (%) of detected linkages of one representative data set is included in the graphs to show relative differences between cell wall polysaccharides between samples.

Chitin synthase strains were generously provided by Dr Hiroyuki Horiuchi in the Department of Biotechnology at the University of Tokyo.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258336