Early Colony Establishment in Neurospora crassa Requires a MAP Kinase Regulatory Network

Abstract

Vegetative fusion is essential for the development of an interconnected colony in many filamentous fungi. In the ascomycete fungus Neurospora crassa, vegetative fusion occurs between germinated conidia (germlings) via specialized structures termed “conidial anastomosis tubes” (CATs) and between hyphae within a mature colony. In N. crassa, both CAT and hyphal fusion are under the regulation of a conserved MAP kinase cascade (NRC1, MEK2, and MAK2). Here we show that the predicted downstream target of the MAK2 kinase pathway, a Ste12-like transcription factor known as PP1, regulates elements required for CAT and hyphal fusion. The PP1 regulatory network was revealed by expression profiling of wild type and the Δpp-1 mutant during conidial germination and colony establishment. To identify targets required for cell fusion more specifically, expression-profiling differences were assessed via inhibition of MAK2 kinase activity during chemotropic interactions and cell fusion. These approaches led to the identification of new targets of the cell fusion pathway that, when mutated, showed alterations in chemotropic signaling and cell fusion. In particular, conidial germlings carrying a deletion of NCU04732 (Δham-11) failed to show chemotropic interactions and cell fusion. However, signaling (as shown by oscillation of MAK2 and SO to CAT tips), chemotropism, and cell fusion were restored in Δham-11 germlings when matched with wild-type partner germlings. These data reveal novel insights into the complex process of self-signaling, germling fusion, and colony establishment in filamentous fungi.

CELL fusion between genetically identical cells is important in development (for example, myoblast fusion during muscle formation) and occurs in many multicellular organisms from simple ascomycete fungi to mammals (Chen et al. 2007; Aguilar et al. 2013). Cell fusion between genetically identical cells can be mediated by cells that have differentiated, but in some cases, also between cells in an identical developmental state, for example, cell fusion between germinating asexual spores (conidia) of filamentous fungi (Pandey et al. 2004; Roca et al. 2005a; Read et al. 2010). In filamentous fungi, these fusions are integral to the formation of an interconnected hyphal network, which mediates genetic mixing and the sharing of resources (Simonin et al. 2012; Roper et al. 2013). How this process is initiated and maintained and what proteins are involved are still mostly unknown.

In filamentous ascomycete fungi, a conserved MAP kinase pathway that is involved in pheromone response and mating in Saccharomyces cerevisiae (Ste11, Ste7, and Fus3) (Bardwell 2005) is required for cell fusion and heterokaryon formation during vegetative growth (Hou et al. 2002; Wei et al. 2003; Pandey et al. 2004; Fu et al. 2011; Jun et al. 2011; Dettmann et al. 2012). This conserved pathway also plays a role in sexual development and secondary metabolism and is required for the virulence of both plant and animal fungal pathogens (Roman et al. 2007; Rispail and Di Pietro 2010; Bayram et al. 2012). In S. cerevisiae, reception of a mating-type-specific pheromone signal results in the activation of Ste11 (MEKK), Ste7 (MEK), and Fus3 (MAPK) (Bardwell 2005). After Fus3 is activated, it enters the nucleus and phosphorylates Ste12, a mating-type-specific transcription factor, as well as two regulators of Ste12, Dig1 and Dig2 (Blackwell et al. 2007). Activated Ste12 regulates the expression of many genes involved in the mating process, both indirectly and directly by binding pheromone response elements in target promoters (Zeitlinger et al. 2003).

Ste12-like proteins have also been studied in filamentous fungi, where they play essential roles in development and pathogenicity (Alspaugh et al. 1998; Vallim et al. 2000; Borneman et al. 2001; Park et al. 2002; Tsuji et al. 2003; Li et al. 2005; Nolting and Poggeler 2006; Ren et al. 2006; Tollot et al. 2009; Rispail and Di Pietro 2010; Wong Sak Hoi and Dumas 2010). Fus3-like proteins in filamentous fungi could directly or indirectly phosphorylate Ste12-like proteins, but this has not yet been shown in any system. However, strains carrying deletions of Fus3-like proteins or Ste12-like proteins often show a similar phenotype in many filamentous fungi, suggesting that Fus3-like kinases positively activate Ste12-like proteins. All Ste12-like proteins contain a divergent homeodomain (STE) near their N terminus, which is involved in DNA binding (Errede and Ammerer 1989). Ste12-like proteins in filamentous fungi also contain two C-terminal C2H2-Zn²⁺ motifs (Vallim et al. 2000; Chang et al. 2004) (Figure 1A) that are absent in S. cerevisiae and other related ascomycete yeast species. In the plant pathogen Magnaporthe grisea, both the STE domain and the C2H2-Zn²⁺ motifs of the STE12 homolog, mst12, are essential for the development of appressorial penetration pegs (Park et al. 2004), while in the human basidiomycete pathogen Cryptococcus neoformans, the STE domain and the C2H2-Zn²⁺ motifs were important for function, although mutagenesis of each domain resulted in strains with different phenotypes (Chang et al. 2004). In S. cerevisiae, Ste12-binding partners include Mcm1 and Tec1 (Primig et al. 1991; Madhani and Fink 1997). In Sordaria macrospora, a physical interaction between STE12 and MCM1 is required for sexual spore formation (Nolting and Poggeler 2006).

Figure 1.

The homeodomain of PP1 is required for function. (A) Schematic representation of PP1 with protein domains depicted. Point mutations resulting in amino acid substitutions in the respective protein domains are indicated. (B) Phenotypic results of introduction of wild-type and mutated pp-1 alleles into a Δpp-1 mutant. Strains with alleles are listed on the left, and phenotypic analyses are depicted or listed on the right.

In Neurospora crassa, proteins orthologous to the Fus3 MAP kinase pathway (NRC1, MEK2, MAK2) regulate vegetative cell fusion (Pandey et al. 2004; Fleissner et al. 2008; Fu et al. 2011; Dettmann et al. 2012). Cell fusion events are particularly frequent in germinating asexual spores (conidia) and are associated with the production of specialized fusion structures termed “conidial anastomosis tubes” (CATs) (Roca et al. 2005b). Chemotropic interactions occur when two neighboring germlings sense each other and redirect the growth of CATs. During chemotropic interactions, NRC1, MEK2, and MAK2 oscillate together from CAT tip to cytoplasm. The oscillation of NRC1/MEK2/MAK2 occurs every 4 min and is perfectly out of phase with a second protein, SO, which also oscillates from CAT tip to cytoplasm in communicating germling pairs (Fleissner et al. 2009b; Dettmann et al. 2012). A STE12 homolog has also been identified in N. crassa (pp-1). A Δpp-1 mutant shows slow growth, is female sterile (they do not develop protoperithecia), and displays ascospore lethality (Li et al. 2005), phenotypes that are very similar to that displayed by Δmak-2 mutants. Previous analyses using a N. crassa partial-genome (13%) microarray identified genes whose expression is altered in a mature colony (7 days old) in the absence of mak-2 or pp-1. These analyses revealed that PP1 is involved in the regulation of sexual development, asexual development, and secondary metabolism (Li et al. 2005).

In this study, we evaluated the role of PP1 during cell fusion of asexual spores and expanded analyses using whole-genome arrays and RNA-seq to examine the function of MAK2 and PP1 during germling fusion and early colony establishment. These analyses revealed MAK2- and PP1-dependent gene networks involved in various aspects of colony development and identified novel genes with important roles in regulating communication and chemotropic attraction between N. crassa germlings.

Materials and Methods

Molecular techniques and strain construction

All strains constructed and used in this study are listed in Table 1 and Table 2. Strains were grown on Vogel’s minimal medium (VMM) (Vogel 1956) (with supplements as required) and were crossed on Westergaard’s medium (Westergaard and Mitchell 1947). Transformations and other N. crassa molecular techniques were performed as described (Colot et al. 2006) or using protocols available at the Neurospora home page at the Fungal Genetics Stock Center (FGSC) (http://www.fgsc.net/neurospora/neurosporaprotocolguide.htm).

Table 1. N. crassa strains used in this study.

Name	Mutant genotype	Source
R13-5	Δpp-1 A	Li et al. (2005)
	Δpp-1 a	This study
FGSC 9537	Dip	FGSC
FGSC 6103	his-3 A	FGSC
	his-3; Δpp-1 A	This study
	his-3+::Pccg1 pp-1W155C a	This study
	his-3+::pp-1W155C a	This study
	his-3+::pp-1W155C, C623S, C626S a	This study
	his-3+::Pccg1 pp-1C623S, C626S a	This study
	his-3+::pp-1C623S, C626S a	This study
	his-3+::Pccg1 pp-1 a	This study
	his-3+::pp-1 a	This study
mal1	his-3+::Pccg1 mak-2Q100G; Δmak-2 A	Fleissner et al. (2009b)
FGSC 9718	Δmus-51::bar a	FGSC
FGSC 988	Oak Ridge WT a	FGSC
FGSC 2489	Oak Ridge WT A	FGSC
R9-08	hex-1 A	Jedd and Chua (2000)
AF-SoT8	his-3+::Pccg1 so-gfp A	Fleissner et al. (2009b)
AF-M512	his-3+::Pccg1 mak-2-gfp A	Fleissner et al. (2009b)
	his-3; Δham-7 A	This study
	his-3+::Pccg1 so-gfp; Δham-7 A	This study
	his-3+::Pccg1 mak-2-gfp; Δham-7 A	This study
	his-3; Δham-11 A	This study
	his-3+::Pccg1 so-gfp; Δham-11 A	This study
	his-3+::Pccg1 mak-2-gfp; Δham-11 A	This study
	his-3+::Ptef1 ham-11-gfp; Δham-11 A	This study
	his-3+::Pccg-1 Prm-1-gfp; ΔPrm-1 A	Fleissner et al. (2009a)

Table 2. Fusion phenotype of strains carrying deletions that showed significant differences in expression level in the Δpp-1 mutant relative to wild type.

NCU ID no.	WT RPKM	Δpp-1 RPKM	Protein description	Deletion origin	Fusion % (WT = 86 ± 4)a
NCU00309b	333	527	Hypothetical	FGSC 19692	WT
NCU00811	1302	630	Hypothetical	This study	WT
NCU00881	145	67	HAM7	13775	0
NCU00995	274	0	Hypothetical	This study	WT
NCU01380	413	1	Hypothetical	This study	WT
NCU01697	130	2	Hypothetical	FGSC 14575	WT
NCU02361	42	117	Formamidase	FGSC 21969	WT
NCU02500	49	642	CCG4	FGSC 14944	WT
NCU03013	368	25	ACW10	FGSC 11222	WT
NCU03960	200	24	HAM12	FGSC 17233	66 ± 6
NCU04122	269	19	Malate dehydrogenase	FGSC 21282	WT
NCU04192	70	42	Vacuolar aspartyl aminopeptidase	FGSC 18884	WT
NCU04732	179	65	HAM11	FGSC 17545 and this study	0
NCU05814	145	48	Hypothetical	FGSC 17770	WT
NCU06698	79	49	Glycogenin	FGSC 12301	WT
NCU07503	77	300	Hypothetical	FGSC 14018	WT
NCU07802	148	1	Hypothetical	FGSC 20487	WT
NCU08332	1422	448	HEX1	Jedd and Chua (2000)	WT
NCU08824	37	73	Molybdopterin-binding domain	FGSC 15980	WT
NCU09560	66	150	Superoxide dismutase	FGSC 21068	WT
NCU09562	501	5	Hypothetical	FGSC 21874	WT
NCU09693	116	5	Hypothetical	This study	WT

^aMutants show fusion frequencies similar to their isogenic wild-type parent (FGSC 2489).

^bGenes in boldface type showed increased expression levels in the Δpp-1 mutant.

To construct the pp-1 site-directed mutant alleles, fusion PCR was performed with the mutation included in the overlapping primer region (Supporting Information, Table S1). Alleles were cloned into his-3 targeting plasmid pMF272 (AY598428) (Margolin et al. 1997; Freitag et al. 2004) and integrated into the his-3 locus of a his-3; Δpp-1 strain. Δpp-1 and Δpp-1 transformant strains showing an ascospore lethal phenotype were mated to the Diploid (Dip) mutant strain (FGSC 9537) as described in Hutchison et al. (2009) to construct homokaryotic strains or strains containing auxotrophic markers. In Dip crosses, ∼2/3 of the ascospores are large and diploid. Streaking Dip progeny onto sorbose plates results in restoration of haploid strains. The pp-1 sequence in all engineered strains was confirmed by DNA sequence analyses.

Deletion strains were obtained from the FGSC (McCluskey 2003) and were generated as part of the N. crassa functional genomics project (Colot et al. 2006; Dunlap et al. 2007). For each deletion strain, both the mating type A and mating type a strains were analyzed, if available. The phenotypes of deletion strains that were available only as a single mating type from the FGSC (McCluskey 2003) were confirmed by segregation analysis. Ascospore progeny were selected from crosses with FGSC 2489 (mating type A) or FGSC 988 (mating type a) and assessed for phenotype and hygromycin resistance. To pass this test, hygromycin resistance had to segregate with fusion phenotype. In cases where deletion strains were not publicly available (genes with locus identification numbers: NCU00811, NCU00995, NCU01380, and NCU09693), we constructed deletion mutants using a mus-51 background as described (Colot et al. 2006). Primers (Table S1) were designed to amplify the upstream and downstream flanks of these four genes, as well as the hygromycin cassette from plasmid pCSN44 (Colot et al. 2006). The two flanks, the hygromycin cassette and digested plasmid pRS426 (EcoRI and XhoI), were transformed into yeast, and recombined plasmids were obtained via plasmid rescue (Colot et al. 2006). Hygromycin transformants were crossed to wild type (WT) (FGSC 2489) to eliminate the mus-51 mutation and to obtain homokaryotic progeny. The genotype of all deletion strains was confirmed by PCR, using a primer for the hygromycin cassette (HygRupflank) paired with a gene-specific primer from the flank region, as well as a gene-specific primer outside the gene and a gene-specific primer within the open reading frame (Table S1).

To construct the internally GFP-tagged ham-11 construct, we amplified ham-11 using primers 04732F and 04732R (Table S1) and cloned it into a pCR-Blunt vector (Invitrogen). We sequenced and digested the amplicon from the vector with BglII and ApaI. The tef1 promoter was removed from plasmid ptef1-mek-2-gfp (Dettmann et al. 2012) with NotI and BamHI. Using a three-point ligation, we cloned the tef1 promoter and ham-11 amplicon into plasmid pmf272 (Freitag et al. 2004), using NotI and ApaI. GFP containing codons for eight glycine linkers and a BamHI restriction site on both ends was amplified from plasmid pMF272 using primers Fgfp+8xgly and Rgfp+8xgly (Table S1) and cloned into a pCR-Blunt vector (Invitrogen). We sequenced the amplicon, digested it with BamHI, and ligated it into the BamHI site in ham-11 situated between the two predicted transmembrane domains. This ham-11 construct with internally tagged gfp was transformed into the Δham-11 his3⁻ strain with selection for His prototrophy. A homokaryotic strain was obtained via microconidia purification (Pandit and Maheshwari 1994).

Phenotypic analyses

Macroscopic growth differences were assessed by growing the different strains (WT, Δham-7, Δham-11, and Δham-12) in flasks for 1 week at 25° and on VMM plates for 2 days at 25°. Growth rate on VMM plates was measured at 25° by putting 1 × 10⁶ conidia in the center of the plates and measuring the radial growth front after 24 and 48 hr. Aerial hyphae extension was determined by inoculating tubes containing 1 ml of liquid VMM with 1 × 10⁶ conidia; aerial hyphae height was measured after 3 days of growth at 25° in constant light. Ten replicates were measured for each strain.

To assess the frequency of communication and fusion between conidia within a deletion strain as compared to wild type, slant tubes containing the strains were grown for 4–6 days or until significant conidiation occurred. Conidia were harvested by vortexing the slant tube with 2 ml ddH₂O and subsequently filtered by pouring over cheesecloth to remove hyphal fragments. Conidia were diluted to a concentration of 3.3 × 10⁷ conidia/ml. For each sample, 300 μl of spore suspension was spread on a 9-cm solid VMM plate. The plates were dried in open air for 20–30 min and incubated for 3–4 hr at 30°. Squares of 1 cm were excised and observed with a Zeiss Axioskop 2 using a ×40 Plan-Neofluor oil immersion objective. For each strain, 50–100 germling pairs were counted in triplicate. The ability to communicate is given as a percentage of pairs that display homing behavior when germinated conidia are within ∼15 μm of each other. To assess the frequency of communication between conidia of a deletion strain with wild-type germlings, mixtures of mak-2-gfp- or so-gfp-expressing strains and nonfluorescing strains were prepared and similarly treated as described above.

Fluorescence microscopy

Molecular techniques, including SO-GFP and MAK2-GFP strain construction, were previously described (Fleissner et al. 2009b). The strain used to cross so-gfp into deletion strains was AF-SoT8, and the strain used to cross mak-2-gfp into deletion strains was AF-M512 (Table 1) (Fleissner et al. 2009b).

Oscillation studies performed with MAK2-GFP and SO-GFP were conducted as described previously (Fleissner et al. 2009b); conidia (0.5 × 10⁷ spores/9-cm plate) from MAK2-GFP or SO-GFP strains were mixed with equal amounts of conidia from the respective deletion mutants and plated on solid VMM, dried, and incubated at 30° for 3–4 hr before being used for microscopy. Images were taken using a Leica SD6000 microscope with a ×100 1.4 N.A. oil-immersion objective equipped with a Yokogawa CSU-X1 spinning disk head and a 488-nm laser controlled by Metamorph software. Multiple pairs of interacting germlings were analyzed per experiment, and representative pairs are shown for each strain.

Microarray analysis and quantitative RT-PCR

For the PP1 microarray experiment, RNA was isolated from wild-type or Δpp-1 conidia grown in constant light at 25° for 3–8 hr in VMM at 25° with constant light. Total RNA from frozen samples was isolated using Zirconia/Silica beads (0.5-mm diameter; Biospec) and a Mini- Beadbeater-96 (Biospec) with 1 ml TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. The total RNA was further purified by digestion with TURBO DNA-free (Ambion) and an RNeasy kit (Qiagen). RNA concentration and integrity was checked by Nanodrop and agarose gel electrophoresis.

For the mak-2^Q100G microarray experiment, strains were grown in liquid VMM for 5 hr at 25° with constant light; 20 min prior to harvesting, cells were treated either with 20 µM 1NM-PP1 in DMSO or with DMSO only.

Microarray hybridization and slide image analysis were as described previously (Tian et al. 2007). We chose to use a closed-circuit design for microarray comparisons, which is a statistically robust method of identifying differentially regulated genes over a time series, and used Bayesian Analysis of Gene Expression Levels (BAGEL) (Meiklejohn and Townsend 2005) to determine normalized gene expression values. Microarray profiling data are deposited at the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) as series record GSE46912 (mak-2 microarray) and GSE49679 (pp-1 microarray). Genes were clustered based on their differential expression in Δpp-1 over the time period using Hierarchical Clustering Explorer 3.0 (de Hoon et al. 2004). The MIPS functional catalog (FunCat) was used to assign functional annotation to genes (Ruepp et al. 2004). Enrichment of FunCat categories in gene sets vs. the whole genome was calculated using hypergeometrix distribution for P-value calculation.

Quantitative RT-PCR was used to confirm trends in the microarray and RNA-seq data, and was performed using the EX-PRESS One-Step SYBR GreenER Kit (Invitrogen) and the StepOnePlus Real-Time PCR System (Applied Biosystems). Reactions were performed in triplicate with a total reaction volume of 20 μl including 400 nM each of forward and reverse primers and 60 ng template RNA. Data analysis was performed by the StepOne Software (Applied Biosystems), using the Relative Quantitation/Comparative CT (ΔΔCT) setting. Data were normalized to the endogenous control actin. The primers used are listed in Table S1.

RNA-seq

RNA was isolated from conidia grown in constant light for 5 hr in VMM at 25°, originally inoculated at a concentration of 1 × 10⁶ cells/ml. RNA-seq methods were as described in Ellison et al. (2011). The protocol used for library construction was modified from the standard Illumina method and has been described previously (Ellison et al. 2011). An Illumina Genome Analyzer (http://qb3.berkeley.edu/qb3/gsl/index.cfm) was used to sequence reads, and the analysis program TopHat (Langmead et al. 2009) was used to process the data. Reads per kilobase mapped (RPKM), a normalized measure of gene expression, were calculated using Cufflinks (Roberts et al. 2011). Fisher’s exact test was used to test the significance of differences between wild type and Δpp-1, with a significance cutoff of P < 0.001. Profiling data (pp-1 RNA-seq) are deposited at the GEO database (http://www.ncbi.nlm.nih.gov/geo/) series record (GSE50446).

Cell fractionation and Western blotting

Harvested conidia (1 × 10⁶/ml) were inoculated in 100-ml VMM in flasks, incubated for 2.5 hr at 30° with shaking at 200 rpm and then an additional 2.5 hr at 30° without shaking. Germlings were harvested by vacuum filtration over a PVDF membrane and frozen in liquid nitrogen. Protein extraction from ground mycelium was performed using lysis buffer as described in Pandey et al. (2004) containing complete protease inhibitors but without phosphatase inhibitor and Triton X-100. Cell fractionation via centrifugation was performed as in Bowman and Bowman (1988). The P1 pellet (vacuoles, mitochondria, and endomembranes) and P2 pellet (plasma membrane) were dissolved in lysis buffer with 1% Triton X-100. The remaining supernatant from P2 was used for immunoprecipitation using Protein G Dynabeads (Invitrogen), according to manufacturer’s instructions, with the following exceptions: mouse anti-GFP antibody (Roche) was covalently bound to the beads using BS³ (Sulfo-DSS, Fisher Scientific) according to manufacturer’s instructions. Supernatant samples were incubated with the beads for 3 hr at room temperature. Protein was removed from the beads by heating at 70° for 10 min, and samples were run on a 4–12% Nu-Page Bis-Tris Gel (NOVEX, Life Technologies). Protein gels were subjected to Western blot analysis using standard methods. Samples for the MAK1 and MAK2 phosphorylation Western blot were treated similarly, except, after protein extraction with 1 ml lysis buffer, 25 µl of protein sample was directly loaded on a 7% NuPage Bis-Tris Gel (NOVEX, Life Technologies). Gels were subjected to Western blot analysis using standard methods, and detection of phosphorylated MAK1 and MAK2 was carried out using anti-phospho p44/42 MAP kinase antibodies (1:3000 dilution) (PhosphoPlus antibody kit; Cell Signaling Technology) as described (Pandey et al. 2004).

Results

The homeodomain is required for PP1 function

In filamentous fungi, Ste12-like proteins possess two C2H2-Zn²⁺ motifs, in addition to the well-conserved homeodomain-like STE domain (Park et al. 2002). The STE domain is essential for function and DNA binding in S. cerevisiae (Yuan and Fields 1991). The function of these two domains in N. crassa was investigated through the construction of alleles containing point mutations in the DNA-binding motifs. In the first pp-1 allele, a point mutation (W155C) was made in the STE domain; this residue is in a highly conserved region known to be crucial for homeodomain–DNA interactions (Yuan and Fields 1991). The cysteine residues within the C2H2-Zn²⁺ motifs are important for DNA-binding activity in zinc-finger transcription factors. We therefore constructed a second pp-1 allele, in which the two cysteine residues of one of the C2H2-Zn²⁺ motifs were mutated to serine (C623S and C626S); similar mutations in a Ste12-like gene in C. neoformans resulted in a strain showing an altered virulence phenotype (Chang et al. 2004). In the third pp-1 allele, the point mutations in both the STE and C2H2-Zn²⁺ motifs were combined to create a double-mutant pp-1 allele (Figure 1A). These alleles, plus a wild-type pp-1 allele, were fused to GFP at the C terminus and transformed into the his-3 locus of a his-3; Δpp-1 mutant; selection was for His prototrophy (Margolin et al. 1997). The wild-type pp-1 and the pp-1^C623S;C626S alleles fully rescued the Δpp-1 phenotype; strains were indistinguishable from wild type when assessing vegetative growth, cell fusion, female fertility, and ascospore viability (Figure 1B). In contrast, strains bearing the pp-1^W155C allele or the pp-1^{W155C;C623S;C626S} allele showed a phenotype identical to the Δpp-1 phenotype. These strains showed no cell fusion and were sterile (Figure 1B). These data indicate that the STE domain is essential for all characterized functions of PP1, while the C2H2-Zn²⁺ motif is apparently dispensable.

To assess the phenotypic consequences of overexpression of the wild-type pp-1 and mutant alleles, the pp-1, pp-1^W155C, and pp-1^C623S;C626S alleles were cloned behind the ccg-1 promoter (McNally and Free 1988) and introduced into the Δpp-1 strain at the his-3 locus. Similar to the pp-1^C623S;C626S strains regulated by the pp-1 promoter, strains carrying the pp-1^C623S;C626S allele under the regulation of the ccg-1 promoter showed a wild-type vegetative, cell fusion, and protoperithecial development phenotype (Figure 1B). However, in contrast to strains bearing the pp-1^W155C allele regulated by the pp-1 promoter, strains bearing the pp-1^W155C allele under the regulation of the ccg-1 promoter also showed a wild-type vegetative growth phenotype, as well as normal cell fusion levels, protoperithecial development, and ascospore germination. These data suggest that the mutated STE domain can still bind DNA at very low efficiency, which is presumably enough to rescue the Δpp-1 phenotype when overexpressed.

PP1-GFP was undetectable when pp-1-gfp expression was driven by the native pp-1 promoter. However, PP1-GFP localized to nuclei during chemotropic interactions and cell fusion when pp-1-gfp was expressed from the ccg-1 promoter (Figure 2A). PP1-GFP driven by the ccg-1 promoter also showed nuclear localization during colony establishment, vegetative growth, and conidiation (data not shown); differential localization of PP1 during growth or reproduction was not apparent. Nuclear localization was also observed for strains bearing the pp-1^W155C and pp-1^C623S;C626S alleles, indicating that these point mutations did not interfere with nuclear localization of PP1.

Figure 2.

The role of pp-1 in hyphal network formation. (A) PP1-GFP shows nuclear localization during chemotropic growth in germlings. (B) The closed-circuit experimental design used for microarray analysis. The 4- to 6-hr subpopulation loop is marked in the dashed box. (C) Phenotypic analyses of conidial germination and hyphal network formation in wild type corresponding to identical time points as the 8-hr microarray time course. Arrows denote germling hyphal fusion events. Bar, 10 μm. (D) Visual representation of clustered gene expression profiles from Hierarchical Clustering Explorer analysis (de Hoon et al. 2004). Across each horizontal row (representing a single gene from the 3- to 8-hr germination time course), red indicates the highest dependence on PP1 for expression and green indicates the lowest dependence.

Transcriptional analysis of PP1 function during early colony development

We used full-genome microarrays (Tian et al. 2007) to monitor gene expression profiles in wild-type and Δpp-1 during germination, germling fusion, and early establishment of a colony. Gene expression differences between wild type and the Δpp-1 mutant were measured 3, 4, 5, 6, and 8 hr following inoculation of conidia into liquid media (Figure 2B). After 3 hr, newly germinated conidia and some chemotropic interactions were observed, which increased over time. Germling fusion events were prevalent after 5 hr, with ∼85% of conidia having undergone fusion events by this time point. Hyphal growth and branching started after 6 hr post-inoculation, and elongated and branched hyphae were abundant at 8 hr (Figure 2C).

A total of 5114 genes showed measurable expression levels throughout the time course, but only 92 genes showed expression at all time points and statistically significant expression differences of at least 1.5-fold between wild type and the Δpp-1 mutant at one or more time points (Table S2). Consistent with its role as a transcriptional activator, >50% of the genes showed decreased expression levels in the Δpp-1 mutant as compared to those with increased expression levels (60 down-regulated vs. 32 up-regulated).

Hierarchical clustering of the 92 genes during the 3- to 8-hr time course revealed four distinct clusters (Figure 2D; Table S2). The largest cluster of genes (cluster 1) showed differential expression levels during the 3- to 5-hr time course, when the process of chemotropic interactions and germling fusion events is most prevalent. A second cluster (cluster 2) showed differential expression specifically at the first time point (3 hr), suggesting that these genes may be important for the conidial germination process. The third and the fourth clusters showed differential expression during the 6- to 8-hr time course, suggesting that these genes are involved in the network formation and hyphal growth processes.

ccg-4, which encodes the mating type A pheromone, showed increased expression levels in a mating type A Δpp-1 mutant as compared to a wild-type mating type A strain (FGSC 2489) (Table S2). Increased expression levels of ccg-4 were also reported using partial genome microarrays and Δpp-1 A and Δmak-2 A strains (Li et al. 2005). We confirmed this observation by quantitative RT-PCR (qRT-PCR) from wild type A and Δpp-1 A strains (Figure 3). However, in mating type a strains as well as in Δpp-1 a strains, expression of ccg-4 was very low. We therefore also evaluated the expression of the mating type a pheromone gene, mfa-1, in wild type A and a strains, as well as in Δpp-1 A and a mating type strains; mfa-1 expression levels were significantly higher in the Δpp-1 a cells as compared to a wild-type a strain (Figure 3). To determine whether mutations in Δpp-1 also affect expression of the pheromone receptors, we assessed expression levels of pre-1 (a pheromone receptor) and pre-2 (A pheromone receptor) by performing qRT-PCR with mating type A and a Δpp-1 mutants as compared to wild-type mating type A and a strains. While expression levels for pre-1 were not significantly different between wild type and either A or a Δpp-1 cells, expression levels for pre-2 showed increased expression levels in the Δpp-1 a mating type strain (Figure 3).

Figure 3.

Increased expression of genes encoding mating-type pheromones and receptors in the Δpp-1 mutant. qRT-PCR was performed on wild-type mating type A (FGSC 2489), wild-type mating type a (FGSC 988), and Δpp-1 A and Δpp-1 a strains, 5 hr post-inoculation.

To identify additional genes whose expression correlated with germling fusion events, we assessed expression differences using a subset of the time course, 4–6 hr, as some genes involved in colony development may not be expressed at either 3 or 8 hr, and our statistical analysis method requires data at all time points (see Materials and Methods). From analyzing data from the 4- to 6-hr loop, we identified an additional 26 genes (for a total of 86 genes) that showed decreased expression levels and 9 genes (for a total of 41 genes) that showed increased expression levels in the Δpp-1 mutant relative to wild type (Figure 4, A and B; Table S2). Together with genes identified from the 3- to 8-hr time course, FunCat analysis showed that most of the genes with lower expression in the Δpp-1 strain belonged to unclassified/hypothetical proteins (Table S3). Additional overrepresented categories included various signaling pathways and vacuolar degradation (P < 0.05; Table S3). Of the genes that showed increased expression in the Δpp-1 strain, enrichment for the functional category of transcription was identified (P < 0.05; Table S4).

Figure 4.

Summary of microarray and RNA-seq profiling data on Δpp-1 and mak-2^Q100G mutants relative to wild type. (A) Venn diagram showing the overlap of genes that demonstrated a decrease in expression level identified by Δpp-1 (M Δpp-1) and mak-2^Q100G (M mak-2^Q100G) microarrays and RNA-seq analysis (Rseq Δpp-1) as compared to their wild-type isogenic parent (FGSC 2489). (B) Venn diagram showing the overlap of genes that demonstrated an increased expression level in the Δpp-1 (M Δpp-1) and mak-2^Q100G (M mak-2^Q100G) and RNA-seq analysis (Rseq Δpp-1) as compared to their isogenic wild-type parent (FGSC 2489). (C) qRT-PCR analysis showing relative expression levels of genes identified as being differentially expressed via microarrays/RNA-seq in the Δpp-1 and Δpp-1 + pp-1^W155C mutants relative to wild type (FGSC 2489).

RNA-seq analyses were performed as a more sensitive method to identify transcriptional differences between wild-type and Δpp-1 cells at the time point with the highest number of germling fusion events, 5 hr post-inoculation. After statistical analysis (Materials and Methods), a total of 548 genes were identified as being significantly differentially expressed by at least 1.5-fold between wild type and Δpp-1; 441 genes showed reduced expression levels in the Δpp-1 mutant while 107 genes showed increased expression levels as compared to wild type (Figure 4, A and B; File S1).

Transcriptional analyses of MAK-2 function during early colony development

A Δmak-2 mutant in N. crassa shows a very similar phenotype to a Δpp-1 mutant, including lack of cell fusion, reduced vegetative growth, female sterility, and ascospore lethality (Pandey et al. 2004; Li et al. 2005). The question of whether MAK2 kinase activity is required for MAK2 oscillation, as well as for the oscillation of a second protein, SO, which is also required for cell fusion (Fleissner et al. 2009b), was tested by constructing a mak-2 allele (mak-2^Q100G) containing a point mutation that allows specific inhibition of kinase activity with an ATP analog (1NM-PP1; Fleissner et al. 2009b). Addition of the inhibitor 1NM-PP1 to mak-2^Q100G cells undergoing communication immediately blocks further oscillation of both MAK2 and SO and abruptly abolishes chemotropic interactions (Fleissner et al. 2009b). As a complement to our transcriptional profiling of the Δpp-1 mutant, and to identify targets more specifically related to the MAK2 kinase cascade, we performed transcriptional analysis of mak-2^Q100G cells at the 5-hr time point, with and without exposure to 1NM-PP1, for the final 20 min of their growth (see Materials and Methods). We identified 392 genes with modified expression (≥1.5-fold) in the 1NM-PP1 mak-2^Q100G-treated cells vs. mak-2^Q100G (DMSO only). Of these, 149 genes showed a decreased expression level vs. 243 genes that showed increased expression levels (Figure 4, A and B; File S2). FunCat analysis of genes showing an increased expression level also showed an overrepresentation (P < 0.01) in categories of metabolism and in categories related to transport and defense (File S2), suggesting that treatment with 1NM-PP1 may elicit defense responses in N. crassa. Within the genes that showed decreased levels of expression was mak-2 itself, suggesting a feedback mechanism initiated as a result of the stalled signaling.

In total, 22 genes were identified as PP1 targets via both the microarray and the RNA-seq analyses, and an additional 6 genes were identified with the inclusion of the mak-2^Q100G microarray data (Figure 4, A and B; Table S4). Altogether, 7 genes were identified in all three experiments, including ham-7 (NCU00881), a previously identified gene required for hyphal fusion (Fu et al. 2011; Maddi et al. 2012); NCU04122 (malate dehydrogenase); and 5 genes encoding the predicted hypothetical proteins NCU00995, NCU01697, NCU03960, NCU04732, and NCU09693. Genes encoding five additional hypothetical proteins were identified in the overlap between the Δpp-1 microarray and RNA-seq data (NCU00811, NCU01380, NCU05814, NCU07802, and NCU09562) as well as an anchored cell-wall protein (acw-10) (Maddi et al. 2009); hex-1, which encodes a protein involved in septal plugging (Jedd and Chua 2000); glycogenin (NCU06698); and vacuolar aspartyl aminopeptidase (NCU04192). In the comparison between the down-regulated genes in the Δpp-1 microarray and the mak-2^Q100G microarray, only one additional gene encoding a hypothetical protein was identified (NCU07439). In the gene set that overlapped between the Δpp-1 RNA-seq data and the mak-2^Q100G array data (decreased expression category), 5 additional genes were identified that, when mutated, resulted in strains blocked in germling/hyphal fusion. These include two components of the NADPH oxidase regulatory system, nox-1 (NCU02110) and nor-1 (NCU07850) (Fleissner et al. 2008; Fu et al. 2011); a transcription factor, adv-1 (NCU07392) (Fu et al. 2011); ham-6 (NCU02767), encoding a predicted transmembrane protein (Fu et al. 2011); and Prm-1. Prm-1 is involved in plasma membrane merger in N. crassa during both vegetative and sexual cell fusion (Fleissner et al. 2009a).

To confirm expression differences between Δpp-1 and wild type, we used qRT-PCR to analyze the expression of six genes, two from the set of seven overlapping genes from all three experiments, NCU04732 and ham-7 (NCU00881); two genes identified from the RNA-seq analysis, NCU02613 (hypothetical) and NCU02246 (hypothetical); and Prm-1 (NCU09337), which was identified in both the RNA-seq and mak-2^Q100G microarray analyses. Our qRT-PCR recapitulated our RNA-seq/microarray analyses (Figure 4C), with the expression of all identified target genes being significantly reduced in both the Δpp-1 and pp-1^W155C mutants relative to wild type.

PP1 and the regulation of genes required for hyphal fusion and network development

Approximately 40 genes have been identified with roles in colony establishment in N. crassa, specifically during germling or hyphal fusion (see review in Read et al. 2010, 2012). By evaluating our expression data, 16 of these genes were dependent on PP1 for their expression (>1.5-fold lower in Δpp-1 as compared to WT, P < 0.05, RNA-seq data) (Table 3). This list included ham-7, Prm-1, ham-6, nox-1, nor-1, and adv-1, as well as 2 genes encoding kinases in the cell-wall integrity pathway (mek-1 and mak-1) (Read et al. 2010); 2 genes encoding predicted components of the striatin complex (pp2A and mob-3) (Maerz et al. 2009; Fu et al. 2011); one transcription factor (rco-1) (Aldabbous et al. 2010); genes for two GPI-anchored cell-wall proteins (gpip-2 and gpip-1) (Bowman et al. 2006); ham-8, encoding a predicted transmembrane protein; ham-9, encoding a protein with pleckstrin homology domain and a sterile α-motif domain (Fu et al. 2011); and so (Fleissner et al. 2005). During germling fusion, SO also oscillates to CAT tips, with opposite dynamics to MAK2 (Fleissner et al. 2009b); correct oscillation of SO requires MAK2 kinase activity.

Table 3. Previously identified cell fusion genes that showed dependence on pp-1 for wild-type expression levels in germlings.

Gene	Locus ID no.	Gene description	Reference
Prm-1	NCU09337	Membrane protein	Fleissner et al. (2009a)
mek-1	NCU06419	MAP kinase kinase	Fu et al. (2011)
ham-6	NCU02767	Small hydrophobic protein	Fu et al. (2011)
so	NCU02794	WW domain protein	Fleissner et al. (2005)
ham-8	NCU02811	Transmembrane protein	Fu et al. (2011)
nox-1	NCU02110	NADPH oxidase	Cano-Dominguez et al. (2008)
ham-9	NCU07389	Plekstrin domain protein	Fu et al. (2011)
ham-7	NCU00881	GPI-anchored protein	Fu et al. (2011); Maddi et al. (2012)
adv-1	NCU07392	Transcription factor	Fu et al. (2011)
gpip-2	NCU07999	GPI-anchored protein	Bowman et al. (2006)
nor-1	NCU07850	NADPH oxidase regulator	Cano-Dominguez et al. (2008)
gpip-1	NCU06663	GPI-anchored protein	Bowman et al. (2006)
mak-1	NCU09842	MAP kinase	Fu et al. (2011); Maddi et al. (2012)
pp2A	NCU06563	Catalytic subunit of protein phosphatase 2A	Fu et al. (2011)
mob-3	NCU07674	Phocein	Maerz et al. (2009)
rco-1	NCU06205	Transcription factor	Aldabbous et al. (2010)

Identification of novel proteins involved in hyphal fusion and network development

A comparison of the Δpp-1 microarray and RNA-seq analyses identified a 16-gene overlap that showed reduced expression levels, and a 6-gene overlap that showed increased expression levels, in the Δpp-1 mutant relative to wild type (Figure 4 and Table 2). We therefore evaluated the germling fusion/colony establishment phenotypes of strains containing deletions of these genes. Of these 22 genes, 18 deletion mutants were available from the FGSC (McCluskey 2003), and we constructed strains carrying deletions for the remaining 4 genes (NCU00811, NCU0995, NCU01380, and NCU09693) (Table 2). Germling fusion frequency of each deletion strain was determined and compared to the wild-type isogenic parent. Of the 22 deletion strains assayed, three mutants were affected in the ability to undergo germling fusion. As expected, ham-7 (NCU00881) mutants were completely unable to undergo germling fusion (Fu et al. 2011), and in addition so was a mutant carrying a deletion of NCU04732, which encodes a hypothetical protein (Figure 5A). Deletion of an additional gene, NCU03960, which also encodes a hypothetical protein, resulted in a strain that was capable of fusion, but with a significantly lower fusion frequency than wild type (Table 2 and Figure 5A).

Figure 5.

Fusion phenotypes of deletion mutants in PP1 target genes. (A) Chemotropic interactions and CAT fusion phenotypes of wild type (FGSC 2489), compared to the isogenic mutant strains Δham-7, Δham-11, and Δham-12. (B) Macroscopic growth phenotype of wild type, Δham-7, Δham-11, and Δham-12 mutants in flasks. Note the shortened aerial hyphae in the Δham-7 mutant. (C) Macroscopic growth phenotype on VMM plates of wild type, Δham-7, Δham-11, and Δham-12 2 days post-inoculation. (D) Radial growth (in mm) of wild type, Δham-7, Δham-11, and Δham-12 mutants over a 48-hr time period.

NCU04732 was named ham-11. HAM11 is a 633-amino-acid protein with two predicted transmembrane domains and is highly conserved among the Sordariomycetes. However, predicted proteins with strong homology in the N- and C-terminal regions to HAM11 were present in the genomes of most filamentous ascomycete species (no homologs were detected in Saccharomycetes or in Basidiomycetes). The Δham-11 mutants were fully fertile as a male or as a female, indicating that HAM11 is specifically required for germling fusion and not for sexual fusion. The Δham-11 growth phenotype differs from other fusion mutants (Δmak-2, Δso, and Δham-7) in that colonies were not “flat” due to reduced aerial hyphae extension (Figure 5, B and C). Aerial hyphae extension in Δham-11 mutants was similar to wild type (17 mm ± 1.7 as compared to 18.2 mm ± 2.6 for wild type), while aerial hyphae extension in Δham-7 mutants was only 1/2 of wild type (11 mm ± 1.3). The growth rate of Δham-11 over 48 hr was slightly slower than for wild type (Figure 5D).

A strain carrying a deletion of NCU03960, hereafter called ham-12, showed a significant (Student’s t-test: P < 0.01) reduction in fusion frequencies as compared to wild type (66 vs. 87%, respectively) (Table 2). ham-12 is predicted to encode a 157-amino-acid protein with one predicted transmembrane domain, but without any annotated function. ham-12 is more highly conserved than ham-11, with homologs identified across filamentous ascomycete fungi (Pezizomycotina). The growth, reproductive phenotype, and aerial hyphae extension of the Δham-12 mutant were identical to wild type (Figure 5, B–D).

Fusion defect of Δham-11 germlings is suppressed in the presence of a wild-type partner cell

When wild-type germlings communicate, MAK2 oscillates from the cytoplasm to CAT tips every 4 min (8 min/cycle; Figure 6A). Similarly, when wild-type germlings communicate, SO oscillates from the CAT tips to the cytoplasm every 4 min (Figure 7A); MAK2 and SO oscillate in perfect opposition during chemotropic interactions (Fleissner et al. 2009b). To determine whether Δham-7 or Δham-11 germlings show an altered oscillation pattern of MAK2, we introduced mak-2-GFP constructs into wild type, Δham-7, and Δham-11 strains. For both Δham-7 and Δham-11 germlings, localization of MAK2-GFP was cytoplasmic, and localization to germling tips, chemotropic interactions, or cell fusion was not observed (Figure 6B and Figure 8). However, when Δham-11 germlings were confronted with wild-type cells, chemotropic interactions and cell fusion were observed. However, the percentage of Δham-11 germlings communicating with wild type was ∼25% lower as compared to wild-type–wild-type interactions, as wild-type–type fusions were overrepresented in mixtures of Δham-11 + wild-type germlings.

Figure 6.

Localization of MAK2-GFP in wild type (FGSC 2489; WT) and Δham-11 during germination and germling interactions. (A) MAK2-GFP oscillates to the tip every 8 min (arrows) in a wild-type cell bearing MAK2-GFP when communicating with an otherwise isogenic wild-type germling (WT mak-2-gfp + wild type). (B) MAK2-GFP shows no oscillation in Δham-11 mak-2-gfp germlings in proximity to otherwise isogenic Δham-11 cells (Δham-11 mak-2-gfp + Δham-11). (C) Chemotropic interactions and MAK2-GFP oscillation are apparent when wild-type germlings bearing MAK2-GFP are in proximity to Δham-11 germlings (WT mak-2-gfp + Δham-11). (D) Oscillation of MAK2-GFP is apparent in Δham-11 germlings in proximity to wild-type cells (Δham-11 mak-2-gfp + WT). Arrows indicate localization of MAK2-GFP to CAT tips when observed. Bright-field pictures on right show both germlings. Bar, 10 µm.

Figure 7.

Localization of SO-GFP in wild-type and Δham-11 germlings. (A) A time course of SO-GFP oscillation during chemotropic interactions between wild-type germlings; oscillation of SO-GFP to CAT tips occurs every 8 min (WT so-gfp gfp + WT). (B) Δham-11 germlings bearing SO-GFP lack oscillation and localization of SO-GFP to CAT tips when in proximity to other Δham-11 germlings (Δham-11 so-gfp-gfp + Δham-11). (C) Wild-type cells bearing SO-GFP show oscillation when in proximity to Δham-11 cells and chemotropic interactions are restored (WT so-gfp + Δham-11). (D) Δham-11 germlings bearing SO-GFP show chemotropic interactions and oscillation of SO-GFP to CAT tips when in proximity to wild-type germlings (Δham-11 so-gfp + WT). Note that the oscillation dynamics of SO-GFP are identical to the WT + WT germling interactions shown in A. Arrows show localization of SO-GFP to CAT tips. Bright-field pictures show germlings. Bar, 10 µm.

Figure 8.

Lack of oscillation of MAK2-GFP and chemotropic interactions in Δham-7 germlings in proximity to wild type. (A) Cytoplasmic localization of MAK2-GFP in wild-type germlings when in proximity to Δham-7 germlings (WT mak-2-gfp + Δham-7). No oscillation of MAK2-GFP was observed during the time course. (B) Cytoplasmic localization of MAK2-GFP and lack of chemotropic interactions when mak-2-gfp; Δham-7 germlings are in proximity with wild-type cells (WT + Δham-7 mak-2-gfp). No localization of MAK2-GFP to CAT-like structures was observed during the time course. (C) Lack of chemotropic interactions between Δham-7 germlings and cytoplasmic localization of MAK2-GFP (Δham-7 + Δham-7 mak-2-gfp). Bright-field pictures show the germling proximity. Bar, 10 µm.

We hypothesized that in wild-type + Δham-11 germling pairs where chemotropic interactions were observed, MAK2 oscillation would be restored. As predicted, oscillation of MAK2-GFP in wild-type cells, when confronted with Δham-11 germlings, was identical to wild-type + wild-type pairings and showed identical MAK2 oscillation dynamics (8-min cycle from CAT tip to cytoplasm) (Figure 6C). Similarly, when Δham-11 germlings bearing MAK2-GFP were in proximity to wild-type cells, oscillation of MAK2 was coincident with chemotropic interactions and cell fusion (Figure 6D). In contrast, Δham-11 mak-2-gfp + Δham-11 germlings showed only cytoplasmic localization of MAK2 (Figure 6B). In contrast to the results obtained with Δham-11 mutants, no chemotropic interactions were observed between wild-type + Δham-7 germlings, and oscillation of MAK2 to CAT tips was not apparent in wild-type cells in proximity to Δham-7 germlings, nor in Δham-7 germlings bearing MAK2-GFP in proximity to wild-type cells (Figure 8).

We hypothesized that SO oscillation would also be restored in wild type + Δham-11 germlings when chemotropic interactions and cell fusion were observed. As predicted, SO oscillated in an 8-min cycle from CAT tip to cytoplasm in wild-type cells bearing SO-GFP communicating with Δham-11 cells (Figure 7C) and also when Δham-11 germlings bearing SO-GFP were undergoing chemotropic interactions with wild-type cells (Figure 7D). The dynamics of SO oscillation in wild type + Δham-11 pairings showed identical dynamics to wild-type + wild-type pairings. The phenotype of the Δham-11 mutant is unique among fusion mutants: communication and fusion are not restored when other fusion mutants are confronted with wild-type strains (mutants are unable to send or receive fusion signals) (Xiang et al. 2002; Pandey et al. 2004; Fleissner et al. 2005; Simonin et al. 2010; Fu et al. 2011).

Some fusion mutants are impaired in proper phosphorylation of the two MAPK kinases, MAK1 or MAK2, which are both required for germling fusion (Dettmann et al. 2012; Maddi et al. 2012). Since the Δham-11 mutant failed to initiate chemotropic interactions, we also evaluated phosphorylation of MAK1 and MAK2 in Δham-11 germlings with the Δham-7 mutant as a control. Conidia from wild type, Δham-7, or Δham-11 were germinated for 5 hr, and total protein was extracted and subjected to Western blot using phospho-specific P22/P24 antibodies, which recognize phosphorylated MAK1 and MAK2. As previously described (Maddi et al. 2012), Δham-7 shows a defect in phosphorylation of MAK1 and, to a modest extent, phosphorylation of MAK2. Surprisingly, Δham-11 germlings showed wild-type levels of phosphorylation of both MAK1 and MAK2 (Figure 9). These data indicate that, even though MAK2 oscillation to CAT tips is defective in Δham-11 germlings, phosphorylation of MAK2 by upstream factors is not.

Figure 9.

MAK1 and MAK2 phosphorylation, fusion phenotype, and Western blot of Δham-11 complemented with ham-11-gfp. (A) Four germling pairs of Δham-11 ham-11-GFP show chemotrophic growth and cell fusion (see arrows), showing full complementation of the Δham-11 phenotype by the ham-11-gfp construct. (B) Western blot of PRM1-GFP (which localizes to endomembranes and the plasma membrane of N. crassa by fluorescence microscopy) (Fleissner et al. 2009a). HAM11-GFP was detected in supernatant, endomembranes (cell fraction P1), and outer membranes (cell fraction P2). Cell fractionation purity was also assessed using antibodies to the plasma membrane ATPase (data not shown), which localized only to the membrane fractions. (C) Activation of MAK1 and MAK2 in wild type, Δham-11, and Δham-7 germlings. Protein samples from Δham-11 (lane 1), Δham-7 (lane 2), and wild-type (lane 3) 5-hr-old germlings from liquid VMM are shown. Phosphorylated MAK1 (51 kDa) and MAK2 (41 kDa) were detected using anti-phospho p44/42 MAP kinase antibodies (Cell Signaling Technology). (Bottom) Shows equal loading for each lane.

We constructed gfp-tagged alleles of ham-11 driven by the tef-1 promoter to determine localization of HAM11 in cells undergoing chemotropic interactions vs. those that are not communicating. The tef-1 gene is more highly expressed in germlings than ccg-1 and has been used to localize NCR1 and MEK2 to CAT tips (Dettmann et al. 2012). GFP was fused within the ham-11 ORF between amino acids 62 and 63, which lie between the two predicted transmembrane domains. This construct was transformed into a his-3; Δham-11 strain. Complementation was assessed in homokaryotic his-3::ham-11-gfp; Δham-11 progeny via evaluating germling fusion frequencies. In contrast to the Δham-11 strain that lacks germling fusion, germlings of the his-3::ham-11-gfp; Δham-11 strain showed fusion frequencies similar to wild-type + wild-type pairings (Figure 9A). However, although HAM11 could be detected via Western analyses, GFP fluorescence in live cells could not be detected above background.

Because we could detect HAM11 via Western blotting, but not by fluorescence microscopy, we resorted to cell fractionation to determine HAM11 localization. As a control, we utilized PRM1-GFP, which localizes to the plasma membrane and to internal membrane-bound structures (Fleissner et al. 2009a). We evaluated the presence of HAM11 and PRM1 during germling fusion (5-hr time point). From the cell lysate, three fractions were obtained: a resolubilized pellet containing endomembranes, a second resolubilized pellet containing plasma membranes, and a supernatant fraction from which cytoplasmic GFP-protein was immunoprecipitated. Antibodies to the N. crassa plasma membrane ATPase, which localizes to the plasma membrane (Bowman and Bowman 1988), were used to assure proper fractionation (data not shown). As expected, PRM1-GFP was observed in fractions containing membranes (endomembranes and plasma membrane), but not in the cytoplasmic fraction (Figure 9B), consistent with fluorescence mircroscopy results (Fleissner et al. 2009a). HAM11-GFP, however, was found in all three cellular fractions (Figure 9B). These data suggest that HAM11 is a protein that can shuttle between cytoplasm and vesicles/membranes in the cell.

Discussion

The MAK2 MAPK pathway in N. crassa is required for chemotropic interactions and cell fusion during vegetative growth. In this study, we show that one of the predicted targets of the MAK2 pathway, the transcription factor PP1, is also required for chemotropic interactions. This strongly supports the role of PP1 as a target of this MAPK pathway during cell fusion. We subsequently utilized a combination of transcriptional approaches during a time of intensive chemotropic interactions and germling fusion events to identify new components of the cell fusion pathway. When comparing the three data sets (Δpp-1 and Δmak-2^Q100G microarrays and Δpp-1 RNA-seq), we identified an overlapping set of seven genes that contained three whose deletion affected germling fusion frequency. These included a previously identified fusion gene of unknown biochemical function (ham-7), which encodes a GPI-anchored protein. Recent data implicate HAM7 in functioning as a cell-wall sensor to regulate activation of the cell-wall integrity MAPK MAK1 (Maddi et al. 2012). Strains carrying a deletion of mak-1 show a pleiotropic phenotype and are also unable to undergo chemotropic interactions and germling fusion (Fu et al. 2011). Of the ∼40 genes that have been previously identified as being required for germling fusion, 15 were identified within our gene set that showed a reduction in expression level in Δpp-1 or mak-2^Q100G strains, particularly in the Δpp-1 RNA-seq data set, reflecting the increased power of expression detection over traditional microarray data. In addition to genes that were coordinately regulated in the mak-2^Q100G and the Δpp-1 data sets (genes up in both mutants or down in both mutants), genes that showed opposite regulation (increased expression level in one mutant was correlated with decreased expression in the second mutant) were also identified. In S. cerevisiae, Fus3p regulates negative effectors of Ste12p (Dig1p and Dig2p), in addition to direct regulation of Ste12p (Elion et al. 1993; Tedford et al. 1997; Blackwell et al. 2007), with activation of Fus3p being correlated with activation of Ste12p. In N. crassa, MAK2 also shows nuclear localization in germlings (Fleissner et al. 2009b), suggesting that MAK2 may influence activity of transcription factors in addition to PP1. Additional intricacies of PP1 regulation by MAK2 may occur in N. crassa as well as in other filamentous fungi, and gene sets that showed opposite expression patterns in the Δmak-2 mutant vs. the Δpp-1 mutant offer a opportunity to explore this issue.

In addition to their role in germling fusion, it is clear that both PP1 and MAK2 are important in coordinating germination and hyphal growth. This aspect is also true for many of the previously identified fusion mutants, such as Δham-2, Δham-3, Δham-4, and Δham-7 (Xiang et al. 2002; Simonin et al. 2010; Fu et al. 2011), which also have pleiotropic phenotypes. Thus, our data may also be useful in the identification of components important for these processes in filamentous fungi, such as N. crassa. For example, in S. cerevisiae, Ste12 is important for invasive growth in haploid cells, while in diploid cells it plays a role in filamentation that occurs in response to nitrogen starvation (Gustin et al. 1998; Gancedo 2001). It is noteworthy that, in all our transcriptional profiling data sets during germination and colony establishment, the category showing the largest enrichment was that of unclassified proteins. This observation underscores our lack of understanding of developmental processes unique to filamentous fungi.

MAPK pathways orthologous to MAK2 have been implicated in sexual development in filamentous fungi; Δpp-1 and Δmak-2 mutants fail to form female reproductive structures but are fully fertile as males (although both display an ascospore germination defect). Surprisingly, we found a mating-type-specific effect on the expression of the mating pheromones, ccg-4 (A pheromone; up in Δpp-1 A) and mfa-1 (a pheromone; up in Δpp-1 a), and to a lesser extent the mating-type pheromone receptor, pre-2 (A pheromone receptor; up in Δpp-1 a). These data suggest interplay between the A and a mating-type-specific transcription factors believed to regulate pheromones and receptors in N. crassa (A-1 and a-1) and PP1. Pheromones (expressed by male cells) and receptors (expressed by trichogynes undergoing chemotropic interactions toward a male cell) are essential for mating in N. crassa (Kim and Borkovich 2004, 2006). In the closely related homothallic ascomycete fungus, S. macrospora, STE12 binds the mating-type protein SMTA-1, a small serine-threonine protein, SIP2, and the MADS box protein MCM1 during formation of reproductive structures (Nolting and Poggeler 2006). However, deletion of ste12 in S. macrospora does not affect vegetative growth or fruit body development. In S. cerevisiae, Ste12 via interaction with Mcm1 induces a-specific genes in response to pheromone activation, while Mcm1, α1, and Ste12 induce α-specific genes (Errede and Ammerer 1989; Yuan et al. 1993). Further experiments are needed to understand the interplay of PP1, the mating-type-specific transcription factors, pheromones, and receptors during both vegetative and sexual reproduction.

Ste12-like transcription factors are highly conserved among filamentous fungi. A comparison of our transcriptional data set with Ste12 chromatin-immunoprecipitation data from S. cerevisiae (Ren et al. 2000; Borneman et al. 2007; Zheng et al. 2010) identified nine conserved genes, including STE12/pp-1, FUS3/mak-2, PHD1/Asm-1 (NCU01414) (transcriptional activator), PRM1/Prm-1 (plasma membrane merger), ACO1/NCU02366 (aconitase), PGU1/NCU02369 (endopolygalacturonase), CHS1/chs-3 (chitin synthase), HAA1/NCU04830 (transcriptional activator), and PCL2/NCU04847 (cyclin) that were conserved. Of these genes, mutations in three affect germling fusion in N. crassa: pp-1, mak-2, and Prm-1. Interestingly, preliminary data show that a strain carrying a deletion of Asm-1 is also a hyphal fusion mutant (data not shown). An Asm-1 mutant shows a vegetative (“flat”) phenotype, female sterility, and an ascospore maturation defect (Aramayo et al. 1996), phenotypes that are reminiscent of Δpp-1. In S. cerevisiae, the Ste12-binding motif is TGAAACA, although not all target genes identified in S. cerevisiae show conservation of these sites (Ren et al. 2000; Borneman et al. 2007; Zheng et al. 2010). In N. crassa, only between 4–6.5% of the genes identified in the 1NM-PP1 MAK2^Q100G and PP-1 RNA-seq experiments have a STE12-like motif present in their promoters at high confidence (data not shown). However, transcription-factor-binding sites are short and often degenerate, making computational detection in promoter regions of genes across an entire genome difficult (Tompa et al. 2005). Future experiments using ChIP-seq analyses would clarify direct target genes of PP1, especially if performed under different growth and developmental scenarios.

Of the seven genes that overlapped from all the expression profiling data sets, a deletion of one of them, ham-11, resulted in a phenotype that is unique among fusion mutants. All fusion mutants identified so far are blind and do not respond to the presence of a wild-type partner. The only reported exception is a N. crassa mutant in a gene identified by genome-wide association studies of germling communication (Palma-Guerrero et al. 2013). In this case, mutations in a predicted secreted protease (spr-7) actually increased fusion frequencies (from ∼86% to ∼97%). However, when the Δspr-7 mutant was mixed with wild-type germlings, fusion frequency was reduced to wild-type levels. In contrast, germlings of Δham-11 showed no fusion at all, but showed normal chemotropic interactions; oscillations of MAK2 and SO and near wild-type levels of cell fusion occur when in proximity to wild-type germlings. These data suggest that HAM11 function is essential for the initiation of chemotropic interactions. As ham-11 was identified as a target of PP1 and the MAK2 pathway, these results suggest a positive feedback mechanism associated with the initiation of signaling. Data showing normal levels of MAK2 phosphorylation in the Δham-11 mutant suggest that the defect is associated with the initiation of MAK2 oscillation to CAT tips and that HAM11 is not involved in the upstream cascade leading to MAK2 activation. Cells of N. crassa do not initiate chemotropic interactions, nor is localization of MAK2 or SO to CAT tips observed when cells are >15 μm apart. Thus, HAM11 might be involved in the regulation of a switch-like transition to signaling behavior when germlings are close enough to each other to initiate chemotropic interactions. HAM11 was not detected in the extracellular fraction (data not shown), but was identified in both the cytoplasmic and membrane fractions, suggesting that it may shuttle between compartments, perhaps depending upon MAK2-signaling status. Supporting this hypothesis is the finding that HAM11 was identified in a phosphoproteomic study to identify MAK2 substrates (A. Leeder and N. L. Glass, unpublished data).

Most of the fusion mutants previously identified affect chemotropic interactions and signaling of MAK2 and SO. However, a strain carrying a deletion of one of the transcriptional targets identified in this study, Prm-1, has normal chemotropic interactions, but is affected in plasma membrane merger (Fleissner et al. 2009a). In S. cerevisiae, PRM1 was identified by assessing selectively expressed genes in response to pheromone treatment (Heiman and Walter 2000) and is a direct target of Ste12 (Ren et al. 2000); prm1Δ mutants are also affected in plasma membrane merger. It is possible that genes encoding additional proteins associated with cell-wall breakdown, plasma membrane, and trafficking can be identified in our data sets. Such proteins that are involved in trafficking and late fusion events may not have been identified by our assays for chemotropic interactions and fusion frequency; ΔPrm-1 mutants show normal chemotropic interactions and oscillation of MAK2 and SO, but have a ∼50% reduction in plasma membrane merger (Fleissner et al. 2009a). In our data sets, we identified a number of genes dependent on PP1 for expression that encode proteins involved in vesicle trafficking, such as a homolog to S. cerevisiae SCN2 (NCU00566), which encodes a v-SNARE involved in the fusion of Golgi-derived secretory vesicles with the plasma membrane. Mutations in a gene encoding a homolog to calcium neuronal sensor-1 (NCU04379; cse-1)—which localizes to the Golgi and is involved in regulated secretion in neurons—reduce, but do not abolish, germling fusion frequency (Palma-Guerrero et al. 2013). Fusion frequencies are not restored when Δcse-1 germlings are paired with wild-type germlings. This phenotype is similar to that of Δham-12 mutants, where no differences in fusion frequencies were observed between Δham-12 + Δham-12 germlings and Δham-12 + wild-type germlings. In the future, epistasis and colocalization studies will provide additional clues for the function of these proteins during the processes of germling communication, chemotropic interactions, and cell fusion.