New Insights Into the Roles of NADPH Oxidases in Sexual Development and Ascospore Germination in Sordaria macrospora

Abstract

NADPH oxidase (NOX)-derived reactive oxygen species (ROS) act as signaling determinants that induce different cellular processes. To characterize NOX function during fungal development, we utilized the genetically tractable ascomycete Sordaria macrospora. Genome sequencing of a sterile mutant led us to identify the NADPH oxidase encoding nox1 as a gene required for fruiting body formation, regular hyphal growth, and hyphal fusion. These phenotypes are shared by ∆nor1, lacking the NOX regulator NOR1. Further phenotypic analyses revealed a high correlation between increased ROS production and hyphal fusion deficiencies in ∆nox1 and other sterile mutants. A genome-wide transcriptional profiling analysis of mycelia and isolated protoperithecia from wild type and ∆nox1 revealed that nox1 inactivation affects the expression of genes related to cytoskeleton remodeling, hyphal fusion, metabolism, and mitochondrial respiration. Genetic analysis of ∆nox2, lacking the NADPH oxidase 2 gene, ∆nor1, and transcription factor deletion mutant ∆ste12, revealed a strict melanin-dependent ascospore germination defect, indicating a common genetic pathway for these three genes. We report that gsa3, encoding a G-protein α-subunit, and sac1, encoding cAMP-generating adenylate cyclase, act in a separate pathway during the germination process. The finding that cAMP inhibits ascospore germination in a melanin-dependent manner supports a model in which cAMP inhibits NOX2 activity, thus suggesting a link between both pathways. Our results expand the current knowledge on the role of NOX enzymes in fungal development and provide a frame to define upstream and downstream components of the NOX signaling pathways in fungi.

DURING sexual reproduction, filamentous fungi generate complex fruiting bodies that contain and protect meiosporangia. We used the ascomycetous model fungus Sordaria macrospora to identify genes directly involved in fruiting body development (Kück et al. 2009; Engh et al. 2010; Kück et al. 2009). Due to its homothallic life style, S. macrospora is able to complete the sexual life cycle without the mating of strains with opposite sex, and therefore, fruiting body-deficient mutants can be recognized directly without the need for crossing experiments. In earlier work, we generated sterile mutants showing a developmental block after formation of young fruiting bodies (protoperithecia), but being unable to generate mature perithecia, and referred to these mutants as pro. Recently, we have applied next-generation genome re-sequencing to identify the genes affected in some of these mutants (Nowrousian et al. 2012). Based on this approach, we now have characterized mutant pro32 and show that it carries a mutation in the nox1 gene encoding NAPDH oxidase 1 (NOX1).

NADPH oxidase (NOX) enzymes are transmembrane proteins that are highly conserved among eukaryotes and produce reactive oxygen species (ROS) through the oxidation of NADPH (Lambeth 2004; Kawahara and Lambeth 2007). ROS have long been recognized as damaging agents due to uncontrolled oxidizing reactions with DNA, RNA, proteins, and lipids (Halliwell and Gutteridge 2007). However, there is increasing evidence that ROS act as signaling determinants that induce different cellular processes (Scott and Eaton 2008; Aguirre and Lambeth 2010; Heller and Tudzynski 2011).

In mammals, seven members of the NOX family (NOX1–5, DUOX1, and DUOX2) are known (Aguirre et al. 2005; Kawahara and Lambeth 2007). The activity of NOX2, the most intensively studied NOX, is regulated by a protein complex containing p22phox, p40phox, p47phox, p67phox, and the small GTPase RAC1. NOX1, NOX3, and NOX4 also require p22phox, whereas the activity of NOX5, DUOX1, and DUOX2 is independent of this regulator (Smith et al. 2012). To date, three members of the NOX family are known in fungi. NOX1 and NOX2 [synonymous (syn.) NOXA and NOXB] are homologs of mammalian NOX2 and have been found in most ascomycetes. In contrast, NOX3, the homolog of mammalian NOX5, has been detected only in some fungi like Aspergillus terreus, Magnaporthe grisea, Podospora anserina, and several Fusarium species (Aguirre et al. 2005; Scott and Eaton 2008; Brun et al. 2009). Fungal NOX1 and NOX2 enzymes are regulated by the p67phox homolog NOR1 (NOX regulating, syn. NOXR) and the small GTPase RAC1 (syn. RacA) (Kawahara and Lambeth 2007; Tanaka et al. 2008). In Epichloë festucae, the putative scaffold protein Bem1 was reported to associate with NOR1 and CDC24 (Takemoto et al. 2011). Another candidate NOX regulatory protein is tetraspanin Pls1. Strains from different ascomycetes, lacking the corresponding gene, have phenotypes similar to nox2 deletion strains (Lambou et al. 2008; Ryder et al. 2013; Siegmund et al. 2013). However, very little is known about the upstream and downstream components of the NOX signaling pathways in fungi.

In this study, we carried out a comprehensive genetic analysis using S. macrospora to elucidate the contribution of NOX enzymes to fungal sexual development. In addition to the characterization of sterile mutant pro32, we present a detailed functional analysis of deletion mutants nox1, nox2, and nor1 in S. macrospora, showing that NOX1 and NOR1 are required for fruiting body development and hyphal fusion. For the first time, we provide RNA-seq analysis of ∆nox1 protoperithecia and show NOX1-dependent gene expression compared to gene expression in protoperithecia of the sterile mutant pro1 and wild-type strains. Furthermore, phenotypic and genetic analysis led us to conclude that NOX2 and NOR1 contribute to a signaling pathway controlling ascospore germination and that transcription factor STE12 is part of this pathway. Finally, we relate cAMP levels to NOX2 function. Our analysis extends the current knowledge on the contribution of NOX enzymes to the regulation of two distinct fungal developmental processes and provides important hints to define the cellular routes regulated by these enyzmes.

Materials and Methods

Strains, media, and growth conditions

S. macrospora strains, listed in Table 1, were grown under standard laboratory conditions on complete medium (CM) or cornmeal malt fructification medium (BMM) media (Esser 1982; Nowrousian et al. 1999). Cultivation for ascospore germination assays and DNA extraction were performed as described previously (Nowrousian and Cebula 2005; Kamerewerd et al. 2008; Teichert et al. 2012). For rescue of fertility, strains were inoculated on filter paper covering solid BMM medium. Two consecutive transfers of the filter paper to fresh solid BMM medium were performed after 3 days. Continuous growth on solid BMM medium with or without filter paper served as control (adapted from Malagnac et al. 2004). Quantification of linear growth was performed using race tube assays. Recombinant plasmids were propagated in Escherichia coli XL1 Blue MRF′ (Stratagene, La Jolla, CA) under standard experimental conditions (Sambrook and Russell 2001).

Table 1. S. macrospora strains used in this study.

Strain	Relevant genotype	Relevant phenotype	Reference source
S91327	Wild type	F	Culture collection of the Department of General and Molecular Botany
S84595	fus, spore color mutant	F	Culture collection of the Department of General and Molecular Botany
S96888	∆ku70::nat	F	Pöggeler and Kück (2006)
DD27, DD1	∆nox1::hph	S, DG, HFD	This study
DD194, DD299	∆nox2::hph/fus	F, GDB	This study
DD118-2	∆nox2::hph/∆ku70::nat	F	This study
DD492, DD574	∆nor1::hph/fus	S, GDB, DG, HFD	This study
S104701	∆gsa3::hph	F, GD, DG	Kamerewerd et al. (2008)
S114583, S114602	∆ste12::hph/fus	F, GDB	Kamerewerd et al. (2008)
S107115	∆gsa3::hph/∆nox2::hph/fus	F, GD	This study
S83812	∆gsa3::hph/∆ste12::hph/r2	S, DG	Kamerewerd et al. (2008)
S144534, S114567	∆nox2::hph/∆ste12::hph/r2	F, GDB	This study
DD1093, DD1161, DD1325	∆nox1::hph/nox1::nat	F, HFD	This study
S106371, S106375, S106755	∆nox2::hph/nox2::nat	F	This study
DD2843, DD2909, DD2958	∆nor1::hph/nor1::nat	F	This study
S69656	∆pro40::hph	S	Engh et al. (2007)
S109348	pro32	S, HFD	This study
DD291-3-1, DD290-4-3	pro32/nox1::nat	F	This study

F, fertile; S, sterile; HFD, hyphal fusion defect; GD, germination defect; GDB, germination defect in black ascospores; DG, decreased growth.

Preparation of nucleic acids

DNA and RNA were extracted using phenol/chloroform, and RNA was selectively precipitated (Pöggeler et al. 1997). For RNA-seq analysis, mycelia were grown in surface cultures for 4 days or directly on slides [Molecular Machines and Industries (MMI)] for fixation and dissection in situ (Teichert et al. 2012).

Genome sequencing of developmental mutant pro32

Mutant pro32 from our laboratory collection was backcrossed several times with wild type or red-spored fus mutant (Nowrousian et al. 2012) and finally crossed with fus. DNA was extracted from 40 sterile and 40 fertile progeny as described previously (Nowrousian et al. 2012). Five micrograms of pooled genomic DNA for pro32 and wild type, respectively, was subjected to 50-bp paired-end Illumina/Solexa sequencing with a HiSeq2000 at GATC Biotech (Constance, Germany). Cleaning of raw data, mapping to the S. macrospora reference genome, and analysis of sequence variants were performed as previously described (Nowrousian et al. 2012). The Burrows Wheeler Alignment tool (Li and Durban 2009) was used for mapping and SAMtools (Li et al. 2009) for SNP calling. Further bioinformatics analysis was done using custom-made Perl scripts.

Generation of Δnox1, Δnox2, and Δnor1 deletion strains

Transformation of S. macrospora was performed as described previously with an enzyme mix of 1 g VinoTaste Pro (Novozymes, Blagsvaerd, Denmark), 0.3 g Caylase (Cayla, Toulouse, France), and 27 U Chitinase (ASA Spezialenzyme, Wolfenbüttel, Germany) (Walz and Kück 1995; Engh et al. 2007). For the generation of transgenic plasmids, homologous recombination in S. cerevisiae PJ69-4a and standard cloning procedures were used (James et al. 1996; Sambrook and Russell 2001; Colot et al. 2006; Bloemendal et al. 2012). To allow homologous recombination of nox knockout constructs in S. macrospora ∆ku70, the flanking regions of the corresponding nox genes together with the hph resistance cassette were inserted into pRS426 (Christianson et al. 1992). The 5′ and 3′ regions flanking the nox genes were amplified with specific oligonucleotides using PCR based on wild-type (S91327) genomic DNA [nox1: 5′ region 05007-5fw/05007-5rv (1039 bp), 3′ region 05007-3fw/05007-3rv (1035 bp); nox2: 5′ region 08741-5fw/08741-5rv (1065 bp), 3′ region 08741-3fw/08741-3rv (808 bp); nor1: 5′ region 02124-5fw/02124-5rv (1064 bp), 3′ region 02124-3fw/02124-3rv (1065 bp)]. The hph resistence cassette was obtained from vector pDrivehph (Nowrousian and Cebula 2005) by EcoRI restriction. Plasmids pKO-nox1, pKO-nox2, and pKO-nor1 were linearized with restriction enzymes (BamHI, pKO-nox1 and pKO-nor1; XhoI, pKO-nox2) and used to transform S. macrospora strain ∆ku70. To verify transformants, genomic DNA was isolated and tested for homologous recombination at the nox locus using PCR [∆nox1: 05007_vp1/d1 (1159 bp) and d2/05007_vp2 (2257 bp); ∆nox2: 08742_vp1/d1 (1111 bp) and d2/08742_vp2 (1144 bp); ∆nor1: 02124_vp1/d1 (1190 bp) and d2/02124_vp2 (1193 bp)]. To verify the deletion of nox genes, the following oligonucleotides were used: 05007-5′fw/05007_comp_5′rv (nox1 5′ fragment, 2017 bp), 05007_comp_3′fw/05007-3′rv (nox1 3′ fragment, 2248 bp), 08741-5′fw/08741_comp_5′rv (nox2 5′ fragment, 2053 bp), 08741_comp_3′fw/08741-3′rv (nox2 3′ fragment, 1910 bp), 02124-5′fw/02124_comp_5′rv (nor1 5′ fragment, 1979 bp), and 02124_comp_3′fw/02124-3′rv (nor1 3′ fragment, 1966 bp). Probes for Southern blot detection of nox genes were generated by restriction of complementing plasmids (nox1: pComp_nox1/XhoI; nox2: pComp_nox2/SphI; nor1: pComp_nor1/KspI). The hph probe was obtained from vector pDrivehph by EcoRI restriction (Nowrousian and Cebula 2005).

Plasmids for complementation of nox deletion strains were generated as follows: the nox ORFs and ∼1000 kb up- and downstream regions were amplified via PCR [nox1: 05007-5fw/05007-3rv (3889 bp); nox2: 08742-5fw/08742-3rv (3709 bp); nor1: 02124-5fw/02124-3rv (3842 bp)]. The amplicons were cloned in pRS426_nat (Klix et al. 2010) using homologous recombination in yeast. The complementation vectors pComp_nox1, pComp_nox2, and pComp_nor1 were transformed into the corresponding deletion strains. To verify ∆nox2 complementation, it was necessary to quantify the ascospore germination of black ascospores. For this, the deletion mutant ∆nox2/fus was transformed with pComp_nox2. Resulting strains carrying a nox2 deletion and an ectopical integration of nox2 were designated as ∆nox2/fus::nox2. To obtain black-spored ∆nox2::nox2, the ∆nox2/fus::nox2 was crossed with wild type, and black ascospores were isolated from recombinant asci. The black-spored ∆nox2::nox2 were verified by PCR. Double deletion strains were obtained via crossing of single deletion mutants followed by the isolation of single spores from recombinant asci as described previously (Esser and Straub 1958). All oligonucleotides and plasmids used in this study are listed in Table 2 and Table 3, respectively.

Table 2. Oligonucleotides used in this study.

Oligonucleotide	Sequence (5′–3′)	Specificity
05007-5fw	gtaacgccagggttttcccagtcacgacgggatccgaacaaacacaataacaccattgcc	5′ nox1 with pRS426 overlap; BamHI
05007-5rv	cgagggcaaaggaatagggttccgttgagggttggcgaccgctgaattcctcctc	5′ nox1 with hph overhang
05007-3fw	gcccaaaaatgctccttcaatatcagttgcgttatacttggcttataactataccc	3′ nox1 with hph overlap
05007-3rv	gcggataacaatttcacacaggaaacagcggatcctgattaggcggtattagttatggttg	3′ nox1 with pRS426 overlap, BamHI
08741-5fw	gtaacgccagggttttcccagtcacgacggtcgacgtccttcggtgatgtgccgagagtgc	5′ nox2 with pRS426 overlap, SalI
08741-5rv	cgagggcaaaggaatagggttccgttgaggcgtgtctgggttgcttctgttgtcgt	5′ nox2 with hph overlap
08741-3fw	gcccaaaaatgctccttcaatatcagttgcacgtctttgtcggaattcccgttta	3′ nox2 with hph overlap
08741-3rv	gcggataacaatttcacacaggaaacagcgtcgacctttgcggttgtcgctcatgcgatt	3′ nox2 with pRS426 overlap; SalI
02124-5fw	gtaacgccagggttttcccagtcacgacgggatccctggatacctctaggtcatcaattg	5′ nor1 with pRS426 overlap, BamHI
02124-5rv	cgagggcaaaggaatagggttccgttgagggttgaggtgttgttagacgtgcgta	5′ nor1 with hph overlap
02124-3fw	gcccaaaaatgctccttcaatatcagttgcaaccagctcgccgtctggttttgg	3′ nor1 with hph overlap
02124-3rv	gcggataacaatttcacacaggaaacagcggatcccgtgaggaagctggttgatcctgag	3′ nor1 with pRS426 overlap; BamHI
05007_vp1	gccctgaggcgatttttgtttatc	∆nox1
05007_vp2	gcttttcgctctcacggtagattc	∆nox1
d1	cgatggctgtgtagaagtactcgc	hph
d2	atccgcctggacgactaaaccaa	hph
08741_vp1	ctaagcactttggtccttttcccc	∆nox2
08741_vp2	gattaggaagctgtagatgctcatggag	∆nox2
02124_vp1	ggacaatttccgaggagctggac	∆nor1
02124_vp2	gcttcatgtcagatcgcttgttcc	∆nor1
05007_comp5′rv	gtgccactggtacttggagacttg	nox1
05007_comp3′fw	gtatcagacagcaatccttcgaaac	nox1
08741_comp5′rv	gttctccttcttgatctggatctcg	nox2
08741_comp3′fw	cacatgttcatcgtcttcttctttttc	nox2
02124_comp5′rv	ctatcatccttgacttccagtttcc	nor1
02124_comp3′fw	gtgctgaaatcaaaaaatgttagtcttg	nor1
5007 cDNA rev	gtgttccttccaaaacctgaaatc	nox1
nox1_RT_fw	ggacatggataccacgcaga	nox1 (qRT-PCR)
nox1_RT_rv	ttccgcatgctctcaaagaa	nox1 (qRT-PCR)
nor1_RT_fw	ctggtatgcaggatttggca	nor1 (qRT-PCR)
nor1_RT_rv	gcctcgtttggtcggtagac	nor1 (qRT-PCR)
nox2_RT_fw_2	ctggttcttttccccgtctg	nox2 (qRT-PCR)
nox2_RT_rv_2	ggaccatgctgtcgtgatgt	nox2 (qRT-PCR)
ste12_RT_fw	gcctttcagtcccagtccac	ste12 (qRT-PCR)
ste12_RT_rv	ctgtcccatgttctgtccca	ste12 (qRT-PCR)
pro1_RT_fw	ttcgatcgattcgcattttg	pro1 (qRT-PCR)
pro1_RT_rv	tgatgaatatttgccgctcg	pro1 (qRT-PCR)
gsa3_RT_fw	tcgaccgaatgagatggatg	gsa3 (qRT-PCR)
gsa3_RT_rv	cacttcttgcgttcgctacg	gsa3 (qRT-PCR)
sac1_RT_fw	aggcttgcacttctcttcgg	sac1 (qRT-PCR)
sac1_RT_rv	ttgagcaggcccgttaatct	sac1 (qRT-PCR)
smta-1_RT_fw	catcgtcgccgaatacaaga	Smta-1 (qRT-PCR)
smta-1_RT_rv	aacgacgacactatcgggct	Smta-1 (qRT-PCR)
smtA-3_RT_fw	tcatgatgatggaatgggga	SmtA-3 (qRT-PCR)
smtA-3_RT_rv	ttgttttggcatccgtcttg	SmtA-3 (qRT-PCR)
smtA-2_RT_fw	agcatgctgcgtcattgagt	SmtA-2 (qRT-PCR)
smtA-2_RT_rv	cacccaacacatgcacctct	SmtA-2 (qRT-PCR)
smtA-1_RT_fw	cacgatccctttcacaacga	SmtA-1 (qRT-PCR)
smtA-1_RT_rv	ggcaagtagttttcgcgacc	SmtA-1 (qRT-PCR)
SMU6905for	ggcatcacggtcaatggtgt	teh (qRT-PCR)
SMU6905rev	tgctcagccatcatcctctca	teh (qRT-PCR)
pre1for	gcattcacgcccacatcaac	pre1 (qRT-PCR)
pre1rev	gttgtgccgaaggtgatgca	pre1 (qRT-PCR)
pre2for	tccacccgttccataccctg	pre2 (qRT-PCR)
pre2rev	tcgatgcaagctagttcgcg	pre2 (qRT-PCR)
ppg1-for	ctccgtgacaccaccttcag	ppg1 (qRT-PCR)
ppg1-rev	ggaggcatagcgcttcca	ppg1 (qRT-PCR)
ppg2for	cggtatctcgcctctcaacgt	ppg2 (qRT-PCR)
ppg2rev	gttgtgctcccattgtgcaga	ppg2 (qRT-PCR)
tap1_RT_fw	tgaccaagttgcatcccaag	tap1 (qRT-PCR)
tap1_RT_rv	caaccgtagccctcaacaca	tap1 (qRT-PCR)
SMU4533Ncofor	ccatggctccctcagtcgatcctaccacc	app (Northern blot)
SMU4533Ncorev	ccatggcctccgacggcttgttatcaaccaa	app (Northern blot)

Table 3. Plasmids used in this study.

Plasmid	Feature	Reference
pRS426	URA3, lacZ_a, T7_promoter, T3_promoter, bla, FRT, hph	Christianson et al. (1992)
pDrivehph	pDrive with hph	Nowrousian and Cebula (2005)
pKO-nox1	pRS426 with 1000 kb 5′ and 3′ region of nox1, hph	This study
pKO-nox2	pRS426 with kb 5′ and 3′ region of nox2, hph	This study
pKO-nor1	pRS426 with 1000 kb 5′ and 3′ region of nor1, hph	This study
pComp-nox1	pRS426 with nox1 gene and 1000 kb 5′ and 3′, nat	This study
pComp-nox2	pRS426 with nox2 gene and 1000 kb 5′ and 3′, nat	This study
pComp-nor1	pRS426 with nor1 gene and 1000 kb 5′ and 3′, nat	This study

Microscopy

To investigate sexual propagation, S. macrospora was grown on slides with a thin layer of BMM medium at 27° in continuous light (Engh et al. 2007). Hyphal fusion assays were performed after 2 days of growth on Sordaria Westergaard's medium overlaid with a cellophane layer (Bio-Rad, München, Germany) (Bloemendal et al. 2012). Light microscopy was performed either with an AxioPhot microscope (Zeiss, Jena, Germany) capturing images with an AxioCam using the AxioVision digital system, or AxioImager microscope (Zeiss) capturing images with a Photometrix Cool SnapHQ camera (Roper Scientific). Processing of images was done with MetaMorph (version 7.7.5.0, Universal Imaging) and Adobe Photoshop (Adobe Systems, Dublin, Ireland). The documentation of sexual propagation of complemented strains was done on BMM medium in Petri dishes using a Stemi 2000-C binocular (Zeiss) capturing images with AxioCam ERc5s.

Detection and quantification of ROS

The detection of ROS by nitroblue tetrazolium (NBT) staining was performed as described previously (Malagnac et al. 2004) using cultures with mycelia covering the whole plate (grown for 4 days on solid BMM media). This assay was performed with at least three independent biological replicates per strain. Before and 30 min after NBT addition, pictures were taken of all plates with defined camera settings to determine the mean tonal range (Adobe Photoshop, histogram). For every plate the mean tonal range values after the NBT assay were normalized to the values before NBT addition. Resulting values for the mutant strains were then normalized to the corresponding value for the wild type.

Ascospore germination assays

In this study, we used three different ascosopore germination assays. In the first assay, we quantified germination of ascospores discharged from perithecia generated by selfing (ascospore germination assay of selfing strains). For this, fertile strains were cultivated on BMM medium for 5–10 days at 27° in continuous light. When the first discharged ascospores were observed, a Petri dish with thin BMM medium and 0.5% sodium acetate (BMM-Ac) was put upside down on the S. macrospora culture to directly catch discharged ascospores. After 5 hr incubation at 27°, a minimum of 500 ascospores were microscopically analyzed for germination.

In our second ascospore germination assay, we investigated the effect of different concentrations of ascorbate (antioxidant) and cAMP on germinating ascospores using cell permeable N⁶, 2′-O-dibutyryl-cAMP (db-cAMP, BioLog). In this assay, we used the fertile wild-type and fus strains, which were cultivated and analyzed as described for the ascospore germination assay of selfing strains, except that ascospores were discharged on BMM-Ac medium supplemented with different concentrations of db-cAMP or ascorbate.

To investigate the influence of a certain gene deletion on ascospore germination in sterile mutants, we performed an ascospore germination assay as a crossing experiment (crossing ascospore germination assay). All the investigated strains carried the fus mutation and were crossed with wild type (black ascospores), leading to recombinant asci containing black and red ascospores. After 11 days, wild type (black) and fus (red) ascospores were isolated from recombinant perithecia and recovered on BMM-Ac medium. Colonies from at least 100 germinated black (wild type) and 100 germinated red (fus) spores were tested for hygromycin B resistance to detect those carrying a gene deletion.

RNA-seq analysis

For RNA-seq analysis, RNA isolated from wild-type and ∆nox1 mycelia and protoperithecia was used. For protoperithecia isolation by laser microdissection (LM), the ∆nox1 mutant was grown directly on slides for fixation and dissection in situ (Teichert et al. 2012). RNA was isolated from protoperithecia with the Arcturus PicoPure kit (Applied Biosystems, Carlsbad, CA) and amplified in two linear amplification rounds using the TargetAmp 2-Round aRNA Amplification kit 2.0 (Epicentre Biotechnologies, Madison, WI) with modifications as described in Teichert et al. (2012). For RNA isolation of total mycelia, strains were precultured for 2 days on solid BMM. From these plates, three 20-ml BMM liquid cultures were inoculated and incubated for an additional 4 days. RNA from mycelia was isolated using phenol/chloroform extraction. RNA from protoperithecia (3.5 µg) and total mycelia (400 µg) was used for library preparation and Illumina/Solexa sequencing at GATC Biotech. cDNA libraries were prepared with the TrueSeq RNA sample preparation kit (Illumina, San Diego, CA) and sequenced with a HiSeq2000. For wild-type total mycelia, RNA from one sample was sequenced, for ∆nox1 mycelia and ∆nox1 protoperithecia, RNA from two independent biological replicates was used for sequencing. Resulting reads were cleaned and mapped to the reference S. macrospora genome v2 (Teichert et al. 2012). Differential expression was evaluated using DESeq (Anders and Huber 2010) and a method called “classical analysis.” In this analysis, genes were grouped into five groups (0 to 4) containing genes that are not differentially expressed (group 0) to genes that are strongly and significantly differentially expressed (group 4) as described in Teichert et al. (2012). Based on the DESseq and classical analysis, a consensus was calculated according to the following criteria for differential expression similar to what was described before (Teichert et al. 2012): a gene is described as differentially regulated if ratios in both DESeq and classical analysis are >4 or <0.25, DESeq adjusted P-value ≤ 0.1, and a gene in groups 1–4 in the classical analysis. For comparisons of mycelia vs. protoperithecia samples, ratio thresholds were set to >8 and <0.125 (Supporting Information, File S1).

Quantitative real-time PCR

Quantitative real-time PCR (qRT-PCR) was performed as described previously (Nowrousian et al. 2005) using the master mix from the Promega GoTaq qPCR kit for SybrGreen. qRT-PCR was performed in a StepOnePlus (Applied Biosystems) using StepOne software v2.2. Sequences of oligonucleotides used are given in Table 2.

Accession numbers

Raw sequence data from sequencing mutant pro32 (pro32/fus) and wild type (wt_3) were submitted to the National Center for Biotechnology Information (NCBI) sequence read archive (accession no. SRP033637). The RNA-seq reads and derived expression ratios were submitted to the Gene Expression Omnibus (GEO) database (accession no. GSE49363).

Results

Genome sequencing of mutant pro32 identifies a point mutation in nox1 encoding NADPH oxidase 1

We have previously generated a collection of sterile pro mutants with a developmental block after protoperithecia formation (Kück et al. 2009). Using next-generation sequencing for efficient and time-saving identification of mutations (Nowrousian et al. 2012), we sequenced the genome of sterile pro32 (Figure 1A). After cleaning of raw sequence data from pro32 and a wild-type reference strain, the majority of reads (>95%) mapped to the reference genome (Table S1). We identified one mutation with 100% penetrance at position 810 of SMAC_05007 encoding a putative NADPH oxidase 1 (NOX1) in pro32, but not in the wild-type sample (Table 4). The point mutation in pro32 results in a transition from G to A at position 810 of the SMAC_05007 ORF, creating an early stop codon. As a consequence, the highly conserved NOX1 ferredoxin reductase-like C-terminal domain is missing from the NOX1 protein of the mutant strain (Figure S1). After crossing of pro32 to spore color mutant fus that carries a mutation in the tih melanin biosynthesis gene (Nowrousian et al. 2012), PCR fragments covering the nox1 gene were obtained from nine fertile and nine sterile ascospore progeny. Sequencing of PCR fragments confirmed the nox1 base pair substitution in all strains with the sterile pro32 phenotype (Figure 1B). NOX1 function in fruiting body development was further verified by transformation of pro32 with a wild-type nox1 gene including 5′ and 3′ flanking regions. The resulting transformants showed a restoration of fertility, as indicated by the formation of mature fruiting bodies (Figure 1A).

Figure 1.

Genome sequencing of pro32 mutants reveals a mutation in the nox1 gene. (A) Sexual phenotypes of indicated strains. Insets show a detailed view of either wild-type black perithecia or mutant light brown protoperithecia. Bar, 100 µm. (B) Sequence comparison of the nox1 gene from wild type and 18 ascospore isolates from a cross between pro32 and fus strains. Sterile strains (underlined and blue) show a G to A transition, which changes nox1 W222 codon to a TGA stop codon (boxed).

Table 4. Summary of small sequence variants detected in the pro32 genome when compared to the reference genome.

Genotype	Sequence sample	No. of small variants with coverage >40%	No. of mutations with 100% penetrance	Location of putative mutations
wt	wt_3	146	—	—
pro32	pro32/fus	142	3	G810A in SMAC_05007 results in stop codon at W222
				C926T in SMAC_05015 does not result in change of amino acid sequence
				C2620T in SMAC_05015 results in V310I

Mutations caused by small sequence variants were identified by screening the sequence data for SNPs and indels (insertions/deletions) of <4 bases with a coverage of at least 40% of the average coverage for that sample. For these putative mutations, it was subsequently checked whether they had 100% penetrance, i.e., all the reads in the strain had the SNP/indel and none of the reads in the sequenced wild-type sample carried this specific mutation.

Deletion mutants lacking nox1 or nor1 show developmental defects

To further investigate NOX1 function in S. macrospora, we generated deletion mutants of nox1 as well as nor1 (SMAC_02124), which encodes the NOX regulator NOR1. These genes show 89 and 94% similarity to the corresponding genes from Neurospora crassa (accession nos. XP_964104 and XM_958018, respectively). The nox1 and nor1 genes were replaced by the hygromycin B resistance cassette through homologous recombination at the 5′ and 3′ regions flanking the target genes in a Δku70 host (Figure S2). The transformants were crossed with fus, a mutant generating red ascospores, to obtain homokaryotic strains carrying the nox1 or nor1 deletions without the ku70 deletion. Correct replacement of nox1 or nor1 with the hph cassette was verified by Southern hybridization using hph-, nox1-, or nor1-specific probes (Figure S3).

∆nox1 and ∆nor1 mutants showed distinctive similar phenotypes when compared to the wild-type strain. Both mutants have a sterile phenotype, being able to generate protoperithecia, but no perithecia (Figure 2A). As shown in Figure 2B, microscopic investigations revealed that both strains, as well as the pro32 mutant, are defective in vegetative cell fusion. In addition, ∆nox1 and ∆nor1 mutants were characterized by a significant reduction in hyphal growth by ∼61 and 21%, respectively (Figure 2C). With these phenotypes, pro32, ∆nox1, and ∆nor1 resemble other pro mutants found in a screen to detect strains having a developmental block after protoperithecia formation (Kück et al. 2009). To complement ∆nox1 and ∆nor1 mutants, we transformed them with full-length copies of nox1 or nor1 genes. In all cases, the reduced growth and sterility phenotypes were rescued in the transformants. However, hyphal fusion was only observed in the ∆nor1::nor1 and pro32::nox1 strains, but not in ∆nox1::nox1 (Figure S4). Further, the ectopic integration of nox1 in the corresponding deletion mutant leads only to a reduced number of perithecia, suggesting dose-dependent effects in the ∆nox1::nox1 complemented strain. Nevertheless, complementation analysis of ∆nor1 and pro32 prove that functional nox1 and nor1 genes are required for vegetative cell fusion, hyphal growth, and proper fruiting body formation.

Figure 2.

Phenotype of nox and nor1 deletion mutants. (A) Strains ∆nox1 and ∆nor1 are unable to develop black perithecia, and thus produce only slightly pigmented protoperithecia. Mutant ∆nox2 shows wild-type-like sexual development. Pictures were taken after 7 days of growth on BMM medium and incubation at 27° in constant light. Bar, 50 µm. (B) ∆nox1, pro32, and ∆nor1, but not ∆nox2, are affected in hyphal fusion. Strains were grown for 2 days on minimal medium on a cellophane layer in 27° in constant light. The assay was performed at least three times for every mutant. Hyphal fusion is indicated by white arrowheads; the lack of hyphal fusion between hyphae in close contact is marked by black arrowheads. Bar, 50 µm. (C) ∆nox1 and ∆nor1, but not ∆nox2 mutants, show decreased linear growth. Growth was followed in race tubes for 6 days in three replicates.

For P. anserina, it has been shown that the fertility defect of a PanoxA deletion mutant is rescued by serial transfers of the fungus to nutrient-rich medium (Malagnac et al. 2004). We observed a similar rescue of the S. macrospora nox1 deletion mutant (Figure S5). After germination of ∆nox1 ascospores from the rescued strain, we again obtained sterile strains that showed only protoperithecia formation. To determine whether this rescue phenotype is a general feature of sterile pro mutants, we performed serial media shifts using ∆nor1 and the unrelated ∆pro40 mutant (Figure S5). In contrast to ∆nox1, these mutants remained sterile even after several transfers.

Hyphal fusion mutants show elevated levels of ROS

NADPH oxidases are known to produce ROS in a highly spatiotemporally regulated fashion (Aguirre and Lambeth 2010). Using the NBT assay for the detection of superoxide, we quantified the level of ROS in pro32, ∆nox1, and ∆nor1, as well as in other mutant strains affected in perithecia development and/or defective in hyphal fusion. Compared to wild type, sterile and hyphal fusion defective pro32, ∆nox1, and ∆nor1 showed enhanced ROS levels in vegetative hyphae (Figure 3A) and protoperithecia (Figure 3B). This enhancement of ROS is also shared by the hyphal fusion mutant ∆pro40 (Engh et al. 2007). In contrast, all investigated mutants with wild-type-like cell fusion showed wild-type-like levels of ROS. As a consequence of the hyphal fusion defect in ∆nox1 and other developmental mutants, we propose conditional nutrient starvation, and thus sterility in these strains.

Figure 3.

Mutants affected in hyphal fusion show increased levels of NBT reduction. (A) Quantitative measurement (fold change) of NBT precipitates after 30 min of incubation in at least three replicates. Gray bars indicate hyphal fusion-deficient strains. Hyphal fusion defect correlates with sterility, except for sterile mutant ∆gsa3∆ste12, which can undergo normal hyphal fusion. Normalization was done in reference to wild type. (B) Detailed view of sexual (black arrowheads) and vegetative (white arrowheads) structures stained by NBT. Bar, 50 µm, 1 = ascospores.

In protoperithecia, NOX1 regulates transcription of genes for hyphal fusion and cytoskeleton remodeling

To obtain information about differentially regulated genes that are dependent on NOX1 activity, we performed RNA-seq analysis using RNA samples obtained from wild-type and ∆nox1 mycelia and protoperithecia (Table S2). Mycelia were isolated from surface liquid cultures, while protoperithecia samples were obtained using a recently developed laser microdissection technique that allows protoperithecia isolation with minimal mycelial contamination (Teichert et al. 2012).

To allow conclusions about gene expression in protoperithecia of different developmental mutants, we included RNA-seq data obtained in a previous study for protoperithecia of the wild-type and the sterile mutant pro1 (File S1) (Teichert et al. 2012). pro1 encodes C₆ zinc finger transcription factor PRO1 (Masloff et al. 1999) and shares the sterile phenotype with ∆nox1 strains. Comparison of genome-wide expression patterns between the different samples (total mycelia from wild type and ∆nox1 as well as protoperithecia from wild type, ∆nox1, and pro1) showed that the protoperithecial samples cluster apart from the mycelial samples, and that ∆nox1 protoperithecia are more similar to pro1 protoperithecia than to wild-type protoperithecia (Figure S6A). This reinforces a previous finding that expression patterns of protoperithecia are distinct from those of nonreproductive mycelia (Teichert et al. 2012). Furthermore, the fact that pro1 and ∆nox1 protoperithecia cluster together indicates that there is a common mutant-specific expression pattern in young fruiting bodies blocked at a similar developmental stage. These findings were confirmed by a comparison of differentially expressed genes in protoperithecia and mycelial samples (Figure 4A) and an analysis of the 500 most strongly expressed genes in each of the samples (“top500” analysis, Figure S6B). A total of 501 genes are differentially expressed in both, ∆nox1 and pro1 protoperithecia, compared to wild-type protoperithecia, whereas only 33 genes are differentially expressed specifically in the comparison of ∆nox1 and wild-type protoperithecia as well as ∆nox1 and wild-type mycelia (Figure 4A). The top500 analysis showed that intersections between wild-type and ∆nox1 mycelia and intersections between the protoperithecial samples contained more genes than intersections between protoperithecial and mycelial samples (Figure S6B).

Figure 4.

Results of RNA-seq-based expression analysis in developmental mutants. (A) Consensus analysis of differentially regulated genes identified by DESeq and classical statistical methods. Differentially regulated genes in relation to the corresponding wild-type tissue are indicated in yellow (∆nox1 mycelia), green (∆nox1 protoperithecia), and gray (pro1 protoperithecia). The numbers of up- (↑, red) and down- (↓, blue) regulated genes are indicated. (B) Illustration of functional categories (FunCat) (Ruepp et al. 2004) of up- (red) and down- (blue) regulated genes in protoperithecia from ∆nox1 mutant compared to wild-type protoperithecia. For each FunCat category, the fold representation compared to its representation among all predicted proteins is given. Asterisks indicate statistically overrepresented functional groups (P-value ≤ 0.05). (C) Heatmap of regulated genes in RNA-seq and qRT-PCR experiments using RNA from protoperithecia (proto.) and mycelium (myc.). Different statistical analyses are indicated by “a” (DESeq) and “b” (classical). Numbers (1 and 2) indicate biological replicates. Up- (red) and down- (blue) regulated genes are indicated.

To identify the possible functions of genes differentially regulated in ∆nox1 compared to wild-type protoperithecia, we performed functional categories (FunCat, P < 0.05) and BLAST analyses, using the N. crassa orthologs for FunCat analysis. Figure 4B illustrates functional categories (FunCat) for all N. crassa orthologs, in total 554 up- (red) and 432 down-regulated (blue) genes. Significantly overrepresented categories correspond to metabolism and cell rescue (down-regulated compared to wild type), as well as cell cycle and protein synthesis (up-regulated compared to wild type).

Furthermore, BLAST analysis of differentially regulated genes in ∆nox1 protoperithecia revealed that several of these genes encode proteins important for cytoskeleton remodeling, like CDC42, profilin, cofilin, and coronin-1 (Table S3). Coronin-1 has been shown to have a major role in actin organization and dynamics in N. crassa (Echauri-Espinosa et al. 2012). In a second group, genes important for hyphal fusion are differentially regulated in ∆nox1 protoperithecia compared to wild-type protoperithecia. Among these is the ham-10 gene, which is essential for hyphal fusion and perithecia formation in N. crassa (Fu et al. 2011) (Table S3, Table S4). The third group contains genes encoding subunits of NADH:oxidoreductases of the respiratory chain and of mitochondrial ATPase. Consistent with this, mutants unable to assemble the respiratory chain complex I are female sterile in N. crassa (Duarte and Videira 2000) (Table S3, Table S4). The enrichment of these three groups together with the FunCat analysis suggests that NOX1 has regulatory functions, e.g., in cytoskeleton remodeling, hyphal fusion, metabolism, and mitochondrial respiration.

Extensive genetic analysis has identified a large number of developmental genes involved in fruiting body formation (Pöggeler et al. 2006; Kück et al. 2009; Engh et al. 2010). Results from the RNA-seq-based differential expression analysis of these genes are given in Table 5. In Figure 4C results from qRT-PCR analysis are shown to verify some of the results of the RNA-seq analysis, and to further compare gene expression patterns in mycelia and protoperithecia from ∆nox1 and other developmental mutants. RNA-seq data show that both pheromone receptor genes pre1 and pre2 and the two mating type genes SmtA-1 and SmtA-3 are up-regulated in ∆nox1 mycelia and protoperithecia. Similarly, genes for melanin biosynthesis (pks, teh, sdh, and tih) and two genes associated with fruiting body maturation, app and tap1, are down-regulated in ∆nox1 mycelia and protoperithecia, similar to the situation in other pro mutants (Nowrousian et al. 2005, 2007; Teichert et al. 2012), whereas they only show minor expression changes in fertile mutants ∆nox2 and ∆ste12 compared to wild type (Figure 4C). These findings are in line with the proposed hypothesis of a common pro mutant-specific expression pattern.

Table 5. Log₂ ratios of relative expression of known developmental genes in Δnox1 mycelium and protoperithecia.

S. macrospora locus tag	Gene	Log2 ratios of gene expression in mycelium of Δnox1 vs. wt		Log2 ratios of gene expression in protoperithecia of Δnox1 vs. wt
		DESeq analysis	Classical analysis	DESeq analysis	Classical analysis
Transcription factor genes
SMAC_00338	pro1	1.011	1.295	0.955	1.787
SMAC_03223	pro44	−0.358	−0.140	−2.059	−1.380
SMAC_05219	mcm1	−0.041	0.081	−1.296	−0.565
SMAC_06479	ste12	0.050	0.330	−2.836	−2.196
Pheromone pathway genes
SMAC_02283	pre1	3.907	4.041	7.720	7.381
SMAC_08994	pre2	2.104	2.231	2.867	3.631
SMAC_05970	ppg1	1.434	1.446	−0.365	0.320
SMAC_12697	ppg2	0.496	0.540	−0.166	0.583
Mating type genes
SMAC_05404	Smta-1	4.052	4.188	a	a
SMAC_05401	SmtA-1	3.331	3.366	3.750	4.562
SMAC_05402	SmtA-2	3.836	3.827	1.440	2.207
SMAC_05403	SmtA-3	4.295	4.303	3.162	3.862
Nox genes
SMAC_05007	nox1	−1.109b	−0.913b	−3.076b	−2.404b
SMAC_08741	nox2	1.210	1.430	−1.006	−0.299
SMAC_02124	nor1	2.532	2.718	−0.376	1.025
MAP kinase cell integrity pathway genes
SMAC_03673	mik1	0.479	0.717	1.284	1.939
SMAC_02183	mek1	0.813	0.967	0.401	1.033
SMAC_05504	mak1	0.546	0.743	−1.545	−0.870
STRIPAK complex genes
SMAC_08794	pro11	0.255	0.474	1.400	2.050
SMAC_02580	pro22	0.230	0.417	−0.985	−0.273
SMAC_00877	mob3	−0.245	−0.084	0.311	0.988
SMAC_01224	pro45	0.372	0.601	−0.210	0.438
SMAC_01919	pp2AA	0.876	0.95	1.152	1.841
SMAC_04678	pp2Ac1	1.234	1.305	1.173	1.827
Melanin biosynthesis genes
SMAC_03130	pks	−4.541	−4.251	−6.916	−6.057
SMAC_05880	teh	−4.829	−4.679	−6.227	−5.523
SMAC_02101	sdh	−3.656	−3.603	−4.134	−2.870
SMAC_05650	tih	−5.819	−5.601	−7.691	−6.990
G-protein alpha subunit genes
SMAC_05328	gsa1	−0.094	0.120	−0.360	0.311
SMAC_06605	gsa2	0.235	0.386	−0.422	0.220
SMAC_07195	gsa3	−0.453	−0.311	−1.325	−0.682
Genes associated with ascospore germination
SMAC_01638	sac1	0.714	0.825	−0.619	0.037
SMAC_09071	rac1	0.235	0.393	0.875	1.729
Genes associated with fruiting body maturation
SMAC_06095	app	−8.505	−8.274	−7.069	−6.419
SMAC_03372	tap1	−8.984	−8.722	−9.364	−8.716
SMAC_00522	fbm1	1.470	1.810	0.181	1.150
Other essential genes for fruiting body development
SMAC_07802	pro4	2.579	2.901	0.536	1.347
SMAC_04848	pro41	2.354	2.511	0.406	1.110
SMAC_06775	acl1	−1.353	−1.174	−0.734	0.100
SMAC_08608	asf1	0.879	0.988	1.141	1.809
SMAC_06539	atg7	−0.355	−0.233	−0.643	0.041
SMAC_04815	pro40	1.225	1.468	1.225	1.468

^aNo reads mapped in wild-type protoperithecia. Therefore no ratios can be calculated.

^bCounted reads map to 5′ and 3′ UTRs of nox1, which were not deleted in the ∆nox1 mutant.

NOX2 is required for the germination of melanized ascospores

Like other ascomycetes, the genome of S. macrospora encodes a second NADPH oxidase, NOX2. The corresponding nox2 gene (SMAC_08741) shows 91% sequence similarity to its N. crassa homolog (accession no. XP_001728356). A homokaryotic ∆nox2/fus strain was generated with the strategy described in Materials and Methods. As can be seen in Figure 2, this strain has no defect in vegetative growth, cell fusion, or sexual development. However, when we performed random ascospore analysis from a cross of the primary transformant ∆ku70∆nox2 with the melanin-deficient fus mutant, we observed in a total of 390 analyzed viable ascospores a strict cosegregation of the hygromycin-B-resistant phenotype, corresponding to the nox2 deletion, with fus. To investigate ascospore germination further, we tested germination of diverse fertile strains generating either black or red spores. For sound statistical analysis, we investigated a minimum of 400 spores in each of three replicate experiments and tested their germination capacity on solid BMM-Ac medium. We analyzed wild type, fus, and the ∆ku70 strain, the latter of which was used as the recipient strain for generating the ∆nox2 mutant. As can be seen in Figure 5A, ascospores generally showed a germination rate between 40 and 70%. Similar results were obtained with the red-spored ∆nox2/fus strain. However, we were unable to isolate any viable (germinating) ascospores from black-spored ∆ku70/∆nox2 mutants, indicating that germination in ∆nox2 is suppressed by the presence of melanin.

Figure 5.

Phenotypes of ascospore germination-defective mutants. (A) Quantification of germination of ascospores discharged from selfed perithecia from wild type, different mutants, and ∆nox2::nox2 strains. A minimum of 400 ascospores per strain were tested in each of three biological replicates. Black bars represent strains with full melanized ascospores; brown bars represent strains with a block in melanin biosynthesis. (B) Ascospore germination under different ascorbate concentrations. A minimum of 400 ascospores from selfed perithecia were tested for each strain and ascorbate concentration. (C) Perithecia formation of ∆gsa3∆nox2/fus and ∆nox2∆ste12/fus mutants after 7 and 14 days incubation, respectively. Bar, 100 µm. (D) Time series of ascospore germination for the indicated mutants. The elapsed time in minutes is indicated. White arrowheads mark ascospores that burst instead of showing a germinaton vesicle. Bar, 50 µm. (E) Wild-type and fus mutant ascospore germination rates in the presence of db-cAMP. Discharged ascospores were collected and incubated on medium containing the indicated db-cAMP concentrations. A minimum of 400 ascospores per strain were tested for each of three biological replicates. Black bars represent strains with full melanized ascospores; brown bars represent strains with a block in melanin biosynthesis.

To verify the strict dependence of the germination defect on melanization, we generated a black-spored ∆nox2::nox2 complemented strain. Primary transformant ∆ku70∆nox2 was not appropriate for complementation experiments since it carries selection markers nat1 (noursethricin resistance) and hph (hygromycin B resistance) replacing ku70 and nox2, respectively. Therefore, we transformed hygromycin-resistant ∆nox2/fus with a wild-type copy of nox2. The red-spored ∆nox2/fus::nox2 was then crossed to wild type, and black ascospores were isolated carrying the ∆nox2 deletion and the ectopically integrated nox2 gene. Three randomly selected strains germinated at a frequency similar to wild type and ∆ku70 (Figure 5A), confirming that germination of black ascospores is strictly dependent on NOX2 function.

To assess whether the germination defect is restricted to ∆nox2, we tested further the ascospore germination phenotype of ∆nox1 and ∆nor1, both having a sterile phenotype. ∆nox1/fus and ∆nor1/fus double mutants were crossed to the wild type, and the germination assays of black and red spores were done as described in Materials and Methods. As can be seen from Table 6, ∆nor1, but not ∆nox1 ascospores, displayed the same melanin-dependent germination defect as ∆nox2 ascospores. This result indicates that NOR1 is required to regulate both NOX1 and NOX2.

Table 6. Frequency of hygromycin B resistance in black- and red-spored progeny from indicated crosses.

Cross	100 germinated black ascospores		100 germinated red ascospores
	HygS	HygR	HygS	HygR
∆nox1/fus × wt	65	35	49	51
∆nox2/fus × wt	100	0	43	57
∆nor1/fus × wt	100	0	48	52
∆gsa3/fus × wt	92	8	81	19
∆sac1/fus × wt	80	20	81	19
∆ste12/fus × wt	100	0	47	53
∆pro40/fus × wt	45	55	43	57

From each cross, 100 black and 100 red colony forming ascospores were tested for hygromycin B resistance, indicating the deletion of the corresponding gene.

These data suggest that ROS is necessary for spore germination. It might be hypothesized that in the wild type, both NOX2 as well as other pathways contribute to ROS production allowing spore germination. The ∆nox2 mutant would then retain residual amounts of ROS generated independently of NOX2; however, the residual amount of ROS in ∆nox2 ascospores might be scavenged by melanin in black ascospores. To further test this hypothesis, we performed germination tests with different ascorbate concentrations in the germination media. Ascorbate is an antioxidant that scavenges ROS. As can be seen in Figure 5B, germination of ascospores on ascorbate-containing media is drastically reduced in strains lacking nox2 in the fus mutant background but less pronounced in the fus reference strain. Thus, the antioxidant ascorbate mimics the effect of melanin in ∆nox2/fus spores.

NOX2, NOR1, and transcription factor STE12 act in a genetic pathway determining ascospore germination

The melanin-dependent ascospore germination defect of ∆nox2 and ∆nor1 prompted us to look for other genes involved in this process. Previously, S. macrospora deletion mutants lacking the gsa3 (alpha subunit 3 of the heterotrimeric G-protein), sac1 (adenylate cyclase), and ste12 genes were shown to have an ascospore germination defect (Nolting and Pöggeler 2006; Kamerewerd et al. 2008). We performed further crossing experiments to test whether these genes act with nox2 and nor1 in the same genetic pathway. ∆gsa3/fus, ∆sac1/fus, and ∆ste12/fus were crossed to a wild-type strain and then 100 black and 100 red germinating ascospores from these crosses were tested for hygromycin B resistance. As can be seen in Table 6, the ste12 deletion, but not the gsa3 or sac1 deletions, strictly cosegregated with the fus mutation. Thus, the germination defect of ∆ste12 resembles the ∆nox2 and ∆nor1 ascospore germination defect. To further verify this result, we generated triple mutant ∆nox2∆ste12/fus, having a fertile phenotype (Figure 5C). From a cross of the triple mutant with wild type, we isolated a total of 146 ascospores. Of these, 68 were red colored and hygromycin B resistant, thus indicating the deletion of nox2 and/or ste12. In contrast, 78 black and red ascospores showed wild-type-like sensitivity to hygromycin B. These data thus verify the strict cosegregation of the nox2 or ste12 deletion with the fus mutation, indicating that NOX2 and STE12 act in the same pathway that controls ascospore germination.

In contrast to ∆nox2, ∆nor1, and ∆ste12, sac1 and gsa3 deletion mutants have a germination defect that is independent of melanin in ascospores (Table 6). To determine whether the gsa3 and the nox2 pathways overlap, we generated triple mutant ∆gsa3∆nox2/fus. Like the single deletion strains, this strain is fertile (Figure 5C); however, the ascospores are unable to generate germination vesicles in contrast to the control strains fus and ∆nox2/fus. Instead, they burst after ∼5 hr on germination medium (Figure 5D, File S2, File S3, File S4). None of the 4000 ascospores analyzed showed germination. Thus, the germination-related phenotypes of ∆nox2 and ∆gsa3 (Kamerewerd et al. 2008) are exacerbated in the ∆gsa3∆nox2/fus mutant. This result clearly shows that NOX2 and GSA3 have a different impact on ascospore germination. Therefore, we propose that ascospore germination is regulated by two pathways, namely the NOX2-NOR1-STE12 and the GSA3-SAC1 pathways (Figure 6).

Figure 6.

Interplay of two pathways regulating ascospore germination in S. macrospora. During ascospore germination, two different signaling pathways, NOX2-NOR1-STE12 and GSA3-SAC1, exist and may be interconnected by cAMP. Work from other studies indicates that the GTPase RAC1 is inhibited by cAMP and thus negatively regulates the NOX2-NOR1 complex (Diebold et al. 2009).

cAMP inhibits germination of melanized ascospores

Recent reports indicate that cAMP, generated by an adenylate cyclase homologous to SAC1, inhibits the activity of a NOX2 homolog in human cells (Diebold et al. 2009; Li et al. 2012). At the molecular level, ROS production by the mammalian NOX2 complex is proposed to be reduced when the NOX2 activator RAC1 is directly inhibited by high cAMP concentrations. In analogy to this model, we hypothesized that increased cAMP levels could inhibit NOX2 activity in S. macrospora and reduce ascospore germination. To test this hypothesis, discharged ascospores from wild-type and fus strains were germinated on BMM-Ac medium supplemented with different concentrations of cell-permeable db-cAMP (Figure 5E). A concentration of 22.5 mM db-cAMP resulted in a reduction of ascospore germination of ∼30% in the wild type, but not in the red-spored fus mutant, which does not require NOX2 activity for germination. At higher db-cAMP concentrations of 62.5 mM and 250 mM, ascospore germination frequency of wild type, but not fus, dropped down by 70%. Thus, cAMP is able to mimic the ascospore germination defect of ∆nox2, ∆nor1, and ∆ste12 in a concentration-dependent manner. Based on these results, we hypothesize that cAMP generated by the GSA3-SAC1 pathway (Kamerewerd et al. 2008) exerts a negative regulatory role on NOX2 activity, which results in decreased production of ROS and failure to establish an ascospore germination vesicle, thus connecting the nox2/ste12 and gsa3/sac1 pathways (Figure 6).

Discussion

Fruiting body development is regulated by NOX1 and NOR1

The deletion of nox1 and nor1 genes led to a sterile phenotype that has been found previously in a set of S. macrospora developmental pro mutants, which can form protoperithecia, but are hampered in developing mature fruiting bodies (Kück et al. 2009). The deletion of nox1 or nor1 in other filamentous fungi leads to similar effects on sexual development. In Aspergillus nidulans, noxA deletion blocks the differentiation of cleistothecia (Lara-Ortiz et al. 2003), and similarly, P. anserina and N. crassa nox1 mutants are female sterile (Malagnac et al. 2004; Cano-Dominguez et al. 2008). Notably, in P. anserina (Malagnac et al. 2004) and S. macrospora, the sterile phenotype can be rescued by serially shifting the mutant to rich medium, indicating that sterility is related to the availability of nutritional factors, or ROS scavenging and signaling molecules. Alternatively, shifts to rich medium might induce the activity of NOX2, thereby bypassing the nox1 deletion.

A novel function of NOX1 and NOR1 was recently detected in N. crassa and Botrytis cinerea (Read et al. 2012; Roca et al. 2012), where a deletion of the corresponding genes abolishes fusion of conidial anastomosis tubes (CATs). S. macrospora does not form any conidia; however, hyphal fusion can be observed within the hyphal network (Rech et al. 2007). Our investigation showed that S. macrospora nox1 and nor1 mutants have a defect in hyphal fusion, as was also shown for corresponding mutants from E. festucae (Kayano et al. 2013). Both phenomena, sterility and hyphal fusion defect, are correlated but not strictly linked with each other. For example, the N. crassa ham-4 mutant is fertile but shows a hyphal fusion defect (Simonin et al. 2010) and the sterile Smatg8 mutant from S. macrospora still shows hyphal fusion (Voigt and Pöggeler 2013).

Since NOX enzymes are known to produce superoxide by the reduction of oxygen (Leto et al. 2009), we determined the levels of ROS using the NBT assay. Interestingly, we detected higher superoxide levels in all mutants tested that were defective in hyphal fusion, including ∆pro40. The superoxide level has been investigated in nox deletion mutants of several ascomycetes and diverging results have emerged. For example, while E. festucae nox1 mutants show decreased superoxide levels (Tanaka et al. 2008), increased levels of superoxide are detected in corresponding mutants of M. oryzae, P. anserina, and B. cinerea (Malagnac et al. 2004; Egan et al. 2007; Siegmund et al. 2013). The latter results are rather unexpected since NOX enzymes are known to generate superoxide. Nevertheless, our results highlight that not only a deletion of the nox1 gene leads to elevated superoxide levels, but also a deletion of pro40. Thus, our data support a general correlation between a hyphal fusion defect and enhanced superoxide levels generated by a yet unknown source in mutant strains, as was proposed previously (Malagnac et al. 2004; Egan et al. 2007; Siegmund et al. 2013).

Transcriptional profiling reveals differentially regulated genes in a nox1 mutant

Recently, transcriptional profiling of mycelia from a P. anserina Panox1 mutant was performed by microarray analysis and showed the deregulation of a large set of genes involved in carbohydrate degradation and secondary metabolism (Bidard et al. 2012). However, the authors analyzed total mycelia after 72 hr of growth, therefore expression patterns specific to sexual structures could not be elucidated in these experiments. Thus, our RNA-seq analysis of protoperithecia as well as mycelia significantly extends the current knowledge about NOX1-dependent gene expression by including sexual structures from a specific developmental stage. The increased levels of ROS in the ∆nox1 mutant might also be responsible for some of the gene expression changes observed.

One finding was that a large set of genes involved in cytoskeleton remodeling and hyphal fusion is differentially regulated in the ∆nox1 mutant. This finding is consistent with recent studies in Magnaporthe oryzae, where the NOX1 complex was shown to be important for the maintenance of the cortical F-actin network during penetration of the host plant (Ryder et al. 2013). These authors showed that ∆nox1 still initated penetration peg formation but was unable to proliferate in the infected plant tissue. Furthermore, Yno1p, the NOX enzyme from yeast, has a regulatory role in actin remodeling (Rinnerthaler et al. 2012). Interestingly, F-actin remodeling is also crucial for CAT fusion in N. crassa (Roca et al. 2010). Further, our RNA-seq analysis revealed that among the differentially expressed genes in ∆nox1 protoperithecia, there are some involved in the establishment of hyphal polarity. For example, cdc42, a gene central for the establishment of polarity, is up-regulated in ∆nox1 protoperithecia compared to wild-type protoperithecia. The activity of CDC42 is regulated by the guanine nucleotide exchange factor (GEF) CDC24 and the scaffold protein BEM1, both of which are able to interact with the NOR1 homolog of E. festucae in a yeast-two-hybrid assay (Takemoto et al. 2011). In N. crassa, Bem1 is actively recruited around the forming fusion pore of germlings, and bem1 deletion leads to a drastic reduction of germling fusion, indicating a cross-talk between polarity establishment and hyphal fusion (Schürg et al. 2012). In N. crassa it was further shown that ∆cdc24 and ∆cdc42 are fusion defective (Read et al. 2012). Taken together, our results support a model in which the differential expression of a set of genes involved in cytoskeleton remodeling and hyphal fusion is responsible for the inability of ∆nox1 mutants to undergo cell fusion, normal polar growth, and perithecia development. The regulated genes are prime candidates for further analysis of NOX1-dependent developmental functions.

Besides NOX enzymes, there are other ROS-producing enzymes and processes within a cell, among them complex I of the respiratory chain that includes NADH:oxidoreductases (Murphy 2009). Our RNA-seq data indicate the up-regulation of subunits of the respiratory chain NADH:oxidoreductase of complex I and Cytochrome c in ∆nox1, whereas two ATPase subunits are down-regulated. Thus, we propose that the enhanced ROS levels observed in ∆nox1 may be due to de-regulation of the mitochondrial respiratory chain, perhaps as a consequence of starvation of hyphae and protoperithecia. This would be consistent with the finding that a serial transfer to fresh media restores fertility of the ∆nox1 mutant strain (Figure S5). A similar phenomenon has been described for multiple human cell lines (Scherz-Shouval and Elazar 2007; Chen and Gibson 2008). There, starvation leads to an up-regulation of mitochondrial ROS, which in turn directly activates an autophagy-inducing pathway (Li et al. 2013).

NOX2, NOR1, and STE12 act in a genetic pathway controlling ascospore germination

Our results show that NOX2 and NOR1 are required for the germination of sexual spores. Similar defects have been observed in corresponding P. anserina (Malagnac et al. 2004) and N. crassa (Cano-Dominguez et al. 2008) mutants. We found that NOX2 is required for germination of melanized ascospores only, since melanin-deficient ascospores from ∆nox2 are able to germinate. Lambou et al. (2008) suggested for P. anserina that the weakened cell wall of pigment-deficient ascospores and the absence of melanin either triggered the germination process or prevented its inhibition. However, the ROS scavenging capacity of melanin (Riley 1997) might also play a role in this process. We propose that ROS is necessary for spore germination and a residual amount in ∆nox2 ascospores is scavenged by melanin, thus preventing germination. On the other hand, such residual ROS would not be scavenged in pigment-deficient ascospores, allowing ascospores to germinate.

The ascospore germination assays provided the novel finding that NOX2, NOR1, and the transcription factor STE12 might act in the same genetic pathway to regulate ascospore germination, while SAC1 and GSA3 act in a parallel but interconnected pathway. Indeed, the inhibitory effect of cAMP on germination of melanized ascospores suggests a link between both pathways. In human cells, an increased level of cAMP inhibits GTPase RAC1, which itself is required for NOX2 activity (Diebold et al. 2009). We propose that cAMP generated by adenylate cyclase SAC1 inhibits NOX2 activity via the RAC1 homolog (SMAC_09071) (Figure 6). The fact that gsa3 and sac1 deletion mutants are fertile suggests that cAMP might not play an important role in NOX1 regulation during growth and perithecia development. In summary, our analyses of ascospore germination in various mutants show that two genetic pathways, NOX2-NOR1-STE12 and GSA3-SAC1, are involved in the regulatory network governing this process.