Accumulation of specific sterol precursors targets a MAP kinase cascade mediating cell–cell recognition and fusion

Significance

Deficiencies in sterol biosynthesis resulting in the accumulation of precursor sterol molecules are commonly associated with cellular malfunctioning and disease, including neurodegenerative and inflammatory disorders. However, the molecular and cellular consequences of the aberrant accumulation of sterol precursors are not understood. In particular, it is unclear whether specific biochemical or signaling pathways are targeted by the precursors and to what extent their specific structures contribute to their disruptive effects. Here we show that the accumulation of ergosterol precursors specifically targets a conserved ERK MAP kinase pathway that mediates fungal cell–cell communication and fusion. This effect is only caused by precursors with a conjugated double bond in their aliphatic side chain, indicating specific structure–function relationships in the mechanism of action.

Abstract

Sterols are vital components of eukaryotic cell membranes. Defects in sterol biosynthesis, which result in the accumulation of precursor molecules, are commonly associated with cellular disorders and disease. However, the effects of these sterol precursors on the metabolism, signaling, and behavior of cells are only poorly understood. In this study, we show that the accumulation of only ergosterol precursors with a conjugated double bond in their aliphatic side chain specifically disrupts cell–cell communication and fusion in the fungus Neurospora crassa. Genetically identical germinating spores of this fungus undergo cell–cell fusion, thereby forming a highly interconnected supracellular network during colony initiation. Before fusion, the cells use an unusual signaling mechanism that involves the coordinated and alternating switching between signal sending and receiving states of the two fusion partners. Accumulation of only ergosterol precursors with a conjugated double bond in their aliphatic side chain disrupts this coordinated cell–cell communication and suppresses cell fusion. These specific sterol precursors target a single ERK-like mitogen-activated protein (MAP) kinase (MAK-1)-signaling cascade, whereas a second MAP kinase pathway (MAK-2), which is also involved in cell fusion, is unaffected. These observations indicate that a minor specific change in sterol structure can exert a strong detrimental effect on a key signaling pathway of the cell, resulting in the absence of cell fusion.

Sterols are essential constituents of eukaryotic cell membranes. They feature prominently in the structure and function of the lipid bilayer by maintaining its microfluidic state. In addition, sterols and their specific interactions with proteins and other lipids are critical for the formation of membrane subdomains, mediating a plethora of biological processes, including cell polarization, signal transduction, subcellular sorting, and pathogen defense (1, 2).

Sterol biosynthesis deficiencies, and more specifically the accumulation of intermediates of the biosynthesis pathway, are commonly associated with cellular malfunctioning and disease. Examples include innate, inheritable diseases (e.g., Smith–Lemli–Opitz syndrome or desmosterolosis), neurodegenerative disorders (e.g., Alzheimer’s, Huntington’s, or Parkinson’s disease), and inflammation-induced disorders (e.g., primary cicatricial alopecia) (3–6). In plants, sterol biosynthesis deficiencies result in developmental defects, and in fungi the lack of the main membrane sterol ergosterol causes resistance against important antifungal drugs of the polyene group (7, 8). The accumulation of a specific ergosterol precursor with a conjugated double bond in its side chain results in cell mating defects in baker’s yeast, emphasizing the importance of the structure–function relationship of membrane sterols (9).

Although the detrimental effects of deficient sterol biosynthesis and the accumulation of sterol precursors become increasingly evident, the molecular consequences of these defects remain poorly understood. In particular, it is unclear whether the observed deficiencies are caused by an unspecific malfunctioning of the cell membrane affecting numerous molecular processes or by specific effects on individual biochemical or signaling pathways.

In recent years, Neurospora crassa has become widely appreciated as the primary model for studies on cell–cell fusion and related membrane-associated signaling events in filamentous fungi (10). Genetically identical germinating spores of this fungus undergo fusion, thereby forming a supracellular network, which develops further into the interconnected mycelial colony. Before fusion, cell pairs use a unique mode of communication, in which the two partners switch between signal sending and receiving in a highly coordinated, alternating manner, thereby establishing a cell–cell dialogue that orchestrates the cell pairs’ growing toward and fusing with each other. These switches between signal sending and receiving involve the highly dynamic assembly and disassembly of distinct membrane-associated protein complexes, containing either the mitogen-activated protein (MAP) kinase (MAK-2)-signaling module or the SO (soft) protein, which associates with a second ERK1/2-like MAP kinase pathway, the MAK-1 cascade, in filamentous fungi (11, 12).

In this study, we have analyzed how the accumulation of structurally different ergosterol precursors influences this highly coordinated cell–cell signaling mechanism. We show that specifically the presence of precursors carrying a conjugated double bond in their aliphatic side chain results in defects in cell–cell communication and fusion. These defects can be pinpointed to the MAK-1 MAP kinase cascade, indicating that the buildup of certain sterols causes very specific rather than general effects on the mechanism of self-signaling.

Results

Ergosterol-2 Deficiencies Result in Cell Fusion Defects.

Similar to most polar growing cells, germ tube tips of N. crassa possess a characteristic apical sterol-rich membrane domain (Fig. 1A). Because the MAK-2 and SO protein complexes mediating cell–cell communication and fusion specifically associate with this membrane sector, we set out to test the contribution of ergosterol to these processes. Ergosterol biosynthesis mutants were analyzed by light microscopy for defects in germling communication and fusion. In an erg-2 (ergosterol-2) mutant (FGSC 2723), interacting cell tips frequently failed to arrest growth after physical contact and curled around each other, giving rise to corkscrew-like structures (Fig. 1B). Although more than 60 N. crassa mutants affected in germling fusion have been identified so far, a comparable phenotype has not yet been described (13). The gene erg-2 (NCU01333) is homologous to erg4 of Saccharomyces cerevisiae and encodes a sterol C-24(28) reductase, mediating the last step of the predicted ergosterol biosynthesis pathway in N. crassa (Fig. 2A). Sequencing of the mutant allele identified a nonsense mutation at position 1523 (G > A), shortening the encoded protein by 77 aa residues. To test the effect of a complete loss of erg-2, a Δerg-2 gene knockout mutant was analyzed. The macroscopic appearance of Δerg-2 cultures was normal, although the linear hyphal extension rate was significantly reduced (Fig. S1A). Quantitative analyses revealed two defects related to germling fusion. First, the directed growth of cell pairs toward each other within a population was reduced by about 30% compared with WT (Fig. 1C). In contrast to the predominantly bidirectional tip-to-tip WT interactions, germ tube attraction was increasingly unidirectional in the mutant, and cells met in a tip-to-side mode (Fig. S2A). Second, interacting Δerg-2 cells failed to arrest growth after contact and formed the unique convoluted structures observed in the classical mutant (Fig. 1 B and D and Movies S1 and S2). In these cell pairs, fusion was reduced by about 90% (Fig. 1E). The mutant phenotype was fully complemented by reintroduction of an N-terminally GFP-tagged ERG-2 construct, confirming that it was fully caused by the lack of erg-2 (Fig. S1 C and E). A C-terminal tagged version complemented only partially, suggesting that the presence of GFP at this terminus impairs ERG-2 functions. Both GFP-tagged constructs localized to the perinuclear endoplasmic reticulum, consistent with the predicted site of ergosterol biosynthesis (Fig. S1D) (14). Together, these data indicate that the C terminus of ERG-2 is important for its functions. The failure of erg-2 fusion pairs to arrest growth suggests that the cells are unable to recognize and to react to their physical contact. In Δerg-2/WT pairings, cell–cell recognition and growth arrest was normal, indicating a cell-autonomous phenotype. However, cell merger within these heterotypic pairs was still significantly impaired and the cell contact zones commonly appeared swollen (Fig. 1 B and E and Fig. S2B).

Fig. 1.

Mutation of erg-2 disturbs cell–cell interactions and fusion. (A) Sterol distribution visualized by filipin staining. (B) (Left) erg-2 and Δerg-2 mutant cells twist around each other after physical contact (asterisks). (Right) Quantification of types of interaction following contact in WT, mutant, and mixed pairings. (C) Mutation of erg-2 significantly reduces the number of germlings involved in directed growth toward each other within a cell population. (D) Whereas WT germlings arrest directed growth after the cells have touched (arrow), Δerg-2 germlings continue to grow (asterisk, contact site). Images are stills from Movies S1 and S2, respectively. (E) Fusion of WT germlings expressing GFP and mCherry results in mixing of the cytoplasm (arrow). Δerg-2 pairs frequently fail to fuse after contact (asterisk). In WT/Δerg-2 germling pairings contact sites appear swollen (arrowhead) and the fusion frequency is increased compared with mutant pairs. All error bars represent SDs of three independent experiments. The number of cells or cell pairs per replicate ranged from 50 to 140. (All scale bars, 5 µm.)

Fig. 2.

Cell–cell interactions are only disturbed by the accumulation of sterols containing a conjugated double bond in the side chain. (A) Hypothetical biosynthesis pathway of ergosterol in N. crassa. ERG-10a and ERG-10b redundantly function as sterol C-5 desaturases. Homologous S. cerevisiae enzymes are depicted in parentheses. (B) GC profiles of sterols extracts. Δerg-2 lacks ergosterol and accumulates the precursor ergosta-5,7,22,24(28)-tetraenol (red curve). Numbers refer to the sterols shown in C. (C) Quantitative distribution of different sterols in WT and Δerg-2. (D) Total amount of sterols detected in WT and Δerg-2 samples, compared with the internal standard cholesterol, does not significantly differ from each other (n = 5–6; Student’s t test: P < 0.05). (E) Appearance of fusion pairs in different mutants (Left) and main sterol produced by the respective strains (Right). Arrows: normal contact points; asterisk: twisting germ tubes. (Scale bars, 5 µm.)

Fig. S1.

Deletion of erg-2, a gene required for the biosynthesis of ergosterol in N. crassa, affects both germling and hyphal fusion events. (A) The mutant Δerg-2 (FGSC 17460) shows a macroscopic phenotype comparable to WT (FGSC 2489), forming aerial hyphae that produce similar amounts of spores in MM slant tubes [WT: (6.1 ± 0.4)·10⁸, Δerg-2: (5.6 ± 0.3)·10⁸; errors represent SDs of three independent experiments]. However, the linear hyphal extension rate is strongly reduced by the deletion of erg-2 [WT: (7.0 ± 0.5) cm/d, Δerg-2: (1.7 ± 0.1) cm/d; errors represent SDs of three independent experiments]. (B) In a WT mycelium, vegetative hyphae interact with each other and establish physical contact (arrows), thereby increasing hyphal interconnectedness. Vegetative hyphae of Δerg-2 are defective in arresting growth after having touched (asterisks) and twist around each other in a way comparable to interacting germlings (Fig. 1). (C) Expression of a gfp-erg-2 (MW_545) construct fully complements the deletion of erg-2 and restores the WT-like behavior of germlings during directed growth and at cell–cell contact (arrows). In contrast, an erg-2-gfp construct (MW_549) only partially rescues the defects of Δerg-2 cells, with many germlings still twisting around each other after physical contact (asterisks). Error bars represent SDs of three independent experiments. The number of cells or cell pairs per replicate ranged from 50 to 100. (D) Both GFP-ERG-2 and ERG-2-GFP localize to the perinuclear endoplasmic reticulum as observed in spores originating from the heterokaryotic strains (MW_545 + CR73-1) and (MW_549 + CR73-1), respectively. (E) Sterol profiles generated by UV-visible spectroscopy. The complemented strain (MW_545) shows an adsorption curve indistinguishable from WT that represents the presence of ergosterol. The profiles of the classical (FGSC 2723) and the deletion mutant of erg-2 as well as the noncomplemented strain (MW_549) show an additional adsorption range at about 225–230 nm typical for the sterol intermediate ergosta-5,7,22,24(28)-tetraenol. (Scale bars: B, 20 µm; C, 5 µm; and D, 2 µm.)

Fig. S2.

Cell–cell interactions require both cells to efficiently communicate with each other. (A) In most WT germling pairs (N3-06 + N3-07), both partner cells clearly interact with each other and establish physical contact at their cell tips (arrow). In contrast, interactions between Δerg-2 germlings (N2-46 + N2-47) occur mainly unidirectionally and in a tip-to-side manner (arrowhead). Error bars represent SDs of three independent experiments. The number of cell pairs per replicate ranged from 100 to 120. (B) In mixed cell pairs consisting of WT (N3-06) and Δerg-2 (N2-47) germlings, the WT cells show a more pronounced way of directed growth than the mutant (arrow). In most cases, the WT germlings are unidirectionally growing toward the mutant. However, the Δerg-2 partner cell is able to form small cell protrusions (arrowhead) after having been touched by a WT germ tube. Error bars represent SDs of three independent experiments. The number of cell pairs per replicate ranged from 100 to 110. (C) The localization pattern of MAK-2-GFP after cell–cell contact was compared between germling pairs consisting of WT (N1-41 or N3-07) and/or Δerg-2 (N2-48 or N2-47). The interaction of Δerg-2 with a WT cell partially rescues the mislocalization of the MAP kinase as observed in homotypic mutant cell pairs. The number of cell pairs ranged from 100 to 120. (D and E) The strong mislocalization of SO-GFP in Δerg-2 (N2-49) germlings before and after cell–cell contact is not influenced by the interaction of the mutant with a WT (N3-07) cell. However, recruitment and localization of the protein are worsened in WT (AF-T8) germlings interacting with Δerg-2 (N2-47) cells. The number of cell pairs ranged from 10 to 40 (D) and 20 to 40 (E), respectively. (All scale bars, 5 µm.)

So far, most described germling fusion mutants were also affected in fusion between hyphae within the mature mycelium, when tested (13). Consistent with this notion, hyphal fusion pairs of Δerg-2 also exhibited the growth arrest failure and hyphae twisted around each other (Fig. S1B). In contrast, the sexual interaction between mating partners was unaffected (Fig. S3 A and B), indicating that Δerg-2 defects are specific for vegetative fusion events. Fruiting body development following fertilization and sexual spore formation were, however, significantly impaired in Δerg-2, indicating postfertilization functions for normal ergosterol biosynthesis (Fig. S3 C–F).

Fig. S3.

Deletion of erg-2 does not impair sexual cell–cell interactions and fusion in N. crassa but affects proper fruiting body development. (A) Homozygous and heterozygous crosses (mat a × mat A) between WT (FGSC 988 and/or N2-12) and Δerg-2 (FGSC 17459 and/or N5-17) were analyzed for the ability of trichogynes, specialized female receptive hyphae, to interact and fuse with male cells of the opposite mating type (black arrows) containing nuclei labeled with H1-GFP (white arrows). After 48 h, trichogynes have made contact with individual microconidia. The loss of nuclear fluorescence in the conidia indicates successful mating by migration of the nuclei through the trichogynes into the protoperithecia, a prerequisite for fertilization. (B) Quantitative analysis of directed growth and cell fusion, respectively, between trichogynes and microconidia reveals no influence of Δerg-2 on these processes. Error bars represent SDs of four to five independent experiments. The number of mating pairs per replicate ranged from 10 to 90. (C) WT (FGSC 988) and Δerg-2 (FGSC 17459) form comparable amounts of morphologically similar protoperithecia. (D) Deletion of erg-2 affects the size of matured fruiting bodies, the perithecia, when used as the female crossing partner. Homozygous crosses of the mutant (FGSC 17459 × FGSC 17460) produce perithecia with a smaller diameter compared with homozygous WT crosses (FGSC 988 × FGSC 2489). About 100 perithecia were analyzed for both crosses. (E) Rosettes of asci, large cells harboring the ascospores within the perithecia, are poorly developed in crosses with Δerg-2 used as the female crossing partner. (F) Accordingly, the number of ascospores produced by these crosses is strongly reduced in comparison with those in which the WT is the female strain. (Scale bars: A, 10 µm and C, D, and E, 200 µm.)

Δerg-2–Like Phenotypes Fully Correlate with the Presence of a Conjugated Double Bond in the Side Chain of the Accumulating Sterols.

Sterol profiling revealed that ergosterol, the main sterol of the WT, is absent in the Δerg-2 mutant. It instead accumulates the precursor ergosta-5,7,22,24(28)-tetraenol (Fig. 2 B and C). The only structural difference between this intermediate and ergosterol is an additional double bond within its aliphatic side chain. The overall sterol amount of WT and mutant are, however, comparable (Fig. 2D), and formation of the sterol-rich apical domain is unaffected in the mutant (Fig. 1A). These observations raised the question of whether the observed Δerg-2 defects are either caused by the absence of the end product ergosterol or by the accumulation of the precursor. We therefore analyzed germling fusion in eight additional single, double, and triple ergosterol biosynthesis mutants and determined their respective sterol profiles. The erg-10a and erg-10b mutants still accumulated ergosterol, suggesting redundant functions of the respective proteins. The double mutant lacked ergosterol and accumulated the precursor ergosta-7,22-dienol. In Δerg-11, ergosta-5,7-dienol was formed instead of ergosterol (Fig. 2E and Fig. S4A). Both ergosterol-lacking isolates still established WT-like cell interactions, indicating that the absence of the sterol end product is not responsible for the cell fusion defects (Fig. 2E and Fig. S4B). Strikingly, the introduction of the Δerg-11 mutation into the Δerg-2 strain rescued the mutant phenotype. The double mutant did not accumulate ergosta-5,7,22,24(28)-tetraenol but an intermediate with only one double bond in its side chain at a different position than in ergosterol (between C-24 and C-28). The only other isolate not accumulating ergosta-5,7,22,24(28)-tetraenol, but exhibiting a Δerg-2–like phenotype, was the triple mutant Δerg-2/Δerg-10a/Δerg-10b. Its main precursor was the only other intermediate identified with a conjugated double bond in its side chain (Fig. 2E). The double mutants Δerg-2/Δerg-10a and Δerg-2/Δerg-10b exhibited a Δerg-2-like sterol profile and phenotype consistent with the redundant function of ERG-10a and ERG-10b (Fig. S4). Together these findings indicate that the observed defects in cell–cell recognition and fusion fully correlate with the presence of this conjugated double bond in the sterol side chain and are independent of other structural features of the sterol molecule, such as the position of a single double bond in the side chain or the ring system of the molecule.

Fig. S4.

Not the lack of ergosterol, but the accumulation of ergosta-5,7,22,24(28)-tetraenol, causes the defects of Δerg-2 germlings during cell–cell interactions. (A) The single gene deletion mutants Δerg-10a (FGSC 20057) and Δerg-10b (N4-30) are still able to produce ergosterol and show a WT-like behavior of germlings (arrows). In contrast, the accumulation of ergosta-5,7,22,24(28)-tetraenol in the double mutants Δerg-2 Δerg-10a (N4-32) and Δerg-2 Δerg-10b (N4-34), respectively, provokes the same deficiencies as observed for Δerg-2 cells. (Scale bars, 5 µm.) (B) Quantitative analyses of cell–cell interactions of single, double, and triple ergosterol biosynthesis mutants reveal two distinct groups. Whereas directed growth between germlings of Δerg-10a, Δerg-10b, Δerg-11 (FGSC 13803), Δerg-2 Δerg-11 (N4-36), and Δerg-10a Δerg-10b (N4-38) is comparable to WT (FGSC 2489), the number of these interactions is reduced for cells of Δerg-2 (FGSC 17460), Δerg-2 Δerg-10a, Δerg-2 Δerg-10b, and Δerg-2 Δerg-10a Δerg-10b (N4-40). Germination of spores, however, is not affected in any of these strains. (C) Mass spectra of the main sterol molecules detected in WT and erg mutants of N. crassa. Sterol extracts were treated with MSTFA to produce TMS-derivatized sterols. Sterols were identified by comparing their mass spectra with data available in the literature.

Recruitment of the SO Protein During the Tropic Interaction Is Strongly Reduced in Δerg-2.

The interaction and fusion defects of germlings accumulating sterols with a conjugated double bond in the side chain suggest deficiencies in the highly orchestrated cell dialogue signaling mechanism, which involves the alternating recruitment of the SO protein and the MAK-2 MAP kinase module to the plasma membrane of the growing tips. Live-cell imaging revealed, however, that the dynamics of MAK-2-GFP in the Δerg-2 mutant and the WT were comparable before the interacting cells achieve physical contact (Fig. 3A). After the tips touched, MAK-2-GFP accumulated in WT at the site of cell merger, where it remained until fusion was completed. In the mutant the protein did not focus at the contact zone but was still recruited to the continuously growing tips (Fig. 4A). In WT/Δerg-2 pairings, tip growth ceased after cell–cell contact and MAK-2 recruitment seemed more focused in the mutant partner cell (Fig. S2C). These findings indicate that MAK-2 signaling and recruitment are generally unaffected in the mutant, but the cellular program fails to switch from directed growth to tip growth arrest and fusion after the cells touch. This observation supports the hypothesis that Δerg-2 cells are unable to recognize or process the cell–cell contact signal. In WT, peaking MAK-2 phosphorylation correlates with the time of maximum fusion within the cell population. In Δerg-2 the onset of phosphorylation increase was comparable to WT. However, the phosphorylation level remained high for a significantly prolonged period, consistent with the extended membrane recruitment (Fig. 4D). Polarization of the cytoskeleton and localization of the polarity factor BEM-1 were normal in Δerg-2, indicating that polarized growth is unaffected in the mutant (Fig. S5 A and B). Taken together, these observations indicate that although the mutant can undergo tropic interactions, albeit with decreased efficiency, it remains locked in the cellular program of directed growth, which is not terminated upon cell–cell contact.

Fig. 3.

Deletion of erg-2 disturbs the recruitment and subcellular localization of the SO protein, but not of the MAP kinase MAK-2. (A) MAK-2-GFP is recruited to the cell tips of WT and Δerg-2 germlings in a comparable oscillatory manner (arrows). (B) SO-GFP localizes dynamically to WT cell tips but is only poorly recruited in Δerg-2 cells (arrowheads). In the mutant, SO-GFP strongly mislocalizes into punctuate complexes spreading over the cell periphery. (All scale bars, 5 µm.)

Fig. 4.

SO and the MAK-1 MAP kinase are not recruited to the cell–cell contact point in Δerg-2. (A) (Left and Center) In WT and Δerg-2 cell pairs MAK-2 concentrates at the site of cell–cell contact (arrows). (Right) In the mutant the kinase localizes to the continuously growing tips (arrowheads). (B) (Right) WT cells that strongly focus SO-GFP at their touching cell tips (arrow). (Center and Left) Δerg-2 (N2-49) germlings fail to cluster SO after physical contact (arrowhead) and during subsequent growth (asterisk). (C) GFP-MAK-1 is transiently recruited to the fusion point (arrow) in WT pairings (Top) but does not accumulate at the contact zone (asterisk) of Δerg-2 pairs (Bottom). (D) Immunoblot analysis testing the phosphorylation of the MAP kinases MAK-1 and MAK-2 in WT and Δerg-2 germlings. (All scale bars, 5 µm.)

Fig. S5.

General cell polarization and dynamic recruitment of the signaling proteins MAK-2 and SO during cell–cell interactions are not affected by Δerg-2. (A) Δerg-2 (N5-13) germlings polarize the actin cytoskeleton (labeled with Lifeact-GFP) at the site of cell–cell contact in a way indistinguishable from WT (NCAB1721) cells (arrows). (B) WT (2208) and Δerg-2 (GN1-64) germling pairs recruit the polarization factor BEM1 (labeled with GFP) to the site of physical contact (arrows). (C) Deletion of erg-2 does not affect the antiphase oscillatory recruitment of the signaling proteins MAK-2 and SO. In interacting germlings originating from heterokaryotic WT (N1-41 + AF-SoR1) and Δerg-2 (N2-48 + MW_400) strains, respectively, MAK-2-GFP and dsRED-SO do not colocalize at the same cell tips (arrows). (Scale bars: 5 µm; insets in C, 2 µm.)

In contrast to MAK-2 dynamics, the recruitment of the SO protein was highly affected during the tropic interactions of Δerg-2 germlings, although the overall amount of SO within the cell was comparable in WT and mutant (Fig. S6B). Only few aggregates of the protein formed at the plasma membrane, and these complexes were not concentrated at the cell tip but appeared at seamingly random locations across the cell cortex (Fig. 3B). Nevertheless, this aberrant complex formation still alternated with MAK-2 recruitment, indicating that the coordination between the two fusing cells remains generally unaffected (Fig. S5C). After cell–cell contact, however, SO did not concentrate at the contact zone, as is typical for WT (Fig. 4B). In heterotypic WT/Δerg-2 fusion pairs, SO recruitment did not improve in the mutant cell and worsened in the WT partner, consistent with the notion that the cell dialogue mechanism requires two fully functional partners (Fig. S2 D and E). In conclusion, MAK-2 signaling is not affected in the presence of the sterol precursor, whereas SO dynamics are. In addition to its role in cell fusion, SO aggregates at septal pores in injured hyphae. This function is unaffected in Δerg-2 (Fig. S6A), highlighting the specific effects of sterol precursor accumulation on cell–cell communication and fusion.

Fig. S6.

Deletion of erg-2 does not affect the role of SO and MAK-1 in response to cell damage and oxidative stress. (A) Mature hyphae were injured with a razor blade. In both the WT (AF-T8) and Δerg-2 (N2-49), loss of cytoplasm from the intact hyphal compartment is evaded by the efficient plugging of the septal pore (black arrows). In both strains, SO-GFP concentrates at the septal plugs (white arrows). Quantitative analysis of SO-GFP recruitment after hyphal damage does not differ between WT and Δerg-2. Error bars represent SDs of three independent experiments. The number of hyphae per replicate was 50. (Scale bars, 10 µm.) (B) The expression of SO-GFP in Δerg-2 (N2-49) is comparable to the control strains (AF-T8 and AF-SoT8). The protein was detected using an anti-GFP antibody. Loading of equal amounts of protein was verified by reprobing the membrane with an antibody specific for β-tubulin. (C) Deletion of erg-2 does not affect general activity of the MAP kinase MAK-1 in response to cellular stress. In both the WT (FGSC 2489) and Δerg-2 (FGSC 17460), MAK-1 is efficiently phosphorylated in the presence of 8 mM of the oxidative stress agent H₂O₂. Phosphorylated MAK-1 was detected by an antiphospho p44/42 antibody. Loading of protein samples was controlled with an anti-β-tubulin antibody after reprobing of the membrane.

The Cell–Cell Communication and Fusion Defects Are Caused by Deficiencies in MAK-1 MAP Kinase Signaling.

A recent study in Sordaria macrospora, a close relative of N. crassa, identified the SO homolog PRO40 as a scaffolding protein of the MAP kinase MAK-1 cell wall integrity pathway (12). We confirmed the physical interaction of SO and the two upstream kinases of the MAK-1 module for N. crassa by yeast two-hybrid analysis (Fig. S7A). In addition, Δmak-1 strains of N. crassa exhibit no cell–cell interactions related to fusion, as observed for the Δso mutant (15, 16). We therefore analyzed the dynamics of GFP-tagged MAK-1 in WT and Δerg-2. During tropic growth of WT germlings, no SO-like recruitment of MAK-1 to the cell tips occurred. However, as soon as the cells touched, the kinase accumulated at the contact zone (Fig. 4C and Movie S3), where it colocalized with SO and remained during fusion pore formation (Fig. S7B). Δerg-2 fusion pairs failed to recruit MAK-1 after cell contact, further corroborating our notion that Δerg-2 cells fail to switch their cellular programming toward cell fusion after physical contact (Fig. 4C). Consistent with these localization data, the MAK-1 phosphorylation level was significantly reduced in the mutant compared with the WT reference strain during the germling fusion period (Fig. 4D). In contrast, MAK-1 activation in response to H₂O₂ stress was comparable in both isolates (Fig. S6C), indicating that specifically the fusion-related functions of MAK-1 are affected in Δerg-2.

Fig. S7.

The signaling protein SO interacts with components of the MAK-1 MAP kinase cascade in N. crassa. (A) Yeast two-hybrid analysis reveals physical interactions of SO with the MAPKKK MIK-1 and the MAPKK MEK-1, two components of the cell wall integrity pathway in N. crassa. The MAP kinase MAK-1 does not directly interact with SO. Yeast strains expressing the plasmids GADT or GBKT, containing either the cDNA sequences of the genes mik-1, mek-1, or mak-1 or so, were first cocultivated under mating conditions and afterward shifted to selective media. In addition, strains were transformed with empty or control vectors. If mating is successful, diploid cells grow on selective medium (−Leu −Trp). However, growth on medium additionally lacking adenine and histidine requires the interaction of the proteins encoded by the cDNA sequences to restore prototrophy. (B) In the heterokaryotic strain (NCAL010-1 + AF-SoR1), MAK-1-GFP and dsRED-SO transiently colocalize during cell–cell fusion of germlings (arrow). After SO has been released from the contact site, the MAP kinase is involved in fusion pore formation and/or maintenance (arrowhead). (Scale bar, 5 µm.)

These observations raised the question of whether the lack of MAK-1 recruitment and activation is a consequence or the cause of the Δerg-2 defects. To address this issue, we combined chemical inhibition with molecular genetics. By site-directed mutagenesis, the gatekeeper residue of the MAK-1 ATP binding pocket was replaced by a glycine residue (E104G) by site-directed mutagenesis, rendering the kinase sensitive to the inhibitor 1NM-PP1. Expression of the mutated mak-1^E104G kinase allele in the Δmak-1 mutant fully complemented the phenotype in the absence of the inhibitor (Fig. S8A). The MAK-1^E104G-GFP fusion protein also exhibited WT-like subcellular dynamics during germling fusion (Fig. S9). In contrast, when germlings grew in the presence of 20 µM 1NM-PP1 the phenotype of the inhibitable mutant was comparable to Δmak-1 (Fig. 5A and Fig. S8B). A strain carrying a WT allele showed no defects under the same conditions (Fig. S8). Together, these data indicate that 1NM-PP1 specifically and efficiently inhibits MAK-1^E104G. Inhibition at different stages of the germling interaction indicated that MAK-1 functions are essential for induction and maintenance of tropic growth, but also for fusion pore formation after cell–cell contact (Fig. S9). With decreasing inhibitor concentrations the cell interaction rate rose in an almost linear manner. Inhibitor concentrations between 0.8 and 8 µM caused an increasing number of germling pairs to exhibit the unique Δerg-2–like phenotype of twisting germ tubes (Fig. 5 A and B). In these pairs, cell merger was also reduced comparable to Δerg-2 (Fig. 5C). As a control the same tests were conducted with a Δmak-2 strain carrying the inhibitable variant MAK-2^Q100G. Consistent with the MAK-2 function during cell–cell communication, germling interactions were inhibited by 1NM-PP1 in a dose-dependent manner (Fig. 5D and Fig. S8G). However, Δerg-2–like germ tube twists were never observed.

Fig. S8.

Chemical inhibition of an ATP-analog-sensitive variant of MAK-1 mimics the Δmak-1 phenotype. (A) Strains were grown in MM slant tubes supplemented with 20 µM of the inhibitor 1NM-PP1 (+) or with 0.02% of DMSO as a control (–). Growth of the WT (FGSC 2489), Δmak-1 (FGSC 11320), and the complemented deletion strains expressing mak-1-gfp (NCAL010-1) and dsRed-mak-1 (GN9-36), respectively, is not affected by the presence of the inhibitor. In contrast, the Δmak-1 strains expressing mak-1^E104G-gfp (NCAL011-2) and dsRed-mak-1^E104G (GN9-39), respectively, adopt a phenotype comparable to the one of the knockout mutant when treated with 1NM-PP1, whereas their growth is normal in the absence of the inhibitor. (B) Several WT germlings interact with each other and establish physical contact (arrows). Deletion of mak-1 completely blocks cell–cell interactions. (C) Germlings of strain NCAL010-1 normally undergo directed growth and form WT like cell–cell contacts (arrows) in the presence of 1NM-PP1. For comparison with strain NCAL011-2, see Fig. 5 A and B. (D) Δmak-1 strains expressing mak-1 and mak-1^E104G, respectively, fused to dsRed show the same behavior compared with those strains containing the corresponding gfp-constructs (see Fig. 5A and C and Fig. S8C). Arrows indicate cell–cell contacts. (E) NCAL010-1/GN6-36 germling pairs normally touch (arrows) and undergo cell–cell fusion in both the absence and presence of 1NM-PP1. For comparison with NCAL011-2/GN9-39 cell pairs, see Fig. 5C. (F) In a population consisting of NCAL011-2 and GN9-39 germlings, treatment with 20 µM of 1NM-PP1 completely blocks cell–cell interactions and fusion. (G) Germlings expressing mak-2-gfp (N1-41) are not sensitive to the inhibitor and normally interact with each other (arrows). For comparison with the inhibitable strain Δmak-2 mak-2^Q100G (MAL-1), see Fig. 5D. All error bars represent SDs of three independent experiments. The number of cells or cell pairs per replicate ranged from 50 to 160. (All scale bars, 5 µm.)

Fig. S9.

Kinase activity is required for massive recruitment of MAK-1 to the fusion site and for fusion pore opening. (A) In strain NCAL011-2, MAK-1^E104G-GFP recruitment in the absence of 1NM-PP1 kinase inhibitor follows WT dynamics, including massive recruitment of the kinase to the fusion site upon tip contact (arrowhead), followed by rapid fusion pore opening and subsequent dispersal. (B) In the presence of 20 µM of 1NM-PP1, MAK-1^E104G-GFP becomes weakly recruited to the cell cortex (arrowheads), but contact-induced enhancement during fusion pore opening never occurs. This blockage occurs independent of whether 1NM-PP1 is added before (C) or after (D) physical contact has been established between interacting cells. Black arrows indicate time points of 1NM-PP1 addition. Spore torque response in D is a clear marker of successful physical attachment between cells. (All scale bars, 2 µm.)

Fig. 5.

Inhibition of MAK-1 phenocopies the specific defects of Δerg-2. (A) (Left) Δmak-1 cells expressing MAK-1^E104G are sensitive to 1NM-PP1. Germlings interact normally in the absence of the inhibitor (arrows). Directed growth decreases in a dose-dependent manner in the presence of 1NM-PP1 and unusual cell–cell contacts appear (asterisks). (Right) Quantitative analysis. (B) (Left) An intermediate concentration of 1NM-PP1 results in the Δerg-2–like twisting of germ tubes (asterisk). (Right) Quantitative analysis. (C) Inhibition of MAK-1 reduces cell fusion in a dose-dependent manner. In the absence of 1NM-PP1, GFP- and dsRED-expressing cells form normal cell–cell contacts (arrow) and fuse. In presence of the inhibitor cell–cell fusion is reduced. (D) Interactions of Δmak-2 cells expressing MAK-2^Q100G are reduced in the presence of 1NM-PP1. Growth arrest is, however, normal. (Left) DIC (differential interference contrast). (Center) Quantification of directed growth. (Right) Quantification of contact types. All error bars represent SDs of three independent experiments. The number of cells or cell pairs per replicate ranged from 50 to 150. (All scale bars, 5 µm.)

To test whether the lack of MAK-1 activity influences ergosterol biosynthesis in a qualitative or quantitative way, the sterol profiles and amounts were determined for Δmak-1 and the WT. No significant differences were detected (Fig. S10A). Together, these results indicate that the specific inhibition of MAK-1 is sufficient to fully recapitulate the unique phenotypic Δerg-2 defects observed during germling fusion.

Fig. S10.

Reduced levels of SO expression provoke Δerg-2–like defects during cell–cell interactions; however, neither the deletion of the gene so nor mak-1 affects ergosterol biosynthesis. (A) Sterols extracted from WT (FGSC 2489) and the knockout mutants Δerg-2 (FGSC 17460), Δmak-1 (FGSC 11320), and Δso (FGSC 11293) were analyzed by UV-visible spectroscopy. Deletion of mak-1 and so, respectively, does not alter the sterol profile compared with WT (FGSC 2489), indicating that both mutants normally synthesize ergosterol and do not have an altered sterol composition as observed in Δerg-2 (FGSC 17460). (B) Expression of so-gfp under the control of Ptcu-1 (MW_600) complements the Δso phenotype in the absence of copper. At 10 µM Cu²⁺ the cells adopt a Δerg-2–like phenotype (asterisk), and at 1,000 µM Cu²⁺ interactions are highly reduced and are reminiscent to the deletion of so. (C) Quantitative analysis of contact types between cells. (D) Quantitative analysis of cell–cell interactions. (E) Cytosolic fluorescence of SO-GFP in MW_600 gradually decreases with increasing concentrations of Cu²⁺. All error bars represent SDs of three independent experiments. The number of cells or cell pairs per replicate ranged from 100 to 300. (All scale bars, 5 µm.)

Because the SO protein was identified as an interaction partner of the MAK-1 MAP kinase module and its plasma membrane recruitment is deficient in Δerg-2, we reasoned that a partial inhibition of SO might also result in the observed phenotype. Because no enzymatic function of SO is known, we repressed the expression of the so gene by putting it under control of the copper repressible promoter Ptcu-1 (17). Full repression resulted in Δso-like phenotypes, whereas partial repression again produced Δerg-2–like deficiencies (Fig. S10 B–E). We conclude that both cell–cell communication and cell–cell fusion defects of mutants accumulating sterols with a conjugated double bond in the aliphatic side chain can be fully attributed to deficiencies in MAK-1 MAP kinase signaling.

Discussion

In this study, we show that the accumulation of sterol precursors results in deficiencies in MAP kinase signaling. This effect is, however, highly specific. It is only caused by sterol molecules carrying a conjugated double bond in their aliphatic side chain and affects specifically certain functions of the MAK-1 cascade. This specificity is underlined by the observation that the general development of the respective mutants is only little affected, the plasma membrane at their cell tips is still enriched in sterols, and other membrane-associated processes, such as the localization of polarity factors or recruitment of the MAK-2 MAP kinase, are normal. The sterol precursors therefore obviously substitute most of the general ergosterol functions. These findings raise the question of how the altered membrane composition specifically affects MAK-1 signaling. We hypothesize that the precursors disturb membrane subdomain formation. Sterols play important roles in establishing these subdomains, which are involved in various signaling processes, including MAP kinase signaling. In mammals, the MAP kinase ERK module assembles together with its upstream activator Ras in nanoclusters at the membrane. Mathematical modeling indicated that decreases in Ras cluster formation result in reduced signaling (18). Different membrane-associated proteins are discussed as potential sensors and upstream activating factors of the MAK-1 module (19, 20). The exact functional relationships are, however, so far not understood. We hypothesize that clustering of such upstream factors might also be essential for full MAK-1 activation. Interestingly, stress-induced activation of MAK-1 is not affected in Δerg-2, suggesting that the upstream components mediating cell–cell interactions and the stress response differ, which further highlights the specificity of the precursor’s effect.

Our findings raise the question as to why the conjugated double bond is having a destructive effect. We consider two— mutually not exclusive—hypotheses. First, the presence of a conjugated double bond in the sterol side chain, which is oriented adjacent to the fatty acid tails of phospholipids, results in its increased rigidity, which might lead to a tighter packaging of sterols and membrane lipids, thereby reducing the fluidity of the membrane (9). This in turn could prevent the efficient formation of microdomains or protein nanoclusters involved in MAK-1 activation. Second, the conjugated double bond should be prone to oxidation, which would severely disturb the molecule structure, as shown for oxidation products of cholesterol precursors (21). As a consequence domain formation might also be deficient. At the tips of growing fusion hyphae, reactive oxygen species are strongly accumulating and seem to play a role in cell–cell signaling (22). This oxidizing environment might render fusion tips specifically prone to the oxidation of the sterol precursor. However, when determining the sterol profiles, the expected oxidation products were not detected; they might, however, be unstable and/or are only produced in small amounts.

Our data indicate that MAK-1 activity is involved in two processes of the cell fusion reaction. First, it is required to initiate and maintain the tropic interaction of the fusion cells, and second, it mediates cell–cell recognition upon physical contact. Both functions are affected in Δerg-2, however to a different extent. Although the tropic interactions still occur, albeit in a reduced frequency, contact recognition is mostly abolished. A potential explanation could be different modes of MAK-1 activation in the two processes. Our localization data indicate that the activation during tropic growth occurs mainly in the cytoplasm, whereas contact sensing involves plasma membrane recruitment of the kinase. In general, the subcellular localization of MAP kinase activation can influence its output. Whereas activation of MAK-1 homologous ERK in the cytoplasm of mammalian cells occurs in a dosage-dependent, linear manner, activation at the membrane is more switch-like and even low signal intensities result in maximum output (23). Our MAK-1 inhibition experiments revealed that inhibitor concentrations that still allow tropic interactions of N. crassa cells fully block growth arrest after their physical contact. We therefore hypothesize that the directed growth relies on lower MAK-1 activation levels achieved in the cytoplasm, whereas cell–cell contact recognition requires a rapid and extensive activation, involving recruitment to the plasma membrane. Future challenges include a clearer understanding of these different modes of MAK-1 activation, including the identification of the upstream activating factors.

In summary, our data identified very specific effects of certain sterol precursors on specific functions of an individual MAP kinase pathway. MAP kinase signaling cascades are highly conserved in eukaryotic organisms. It will therefore be of great interest to test whether deficiencies in other organisms that are caused by or correlate with the accumulation of specific sterol precursors also include MAP kinase signaling deficiencies. Elucidating the exact relationship between sterol structure and the activity of individual, specific signaling pathways will be a future challenge furthering our general understanding of membrane-associated signal transduction processes and their role in growth, development, and disease.

Materials and Methods

A detailed description of the materials and methods used in this study is provided in SI Materials and Methods. Strains used in this study are listed in Table S1. Mutants were constructed via transformation (24) and/or crossing (25), purified into homokaryotic strains, and confirmed via genotyping with specific oligonucleotides (Table S2). Fungal cultures were routinely grown on Vogel’s minimal medium (MM) (26).

Table S1.

Strains constructed and used in this study

Strain	Genotype	Source
2208	Δbem1 his-3+::Pccg-1-bem1-gfp Δmus-52::bar+ mat A	27
AF-SoR1	Δso::hph his-3+::Pccg-1-dsRed-so mat A	11
AF-SoT8	Δso::hph his-3+::Pccg-1-so-gfp mat A	31
AF-T8	his-3+::Pccg-1-so-gfp mat A	31
CR73-1	his-3+::Pccg-1-h1-dsRed rid-1 mat A	Gift from N. Louise Glass, University of California, Berkeley, CA
FGSC 988	mat a	FGSC
FGSC 2489	mat A	FGSC
FGSC 2723	erg-2 mat a	FGSC
FGSC 6103	his-3 mat A	FGSC
FGSC 9716	his-3 mat a	FGSC
FGSC 11293	Δso::hph mat A	FGSC
FGSC 11320	Δmak-1::hph mat A	FGSC
FGSC 11321	Δmak-1::hph mat a	FGSC
FGSC 13802	Δerg-11::hph mat a	FGSC
FGSC 13803	Δerg-11::hph mat A	FGSC
FGSC 13983	Δerg-10b::hph mat a	FGSC
FGSC 17459	Δerg-2::hph mat a	FGSC
FGSC 17460	Δerg-2::hph mat A	FGSC
FGSC 20056	Δerg-10a::hph mat a	FGSC
FGSC 20057	Δerg-10a::hph mat A	FGSC
GN1-64	Δerg-2 his-3+::Pccg-1-bem1-gfp	This study
GN4-33	Δmak-1::hph his-3+::Pccg-1-gfp-mak-1 mat A	35
GN6-51	Δso::hph his-3 mat a	This study
GN9-36	Δmak-1::hph his-3+::Pccg-1-dsRed-mak-1 mat A	This study
GN9-39	Δmak-1::hph his-3+::Pccg-1-dsRed-mak-1 E104G mat A	This study
MAL-1	Δmak-2::hph his-3+::Pccg-1-mak-2Q100G al mat A	11
MW_374	Δmak-1::hph Δerg-2::hph his-3+::Pccg-1-gfp-mak-1 mat A	This study
MW_400	Δerg-2::hph his-3+::Pccg-1-dsRed-so mat A	This study
MW_486	Δmak-1::hph his-3 mat A	This study
MW_545	Δerg-2::hph his-3+::P ccg-1-gfp-erg-2 mat A	This study
MW_549	Δerg-2::hph his-3+::Pccg-1-erg-2-gfp mat A	This study
MW_600	Δso::hph his-3+::Ptcu-1-so-gfp mat a	This study
N1-41	his-3+::Pccg-1-mak-2-gfp mat A	This study
N2-12	his-3+::Pccg-1-h1-gfp mat A	Gift from D. J. Jacobson, Stanford University, Stanford, CA
N2-45	Δerg-2::hph his-3 mat A	This study
N2-46	Δerg-2::hph his-3+::Pccg-1-gfp mat A	This study
N2-47	Δerg-2::hph his-3+::Pccg-1-cherry mat A	This study
N2-48	Δerg-2::hph his-3+::Pccg-1-mak-2-gfp mat A	This study
N2-49	Δerg-2::hph his-3+::Pccg-1-so-gfp mat A	This study
N3-06	his-3+::Pccg-1-gfp mat A	28
N3-07	his-3+::Pccg-1-cherry mat A	27
N4-30	Δerg-10b::hph mat A	This study
N4-31	Δerg-2::hph Δerg-10a::hph mat a	This study
N4-32	Δerg-2::hph Δerg-10a::hph mat A	This study
N4-34	Δerg-2::hph Δerg-10b::hph mat A	This study
N4-36	Δerg-2::hph Δerg-11::hph mat A	This study
N4-38	Δerg-10a::hph Δerg-10b::hph mat A	This study
N4-40	Δerg-2::hph Δerg-10a::hph Δerg-10b::hph mat A	This study
N5-13	Δerg-2::hph Δmus-51::bar+ his-3+::Pccg-1-Lifeact-gfp mat A	This study
N5-17	Δerg-2::hph his-3+::Pccg-1-h1-gfp mat A	This study
NCAB1721	Δmus-51::bar+ his-3+::Pccg-1-Lifeact-gfp mat A	32
NCAL010-1	Δmak-1::hph Pccg-1-mak-1-gfp (EC) mat A	This study
NCAL011-2	Δmak-1::hph Pccg-1-mak-1E104G-gfp (EC) mat A	This study

Table S2.

Oligonucleotides used in this study

No.	Name	Nucleotide sequence (5′–3′)
21	His3-F	CTTGCAGTCTTGCACGTTG
22	His3-R	CTCTCGAGTCCCGTTATTGC
52	3ˋR2 soft	GGAGACTTCGTGGGTGGATTT
62	ERG-2-R	ATCATCATTAATTAAAATGACGTACTAGAGAAAATCAA
68	ERG-2-FII	ATTATTATCTAGAATGTCGTCGTCAAGGTACTCGC
253	soft cDNA F	AATGTACATATGTCTCGATCCCGCGGTGTTC
254	soft cDNA R	AATGTAGAATTCCTAATGCCCATACTCCAAATGC
265	soft3for	GTCGCGGGCCGGTATGAAAC
313	ERG-2-TEST-F	ATGGATGGAATGGGGATGCTCGC
315	ERG-10a-TEST-F	TTTCTGTGGGAGGGGAGATGTCTGG
316	ERG-11-TEST-F	TCTGTAGGATGCTGGATGGTTTCGG
317	HPH-TEST-R	TCGTCCGAGGGCAAAGGAATAGAG
340	erg-2-f	GATCTTCTTTCCCATCCTCATGTGG
341	erg-2-r	CACCCTCCTCTCATACTCCTTCC
344	erg-10a-f	ACGACCTGCTCTCTATGATTTCCC
345	erg-10a-r	CTCCTTGACAATCTTCTCCATCTCG
346	erg-11-f	GACAGGAGGAGGTGTACAACCG
347	erg-11-r	CCAGTCGAGGAGAAGAGAAGCC
386	ERG-10b-TEST-F	GTAGAGCGATATGTAGGTGCTGGCC
387	erg-10b-f	GCTCCTATACTGGGTGTTTAGCG
388	erg-10b-r	CATGTGATGCAGTGTGTGGTGGG
420	soft TEST-F	GAACACAGCATCACCGTGG
670	erg-2-f-seq	TTCTCACCGCCCGTGACTTTAGGC
671	erg-2-r-seq	GTCCCCAAGGAGTGCTTTCCAAAGG
706	erg2 rev 214	GCGCCATCTTCCGAGCACAGA
709	gfp-erg-2-BglII-for	ATCATCAGATCTATGTCGTCGTCAAGGTACTC
710	gfp-erg-2-XbaI-rev	ATCATCTCTAGAACCCTTCAGCTCTTCCCATT
778	mak-1-f	GTTCCTGCTTAGCCGCTTGTGC
779	mak-1-r	CATCTCGCCGACATCATCGACAACC
780	mak-1-hph-test-f	GGAAGTCCAGGAACGACCAGG
783	mak-1-oriloci-F	CCACCTCGTTTAAGCAGCAAGC
852	mak-1-YRC-XbaI-F	GTAACGCCAGGGTTTTCCCAGTCACGACGTCTAGAATGGCTGATCTCGTGGGTCGCA
853	mak-1-YRC-EcoRI-R	GCGGATAACAATTTCACACAGGAAACAGCGAATTCGAAGGTCATGATGGTTGGAAGG
856	mak-1-E104G-GGC-F	CGAGACCTATCTCTACGGCGGTACGTCATGGCTCATGCCTCACG
857	mak-1-E104G-GGC-R	CGTGAGGCATGAGCCATGACGTACCGCCGTAGAGATAGGTCTCG
891	Ptcu-1-NotI-F	AATGTAGCGGCCGCGATGGGATAGAGAGAATGGC
892	Ptcu-1-Xbal-R	AATGTATCTAGAGGTTGGGGATGTGTGTGC
929	Mik cDNA NdeI F	ATCATCCATATGTACCAAAATGGCCAGAGGGAC
930	Mik cDNA ClaI R	ATCATCATCGATCTAATAAGTACCCCTGATCT
931	Mek cDNA Broad NdeI F	ATCATCCATATGGCCGATCACCAAGGTCAAAGC
933	Mek cDNA EcoRI R	ATCATCGAATTCCTATGCCTCTTCCTTTCCAT
934	Mak-1 cDNA NdeI F	ATCATCCATATGGCTGATCTCGTGGGTCGCAAG
935	Mak-1 cDNA ClaI R	ATCATCATCGATTTACATGCCACCAAGCTCGG
1142	dsred-mak-1_AscI_F	AATGTAGGCGCGCCAATGGCTGATCTCGTGGGTCGCA
1143	dsred-mak-1_XbaI_R	AATGTATCTAGAGAAGGTCATGATGGTTGGAAGG
AL1	mak1_BamHI_fw	GATCGGATCCATTCGCCATGGCTGATCTCGTG
AL2	mak1_XmaI_rv	GATCCCCGGGATTCGCCATGGCTGATCTCGTG
AL3	MAK1_E104G_GGC_fw	CGAGACCTATCTCTACGGCGAGTTGATGGAGTGCGATCTTGC
AL4	MAK1_E104G_GGC_rv	GCAAGATCGCACTCCATCAACTCGCCGTAGAGATAGGTCTCG

Germling interaction and fusion assays were conducted as described previously (27). Directed growth between germinating spores was calculated from microscopic images with comparable cell densities. Cell–cell contact sites were classified as normal (narrow contact surface), swollen (broadened contact surface), or twisted (germ tubes unable to arrest growth). Quantitative cell–cell fusion assays were performed as described before (28).

For live-cell imaging, samples were analyzed by fluorescence or deconvolution microscopy (27). Membrane sterols were stained with a solution of 100 µg/mL of filipin III in 1% (vol/vol) DMSO.

To construct an ATP-analog-sensitive variant of MAK-1, the gatekeeper amino acid residue E104 was mutated into glycine (29). For the dose-dependent inhibition of MAP kinase activities, agar blocks were cut after 2 h of incubation and treated with 0.8–40 µM of 1NM-PP1 or 0.2% (vol/vol) of DMSO as a control. After incubation for an additional 2 h in a humidity chamber at 30 °C, samples were analyzed for directed growth, cell–cell contacts, and fusion.

Immunoblot analysis of phosphorylated levels of MAP kinases in interacting germlings was performed as described in ref. 30.

For sterol extraction and analysis, mycelia were harvested from shaking cultures. In short, sterols were obtained from the biomasses by saponification with alcoholic KOH solutions and subsequent extraction of the samples with n-hexane. The organic phases were derivatized with N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA) and analyzed by GC/MS.

SI Materials and Methods

Strains and General Growth Conditions.

Fungal strains used and produced in this study are listed in Table S1. Strains were routinely grown in agar slant tubes containing MM (26) as previously described (27). WT strains and knockout mutants lacking specific sterol biosynthesis genes were obtained from the Fungal Genetics Stocks Center (FGSC, www.fgsc.net/). Double and triple erg gene knockout strains were generated by crossing the corresponding single and double gene deletion mutants on Westergaard’s medium (25) and were subsequently confirmed via PCR genotyping analysis using oligonucleotides specific for WT and gene deletion loci, respectively (Table S2).

Transformation of N. crassa was performed as described previously (31). Heterokaryotic primary transformants were purified into homokaryotic strains by crossing or repeated single-spore isolation on selection medium. Genotypes were confirmed via PCR analysis.

Linear hyphal growth rates were determined as described previously (16). To measure sporulation, strains were grown in MM slant tubes for 3 d at 30 °C followed by 5 d at room temperature. Conidia were collected in sterile water and counted with a hemocytometer.

Complementation of Δerg-2.

To construct the Pccg-1-gfp-erg-2 expression fragment, the erg-2 locus was PCR-amplified from purified genomic WT DNA (FGSC 2489) using the oligonucleotides 709 and 710 (Table S2) and cloned into plasmid pMF334-gfp, a gift from E. Hutchison, University of California, Berkeley, CA, at the BglII and XbaI restriction sites. Similarly, the Pccg-1-erg-2-gfp fragment was constructed by amplifying erg-2 with primers 62 and 68 (Table S2) and subsequently cloning it into pMF272 at the XbaI und PacI restriction sites. The resulting vectors were verified by sequencing and finally transformed into strain N2-45.

Germling Interaction and Fusion Assays.

For germling interaction and fusion assays, MM agar plates were inoculated with aqueous suspensions of conidia as described in ref. 27 and incubated at 30 °C for 4 h. Agar blocks of about 10 × 10 mm were cut from the cultures and analyzed using light and/or fluorescence microscopy as described previously (27). The number of germlings showing directed growth toward each other was calculated from microscopic images with comparable cell densities. Conidial germination rates were routinely determined to check consistency between strains and growth conditions. Cell–cell contacts between interacting germlings were classified as normal (WT-like, narrow contact surface), swollen (broadened contact surface), or twisted (germ tubes unable to arrest growth like those of Δerg-2). Quantitative germling fusion assays between cell pairs expressing green and red fluorescing proteins, respectively, were performed as described before (28).

Analysis of Hyphal Fusion and Septal Plugging.

To examine fusion events between mature vegetative hyphae, MM plates were inoculated with spore suspensions and grown for 18–22 h at 30 °C. Agar blocks were cut from the interior of the mycelium, covered with a droplet of water and a coverslip, and analyzed by light microscopy as described in ref. 27. Septal plugging after hyphal injury was analyzed as described previously (31).

Fluorescence Microscopy.

For live-cell imaging of GFP- and dsRED-fusion proteins, samples were prepared as described above and analyzed by fluorescence or deconvolution microscopy (27). If required, images were captured as stacks and deconvolved using Huygens deconvolution software (SVI).

To stain membrane sterols in germlings, 10 µL of a solution of 100 µg/mL of filipin III in 1% (vol/vol) DMSO were added to the agar blocks. Cells were immediately analyzed by fluorescence microscopy using a DAPI filter setup.

MAK-1-GFP Cloning.

To express the MAK-1-GFP fusion protein in N. crassa, the plasmid pAL1-MAK-1 was constructed by first generating the GFP expression vector pAL1 through subcloning the gfp coding region from pMF272 into pBARGRG1 using BamHI/EcoRI restriction/ligation, and subsequently inserting the mak-1 gene amplified from WT cDNA (FGSC 2489) using oligonucleotides AL1 and AL2 into BamHI/XmaI linearized pAL1 in-frame to gfp while omitting the mak-1 stop codon. Recovered pAL1-MAK-1 plasmids were verified by sequencing and transformed into ∆mak-1 (FGSC 11320) using electroporation. Transformants were selected on MM supplemented with the selection marker Ignite. Expression of the MAK-1-GFP fusion protein was verified by fluorescence microscopy.

MAK-1-GFP Localization.

For 2D time courses of cell fusion (Movie S3) the “inverted agar method” was used to stably sandwich germlings between agar and glass cover slide (32). Confocal microscopy was performed using a Bio-Rad Radiance 2100 system mounted on a Nikon TE2000-U Eclipse inverted microscope. MAK-1-GFP was excited with the 488-nm laser line of a 40-mW argon ion laser set to ≤15% and fluorescence detected at 510–560 nm using a Nikon Plan Apo 60× 1.2 N.A. water immersion lens. Laser intensity and laser dwell time on the cells were kept to a minimum to reduce photobleaching and phototoxic effects. Simultaneous bright-field images were captured with a transmitted light detector. Time-lapse imaging was performed at 1- to 2-min intervals for periods of up to 30 min, or at longer time intervals for periods of several hours. Images were captured with a laser scan speed of 166 lines per s and a resolution of 1,024 × 1,024 pixels using Lasersharp 2000 software (version 5.1; Bio-Rad Microscience) and stored as TIFF files.

Liquid cell cultures were used to allow the application of biochemical inhibitors, such as 1NM-PP1, at any time point of the cell fusion process (Fig. S9). For this, 20 μL of a 1 × 10⁷ spores per mL stock solution were added to 180 µL of liquid MM into each well of an eight-well chambered cover slide (Nalge Nunc International) and incubated at 30 °C. A DeltaVision RT system (Applied Precision) based on an Olympus IX70 equipped with an Olympus Plan-Apo 100× 1.4 N.A. oil immersion objective, 75W HBO illuminator, Chroma Sedat Quad ET filter set (for GFP: excitation 490/20 nm, emission 528/38 nm; Chroma Technology Corp.) and a CoolSnap HQ EM-CCD camera (Photometrics) was used to capture the fusion process at desired time points. Exposure times ranged from 200 to 400 ms. To acquire 3D (x, y, z) images, 30–40 optical sections were obtained at 0.2-µm steps. For 4D imaging (x, y, z, and t), 10–15 optical sections were obtained at 0.4- or 0.5-µm steps and 30- to 120-s intervals. Images were processed through 10 iterative deconvolutions using implemented SoftWorx (Applied Precision) image processing and analysis software and stored as TIFF files.

Independent of the acquisition technique, projections and further image processing were carried out with the MacBiophotonics package ImageJ 1.43b software (https://imagej.nih.gov/ij/).

Live-Cell Imaging of Actin.

To visualize the actin cytoskeleton, conidia of the Lifeact-GFP expressing strains NCAB1721 (32) and N5-13 were spread on MM plates and grown as described above. Interacting germling pairs were analyzed by fluorescence microcopy.

Chemical Genetics.

To construct an ATP-analog-sensitive variant of MAK-1, multiple sequence alignment was used to identify E104 of this MAP kinase as the gatekeeper amino acid homologous to the previously described residues Q93 in S. cerevisiae Fus3p (29) and Q100 in N. crassa MAK-2 (11), respectively. In two separate approaches, site-directed mutagenesis and yeast recombinational cloning, respectively, were used to mutate codon GAG (E) into GGC (G), rendering the MAP kinase sensitive to 1NM-PP1 (29).

To generate the mak-1^E104G-gfp allele by site-directed mutagenesis, primers AL3 and AL4 were used to introduce the E104G exchange into plasmid pAL1-MAK-1. The resulting plasmid pAL1-MAK-1^E104G was verified by sequencing and transformed into ∆mak-1 (FGSC 11320). Positive transformants were selected as described above.

To generate the dsRed-mak-1 construct, the mak-1 gene (NCU09842.7) was amplified from WT DNA (FGSC 2489) using primers 1142/1143 and cloned into vector pMF334 at the restriction sites AscI and XbaI. To create the dsRed-mak-1^E104G allele, a 5′ and a 3′ fragment of mak-1 were amplified with primer pairs 852/857 and 853/856, respectively, and cotransformed with the EcoRI/XhoI linearized plasmid pRS426 into yeast strain FY834. The modified mak-1 gene was then amplified by PCR from the yeast plasmid using primers 852/853 and subcloned into pMF272 via XbaI/EcoRI. After PCR amplification with primers 1142/1143, the mak-1^E104G mutated allele was cloned via AscI and XbaI into pMF334. All plasmids were sequenced and transformed into strain MW_486.

Functionality of all fluorescently labeled MAK-1 fusion constructs was confirmed by full complementation of the mak-1 deletion in the absence of the inhibitor, whereas 20 µM of 1NM-PP1 were used to mimic the Δmak-1 phenotype in those strains expressing the modified MAK-1^E104G. For the dose-dependent inhibition of MAK-1 and MAK-2 activities during germling interaction and fusion, agar blocks were cut after 2 h of incubation and treated with 0.8–40 µM of 1NM-PP1 or 0.2% (vol/vol) DMSO as a control. After incubation for an additional 2 h in a humidity chamber at 30 °C, samples were analyzed for directed growth, cell–cell contacts, and fusion between germlings as described above.

Expression of so Under the Control of the tcu-1 Promoter.

To control the expression of so, the gene was set under the control of the tcu-1 promoter, which represses gene expression in the presence of Cu²⁺ (17). The tcu-1 promoter region was amplified from WT genomic DNA (FGSC 2489) using primer pair 891/892 and cloned via the restriction sites NotI and XbaI into vector pMF272, resulting in plasmid Ptcu-1-gfp (number 517). The so-gfp sequence was taken from vector pSO8 (31) and cloned into plasmid 517 via the restriction sites XbaI and EcoRI, resulting in Ptcu-1-so-gfp (number 663). All plasmids were verified by sequencing. Plasmid 663 was transformed into strain GN6-51. Primary transformants were purified into homokaryons, resulting in strain MW_600.

To test the influence of high and low expression of so-gfp on the behavior of germlings, MW_600 was first grown in MM slant tubes supplemented with ultrapure agarose instead of normal agar-agar. Defined concentrations of copper sulfate were added to the media, ranging from 0 to 1000 µM. The same amounts of Cu²⁺ were used in the MM plates on which germling assays were performed. Samples were again analyzed for directed growth and cell–cell contacts between interacting cells as described above. SO-GFP levels were analyzed by fluorescence microscopy.

Immunoblot Analyses.

Immunoblot analysis of phosphorylated levels of MAP kinases in interacting germlings was performed as described in ref. 30 with minor modifications. Cellophane sheets covering MM plates were inoculated with 300 µL of conidial suspensions (5 × 10⁷ spores per mL) from 7-d-old cultures, grown at 30 °C, harvested at various time points, and ground in liquid nitrogen. One milliliter of frozen samples was extracted with 700 µL of extraction buffer [50 mM Hepes (pH 7.5), 2 mM EGTA (pH 8.0), 2 mM EDTA (pH 8.0), 100 mM NaCl, 1% Triton X, 10% glycerol, 60 mM β-glycerol phosphate, 15 mM p-nitrophenyl phosphate, 10 mM sodium fluoride, 1 mM sodium orthovanadate, and Roche protease inhibitor mixture]. After incubation on ice for 30 min, mixing every 10 min and centrifugation at 16,000 × g for 8 min at 4 °C, supernatants were recovered to prepare protein samples with Laemmli loading buffer. Proteins were separated by SDS/PAGE on 8% polyacrylamide separating gels and transferred on PVDF membranes by wet electroblotting. Membranes were reversibly stained with 0.1% (wt/vol) Ponceau S in 5% (vol/vol) acetic acid to ensure protein transfer and equal loading of samples and subsequently blocked in Tris-buffered saline [TBS; 50 mM Tris⋅HCl (pH 7.5), 150 mM NaCl, and 0.05% Tween 20] with 5% BSA for 8 h at room temperature. At 4 °C, membranes were hybridized overnight with an anti-phospho p44/42 MAP kinase polyclonal antibody, which recognizes both phosphorylated MAK-1 and MAK-2 (9101, rabbit, 1:1,000 dilution in blocking buffer; Cell Signaling Technology), washed six times for 10 min in TBS, incubated for 90 min with a horseradish peroxidase-conjugated anti-rabbit secondary antibody (7074, goat, 1:120,000 dilution in blocking buffer; Cell Signaling Technology), and washed again. Proteins were detected by chemoluminescence using the SuperSignal West Femto kit (Thermo Scientific) and exposure of the membranes to X-ray films.

To detect MAK-1 phosphorylation under stress conditions, liquid MM cultures in polypropylene (PP) flasks were inoculated with 2 × 10⁴ spores per mL and incubated for 24 h at 30 °C and 100 rpm. Eight millimolar H₂O₂ was added 15 min before harvest of the mycelia. After protein extraction, 30 µg of protein per sample were analyzed by the antiphospho immunoblotting procedure described above.

To determine SO-GFP expression, liquid cultures in PP flasks containing 1 × 10⁴ spores per mL were incubated for 19 h at 30 °C and 100 rpm. PVDF membranes blotted with 30 µg of protein per sample were blocked for 1 h at room temperature in TBS (0.1% Tween) with 5% BSA, hybridized for 1.5 h with an anti-GFP monoclonal antibody (clones 7.1 and 13.1, mouse, 1:2,000 in blocking buffer; Roche), washed four times for 5 min in TBS, incubated for 30 min with a peroxidase-conjugated anti-mouse secondary antibody (goat, 1:4,000 dilution in blocking buffer; Invitrogen), and washed again. Signals on exposed X-ray films were detected with the SuperSignal West Pico chemoluminescent substrate (Thermo Scientific).

To control protein loading, membranes were first incubated with stripping buffer [50 mM Tris⋅HCl (pH 7.5), 150 mM NaCl, 6 M guanidine hydrochloride, and 0.2 M acetic acid] for 1 h at room temperature, washed three times for 5 min in TBS (0.1% Tween), and blocked overnight at 4 °C. At room temperature, membranes were subsequently reprobed for 1.5 h with an anti–β-tubulin monoclonal antibody (clone TU27, mouse, 1:2,000 dilution in blocking buffer; Covance) and washed four times for 10 min. Secondary antibody hybridization and signal detection were performed as described for the anti-GFP procedure.

Sterol Extraction and Analyses.

For sterol extraction and analysis, PP flasks with liquid MM were inoculated with 0.5 × 10⁶ of spores per mL and incubated for 24 h at 30 °C and 100 rpm. Mycelia were harvested by filtration through Miracloth (Calbiochem), washed with double-distilled H₂O, and ground under liquid nitrogen. To obtain a general overview of the sterol composition of a strain, a detailed sterol extraction procedure was chosen. About 1 g of frozen biomass was extracted with 20 mL of a 3:1 (vol/vol) mixture of CHCl₃/MeOH. After solvent evaporation, residues were saponified with 20 mL of a solution of 10% (wt/vol) KOH in MeOH for about 4 h at 60 °C under continuous stirring. Chilled samples were extracted with 20 mL of a 1:1 mixture of water and n-hexane. The aqueous phase was reextracted twice with each 10 mL of n-hexane. All organic phases were combined, concentrated to a final volume of 1 mL in n-hexane, and stored at −20 °C. Sterol extracts were derivatized with MSTFA and analyzed by GC-MS as described in ref. 33. Sterols were identified by comparison of mass spectra with data available from databases.

For quantitative analyses of multiple sterol samples, a short extraction protocol was used. Frozen biomass (0.2 g) was saponified with 3 mL of 10% (wt/vol) KOH in 60% (vol/vol) EtOH for 1 h at 70 °C under continuous stirring. After cooling, the suspension was extracted with 1 mL of water and 3 mL of n-hexane. The organic phase was collected and stored at −20 °C. As an internal standard for quantification, 1 µL of a solution of cholesterol (1 mg/mL in n-hexane) was added to 100 µL of the extract. The mixture was treated with 30 µL of MSTFA for 60 min at 60 °C. Silylated samples were concentrated under a stream of nitrogen and diluted with 50 µL of CH₂Cl₂. GC/MS was performed as described in ref. 34, with the exception that the gas chromatograph was programmed as follows: 10 min at 50 °C, then increasing with 10 °C/min to 320 °C. The gas chromatograms were used to calculate the total amount of sterols per sample from the area of signals present in the extract in comparison with the signal area of the internal standard.

For UV-visible spectrometry, sterol samples were obtained from 0.1 g of biomass according to the short extraction procedure described above. The extracts were analyzed by absorption spectroscopy at wavelengths between 200 and 350 nm with reference to the solvent n-hexane.

Analysis of Sexual Reproductive Structures and Mating.

Crosses were performed on plates containing Westergaard’s medium (25). Images were captured with a Leica M60 stereomicroscope coupled to a DFC295 camera using the Leica Application Suite software. Ascospores were washed off from the lids of the Petri dishes and suspended in 2 mL of water. Trichogyne–microconidium fusion assays were performed as described in ref. 16.

Yeast Two-Hybrid Analysis.

The yeast two-hybrid experiments were performed using the BD Matchmaker Library Construction & Screening Kits (BD Biosciences Clontech). mRNA collected from WT (FGSC 2489) was used as a template for reverse transcription into cDNA using the primer pairs 253/254, 934/935, 931/933, and 929/930 (Table S2) for the N. crassa genes so (NCU02794.7), mak-1 (NCU09842.7), mek-1 (NCU06419.7), and mik-1 (NCU02234.7). The so amplicon was cloned into vector pGBKT7 (GAL4 DNA binding domain), and the other products were cloned into vector pGADT7-Rec (GAL4 activation domain). Yeast matings, selection, and controls were performed according to the manufacturer’s instructions.