Abstract

Cell-to-cell fusion is crucial for the development and propagation of most eukaryotic organisms. Despite this importance, the molecular mechanisms mediating this process are only poorly understood in biological systems. In particular, the step of plasma membrane merger and the contributing proteins and physicochemical factors remain mostly unknown. Earlier studies provided the first evidence of a role of membrane sterols in cell-to-cell fusion. By characterizing different ergosterol biosynthesis mutants of the fungus Neurospora crassa, which accumulate different ergosterol precursors, we show that the structure of the sterol ring system specifically affects plasma membrane merger during the fusion of vegetative spore germlings. Genetic analyses pinpoint this defect to an event prior to engagement of the fusion machinery. Strikingly, this effect is not observed during sexual fusion, suggesting that the specific sterol precursors do not generally block membrane merger, but rather impair subcellular processes exclusively mediating fusion of vegetative cells. At a colony-wide level, the altered structure of the sterol ring system affects a subset of differentiation processes, including vegetative sporulation and steps before and after fertilization during sexual propagation. Together, these observations corroborate the notion that the accumulation of particular sterol precursors has very specific effects on defined cellular processes rather than nonspecifically disturbing membrane functioning. Given the phenotypic similarities of the ergosterol biosynthesis mutants of N. crassa during vegetative fusion and of Saccharomyces cerevisiae cells undergoing mating, our data support the idea that yeast mating is evolutionarily and mechanistically more closely related to vegetative than sexual fusion of filamentous fungi.

Fusion between genetically and developmentally identical or different cells is a common process in the growth and development, of eukaryotic organisms. It serves a plethora of biological functions including gamete merger during sexual propagation, muscle and organ development, neuronal repair, or the formation of multinucleate cells, such as macrophage-derived osteoclasts or giant cells. In filamentous fungi, cell-to-cell fusion promotes multicellular growth and the formation and functioning of mycelial colonies. In many ascomycete species, the cell tips of germinating vegetative spores (conidia) mutually attract each other during colony establishment and fuse into supracellular networks (germling fusion) that further develop into the hyphal colony (Roca et al. 2005). This cooperation of genetically identical individuals likely promotes competitiveness and allows the coordinated conquering of a specific habitat by clonal spores (Richard et al. 2012). Within mature mycelial colonies, fusion between hyphal branches (hyphal fusion) increases the interconnectivity within the cellular network, promoting the exchange of water and nutrients throughout the colony, and supporting its coordinated growth and propagation.

In addition, the sexual reproduction of outcrossing fungal species involves the merger of two genetically different mating partners. Fertilization is typically achieved through the fusion of specialized reproduction structures or cells to enable subsequent nuclear fusion and genetic recombination (Bistis 1981).

Despite the broad occurrence and high relevance of cell-to-cell fusion, the molecular mechanisms mediating this process remain only poorly understood in any biological system. In particular, the last step of the cell-to-cell fusion process, the merger of the plasma membranes, remains enigmatic. In recent years, the red bread mold Neurospora crassa has advanced as a model system for studying the molecular mechanisms of eukaryotic cell-to-cell fusion in general, and fungal colony initiation and development in particular (Herzog et al. 2015; Fischer and Glass 2019). In vegetative germling fusion, N. crassa employs an unusual communication mechanism, in which the two fusion partners take turns in signal sending and receiving. As a result of this cellular dialog, the germlings direct their growth toward each other until physical contact is achieved (Fleissner et al. 2009a; Fleißner and Herzog 2016). While the molecular factors governing these cellular interactions are starting to unfold, the mechanisms mediating the fusion process after cell-to-cell contact remain mostly elusive. Fusion pore formation requires cell wall remodeling, followed by adherence and subsequent fusion of the plasma membranes. So far, no bona fide fusogen, a protein required and sufficient for plasma membrane merger, has been identified in fungi (Aguilar et al. 2013). However, transmembrane proteins promoting plasma membrane fusion during vegetative and sexual cell merger have been described in Saccharomyces cerevisiae, Schizosaccharomyces pombe, and N. crassa. In all species, a significant fraction of fusion pairs lacking the PRM1 protein fail to merge the plasma membranes and remain with tightly adhered lipid bilayers (Heiman and Walter 2000; Fleissner et al. 2009b; Curto et al. 2014). A similar but less-severe defect is observed in Δlfd-1 mutants of N. crassa, in which about one-fifth of fusion pairs fail in membrane merger (Palma-Guerrero et al. 2014). PRM1 and LFD-1 mediate both vegetative and sexual fusion, suggesting that both are part of a general fusion machinery. In addition to the membrane merger defect, all of the described S. cerevisiae and N. crassa mutants exhibit significantly increased lysis of fusion pairs. The onset of lysis coincides with the moment of cytoplasmic mixture, suggesting that it is caused by the engagement but aberrant functioning of the fusion machinery (Aguilar et al. 2007; Palma-Guerrero et al. 2014).

In addition to fusion-mediating proteins, plasma membrane merger is also influenced by the composition of the lipid bilayer itself. Crucial lipid constituents of eukaryotic cell membranes are sterols, which have essential roles for membrane properties and the functioning of membrane-associated proteins (Dufourc 2008). Different S. cerevisiae and N. crassa mutants deficient in the biosynthesis of ergosterol, the prevalent sterol in fungal cell membranes, are affected in cell fusion-related cellular communication or plasma membrane merger (Jin et al. 2008; Aguilar et al. 2010; Weichert et al. 2016). Strikingly, specific phenotypic deficiencies of some ergosterol biosynthesis (erg) mutants correlate with specific structural features of the accumulating sterol precursors. For example, a defect in the last step of ergosterol biosynthesis mediated by the sterol C-24(28) reductase ERG-2 results in the accumulation of a precursor carrying a conjugated double bond in its aliphatic side chain. As a consequence, signaling through the MAK-1 mitogen-activated protein kinase (MAPK) cascade, which is essential for germling communication and cell-to-cell contact recognition in N. crassa, is strongly affected, while other related membrane-associated signaling processes, including a second MAPK pathway (MAK-2), remain functional (Weichert et al. 2016). These observations indicate that changes in the sterol structure have specific rather than general effects on cell functioning.

In S. cerevisiae, mutants accumulating ergosterol precursors with an altered double-bond configuration in the B ring (Δerg2 or Δerg3, Figure 1A) exhibit a Δprm1-like phenotype, with a significant percentage of mating pairs remaining in the stage of closely adhered plasma membranes. Despite the morphological similarities of the Δprm1 and erg mutants, both defects act independently, indicated by additive effects in the respective double mutants (Jin et al. 2008). However, so far the molecular processes affected by the accumulation of intermediates of ergosterol biosynthesis in these erg mutants remain unknown.

Figure 1

The sterol C-5 desaturases ERG-10a and ERG-10b have a redundant function in ergosterol biosynthesis in N. crassa. (A) Structure of ergosterol. Enzymes and their functions mediating double-bond arrangements in ergosterol precursors are depicted (S. cerevisiae/N. crassa). Numbers along ergosterol denote the position of selected carbon atoms, and letters indicate the carbon rings. (B) Late steps in the hypothetical biosynthesis pathway of ergosterol in N. crassa showing the reactions catalyzed by the enzymes presented in (A). ERG-10a and ERG-10b mediate the same reaction, whose block results in the accumulation of a sterol intermediate with an altered ring structure (gray arrow).

Here, we show that structural changes in the ring system of ergosterol lead to membrane fusion deficiencies during vegetative cell-to-cell fusion in N. crassa, indicating functional conservation between yeast mating and vegetative fusion in the filamentous fungus. However, sexual cell-to-cell merger remains unaffected in the erg mutants of N. crassa, indicating that the altered membrane structure does not compromise the general basic fusion machinery. Consistent with these findings, we could pinpoint the sterol-dependent defect in plasma membrane merger to a step prior to the engagement of the fusion machinery, suggesting a novel role of membrane structure in the signaling events preceding the initiation of plasma membrane fusion.

Materials and MethodsStrains and growth conditions

The strains of N. crassa used and generated in this study are listed in Supplemental Material, Table S1. Strains were routinely grown at 30° in agar slant tubes containing Vogel’s minimal medium (MM) (Vogel 1956) as previously described in Schürg et al. (2012). Crosses were conducted on Westergaard’s medium (Westergaard and Mitchell 1947) as described before (Serrano et al. 2018). Spores were obtained from 1-week-old cultures through harvesting in sterile distilled water. Spore suspensions were filtered through Miracloth (Calbiochem, San Diego, CA) and counted with a hemocytometer.

Linear hyphal growth rates were determined as described in Fleissner et al. (2005). To quantify sporulation, the strains were first grown in MM slant tubes for 7 days at 26° with a 12-hr light–dark rhythm. Conidia were harvested in 2 ml of sterile distilled water and purified from hyphal fragments by filtration through Miracloth. The spore numbers per tube were determined with a hemocytometer. To determine the lengths of aerial hyphae, 2 ml of liquid MM in a long reaction tube inoculated with 104 spores were incubated for 5 days at 30° in the dark.

To compare the sensitivity of strains to nystatin, agar plates with sorbose-rich medium, which induces colonial growth in N. crassa (Tatum et al. 1949), were spotted with 5-µl droplets of 10-fold serial dilutions of spore suspensions, resulting in 105–10 conidia per spot. The media were supplemented with freshly prepared solutions of nystatin dihydrate (100 µg/ml in DMSO; Roth). Photos of the plates were captured after incubation at 30° for 72 hr.

Strain constructionComplementation of Δerg-10a/Δerg-10b:

To complement the double mutant Δerg-10a/Δerg-10b, strain MW_126 was transformed with plasmids containing either the gene erg-10a (NCU06207) or erg-10b (NCU04983), including 1000 bp upstream and downstream of these sequences. To construct the plasmids, both genes were PCR-amplified from genomic DNA of the wild-type (WT) strain FGSC (Fungal Genetics Stock Center) 2489 using the primer pairs 1445/1446 and 1447/1448, respectively (Table S2). The PCR products were cloned via PacI and NotI into vector 728 (a pMF272 [Freitag et al. 2004] derivative lacking the gfp sequence), enabling the site-specific integration at the his-3 locus of the N. crassa genome (Freitag and Selker 2005). The resulting plasmids 745 (erg-10a) and 746 (erg-10b) were transformed into the histidine-auxotrophic double mutant MW_126 as described in Margolin et al. (1997).

Construction of erg-1 gene deletion strains:

For the deletion of erg-1 in N. crassa, a gene replacement cassette for homologous recombination was constructed using yeast recombinational cloning (Raymond et al. 1999) as described in Colot et al. (2006). The 5′ and 3′ untranslated regions of the erg-1 gene were PCR-amplified from genomic DNA of strain FGSC 9719 with primer pairs 103/104 and 105/106, respectively (Table S2). Using the S. cerevisiae strain FY834 for the yeast transformation (Winston et al. 1995), these flanks were assembled with the hygromycin resistance gene amplified with primers 82/83 from plasmid pCSN44 (Staben et al. 1989) and the EcoRI/XhoI-linearized yeast vector pRS426 (Christianson et al. 1992) as described before (Colot et al. 2006). The plasmid DNA extracted from the yeast colonies was used as a template in a PCR with primers 103 and 106 to amplify the assembled erg-1 gene deletion cassette. Transformation of N. crassa with the gene knockout cassette was performed as previously described (Fleissner and Glass 2007), using the mus-52::bar deletion mutant FGSC 9719 as the recipient strain for efficient homologous recombination (Colot et al. 2006). Transformants were confirmed by Southern blot analysis and purified by backcrossing with the WT strain. To generate Δerg-1 strains expressing cytosolic GFP and mCherry, a mutant auxotrophic for histidine (MW_347) was transformed with plasmids pMF272 (Freitag et al. 2004) and pMFcherry (Schürg et al. 2012), respectively. Accordingly, a histidine-auxotrophic double deletion strain lacking both erg-1 and Prm1 (MW_450) was transformed in the same way to generate ΔPrm1/Δerg-1 strains expressing GFP and mCherry, respectively.

Construction of erg mutants lacking Prm1:

Mutants lacking erg genes and Prm1 were obtained via sexual crosses (data not shown). The strains lacking erg-10a, erg-10b, and Prm1, and expressing cytosolic GFP and mCherry, respectively, were constructed by transforming a histidine-auxotrophic triple mutant (MW_277.1) with the vectors pMF272 or pMFcherry.

Cell-to-cell fusion assays

Vegetative spore germling fusion assays were conducted as previously described (Fleissner et al. 2009b). In brief, conidial suspensions were spread on MM agar plates, incubated at 30° for 4 hr and analyzed by brightfield and fluorescence microscopy. Agar blocks of ∼10 × 10 mm were cut from the plates, covered with coverslips, and analyzed by brightfield and fluorescence microscopy. The number of germlings undergoing directed growth toward each other was determined from microscopic images with comparable cell densities as previously described (Weichert et al. 2016). The quantitative analysis of cell-to-cell fusion between germlings expressing green- and red-fluorescing proteins was performed as described in Fleissner et al. (2009b). For live-cell imaging of SO-GFP and MAK-2-GFP in interacting germlings, samples were prepared as described above and analyzed by fluorescence microscopy (Fleissner et al. 2009a). Sexual fusion between trichogynes and microconidia was assessed as described in Fleissner et al. (2005).

Analysis of fruiting body development

The phenotypic characterization of fruiting bodies and progeny was performed with a stereomicroscope as described in Weichert et al. (2016). To determine the number of protoperithecia per area unit, strains were grown on crossing plates for 7 days as described above. The quantification of the protoperithecia was based on representative images captured from at least three independent cultures. The average number of protoperithecia per 1 mm2 was determined using the software Fiji (https://fiji.sc/). Images of perithecia were taken 10–14 days after fertilization, either as viewed from above the crossing plates or at an angle of ∼45° relative to the camera. Rosettes of asci from crosses were prepared by placing isolated perithecia into a droplet of water on a glass slide and gently squeezing them under a coverslip. At least 1 month after fertilization, ejected ascospores were harvested in 2 ml of sterile distilled water from the lids of the crossing plates, heat-activated at 59° for 20 min, spread on MM plates, and incubated at room temperature for 16–20 hr. Germinating ascospores were analyzed with a stereomicroscope and isolated in MM slant tubes for growth into vegetative mycelia.

Staining of the plasma membrane and cell wall of germling pairs

The staining of cell membranes with FM4-64 was performed as described in Hickey et al. (2002). FM4-64 was prepared as a solution of 100 µg in 1 ml of sterile distilled water, which was further diluted to working solutions of 25 µM.

To simultaneously stain the cell membrane and the cell wall in germling pairs, samples were treated with 10 µl of a mixture of FM4-64 and calcofluor white (CFW, fluorescence brightener 28). First, CFW was prepared as a stock solution of 7 mg/ml in DMSO and then diluted 1:100 in sterile distilled water. Next, one volume of the working solution of CFW was mixed with nine volumes of the working solution of FM4-64, which were immediately used for fluorescence microscopy of the double-stained germling pairs.

Analysis of cell lysis in germling pairs

Fusion-induced cell lysis was quantified as previously described in Schumann et al. (2019), using MM with normal and reduced Ca2+ concentrations (0.68 and 0.34 mM, respectively).

Visualization of sterol distribution in germlings and trichogynes

To determine the distribution of plasma membrane sterols in interacting germlings, samples were prepared as described above and treated with 10 μl containing 100 μg/ml of filipin III (Sigma [Sigma-Aldrich], St. Louis, MO) in 1% (v/v) DMSO. Trichogynes were obtained from crosses as described in Fleissner et al. (2005), followed by the treatment of samples (agar blocks of 5 × 5 mm) with 5 µl of the filipin III solution between 24 and 48 hr after incubation. Cells were analyzed within 15 min after filipin staining by fluorescence microscopy using a DAPI filter setup.

Brightfield and fluorescence microscopy

Images of germlings were captured with an Axiophot 2 microscope (Zeiss [Carl Zeiss], Jena, Germany) coupled to a pixelfly 270 XS (pco) camera, using 63×/1,40 oil plan-apochromat and 100×/1,30 oil plan-neofluar objectives and DIC optics. The double-staining of germlings with FM4-64 and CFW was performed using an Axio Observer.Z1 inverse microscope (Zeiss) using light-emitting diodes for the excitation light. Brightfield images of fruiting bodies and asci were captured with an M60 stereomicroscope (Leica) coupled to a DFC295 camera.

Sterol extraction and analysis

The extraction and analysis of sterols was performed as described in our previous study (Weichert et al. 2016). In brief, sterols were isolated from frozen, ground mycelia using a 3:1 (v/v) mixture of chloroform and methanol, subsequent saponification with 10% (w/v) potassium hydroxide in methanol, followed by extraction in n-hexane. Sterol extracts were derivatized with 2,2,2-trifluoro-N-methyl-N-(trimethylsilyl) acetamide and analyzed by gas chromatography/mass spectrometry (GC/MS) as described in (Nawrath et al. 2010). The sterol biosynthesis intermediates were identified by comparing their mass spectra with data available from databases and the literature.

Statistical analysis

Statistical data analysis was performed with GraphPad Prism (version 8.3.1) using analysis of variance (ANOVA) with Tukey’s or Dunnett’s multiple comparison tests, or unpaired, two-tailed Student’s t-tests, to evaluate statistically significant differences in fungal growth, cell-to-cell fusion, and cell lysis.

Data availability

Strains and plasmids are available upon request. The authors affirm that all reagents, software, and data that are necessary for drawing the conclusions presented in this study are fully represented within this article, and its tables and figures. Supplemental material available at figshare: https://doi.org/10.25386/genetics.13050176.

ResultsThe absence of the redundant sterol C-5 desaturases ERG-10a and ERG-10b affects specific differentiation steps in the life cycle of N. crassa

In earlier studies, we reported that the sterol C-5 desaturases ERG-10a and ERG-10b, encoded by the gene loci NCU06207 and NCU04983, respectively, have a redundant function in ergosterol biosynthesis in N. crassa (Weichert et al. 2016; Herzog et al. 2020). They introduce a double bond between C-5 and C-6 of the B ring in the sterol molecule (Figure 1A). Their absence results in the accumulation of ergosta-7,22-dienol, which differs from ergosterol only by the lack of this specific double bond (Figure 1B).

To test if this metabolic defect influences growth and development of the fungus, we characterized respective gene knockout mutants. Consistent with the redundant function of ERG-10a and ERG-10b, mutants lacking the individual genes (FGSC 20057 and N4-30, respectively) were macroscopically indistinguishable from the WT reference strain FGSC 2489 (Figure 2A). However, a Δerg-10a/Δerg-10b double mutant (N4-38) showed a strong decrease in asexual spore differentiation, whereas the overall growth rate was only slightly affected (Figure 2B) and aerial hyphae formation was WT-like (data not shown). In contrast to the WT or single mutants, cultures of the double mutant also accumulated a brownish pigment (Figure 2A). Reintegration of either the erg-10a or erg-10b gene into the double mutant restored the WT phenotype, confirming the redundant functions of the encoded enzymes (Figure 2, A and B).

Figure 2

Impact of Δerg-10a and Δerg-10b on vegetative growth of N. crassa. (A) Growth of WT (FGSC 2489), Δerg-10a (FGSC 20057), Δerg-10b (N4-30), Δerg-10a/Δerg-10b (N4-38), Δerg-10a/Δerg-10b + erg-10a (MW_630), and Δerg-10a/Δerg-10b + erg-10b (MW_631) in MM tubes. Insets: in contrast to WT (i), Δerg-10a/Δerg-10b (ii) typically produces dark pigments along the glass wall of the tube. (B) Quantification of linear hyphal extension rates and sporulation. Values represent the mean ± SD from at least three independent experiments per strain (****P < 0.0001, ANOVA with Tukey’s post hoc test). (C) Ten-fold serial dilutions of conidia of the indicated strains were spotted onto sorbose-rich medium in the absence or presence of the polyene antifungal compound nystatin. FGSC, Fungal Genetics Stock Center; MM, minimal medium; WT, wild-type.

Due to their lack of ergosterol, classical erg mutants of N. crassa generated by UV mutagenesis show increased resistance to ergosterol-binding antifungals called polyenes (Grindle 1973, 1974). The Δerg-10a/Δerg-10b isolate was also less sensitive to the polyene nystatin than the WT, the single-gene deletion mutants, or the complemented strains (Figure 2C), consistent with a block in ergosterol biosynthesis in the double mutant. Taken together, the presence of either ERG-10a or ERG-10b sufficiently supports WT-like growth of N. crassa, while the absence of both sterol C-5 desaturases results in a specific deficiency in spore differentiation.

The lack of ERG-10a and ERG-10b results in plasma membrane fusion deficiencies during vegetative germling fusion

Our earlier study revealed that Δerg-2 germlings interact at a significantly reduced frequency and are unable to arrest growth after cell-to-cell contact. In contrast, directed growth of Δerg-10a/Δerg-10b germlings is comparable to WT (Weichert et al. 2016). The interaction of WT germlings includes membrane recruitment of the MAPK MAK-2 and the fungal-specific signaling factor SO (Fleissner et al. 2009a). As expected, the membrane recruitment of both proteins was also unaffected in the Δerg-10a/Δerg-10b double mutant (Figure S1). The structural differences of ergosta-7,22-dienol and ergosterol therefore have no effect on the membrane-associated signaling processes preceding cell-to-cell contact, in contrast to the changed aliphatic side chain of the sterol accumulating in Δerg-2.

To test for potential fusion defects after cell-to-cell contact, germlings from WT or Δerg-10a/Δerg-10b strains expressing either cytosolic GFP or mCherry were mixed, and incubated to allow cell-to-cell interactions. Cell pairs consisting of a red and a green cell were analyzed by fluorescence microscopy. WT cell pairs mostly exhibited both fluorescent markers in either fusion partner, indicating cytoplasmic mixing as a result of cell-to-cell fusion (Figure 3A). In contrast, >30% of Δerg-10a/Δerg-10b pairs failed to exchange their cytoplasm, although the cells were tightly adhered to one another (Figure 3, B and C). We also analyzed the Δerg-11 and Δerg-2/Δerg-11 mutants, which interact in a WT-like manner (Weichert et al. 2016). These isolates were not affected in their fusion frequencies (88.2 ± 0.8% and 87.4 ± 1.5%, respectively), indicating that the Δerg-10a/Δerg-10b fusion deficiency represents a novel phenotype among all of the erg mutants of N. crassa tested so far.

Figure 3

Germlings of ∆erg-10a/∆erg-10b are impaired in cell-to-cell fusion. (A) Microscopic analysis of cell-to-cell fusion in a pair of WT germlings expressing cytosolic GFP and Cherry Red (N3-06 + N3-07) after physical contact (arrowhead), resulting in mixing of green and red fluorescence. (B) A pair of Δerg-10a/Δerg-10b germlings expressing cytosolic GFP or Cherry Red (MW_175 + MW_179). Note the lack of cytoplasmic mixing and the extrusion extending from one cell into the other (arrow). Bar in (A and B), 5 μm. (C) Quantification of germling fusion in cell pairings of WT (N3-06 + N3-07), Δerg-10a (N5-01 + N5-02), Δerg-10b (MW_103 + MW_105), and Δerg-10a/Δerg-10b (MW_175 + MW_179). Values represent the mean ± SD from at least three independent cell populations with 87–109 germling pairs per replicate (****P < 0.0001, ANOVA with Tukey’s post hoc test). WT, wild-type.

In nonfused Δerg-10a/Δerg-10b cell pairs, one cell often formed a finger-like structure, which reached into the cytoplasm of the opposing cell (Figure 3B). This phenotype strongly resembled the earlier reported membrane fusion defect of S. cerevisiae or N. crassa Δprm1 mutants. In both species, 50% of mutant cell pairs fail to merge their tightly opposed plasma membranes after successful cell wall deconstruction (Heiman and Walter 2000; Fleissner et al. 2009b). The comparison of green- and red-fluorescing cell pairs of Δerg-10a/Δerg-10b and ΔPrm1 readily revealed comparable finger- or bubble-like structures in both mutants (Figure S2). This striking phenotypic similarity suggested that the Δerg-10a/Δerg-10b mutant is also defective in plasma membrane merger after cell wall remodeling. Costaining of the cell wall and the plasma membrane confirmed that unfused ΔPrm1 and Δerg-10a/Δerg-10b germling pairs both showed a WT-like opening in the cell wall, but no subsequent fusion pore formation (Figure 4A). Complementation of the double mutant with either the erg-10a or erg-10b gene fully restored WT-like fusion pore formation (Figure S3). These observations indicate that the altered sterol composition in the Δerg-10a/Δerg-10b mutant specifically affects plasma membrane fusion after cell wall degradation.

Figure 4

Germlings of the mutants Δerg-10a Δerg-10b and Δerg-1 are defective in plasma membrane merger during cell-to-cell fusion. (A) Costaining with CFW and FM4-64 to visualize the cell wall and the plasma membrane, respectively. Black arrowheads indicate cell contacts. Note that WT (FGSC 2489) pairings form both an opening in the cell wall (white arrowhead) and a fusion pore in the membrane (asterisk). The mutants ΔPrm1 (A32), Δerg-10a/Δerg-10b (N4-38), and Δerg-1 (MW_308) form membrane invaginations (arrows) that stretch through the cell wall opening from one into the other cell (white arrowheads). Bar, 5 μm. (B) Quantification of germling fusion of erg mutants carrying the additional deletion of Prm1. Values represent the mean ± SD from three independent experiments per set of strains with a range of 40–211 germling pairs per population (**P < 0.01, ***P < 0.001, ****P < 0.0001; ANOVA with Tukey’s post hoc test). See Figure S2 for images of representative germling pairs from these experiments. CFW, calcofluor white; ns, not significant; WT, wild-type.

The germling fusion defects of Δerg-10a/Δerg-10b are independent of PRM1 and LFD-1

The striking resemblance of the membrane fusion defects of the ΔPrm1 and Δerg-10a/Δerg-10b mutants suggested that accumulation of the ergosterol precursor could specifically impair the function of the PRM1 protein. To test this hypothesis, we employed classical genetics and analyzed membrane fusion in a ΔPrm1/Δerg-10a/Δerg-10b triple mutant. As expected, the deletion of all three genes also caused membrane invaginations similar to those found in cell pairs of ΔPrm1 or Δerg-10a/Δerg-10b (Figure S2). However, the rate of successful fusion dropped in the triple mutant to <10% compared with ∼50% and 60% in ΔPrm1 and Δerg-10a/Δerg-10b, respectively (Figure 4B). These data indicate that the fusion defects of ΔPrm1 and Δerg-10a/Δerg-10b are independent, suggesting that the altered sterol composition of this erg mutant does not impact PRM1 functions.

To further investigate this question, we quantified fusion in heterotypic germling pairs. Consistent with previous observations (Heiman and Walter 2000; Fleissner et al. 2009b), the fusion frequencies of ΔPrm1 and Δerg-10a/Δerg-10b cells approached almost normal levels (∼85%) when one of the interaction partners was WT (Figure 4B). Similarly, mixed germling pairs consisting of a Δerg-10a/Δerg-10b and a ΔPrm1 cell exhibited 80% fusion success, indicating that the mutants can mutually compensate their deficiencies.

In N. crassa, the LFD-1 protein possesses a redundant role that is independent of PRM1 during plasma membrane merger (Palma-Guerrero et al. 2014). Consistent with these earlier studies, we also observed the formation of membrane invaginations after cell-to-cell contact with a relatively mild effect on membrane fusion in a Δlfd-1 mutant (Figure S4, A–C). A triple mutant lacking lfd-1, erg-10a, and erg-10b showed slightly further reduced fusion frequencies compared to Δerg-10a/Δerg-10b (Figure S4C). These observations suggest that LFD-1 is rather a supporting factor during plasma membrane merger that, similarly to PRM1, acts independently of the altered sterol composition of Δerg-10a/Δerg-10b.

Structural alterations in the sterol ring system in the Δerg-1 mutant also disturb plasma membrane fusion between germlings

Our earlier study revealed that the structural features of the main sterol accumulating in erg mutants can very specifically correlate with distinct developmental and cell biological defects (Weichert et al. 2016). Since the membrane fusion defect of Δerg-10a/Δerg-10b germlings was unique among all of the N. crassa ergosterol biosynthesis mutants tested so far, we hypothesized that this phenotype correlates with specific structural features of the sterol precursor formed by this mutant. Of all previously described erg mutants, Δerg-10a/Δerg-10b is the only one to accumulate ergosta-7,22-dienol (Weichert et al. 2016), which differs from ergosterol only by a missing double bond in the ring system (Figure 1, A and B). Therefore, we decided to analyze mutants of the gene erg-1 (NCU04156), which encodes a sterol C-8 isomerase homologous to Erg2 of S. cerevisiae, that contributes to the structure of the sterol ring system (Figure 1A). Earlier studies indicated that a classical erg-1 mutant obtained by random mutagenesis (FGSC 2721) accumulates sterols with a double bond between positions C-8 and C-9, instead of C-7 and C-8 (Grindle 1973, 1974; Grindle and Farrow 1978). We confirmed this mutant by sequencing of the erg-1 gene, which revealed a single point mutation that results in a Gly to Asp substitution at position 66 of the protein. Interestingly, the macroscopic appearance of the erg-1 mutant resembled the Δerg-10a/Δerg-10b strain, including the characteristic pigment production (Figure 2A and Figure S5A). Germling fusion pairs of erg-1 also showed membrane invaginations at the cell-to-cell contact zone similar to those observed in Δerg-10a/Δerg-10b (Figures S3A and S5B). For further analysis, we created a Δerg-1 gene knockout strain (Figure S5, C and D). Backcrossing of a primary transformant with WT revealed that the deletion of erg-1 significantly delayed growth and development of the mutant progeny (Figure S6). Mycelia of Δerg-1 exhibited developmental defects similar to the classical erg-1 mutant and the Δerg-10a/Δerg-10b isolate (Figure 2, A and B and Figure S5D). Consistent with the sterol composition reported for the classical erg-1 mutant and an independent deletion strain (Grindle and Farrow 1978; Hu et al. 2018), our GC/MS analysis of Δerg-1 not only confirmed a block in ergosterol biosynthesis, but also the accumulation of mainly three sterol intermediates that all contained a double bond between positions C-8 and C-9, instead of C-7 and C-8, of the ring system (Figure 1A and Figure S7, A and B). Two of these intermediates (ergosta-8,22-dienol and fecosterol) also lacked the double bond between C-5 and C-6 (Figure S7B), comparable with the sterol accumulating in Δerg-10a/Δerg-10b (Figure 1B) (Weichert et al. 2016). In summary, mutations in the genes erg-1 and erg-10a/erg-10b cause altered double bond arrangements in the B ring of the sterol core and result in comparable developmental defects.

While the frequency of interactions between Δerg-1 germlings (MW_308, 81.7 ± 0.9%) was comparable to WT (Weichert et al. 2016), cell pairs of this mutant typically formed membrane invaginations after physical contact reminiscent of Δerg-10a/Δerg-10b, ΔPrm1, and Δlfd-1 (Figure 4A and Figure S4A). Similar to Δerg-10a/Δerg-10b, the fusion frequency in Δerg-1 germlings was reduced by ∼50% (Figure 4B). Knockout of the Prm1 gene in Δerg-1 further reduced this number by another 50%, supporting the notion that the fusion defects caused by an altered sterol composition are independent of PRM1 (Figure 4B). While the triple mutant Δlfd-1/Δerg-10a/Δerg-10b had also shown an additive fusion defect, the frequency of germling fusion in the Δlfd-1/Δerg-1 double mutant (62%) was surprisingly in between those of the corresponding Δerg-1 (39%) and Δlfd-1 (87%) strains (Figure S4C). Although these observations show differences between both erg mutants with regard to LFD-1, their phenotypic similarities and additive fusion defects with ΔPrm1 indicate that their altered sterol composition affects similar molecular processes. Since PRM1 and LFD-1 have independent and redundant roles during plasma membrane merger in N. crassa (Palma-Guerrero et al. 2014), both transmembrane proteins are probably differentially affected by an altered membrane composition. In conclusion, the structural similarities of the accumulating sterols in Δerg-10a/Δerg-10b and Δerg-1 (Figure 1B and Figure S7B) cause comparable specific macroscopic and microscopic phenotypes, but appear to differentially impact membrane-associated proteins.

Membrane fusion in Δerg-10a/Δerg-10b and Δerg-1 is affected before engagement of the fusion machinery

Cells of S. cerevisiae and N. crassa lacking PRM1 homologs are prone to lysis during mating and germling fusion, respectively. Lysis correlates with cytoplasmic mixture, indicating that aberrant engagement of the fusion machinery results in membrane rupture and cell death (Jin et al. 2004; Aguilar et al. 2007). In both fungi, fusion-induced lysis is enhanced by low levels of extracellular calcium ions (Ca2+), suggesting that Ca2+-dependent repair mechanisms counteract this deficiency (Aguilar et al. 2007; Palma-Guerrero et al. 2014). Since germling pairs of Δerg-10a/Δerg-10b and Δerg-1 have a ΔPrm1-like defect in plasma membrane fusion (Figure 4A), we tested whether these erg mutants also undergo fusion-induced cell lysis. Consistent with previous reports (Palma-Guerrero et al. 2014), ∼12% of ΔPrm1 cell pairs and 3% of WT pairings underwent lysis on regular MM (Figure 5). This frequency increased to ∼20% for ΔPrm1 germling pairs when the Ca2+ content was reduced by one-half, compared with no significant increase in WT. In contrast to ΔPrm1 and similar to WT, cell integrity of Δerg-10a/Δerg-10b and Δerg-1 fusion pairs was not compromised on regular medium, and only a slight increase in lysis was observed on reduced Ca2+. Interestingly, introducing the deletion of erg-10a/erg-10b or erg-1 into the ΔPrm1 strain partially suppressed its lysis phenotype to a level of 8% on regular medium and 10% on reduced Ca2+ (Figure 5). Taken together, these data indicate that in the Δerg-10a/Δerg-10b and Δerg-1 mutants, membrane fusion is blocked at a stage before PRM1 is required and therefore before engagement of the fusion machinery.

Figure 5

The various plasma membrane fusion mutants differ in the frequency of fusion-induced cell lysis. Quantitative analysis of cell lysis in germling pairs of WT (FGSC 2489) and mutants lacking erg genes and/or Prm1 on MM, with normal and reduced levels of Ca2+. Values represent the mean ± SD from three independent populations of 100–125 germling pairs per replicate, strain, and growth condition. Statistical analysis with ANOVA and Tukey’s post hoc test: **P < 0.01, ***P < 0.001, ****P < 0.0001 for comparisons with ΔPrm1 at normal Ca2+ concentration; horizontal lines denote additional comparisons between strains (***P < 0.001); # denotes a significant difference (P < 0.001) for comparisons of ΔPrm1 between normal and half Ca