Characterization of the Neurospora crassa Cell Fusion Proteins, HAM-6, HAM-7, HAM-8, HAM-9, HAM-10, AMPH-1 and WHI-2

Intercellular communication of vegetative cells and their subsequent cell fusion is vital for different aspects of growth, fitness, and differentiation of filamentous fungi. Cell fusion between germinating spores is important for early colony establishment, while hyphal fusion in the mature colony facilitates the movement of resources and organelles throughout an established colony. Approximately 50 proteins have been shown to be important for somatic cell-cell communication and fusion in the model filamentous fungus Neurospora crassa. Genetic, biochemical, and microscopic techniques were used to characterize the functions of seven previously poorly characterized cell fusion proteins. HAM-6, HAM-7 and HAM-8 share functional characteristics and are proposed to function in the same signaling network. Our data suggest that these proteins may form a sensor complex at the cell wall/plasma membrane for the MAK-1 cell wall integrity mitogen-activated protein kinase (MAPK) pathway. We also demonstrate that HAM-9, HAM-10, AMPH-1 and WHI-2 have more general functions and are required for normal growth and development. The activation status of the MAK-1 and MAK-2 MAPK pathways are altered in mutants lacking these proteins. We propose that these proteins may function to coordinate the activities of the two MAPK modules with other signaling pathways during cell fusion.


Introduction
Cell-to-cell fusion between vegetative cells plays a critical role in the life cycles of the filamentous fungi. The fusion between germinating conidia allows the cells to share resources and helps them to establish a colony [1][2][3][4]. As the fungal colony matures, cell fusion is important for the movement of resources throughout the colony, a prerequisite for asexual and sexual development. In the model filamentous fungus, Neurospora crassa, cell-to-cell fusion plays an important role during colony establishment, as well as during conidiation (asexual development) and protoperithecium formation (sexual development) [5][6][7]. During colony establishment, fusion between germinating conidia occurs between specialized cells called conidial anastomosis tubes (CATs), which are morphologically and physiologically distinct from germ tubes [7,8]. Germ tubes are wider and exhibit negative chemotrophic interactions, while CATs are thinner and exhibit chemotrophic attraction towards each other [8]. Mutants that are defective in cell fusion can't form an interconnected hyphal network to support nutrient transport within the colony [1,9]. During the N. crassa asexual life cycle, wild type colonies transport nutrients from a vegetative hyphal network into the growing aerial hyphae, which generate conidia (asexual spores). Cell fusion mutants are defective in producing the long aerial hyphae typical of wild type cells. They produce short aerial hyphae, which give a ''flat'' carpet-like conidiation phenotype. [10]. Cell fusion is also important for the N. crassa sexual life cycle. Cell fusion mutants are female sterile, and this may be because the efficient transport of amino acids and other nutrients from a vegetative hyphal network into the developing protoperithecia is needed to support sexual development.
Various groups have defined approximately 50 genes required for cell-cell communication and fusion in N. crassa [9][10][11][12][13][14][15][16][17]. Many of these cell fusion genes encode components of the MAK-1 and MAK-2 mitogen-activated protein kinase (MAPK) signal transduction pathways [1,[18][19][20][21][22][23][24], which are homologous to the yeast cell wall integrity (CWI) and pheromone response cascades, respectively [25][26][27]. MAK-2 and HAM-1/SO, a protein of unknown molecular function, display oscillatory recruitment to opposing cell tips during CAT communication, suggesting that the chemotrophic interactions between two CATs are coordinated by the MAK-2/SO Ping-Pong signaling behavior [23,28]. NRC-1 and MEK-2, the upstream MAPKKK and MAPKK in MAK-2 pathway, were also found to have the oscillatory signaling behavior during cell fusion [24]. The N. crassa MAK-1 CWI pathway initiates through a set of transmembrane sensors. Signals are integrated by the small GTPase RHO1, which activates a conserved mitogen-activated protein kinase (MAPK) cascade through its interaction with protein kinase C [29][30][31][32]. In addition to its function in cell wall stress integration, the CWI pathway is also a central component of the cell-cell communication machinery. The functional relationship between the two signaling pathways during intercellular communication is poorly understood, but evidence for cross-talk between MAK-1 and MAK-2 is provided by work on the striatin interacting phosphatase and kinase (STRIPAK) complex [33,34]. Its subunits, HAM-2, HAM-3, HAM-4, MOB-3, PP2A and PPG-1, are all required for cell-cell communication [9,13,35,36]. Phosphorylation of MOB-3 by MAK-2 is required for nuclear localization of MAK-1 in vegetative hyphae, suggesting that MAK-1-dependent expression of cell fusion genes may be required for establishing cell-cell communication competence. Among the identified cell fusion genes, there were six genes whose functions were largely uncharacterized. To better understand how cell-to-cell fusion is regulated, these six genes, ham-6, ham-8, ham-9, ham-10, amph-1, and whi-2 have been further characterized. The expression patterns, intracellular locations, and how the loss of these genes affects the activation status of the MAK-1 and MAK-2 pathway were examined. A seventh gene, ham-7, which has been shown to function as a sensor for the MAK-1 pathway [31], was included in the analysis to examine the cell type expression pattern and cellular location of the HAM-7 sensor. In further characterizing these genes, we demonstrate here that HAM-6, HAM-7 and HAM-8 function as upstream elements in the pathway regulating MAK-1 kinase activity during cell fusion. We further demonstrate that HAM-9, HAM-10, AMPH-1 and WHI-2 are proteins with general functions in regulating N. crassa growth, and we suggest that HAM-9, HAM-10, and WHI-2 may provide cross-talk between the two MAP kinase pathways and other signaling cascades during cell fusion.

Plasmid construction and expression of tagged proteins
Plasmids used in this study are listed in Table S1. The primers used to construct GFP-tagged, RFP-tagged and HA-tagged protein constructs are listed in Table S2. GFP and dsRed RFP fusion proteins were generated using the pMF272 and pMF334 vectors [37,41] and were expressed under the control of ccg-1 promoter at the his-3 locus [42].
HA-tagged proteins (HA-HAM-6, HA-HAM-8, HA-HAM-9, HA-AMPH-1 and HA-WHI-2) were cloned using the vectors pBM60 and pBM61 [43][44][45] and were expressed under the control of their own promoters at the his-3 locus. To identify sites for HA tagging, the protein sequences for HAM-6, HAM-8, HAM-9, AMPH-1 and WHI-2 were used to generate multiple sequence alignments with homologous proteins from other fungi [46]. Protein sequences were also analyzed by the online program Globplot to predict protein structural information [47]. Nonconserved protein sequence regions, which were predicted to be exposed, were chosen for HA tagging. For ham-6, ham-8 and ham-9, the HA tag coding sequence was inserted immediately before the stop codon. For amph-1, the HA tag coding sequence was inserted between the third and the fourth amino acid codons. For whi-2, the peptide sequence VPVDPASGA (amino acid 118-126) was replaced with the HA tag. The cloning of these HA-tagged protein constructs involved two PCR amplification steps. PCR primers were used to amplify approximately 1,500 bp of 59 UTR sequence, the coding sequence upstream of the HA tag insertion site, and the HA tag coding sequence. A second set of primers were used to amplify the HA tag coding sequence, the rest of the coding sequence of the gene, and the approximately 500 bp 39 UTR sequence. The two PCR products were then mixed together and used as templates to amplify the entire HA-tagged gene. The primers designed for the two ends of the genes contained an added restriction enzyme site to allow the insertion of the amplified DNA into the multicloning sites of pBM60 or pBM61. An HA-tagged version of HAM-7 with its endogenous promoter was generated by Retrogen Inc. (San Diego, CA). The DNA sequence encoding amino acids 190-198 (YTINILESG), which are located immediately in front of the GPI anchor addition site, were replaced by the HA tag sequence (YPYDVPDYA) in the HA-tagged HAM-7. Plasmids pgrp-GFP, pRFP-vps-52, parg-4-GFP and pRFP-vam-3 were obtained from the FGSC as fluorescent protein markers for ER, Golgi, mitochondria, and vacuoles respectively [48], and used in co-localization experiments. Plasmids pso-GFP and pmak-2-GFP were kind gifts from Dr. Louise Glass's lab, and were used to study the MAK-2 signal transduction pathway [23].

Analysis of HA-tagged protein expression
To examine the expression of HA-tagged proteins in vegetative hyphae, cultures were grown in 100 ml liquid Vogel's sucrose medium in shaking Erlenmeyer flasks at room temperature for 36 to 48 hours. To examine the expression of HA-tagged proteins in germ tubes and CATs, conidia were used to inoculate 16 ml of Vogel's sucrose medium at a titer of 10 6 per ml and grown in 100615 mm Petri dishes at 34uC without agitation for 4 hours to allow the formation of germ tubes and CATs [8]. Cells were harvested by filtration using a Büchner funnel and ground in liquid nitrogen. Protein extraction buffer [100 mM Tris/HCL pH 7.4, 1% (w/v) SDS; supplemented with 1X protease cocktail (P-8340 Sigma-Aldrich, St. Louis, MO)] was added, and the cell extracts were collected after centrifugation.
For Western blot analysis of HA-tagged proteins, the protein concentration of the cell extracts was determined by using the DC protein assay kit (BioRad, Hercules, CA). Samples containing 60 mg of protein were subjected to SDS-PAGE and transferred to nitrocellulose membrane. The nitrocellulose membranes were then subjected to Ponceau S red (Sigma-Aldrich, St. Louis, MO) staining to verify equal loading of different protein samples. Western blot experiments with mouse monoclonal anti-HA (Covance, Princeton, NJ) and rabbit anti-mouse IgG-HRP (Sigma-Aldrich, St. Louis, MO) were used to assess the level of protein expression. Chemiluminescent signal was detected by using a ChemiDoc XRS+ System and images were analyzed with Image Lab Software (BioRad, Hercules, CA).

Evaluation of MAK-1 and MAK-2 phosphorylation status
Polyclonal antibody directed against the phosphorylated activation site in common on MAK-1 and MAK-2 was used to evaluate their activation status. To examine the status of the pathways in vegetative hyphae, liquid N. crassa cultures were grown at room temperature and harvested by filtration using a Büchner funnel [31]. In some experiments, vegetative cultures were subjected to 8 mM H 2 O 2 for 10 minutes just prior to being harvested to determine if the pathways could be activated by oxidative stress. To study the MAK-1 and MAK-2 phosphoryla-tion status in germ tubes and CATs, conidia were grown for four hours under the conditions described above to allow germ tube and CAT formation. Germlings were harvested by gently scraping them off the culture dishes and collected on a Büchner funnel. The harvested vegetative hyphae samples and germ tubes/CATs samples were ground to a fine powder in a mortar and pestle in liquid nitrogen. Protein extraction for the analysis of the MAK-1 and MAK-2 phosphorylation status was performed as described by Maddi et al. [31].

Preparation of cells expressing HA-tagged proteins for immunolocalization
For each HA-tagged strain, six Petri dishes containing 16 ml of conidial suspension (10 6 conidia/ml) were grown at 34uC for 4 hours to allow germ tube and CAT formation. Cells were harvested by gently scraping them off Petri dishes, and were collected by centrifugation at 60006g for 5 minutes. Cells were transferred into a microcentrifuge tube and fixed in PBS (phosphate buffered saline) with 3.7% formaldehyde for 15 minutes. After washing with PBS, the samples were incubated for 30 minutes in PBS containing 1 mg/ml of Novozyme 234 (InterSpex Products Inc., Foster City, CA) and 1% bovine serum albumin to digest the cell wall [49,50]. Following the cell wall digestion step, the cells were collected by centrifugation and washed with PBS. Cell samples were then incubated in permeabilization buffer (PBS with 1% BSA and 0.5% Triton-X 100) for 5 minutes to permeabilize the membrane. After the membrane permeabilization step, the cells were collected and washed with PBS. Cell samples were then incubated overnight at 4uC in mouse monoclonal anti-HA primary antibody (Covance, Princeton, NJ) used at a 1:200 dilution in PBS with 1% BSA. After the primary antibody incubation, samples were washed in PBS three times and then incubated for 2 hours in Alexa Fluor 488conjugated goat anti-mouse secondary antibody (Life Technologies, Carlsbad, CA) used at a 1:100 dilution in PBS with 1% BSA. After three washes, cell samples were used for microscopic observation. All of the incubation buffers and washing buffers used in this assay were supplemented with 1X protease inhibitor cocktail and 1 mM PMSF (Sigma Aldrich, St. Louis, MO).

Live cell imaging for the quantification of CATs
Live-cell imaging was performed using an inverted microscope Diaphot-TMD inverted microscope (Nikon, Japan) to quantify CAT fusion activity for wild type and mutant strains as previously described [8,51]. In brief, 1 ml of fresh conidia at a density of 10 6 per ml were grown in Petri dishes (35610 mm) at 34uC. Benomyl was added to distinguish CATs from germ tubes. This is because benomyl inhibits the formation of long germ tubes but not CAT formation or fusion [52,53]. After 4 hours of incubation, the cells were examined under the microscope for the presence of CATs and CAT fusion. For each sample, five random images were captured for later quantification of CAT fusion activity. The frequency of cell fusion observed in the mutant samples was compared to the wild type cell fusion frequency to obtain a relative CAT fusion activity for each of the mutants.

Sample preparation and confocal microscopy
For live-cell imaging of CAT formation, an inverted agar block method was adapted for this study [54]. A 1 ml aliquot of conidia expressing GFP-tagged protein or RFP-tagged protein (10 6 conidia per ml) was placed in a fluorodish (World Precision Instruments, FL). A thick agar block was then placed in the middle of the fluorodish to facilitate attachment of the conidia to the surface of the cover glass bottom. Benomyl was added in both the liquid medium and the agar block to inhibit germ tube formation. The cells were observed between 3 and 6 hours to follow the formation of CATs.
For imaging of germ tubes and hyphal cells expressing fluorescent protein constructs, conidia (10 5 conidia per ml) were grown in Vogel's sucrose liquid medium for 6 to 8 hours. Cell samples were placed on slides for immediate microscopic observation.
To examine the nuclear localization of MAK-1-GFP and MAK-2-GFP in germ tubes/CATs, conidia (10 6 per ml) were grown in Vogel's sucrose liquid medium at 34uC for 4 hours to allow conidia germination and CAT formation. Cell samples were then collected, fixed, digested with Novozyme, and permeabilized following the procedure described above. After membrane permeabilization, cells were treated with 1 mg/ml RNAase for 30 minutes at room temperature to digest RNA. Cells were then washed once with PBS and stained with propidium iodide (5 mg/ ml propidium iodide in PBS) for 20 minutes [55]. After three washing steps, cells were placed on slides for microscopic observation.
Confocal laser scanning microscopy was performed using a Zeiss LSM 710 Confocal Microscope (Carl Zeiss, Inc., Thornwood, NY). Plan-Apochromat 636/1.40 Oil DIC M27 objective or Plan-Apochromat 406/1.3 Oil DIC M27 objective lenses were used for imaging. Excitation wavelength and detection wavelength were set up either according to references [37,41,56] or by using the smart setup program of the ZEN2012 image capture software. GFP images were collected at 493 to 598 nm with excitation at 488 nm. RFP images were collected at .570 nm with excitation at 558 nm. When imaging samples with both GFP and RFP signals, the images were collected at 493 to 539 nm and at 554 to 703 nm with excitation at 490 nm and 514 nm. When examining GFP protein samples that have been stained with propidium iodide, the images were collected at 499 to 560 nm and at 572 to 719 nm with excitation at 488 nm and 535 nm. The images were collected sequentially using either a line sequential scanning mode or plane scanning mode. Bright-field images were captured with a transmitted light detector. Time-lapse imaging was performed to evaluate Ping-Pong signaling at time intervals of 10 s to 60 s for periods up to 10 minutes. Images were analyzed with image processing software ZEN lite and Image J.
For the detection of HA-tagged proteins, anti-HA primary antibody was used in conjunction with Alexa Fluor 488 conjugated secondary antibody. Immunofluorescent images were collected at 497 to 622 nm with excitation at 488 nm.

Characterization of cell fusion mutants reveals two functional mutant groups
A screening of plates 110 to 120 from the single gene deletion library identified a new cell fusion gene, whi-2 (NCU10518), which is a predicted homolog of yeast stress response factor protein Whi2p [57,58]. We confirmed that the cell-fusion defect and hygromycin resistance co-segregated. Moreover, ectopic expression of HA-WHI-2 at the his-3 locus complemented the Dwhi-2 mutant phenotypes (Figure 1 and 2), demonstrating that the loss of whi-2 was responsible for the mutant defects. An examination of the conidial morphology of Dwhi-2 revealed a conidial phenotype similar to that of the Dham-10 and Damph-1 mutants ( Figure S1). Instead of making mature macroconidia, Dham-10, Damph-1 and Dwhi-2 produced chains of macroconidia that stopped development at the major constriction stage ( Figure S1) [59]. Dham-10 produced fewer conidia than Damph-1 and Dwhi-2. In contrast, Dham-6, Dham-7, Dham-8 and Dham-9 generated normal macroconidia ( Figure S1).
All of the cell fusion mutants had a flat conidiation phenotype, which is due to a defect in the generation of long aerial hyphae ( Figure 1). The phenotypic differences between the Dham-6, Dham-7, Dham-8 and Dham-9 group of mutants, and the Dham-10, Damph-1 and Dwhi-2 mutants suggest that there are functional differences between the proteins encoded by these two groups of cell fusion genes.

HAM-6, HAM-7 and HAM-8 are specifically expressed in germ tubes/CATs
To examine the cell type expression pattern for these cell fusion proteins, we generated HA-tagged versions that were expressed at the his-3 locus under the control of their own promoters. The HAtagged proteins fully rescued the mutant developmental and CAT fusion defects of Dham-6, Dham-7, Dham-8, Dham-9 and Dwhi-2 (Figures 1 and 2) [9]. The HA-tagged version of AMPH-1 provided only a partial rescue of Damph-1 (32.3% of the wild type cell fusion level) (Figures 1 and 2). Western blot analysis was performed to examine the size and expression patterns of the HAtagged proteins (Figure 3). The predicted MW (molecular weight) for HAM-6, HAM-7, HAM-8, HAM-9, AMPH-1 and WHI-2 are 15.9 KD, 24.4 KD, 57.7 KD, 96.7 KD, 29.9 KD and 32.2 KD respectively. HAM-9, AMPH-1 and WHI-2 are predicted to be cytosolic proteins, and their HA-tagged proteins gave MWs very close to the predicted MWs. HAM-6 and HAM-8 contain three and four predicted TM (transmembrane) domains respectively, and the measured MWs of their HA-tagged proteins were also very close to their predicted MWs. HAM-7 has been shown to be a GPI-anchored cell wall protein [31]. The measured MW for HA-HAM-7 was 42 KD, 18 KD larger than the predicted MW, which suggests the GPI-anchored cell wall protein is heavily glycosylated.
HA-HAM-6 and HA-HAM-8 displayed a germ tubes/CATsspecific expression pattern, with only a trace amount of expression in vegetative hyphae ( Figure 3). HA-HAM-7 was expressed at a very high level in germ tubes and CATs, and at a 5-fold reduced level in vegetative hyphae. In contrast, we determined that HA-HAM-9 and HA-WHI-2 were expressed at about equal level in the germ tubes/CATs and hyphae samples, while HA-AMPH-1 was expressed at higher level in hyphae than in germ tubes/CATs (Figure 3). In summary, these expression experiments support our phenotypic classification of the mutants, and suggest that HAM-6, HAM-7, and HAM-8 form a group of proteins that primarily functions in germlings and during CAT fusion, while HAM-9, WHI-2 and AMPH-1 have general functions during growth and differentiation.

HAM-7 and HAM-8 are found in a punctate pattern near the tips of germ tubes and CATs
In order to determine the location of the cell fusion proteins, we expressed them as GFP-and dsRed RFP-tagged constructs under the control of the ccg-1 promoter in their respective mutant backgrounds (Figure 1). The HAM-9-GFP, RFP-HAM-9, HAM-10-GFP and AMPH-1-GFP fusion proteins failed to rescue the mutant phenotypes. We were unable to detect the GFP and RFP signals from these tagged proteins, suggesting that the tagged proteins were rapidly degraded. HAM-8-GFP and RFP-HAM-8 rescued the Dham-8 conidiation defects, but failed to restore CAT fusion activity, indicating that the tagged HAM-8 proteins were only partially functional (Figures 1 and 2). Moreover, some of the GFP and RFP signals were detected at large vacuolar-like structures, suggesting that the tagged HAM-8 proteins might have been targeted to the vacuole for degradation, and may not reflect the normal localization for HAM-8 ( Figure S2A). Co-localization experiments with marker proteins showed that HAM-8-GFP and the vacuolar marker RFP-VAM-3 showed co-localization ( Figure  S2B). RFP-AMPH-1 gave a partial rescue on both conidiation phenotype and CAT fusion activity (18.5% of wild type cell fusion level) (Figures 1 and 2), and localized in a punctate pattern, suggestive of being associated with small vesicles (Figure 4). We also detected RFP signal in larger vacuolar-like structures, which may represent RFP entering into vacuoles and being degraded (see below). RFP-HAM-10 was the only fluorescent fusion protein that fully rescued both mutant conidiation phenotype and CAT fusion activity (Figures 1 and 2). RFP-HAM-10 was found in vesicular or vacuolar-like structure in germ tubes and CATs (Figures 4 and  S3). Because MAK-2 and SO have been detected in association with small vesicles near the tips of CATs, we asked whether the RFP-HAM-10 and RFP-AMPH-1 co-localized with MAK-2-GFP or SO-GFP ( Figures S3 and S4). We did not see evidence for the co-localization of RFP-HAM-10 or RFP-AMPH-1 with either MAK-2-GFP or SO-GFP near the tips of CATs during cell fusion ( Figures S3 and S4).
The HA-tagged versions of HAM-6, HAM-7, HAM-8, HAM-9, AMPH-1, and WHI-2, expressed under the control of their endogenous promoters, provided an alternative opportunity to examine the location of these proteins in fixed cells. Except for HA-AMPH-1, which was only partially functional, the HA-tagged version of these proteins fully rescued the mutant defects (Figures 1  and 2). We were unable to get immunolocalization data for HA-HAM-6 and HA-HAM-9, which was not surprising because these two proteins were expressed at very low levels ( Figure 3). HA-HAM-7 and HA-HAM-8 were localized in a punctate pattern, suggestive of being found in small vesicles or vacuoles ( Figures 5A  and 5B). The intensity of the fluorescent signal for HA-HAM-8 near the tip region of the germ tubes/CATs was consistently found to be significantly higher than the signal in the rest of the cell.
HA-AMPH-1 was localized in a punctate pattern, suggestive of being associated with small vesicle, and in the cytosol ( Figure 5C), consistent with the localization pattern of RFP-AMPH-1 (Figure 4). Significantly, we did not detect HA-AMPH-1 in large vacuolar-like structures, underscoring our suggestion that the GFP and RFP-tagged fusion proteins detected in large vacuolar-like structures have been targeted to the vacuoles for degradation. HA-WHI-2, which is predicted to be a cytosolic protein, also showed  localization to what appear to be small vesicles or vacuoles as well as to the cytosol. ( Figure 5D).

Oscillatory recruitment of MAK-2 and SO to cell tips is abolished in cell fusion mutants
MAK-2 and SO have been shown to be recruited to the cell tips in an oscillatory fashion during CAT fusion [23]. In an effort to identify whether ''Ping-Pong'' signaling is disrupted in our cell fusion mutants, we generated MAK-2-GFP-expressing and SO-GFP-expressing Dham-6, Dham-7, Dham-8, Dham-9, Dham-10, Damph-1 and Dwhi-2 isolates by mating cell fusion mutants with MAK-2-GFP-expressing and SO-GFP-expressing wild type strains of the opposite mating type. Germinating conidia of these mutant strains made CAT-like structures in non-cell fusion contexts at very low frequency. Germ tube germination was not affected in cell fusion mutants ( Figure 6). Microscopic examination of these strains showed that MAK-2-GFP and SO-GFP never localized at the tip of germ tubes or CAT-like structures ( Figure 6).
In separate experiments, mutant conidia expressing MAK-2-GFP and SO-GFP were mixed with an equal number of wild type conidia expressing cytosolic RFP to determine whether the mutant conidia could respond to wild type signals, participate in MAK-2/ SO Ping-Pong signaling, and fuse with wild type conidia. We observed that Damph-1 and Dwhi-2 conidia were able to fuse with wild type at a very low frequency while the remaining mutants never fused with wild type. Interestingly, the few Damph-1 and Dwhi-2 conidia that participated in cell fusion with wild type displayed normal macroconidial morphologies (Figure 7), while the typical, abnormally-shaped mutant conidia did not (see Figure  S1). Because fusions between wild type and Damph-1 or Dwhi-2 conidia were very rare, we were unable to determine whether the oscillatory recruitment of MAK-2-GFP and SO-GFP to the CAT tip occurred in these germling pairs.

MAK-1 and MAK-2 phosphorylation status is affected in mutant germ tubes/CATs and vegetative hyphae
The phosphorylation of MAK-1 and MAK-2 activates the MAP kinases and is required for cell-cell communication and cell fusion. In order to determine whether HAM-6, HAM-7, HAM-8, HAM-9, HAM-10, AMPH-1 and WHI-2 influence the activity of MAK-1 and MAK-2, we looked at the phosphorylation status of both MAPKs in mutant germlings during CAT-inducing conditions ( Figure 8). MAK-1 phosphorylation was dramatically reduced in Dham-6, Dham-7, Dham-8, and Dwhi-2, and slightly reduced in Dham-10. MAK-1 phosphorylation was not reduced in Dham-9 and Damph-1. These results demonstrate that HAM-6, HAM-7, HAM-8, WHI-2, and perhaps HAM-10 are required for activation of the MAK-1 pathway (Figures 8 and S5). In contrast, MAK-2 phosphorylation was reduced in all cell fusion mutants, which may contribute to the MAK-2/SO signaling defect observed in all of the mutants.
We were also interested in assessing the ability of the mutants to activate the MAPK pathways in response to stress-inducing conditions during vegetative growth. Peroxidase treatment has been used as one way to activate MAK-1 and MAK-2 in vegetative cells, and we used peroxidase treatment to evaluate the activation of the MAP kinase pathways in our mutants. Before peroxidase treatment, the MAK-1 and MAK-2 phosphorylation status in mutant vegetative hyphae was similar to the MAK-1 and MAK-2 phosphorylation status in mutant germ tubes/CATs (Figures 8, 9A and 9B) for Dham-6, Dham-7, Damph-1 and Dwhi-2. In Dham-8 and Dham-9, the MAK-1 and MAK-2 phosphorylation levels in vegetative hyphae were below the threshold for In these experiments with vegetative hyphae, MAK-2 was activated to much lower levels in Dham-6 and Dham-9 than in wild type hyphae ( Figure 9B). Dham-8, Dham-10, Damph-1 and Dwhi-   Figure 9B). The differences in MAK-1 and MAK-2 phosphorylation status between germ tubes/ CATs and vegetative hyphae for some of the mutants suggest that there may be differences in how the two MAP kinase pathways are being regulated during the various stages of the N. crassa life cycle.

MAK-1 and MAK-2 nuclear accumulation is normal in mutant germ tubes/CATs
Mutants of the STRIPAK complex have been shown to have a defect in MAK-1 nuclear accumulation, which is regulated by MAK-2 phosphorylation of MOB-3 [34]. MAK-2 nuclear localization is also required during cell fusion [23]. In order to examine if MAK-1 or MAK-2 nuclear accumulation is compromised in the cell fusion mutants, propidium iodide was used to label nuclei in MAK-1-GFP and MAK-2-GFP-expressing mutant strains (Figures S6 and S7). We found that nuclear accumulation

Discussion
Our screen of approximately 11,000 single gene deletion strains from the first 120 plates of the N. crassa single gene deletion library identified 25 genes required for cell-to-cell fusion. This screen identified the MAK-1 and MAK-2 MAPK pathways as well as the STRIPAK complex as three major signaling modules regulating cell fusion [1,6,7,24,34,60]. In this report, we focused our research on seven cell fusion genes whose functions were less well-characterized. HAM-6, HAM-7 and HAM-8 are highly conserved proteins present in all filamentous ascomycetes. The corresponding mutants share the same protoperithecium-deficient, flat conidiation phenotype and are morphologically indistinguishable from each other (Figure 1). They produce abundant macroconidia, but the germinating macroconidia rarely produce CAT-like structures under CAT induction conditions. The ham-6 gene encodes a 145amino-acid protein, the ham-7 gene encodes a 230-amino-acid protein, and the ham-8 gene encodes a 597-amino-acid protein. HAM-6 and HAM-8 were predicted to be membrane proteins with 3 and 4 transmembrane domains respectively. HAM-7 is a GPI-anchored cell wall protein, and we have previously shown that it functions as a MAK-1 pathway sensor during hyphal fusion  [31]. We found that the three proteins were expressed at much higher levels in germ tubes/CATs than in vegetative hyphae ( Figure 3). HAM-7 and HAM-8 were found to be localized in punctate pattern, suggestive of small vesicular or vacuolar structures ( Figures 5A and 5B). The HAM-8 containing structures were found to be concentrated near the tip of the germlings and CATs ( Figure 5B). Although HAM-7 has been shown to be a GPIanchored cell wall protein, we did not see immunolocalization of HAM-7 at the plasma membrane/cell wall boundary. We attribute this to the heavily glycosylated status of HAM-7, which could block to interaction between the glycosylated HAM-7 and the antibody used for immunolocalization. The HAM-7 observed in our localization studies ( Figure 5A) may well represent newly synthesized HAM-7 in transit through the secretory pathway that hasn't been fully glycosylated. Tip localization of MAK-2-GFP and SO-GFP was missing in the few CAT-like structures formed by these mutants and the mutants failed to fuse with wild type conidia (Figures 6 and 7). We found that HAM-6, HAM-7 and HAM-8 are required for MAK-1 kinase activation during conidial germination and CAT formation (Figure 8). Despite the lower levels of expression for HAM-6, HAM-7 and HAM-8 in vegetative cells, MAK-1 phosphorylation was dramatically reduced in Dham-6, Dham-7, and Dham-8 during vegetative hyphal growth. Leeder et al. [1] determined that the expression of the three genes is coregulated and controlled by the MAK-2 pathway-dependent transcription factor PP-1. In summary, we propose that the GPIanchored cell wall HAM-7 and the two transmembrane proteins, HAM-6 and HAM-8, function together to regulate the MAK-1 pathway. Given the cell wall/plasma membrane location for the GPI-anchored protein HAM-7, we suggest that the three proteins might participate in a signaling complex at the cell wall/plasma membrane boundary, but our data would also be consistent with a signaling complex localized to intracellular membranes.
The ham-9 gene encodes an 869-amino-acid protein containing a SAM domain and two PH domains. The SAM domain has been identified in yeast Ste11p (S. cerevisae homolog of N. crassa NRC-1) [61], and the PH domains have been suggested to play a role in targeting signal transduction proteins to intracellular membrane in signaling events [62]. The C-terminal GFP-tagged and N-terminal RFP-tagged HAM-9 fusion proteins were not functional, precluding any live-imaging analysis. HA-HAM-9 was expressed in both vegetative hyphae and germ tubes/CATs (Figure 3), but its expression level was too low for immunolocalization. The requirement of HAM-9 for both MAK-1 and MAK-2 activation in vegetative hyphae ( Figures 9A and 9B), may suggest that HAM-9 regulates cross-communication of the two MAPK pathways during vegetative growth.
The amph-1 gene encodes a 262-amino-acid protein containing a bar domain, a domain frequently involved in protein-protein interaction and regulation of membrane curvature [63]. N. crassa AMPH-1 is a homolog of the yeast Rvs161p and Rvs167p proteins. Rvs161p and Rvs167p are required for endocytosis and cell fusion during yeast mating [63,64]. HA-AMPH-1 and RFP-AMPH-1 localized to small vesicles in the germ tubes/CATs, and some of the vesicles appeared to be associated with the plasma membrane (Figures 4 and 5C), suggesting that N. crassa AMPH-1 plays a role in vesicular trafficking and endocytosis. MAK-1 and MAK-2 activity in Damph-1 germlings were similar to the wild type control, indicating that AMPH-1 does not affect cell fusion by regulating these pathways. This is consistent with our observation that a few Damph-1 conidia having wild type morphology were able to participate in cell fusion with wild type conidia (Figure 7). HA-AMPH-1 was expressed in both vegetative hyphae and germlings, suggesting it is a general factor required for all stages of N. crassa life cycle. In summary, we suggest that AMPH-1 functions during vesicular trafficking and endocytosis.
The ham-10 gene encodes a 1,422-amino-acid protein containing a C2 domain near the C terminus. C2 domains function as calcium-dependent lipid-binding domains and are thought to be involved in vesicular trafficking, exocytosis, and signal transduction [65]. HAM-10 tagged with RFP at its N terminus fully rescued Dham-10, but HAM-10 tagged with GFP at the C terminus did not (Figures 1 and 2), suggesting that modification near the C terminal C2 domain may affect the function and stability of HAM-10. RFP-HAM-10 localized in the cytosol and in a punctate pattern, suggestive of a vesicular or vacuolar network location ( Figure 4). However, our proposed localization of HAM-10 should be considered as a tentative assignment. We did not demonstrate that the RFP-tag remained attached to HAM-10, nor have we carried out extensive co-localization studies with known vesicle and vacuolar marker proteins to definitively demonstrate co-localization of the RFP-HAM-10 with organelle-specific markers. Our results demonstrate that HAM-10 was not enriched at CAT tips during cell fusion (Figures 4 and S3). The requirement of HAM-10 in both MAPK pathways during different development stages suggests HAM-10 could be a general factor in regulating cell growth.
The whi-2 gene encodes a 298-amino-acid protein with homology to the yeast general stress response protein Whi2p, which has been shown to activate autophagy and mitophagy under nutrient starvation conditions [66,67]. N. crassa Dwhi-2 displayed a conidial development defect ( Figure S1). HA-WHI-2 was expressed in both vegetative hyphae and germ tubes/CATs and was localized in cytosol and in a punctate pattern suggestive of small vesicles or vacuoles (Figures 3 and 5D). MAK-2/SO signaling was abolished in Dwhi-2 ( Figure 6), but we found that a few mutant macroconidia with wild type morphology were able to participate in cell fusion with wild type conidia (Figure 7). Interestingly, the phosphorylation levels of both MAK-1 and MAK-2 were reduced in germlings and during vegetative hyphal growth (Figures 8 and 9), suggesting WHI-2 functions as a general stress response factor regulating both MAPK pathways.
In summary, our studies on the seven cell fusion genes confirmed that the MAK-1 and the MAK-2 pathways play critical roles during conidia germination and CAT fusion. Figure 10 shows a diagrammatic representation of a CAT tip with many of the proteins we have discussed. The phenotypic characteristics, cell type-specific expression patterns, cellular locations, and MAP kinase activity status of HAM-6, HAM-7 and HAM-8 suggest that the three proteins may form a multimeric sensor complex at the cell wall/plasma membrane or on intracellular vesicles and regulate MAK-1 activation during CAT fusion. Our studies on HAM-9, HAM-10, AMPH-1 and WHI-2 suggest that cell fusion is also affected in mutants lacking proteins with general functions in growth and development. HAM-10, AMPH-1 and WHI-2 clearly play a role in conidial development as well as during CAT fusion. The importance of HAM-9, HAM-10 and WHI-2 for both MAK-1 and MAK-2 signaling may provide opportunities to study crosstalk regulation between MAP kinase pathways and other signaling modules.  Figure S5 Ponceau stain for MAK-1 and MAK-2 phosphorylation status in mutant germ tubes/CATs. Ponceau stain image is shown below the Western blot image for the MAK-1 and MAK-2 phosphorylation status in wild type (WT) and mutant germ tubes/CATs. The Western blot image is found as Figure 8 in the manuscript and the Ponceau stain image is given here to demonstrate equal loading of the samples used in the Western blot. (TIF) Figure S6 Nuclear localization of MAK-1-GFP in mutant germ tubes. Propidium iodide was used to stain nuclei in MAK-1-GFP-expressing wild type (WT) and mutant germ tubes. The figure shows DIC images, GFP fluorescent images, propidium iodide red fluorescent images, and merged images (from left to right respectively). Images are shown for wild type (WT) (row 1), Dham-6 (row 2), Dham-7 (row 3), Dham-8 (row 4), Dham-9 (row 5), Dham-10 (row 6), Damph-1 (row 7), and Dwhi-2 (row 8) germ tubes. (TIF) Figure S7 Nuclear localization of MAK-2-GFP in mutant germ tubes. Propidium iodide was used to stain nuclei in MAK-2-GFP-expressing wild type (WT) and mutant germ tubes. The figure shows DIC images, GFP fluorescent images, propidium iodide red fluorescent images, and merged images (from left to right respectively). Images are shown for wild type (WT) (row 1), Dham-6 (row 2), Dham-7 (row 3), Dham-8 (row 4), Dham-9 (row 5), Dham-10 (row 6), Damph-1 (row 7), and Dwhi-2 (row 8) germ tubes. (TIF)