All N. crassa strains used in this study, including Fungal Genetics Stock Center (FGSC)#4200 (wild type, a), FGSC#2225 (wild type, A), FGSC#11482 (Δmak-2, a), FGSC#11321 (Δmak-1, a), and knockout mutants for phenotypic analysis were obtained from the Fungal Genetics Stock Center¹ (University of Kansas Medical Center). Strains were maintained on Vogel’s slant (1× Vogel’s salts, 2% sucrose, and 1.5% Bacto Agar), Vogel’s plate (1× Vogel’s salts, 2% glucose, and 0.75% Bacto Agar), and liquid Vogel’s medium (1× Vogel’s salts, and 2% glucose). Synthetic crossing (SC) medium with 0.1% sucrose and 0.025 SC medium with 2% sucrose were used for protoperithecial production and evaluation, respectively (Westergaard and Mitchell, 1947). Strains were grown at 28°C for vegetative growth and asexual sporulation and at 25°C for sexual development.

Genome-wide transcriptional profiles of the Δmak-1 mutant, the Δmak-2 mutant, and the wild-type strain at the vegetative growth stage and the initial stage of protoperithecium formation were obtained by RNA sequencing. To obtain RNA at the vegetative growth stage, strains were cultured in liquid Vogel’s medium at 28°C for 22 h with orbital shaking, and the mycelia were harvested by vacuum filtration. To obtain RNA in the initial stage of protoperithecium formation, strains were inoculated on a solid synthetic crossing medium with 0.1% sucrose, on which sterilized cellophane was overlaid. Plates were incubated under constant dark at 25°C for 3.5 and 5.5 days.

Mycelia were immediately frozen after harvest and ground into fine powder in liquid nitrogen. Total RNA was extracted according to the standard TRIzol protocol (Invitrogen Corporation, Carlsbad, CA, United States). For transcriptomic profiling analysis, mixed RNA from three repeats were sent to Beijing Genomics Institute for RNA-seq analysis using the Illumina Hiseq2000 with a single-end module (Illumina, San Diego, CA, United States). The raw data have been deposited in the National Center for Biotechnology Information with the accession number GSE184024. The data analysis followed the method described by Sun et al. (2019). In this study, genes with a transcriptional change for more than twofold between two samples and genes with reads per kilobase of exon model per million mapped reads value of more than 10 in at least one sample were defined as differentially expressed genes.

For quantitative reverse transcription-polymerase chain reaction (RT-qPCR) analysis, RNA extraction, DNase I treatment, complementary DNA synthesis, and qPCR analysis followed the methods described previously (Sun et al., 2019). Each complementary DNA sample was analyzed in triplicate, and the average threshold cycle was calculated. Relative expression levels were calculated using the 2^–ΔΔCt method and normalized to the expression of β-tubulin (Livak and Schmittgen, 2001). Primer pairs used for RT-qPCR assay are shown in Supplementary Table 1.

Protoperithecium formation in knockout mutants for candidate genes co-regulated by the MAK-1 and MAK-2 proteins was observed, and mutants defective in protoperithecium formation or female fertility were screened. First, the wild-type and single-gene deletion mutants were individually inoculated on Vogel’s slants and grown at 28°C for 7 days. The fresh conidia were then harvested into distilled water, and the final concentration of conidial suspensions was adjusted to 10⁷ conidia/ml. Three microliters of the conidial suspension were inoculated on cellophane covered on an SC plate with 0.1% sucrose and cultured at 25°C. After 3.5, 5.5, and 7.5 days, protoperithecium was observed under a microscope.

To evaluate the female fertility of mutants, strains were inoculated onto an autoclaved filter paper covered on a sugar-free SC medium plate and incubated at 25°C in constant darkness for 5 days. Then, the conidial suspension of wild type with opposite mating type, a (#4200) or A (#2225), was inoculated on the colony of the female strain and cultured in constant darkness for 7 days, followed by continuous illumination for more than 1 week. Perithecium formation and ascospore ejection were observed by microscope.

As many of the isolates in the N. crassa single-gene deletion library have mutations in addition to the targeted deletion (Fu et al., 2011), these mutants with defects in protoperithecium or perithecium formation were subjected to check if their phenotype co-segregate with hygromycin resistance according to the co-segregation database constructed by Stephen Free’s group² (Chinnici et al., 2014). Also, the one that did not co-segregate was removed from the final data.

The sample preparation for protein extraction was the same as RNA-seq. After harvest, the mycelium of each sample was ground in liquid nitrogen and mixed with the proper amount of protein extraction buffer [50-mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.4; 137-mM sodium chloride; 10% glycerol] with phosphatase inhibitor mixture (All-in-One, APPLYGEN, China). After centrifugation (4°C, 12,000 rpm, 30 min), the supernatants were transferred to new Eppendorf tubes and centrifuged again if necessary. The protein concentration was determined using a Nanodrop spectrophotometer (BioSpec-nano, SHIMADZU Biotech). A proper volume of protein extracts with 30-μg total protein per lane was applied to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and subsequent Western blotting. To analyze the phosphorylation status of the MAK-1 and MAK-2 proteins, a polyclonal rabbit phospho-p44/42 MAP kinase antiserum (Cell Signaling Technology, United States) was used as a primary antibody, and AffiniPure goat anti-rabbit immunoglobulin G (IgG) (H+L) (APPLYGEN) was used as the secondary antibody (Roux and Blenis, 2004; Baccarini, 2005). The polyclonal rabbit Kss1 antiserum (y-50) (Santa Cruz Biotechnology, Inc.) (Chapman and Asthagiri, 2009), polyclonal goat Fus3 antiserum (yC-19) (Santa Cruz Biotechnology, Inc.) (Rodríguez-Escudero et al., 2006), and anti-β-tubulin mouse monoclonal antibody (CWBIO) were used as primary antibody to detect protein levels of MAK-1, MAK-2, and β-tubulin, respectively. AffiniPure goat anti-rabbit IgG (H+L) (APPLYGEN), horse anti-goat (H+L) (Dogesce), and goat anti-mouse IgG (H+L) (Multisciences) were used as secondary antibody.

All three MAP kinase cascades are required for growth and development in N. crassa (Lichius et al., 2012). More interestingly, deletion mutants of genes mak-2 and mak-1, which, respectively, encode the most downstream kinases of PR- and CWI-MAP kinase pathways, shared some phenotypic similarities. As shown in Supplementary Figure 1A, when grown in Vogel’s agar tubes, both the Δmak-2 and Δmak-1 mutants produced short aerial hyphae and fewer conidia, as compared with wild type. On synthetic crossing medium, N. crassa wild type formed regular black subspherical protoperithecia with trichogyne. However, only several abnormal mycelial knots were found on the colony surface of the Δmak-1 and Δmak-2 mutants (Supplementary Figure 1B). After mating, protoperithecia in the wild-type strain further developed into perithecium, whereas no similar structures were found in either the Δmak-1 mutant or the Δmak-2 mutant (Supplementary Figure 1C). These results are consistent with a previous study (Pandey et al., 2004). Phenotypic similarities in hyphal growth and asexual and sexual development between the Δmak-1 and Δmak-2 mutants suggest that PR- and CWI-MAP kinase pathways have some functional overlap.

However, these two mutants also displayed obvious differences in colony growth. The Δmak-1 mutant formed a rosette-like colony, which was not seen in the Δmak-2 mutant, and its colony growth was much slower than that of the Δmak-2 mutant (Supplementary Figure 1A; Pandey et al., 2004; Li et al., 2005; Maerz et al., 2008), suggesting that the MAK-1 and MAK-2 proteins might involve in similar biological processes but have some different functions.

To gain an insight into the molecular mechanisms of the MAK-1 and MAK-2 proteins on the regulation of protoperithecium development of N. crassa, we examined genome-wide transcriptional responses to single-gene deletion of mak-1 and mak-2. Samples of wild type, the Δmak-1 and Δmak-2 mutants from two developmental stages, including the early stage of protoperithecium formation (3.5 days on SC medium) and the vegetative growth stage, were used for RNA sequencing (Figure 1A). A total of 9,873 genes were mapped onto the genome. In wild type, 2,894 genes were upregulated, and 655 genes were downregulated in the early stage of protoperithecium formation as compared with vegetative growth stage (Figure 1B, Supplementary Data 1, and Supplementary Table 2). These results indicate that N. crassa has a genome-wide transcriptional response to the induction of protoperithecium development.

The RNA-seq results also revealed that the transcriptional responses of some genes were affected by the deletion of the mak-1 or mak-2 gene. Among the 2,892 upregulated genes during the protoperithecium formation in wild type, the transcript levels of 412 genes in the Δmak-1 mutant and 549 genes in the Δmak-2 mutant were dramatically lower than those in wild type (Figure 1C). Among the 622 downregulated genes during the protoperithecium formation in wild type, transcript levels of 141 genes in the Δmak-1 mutant and 150 genes in the Δmak-2 mutant were dramatically higher than those in wild type (Figure 1D). Interestingly, 328 genes were downregulated in both the Δmak-1 and Δmak-2 mutants during protoperithecium formation, and they account for 79.6 and 59.7% of genes whose expressions were downregulated in the Δmak-1 and Δmak-2 mutants (Figure 1C and Supplementary Data 2). Also, 102 genes were upregulated in both the Δmak-1 and Δmak-2 mutants, accounting for 72.3 and 68.0% of genes whose expression was upregulated in the Δmak-1 and Δmak-2 mutants, respectively (Figure 1D and Supplementary Data 2). These results indicate that there is a large overlap in the regulatory functions between the MAK-1 and MAK-2 proteins. GO enrichment analysis revealed that, among the 430 genes affected by both the mak-1 and mak-2 gene deletions, some are related to oxidoreductase activity, whereas some encode proteins belonging to integral components of a membrane (p < 0.01) (Supplementary Table 3).

Among these genes, mkr-2, mkr-3, and mkr-5, three downstream genes in the PR-MAP kinase pathway, are specifically expressed during sexual development (Li et al., 2005). RT-qPCR further verified the regulation of these three genes by MAK-1 and MAK-2. As shown in Figure 2, the transcript levels of these genes dramatically increased during the protoperithecium formation and peaked after growth for 3.5 days in wild type. However, in both the Δmak-2 mutant and the Δmak-1 mutant, the transcript levels of mkr-2, mkr-3, and mkr-5 were dramatically lower than those in wild type, and the transcript levels of mkr-2 and mkr-3 were almost undetectable (Figure 2).

In addition, some sexual development essential genes, such as poi-2, app, and fsd-1, and several genes with unknown functions, such as NCU07743, NCU09716, NCU05683, and NCU05789, were also selected for RT-qPCR analysis. In wild type, the transcript levels of these genes were dramatically increased during protoperithecium formation, but these increases were abolished or greatly reduced in both the Δmak-1 and Δmak-2 mutants (Figure 2). All these results demonstrate that protoperithecium development is co-regulated by the MAK-1 and MAK-2 proteins.

To examine whether these genes are regulated by the known transcription factors (ADV-1 and PP-1) downstream of MAK-1 and MAK-2 proteins, we analyzed the ADV-1 and PP-1 binding sites of these co-regulated genes in their promoters. Among the 430 genes co-regulated by the MAK-1 and MAK-2 proteins during protoperithecium development, a total of 167 genes have ADV-1 and/or PP-1 binding sites in their promoters, including several transcription factors (Supplementary Data 2). In addition, among the 430 co-regulated genes, 26 were regulated by ADV-1 and/or PP-1 during hyphal growth (Fischer et al., 2018). These results indicate that the MAK-1 and MAK-2 proteins regulated protoperithecium development may be partially through ADV-1 and/or PP-1, and there may be other transcription factors downstream of the MAK-1 and MAK-2 proteins involved in protoperithecium development.

Among the 430 genes co-regulated by the MAK-1 and MAK-2 proteins in the early stage of protoperithecium development, roles of 61 genes in sexual development were previously assayed (Kim and Nelson, 2005; Nowrousian et al., 2007; Fu et al., 2011; Park et al., 2011; Chinnici et al., 2014; Carrillo et al., 2017; Fischer and Glass, 2019), including app, poi-2, stk-17, fsd-1, vsd-8, and NCU03868 (Table 1 and Supplementary Data 2). The gene app (abundant perithecial protein) is an indicator of sexual development, and its transcripts can be detected after the onset of protoperithecium development (Nowrousian et al., 2007). The remaining five genes, poi-2, stk-17, fsd-1, vsd-8, and NCU03868, are essential for the normal development of protoperithecium or perithecium formation (Kim and Nelson, 2005; Park et al., 2011; Carrillo et al., 2017).

RNA-seq results showed that the transcript levels of these six genes increased during protoperithecium development in wild type. However, their transcriptional responses to protoperithecium development induction were dramatically impaired in both the Δmak-2 and Δmak-1 mutants. After 3.5 days of sexual development induction, the transcript levels of app, poi-2, stk-17, fsd-1, and vsd-8 and NCU03868 in the Δmak-2 mutant were only 0.15, 0.12, 43.35, 16.46, 23.63, and 22.54%, respectively, of that in wild type, and the transcript levels of them in the Δmak-1 mutant were only 0.31, 0.19, 47.56, 41.9, 15.52, and 20.07%, respectively, of that in wild type (Table 1 and Supplementary Data 2).

Melanin is involved in the pigmentation of sexual structures in several fungi (Engh et al., 2007; Ao et al., 2019). According to the RNA-seq results, six of seven melanin synthesis genes, per-1, pkh-1, pkh-2, pkr-2/tnr-2/4hnr, pkr-1/tnr-1/3hnr, and scy-1/sd (Ao et al., 2019), as well as a transcriptional regulator encoding gene mld-1 that is required for melanin synthesis (Carrillo et al., 2017), were co-regulated by the MAK-1 and MAK-2 proteins (Table 1). To verify these results, tnr-1, tnr-2, scy-1, and mld-1 genes, which are involved in the last steps of melanin biosynthesis and its regulation (Carrillo et al., 2017; Ao et al., 2019), were selected for RT-qPCR analysis. As shown in Figure 3A, expression of these four genes was remarkably increased during protoperithecium formation and reached their peak at 5.5 days when protoperithecia became mature and darkened in the wild type. The expression levels of all these genes in both the Δmak-2 and Δmak-1 mutants were dramatically lower than those of wild type, verifying that MAK-2 and MAK-1 positively regulate these genes. Moreover, the transcriptional induction of tnr-2, tnr-1, and scy-1 during protoperithecium development was dramatically lower in the Δmld-1 mutant than wild type (Figure 3A). Thus, the MLD-1 protein may function downstream of the MAK-1 and MAK-2 proteins in regulating the melanin synthesis genes.

Interestingly, compared with wild type, the transcript levels of tnr-2, tnr-1, scy-1, and/or mld-1 were also lower in the asexual stage in the Δmld-1, Δmak-2, and Δmak-1 mutants (Figure 3A), indicating that the basal expression of melanin synthesis genes was also regulated by the MLD-1, MAK-1, and MAK-2 proteins.

The phenotypes of the knockout mutants of the four genes mentioned earlier were also analyzed. All Δtnr-2, Δtnr-1, Δscy-1, and Δmld-1 mutants can form protoperithecia and perithecia as the female parent. However, the perithecia of the Δscy-1 and Δtnr-1 mutants were more fragile and easier to disintegrate in the smashing process, and the color of the smashed perithecia is slightly lighter, especially in the Δtnr-1 mutant, compared with the dark brown or black in the wild type (Figure 3B). Moreover, due to the secreted melanin, the filter papers covering the SC plates were dark brown after the perithecia of the wild-type strain as the female parent matures. In contrast, the filter papers were brown-orange of the deletion mutants of tnr-2, tnr-1, or scy-1 after the perithecia formation or maturation and completely white of the Δmld-1 mutant, indicating abnormal melanin synthesis and secretion in these mutants during the crossing. More interestingly, we found that the Δmld-1 and Δtnr-2 mutants formed mature asci with eight ascospores, whereas the Δtnr-1 and Δscy-1 mutants did not. The Δtnr-1 mutant cannot detect any mature asci, whereas the Δscy-1 mutants can form very few and immature asci with ascospores (Figure 3B). These results indicate that the synthesis of the secreted melanin is not essential for perithecia formation and maturation, but some genes for melanin synthesis are critical for the development of perithecia, asci, and ascospores. Taken together, the melanin synthesis is required for the production of darkened sexual structures, and this process can be co-regulated by the MAK-1 and MAK-2 proteins.

Among the 430 genes co-regulated by the MAK-1 and MAK-2 proteins (Figure 1), only 61 have been analyzed for their function in sexual development (Kim and Nelson, 2005; Nowrousian et al., 2007; Fu et al., 2011; Park et al., 2011; Chinnici et al., 2014; Carrillo et al., 2017; Fischer and Glass, 2019).

To do a systematic analysis of the roles played by the MAK-1 and MAK-2 proteins during sexual development, a total of 275 gene deletion mutants were evaluated for their ability to form protoperithecia and perithecia. Two hundred seven of these genes were positively regulated, and 68 genes were negatively regulated by both the MAK-1 and MAK-2 proteins. Mutants of the remaining 94 genes are lethal and/or not available from FGSC. A total of 244 gene deletion mutants showed no defects in protoperithecium and perithecium formation (Supplementary Data 2), but 31 gene deletion mutants showed defects in protoperithecium or perithecium formation to varying degrees (Table 1 and Figure 4). Specifically, single-gene deletion mutants of 12 genes [NCU07743, NCU02250 (oli), NCU05948, NCU09782, NCU09248 (tcf-27), NCU04930 (cea-6), NCU04474 (mol-1), NCU00236 (fpo-1), NCU07776 (acw-5), NCU04923 (gcy-1), NCU04936 (gld-1), and NCU08457 (eas)] displayed female sterility, as no perithecium formed after mating (Table 1 and Figure 4A). Deletion mutants of NCU02209 (ods, fatty acid metabolism-4), NCU03422 [named as reduced perithecia -1 (rp-1)], NCU04489 [named as reduced perithecia -2 (rp-2)], NCU09199 (ty-6, tyrosinase-6), and NCU04952 (gh3-4, beta-D-glucoside glucohydrolase) produced less perithecia as compared with wild type (Table 1 and Figure 4B), and deletion mutants of NCU04496, NCU04949, NCU09976 (ce12-1, rhamnogalacturonan acetylesterase), NCU00830 (tcu-1, ctr copper transporter), NCU07053 (ahd-4, alhehyde dehydrogenase-4), and NCU09415 showed delayed formation of protoperithecia (Table 1).

However, the mutant phenotype of NCU09716, NCU05104, NCU01383, NCU03530, NCU05789, NCU05683, NCU08131, and NCU07180 did not co-segregate well with hygromycin resistance (Chinnici et al., 2014), indicating that additional genetic modifications are likely the cause of the observed deficiencies. Therefore, 12 mutants with female sterility phenotypes can be ascribed to a single gene deletion.

Among the 12 mutants that were female sterile, only deletion mutants for NCU07743, NCU02250 (oli), and NCU05948 did not form protoperithecium (Figure 4A, up panel). NCU02250 (oli) encodes an ATP synthase subunit ATP9, NCU07743 encodes a TauD domain-containing hypothetical protein, and NCU05948 encodes a hypothetical protein. Therefore, we named NCU07743 and NCU05948 as protoperithecium formation defect -1 (pfd-1) and pfd-2, respectively. Interestingly, similar to the △mak-2 and Δmak-1 mutants, deletion of NCU02250 led to short aerial hyphae (Figure 4C and Supplementary Figure 1A), suggesting a functional association between the gene NCU02250 and the MAK-1 and MAK-2 kinases.

The remaining nine mutants among the 12 female sterile mutants produced protoperithecia. Deletion mutants of NCU09782, NCU09248 (tcf-27), NCU04930 (cea-6), and NCU04474 (mol-1) produced less protoperithecia (Figure 4A, middle panel), whereas mutants of NCU00236 (fpo-1), NCU07776 (acw-5), NCU04923 (gcy-1), NCU04936 (gld-1), and NCU08457 (eas) produced a normal amount of protoperithecia (Figure 4A, last panel). These mutants may have defects in protoperithecial maturation, mating, and/or perithecium development, leading to no perithecium formation. Thus, the MAK-1 and MAK-2 proteins can regulate protoperithecial abundance and/or maturation, mating and/or perithecium development.

As described earlier, our RNA-seq data revealed a large regulatory overlap between MAK-1 and MAK-2, suggesting cross-talk between these two pathways. All three genes in the PR-MAP kinase pathway (nrc-1, mek-2, and mak-2) have increased transcription during protoperithecium development. Specifically, transcription of the mak-2 gene increased by 7.6-fold in protoperithecia compared with vegetative hyphae (Table 2). For the CWI-MAP kinase pathway, only one gene was upregulated during protoperithecium development. The transcription of the mak-1 gene in protoperithecia was 2.5-fold greater than in vegetative hyphae (Table 2). Deletion of mak-2 did not affect the transcription of genes in the CWI-MAP kinase pathway. However, the deletion of mak-1 affected the expression of the nrc-1 and mak-2 genes in the PR-MAP kinase pathway, whose transcript levels after 3.5 days of protoperithecium development induction were only 63 and 52% of that in wild type, respectively (Table 2). These results indicate that the transcriptional response of genes in the PR-MAP kinase pathway to sexual development signal may be regulated by MAK-1.

We next examined the protein levels and the phosphorylation status of the MAK-1 and MAK-2 proteins in all the deletion mutants of kinases (except for mik-1) belonging to the PR and CWI pathways by Western blotting. In wild-type cells, both the MAK-1 protein and phosphorylated MAK-1 were undetectable in vegetative hyphae but were strongly detected at both 3.5 and 5.5 days after protoperithecium induction (Figure 5). Neither the MAK-1 protein nor phosphorylated MAK-1 was detected in any of the tested MAPK deletion mutants. The MAK-2 protein was detected in both wild-type and mutant (except the Δmak-2 mutant) cells at roughly equivalent levels in both vegetative hyphae and protoperithecia. Notably, we observed a slight reduction in the amount of MAK-2 protein in the Δmak-1 mutant protoperithecia, which is consistent with our RNA-seq results (Table 2). Phosphorylated MAK-2 was strongly detected in wild-type hyphae and at 3.5 days after protoperithecium induction but was barely detectable at 5.5 days after protoperithecium induction (Figure 5). In contrast, phosphorylated MAK-2 was detected at equivalent levels at every developmental stage in the CWI-MAPK pathway mutants (Figure 5). Phosphorylated MAK-2 was not detected in any of the PR-MAPK pathway mutants.