Adaptive Spo11 RNA editing gate optimizes meiosis I pace and mitotic proliferation while preserving ascospore formation Authors: Mengchun Wu, Junfeng Liu, Jiahui Han, Junqi Huang, Chanjing Feng, Cong Jiang, Jin-Rong Xu, Qinhu Wang, Huiquan Liu Published: 2026 ABSTRACT Spo11-mediated DNA double-strand breaks (DSBs) are essential for meiotic recombination, yet how Spo11 activity is temporally regulated during mitosis and fungal development remains unclear. In the fungal plant pathogen Fusarium graminearum, we found that FgSpo11 has a DSB-independent role delaying meiosis I and a DSB-dependent role critical for postmeiotic mitoses during ascosporogenesis. Loss of FgSpo11 accelerates meiosis I and causes excessive postmeiotic divisions, ultimately causing aborted ascospores. A premature stop codon (TAG) is corrected to tryptophan (TGG) by adenosine-to-inosine RNA editing exclusively during sexual reproduction, enabling full-length protein synthesis. A genomically “corrected” allele bypassing this editing preserves ascospore formation but causes meiotic and vegetative mitotic defects. Beyond its on-switch function, this editing acts as a tunable rheostat fine-tuning FgSpo11 dosage during meiosis. Evolutionary analyses reveal recurrent gain and loss of this editing, highlighting adaptive modulation of Spo11 deployment. This study uncovers a single-site RNA editing gate controlling a key meiotic regulator and illustrates transcriptome plasticity in reconciling life cycle demands in eukaryotic pathogens. INTRODUCTION Meiosis is a fundamental biological process in sexually reproducing organisms, in which diploid progenitor cells undergo two successive divisions to produce haploid gametes (1). A hallmark of meiosis is homologous recombination, which plays a central role in promoting genetic diversity and maintaining genomic integrity (2). Meiotic recombination fulfills at least two essential functions. First, it facilitates the formation of crossovers between homologous chromosomes, resulting in chiasmata that physically link homologs and ensure their accurate segregation during the first meiotic division (meiosis I). Second, by reshuffling alleles between homologous chromosomes, meiotic recombination generates novel genetic combinations among gametes, thereby driving evolutionary adaptation and species diversity (3, 4). Fusarium graminearum is a major causal agent of Fusarium head blight (FHB), a devastating disease of cereal crops worldwide (5, 6). In addition to causing yield losses, this pathogen produces mycotoxins such as deoxynivalenol (DON), which are harmful to humans and animals (7). Sexual reproduction is central to its life cycle, as ascospores discharged from perithecia (sexual fruiting bodies) serve as the primary inoculum for FHB (8). Genetic recombination enhances pathogen population diversity, facilitating the emergence of new variants or lineages that are better adapted to their hosts and environments (9). In F. graminearum, meiotic recombination occurs at a high frequency, contributing to its substantial evolutionary potential and the prevalence of admixed isolates (10, 11). High recombination rates are observed in active genomic regions, promoting the generation of genetic diversity (12, 13). This diversity enables F. graminearum to rapidly acquire fungicide resistance and evolve more aggressive, pathogenic populations in agricultural fields (14). F. graminearum is homothallic and readily produces perithecia under laboratory conditions, making it an excellent model for investigating the molecular mechanisms underlying sexual reproduction (8). Recently, adenosine-to-inosine (A-to-I) mRNA editing has been identified as an epitranscriptomic regulatory mechanism in F. graminearum and other filamentous ascomycetes, occurring specifically during sexual reproduction (15–17). This process is mediated by the Tad2 (tRNA-specific adenosine deaminase)-Tad3-Ame1 (activator of mRNA editing) ternary complex, which represents a conserved feature among Sordariomycetes (18, 19). In this complex, the Tad2-Tad3 deaminase catalyzes A-to-I editing at the wobble position (A34) of tRNA anticodons—a function conserved across eukaryotes (20). The fungal-specific cofactor Ame1, which is expressed exclusively during sexual development, interacts with Tad3 and broadens the substrate specificity of the complex to include mRNA (18). Although the editing machinery differs, A-to-I mRNA editing is also a widespread phenomenon in animals and bacteria (19, 21, 22). Inosine is interpreted by the cellular machinery as guanosine (G), making A-to-I editing functionally equivalent to A-to-G mutations at the RNA level and potentially leading to protein recoding. More than 40,000 A-to-I mRNA editing sites have been identified in the perithecia of both F. graminearum and the model filamentous fungus Neurospora crassa, with most sites located in coding regions (23, 24). While Ame1 and several A-to-I editing sites have been shown to play critical roles in fungal sexual development (18, 22, 25, 26), identifying which individual editing sites are functionally important during sexual reproduction remains a challenge. A long-standing question in the field of RNA biology is why evolution has favored RNA editing over direct DNA mutations—specifically, what adaptive advantages RNA editing provides compared to permanent genomic alterations (27, 28). Unlike DNA mutations, which affect all transcripts from a gene, RNA editing is often partial, resulting in the simultaneous production of both edited and unedited isoforms from a single gene (15, 22). Moreover, the expression of RNA editing enzymes is developmentally regulated, leading to variable editing levels across tissues and developmental stages (24, 29). This dynamic regulation may allow organisms to fine-tune gene function by adjusting the proportion of edited versus unedited isoforms according to spatial and temporal requirements. However, the extent to which A-to-I RNA editing serves as a versatile mechanism for adaptive evolution and whether it contributes to genetic diversity remains to be fully elucidated. Meiotic recombination is initiated by programmed double-strand breaks (DSBs) catalyzed by the evolutionarily conserved enzyme Spo11, which is essential for meiosis in plants, animals, and fungi (30). Spo11-mediated DSBs are essential for meiotic recombination, but their formation poses inherent risks, such as mutagenesis and compromised gamete viability (31–34). To maintain genomic stability, the generation and repair of these breaks are subject to strict spatiotemporal regulation. This ensures that DSBs occur at the right time and place, maximizing their benefits for genetic diversity and accurate chromosome segregation while minimizing the potential for chromosomal damage (30). A key aspect of this regulation is Spo11’s inherently weak dimer interface, which limits its activity and necessitates the involvement of auxiliary factors to control the timing and location of DNA cleavage (30, 35–37). Here, we show that beyond its canonical role in meiotic recombination, the FgSPO11 gene in F. graminearum also negatively regulates meiosis I progression independently of its DSB-forming activity. Furthermore, we identify an unexpected DSB-dependent function for FgSPO11 in postmeiotic mitoses during ascosporogenesis. In addition, we identify a developmentally timed regulatory mechanism in which an A-to-I RNA editing event recodes a premature stop codon (PSC) in FgSPO11 during sexual reproduction, enabling production of the full-length functional protein. These findings provide insights into the interplay between epigenetic and genetic regulation and underscore a broader role for RNA editing in facilitating rapid adaptation in eukaryotes. RESULTS Accurate annotation of FgSPO11 reveals critical RNA editing sites The F. graminearum genome contains a single ortholog of the yeast SPO11 gene, designated as FgSPO11. Notably, RNA sequencing (RNA-seq) data revealed that the fourth and sixth introns of FgSPO11 were incorrectly spliced in previous gene models (FGRAMPH1_01T19143 or FGSG_05949) derived from computational predictions but were accurately spliced in our updated gene model (FG3G18670) based on full-length transcriptome sequencing (Fig. 1A) (38). According to this corrected gene model, FgSPO11 harbors a PSC (TAG) in the fifth exon, resulting in the production of a truncated protein (209 amino acids) that lacks the entire DNA topoisomerase domain (TOP6A-Spo11_Toprim, PF21180) (Fig. 1A). Role of FgSPO11 in sexual development. (A) Schematic of FgSPO11 gene structure, A-to-I editing sites, and protein domains. RNA-seq read coverage [0 to 59] from 8-days postfertilization perithecia (Sex8d) in a developmental time series indicates the transcription region and editing levels at three coding A-to-I sites. Green dashed lines mark two introns corrected in the updated gene model (FG3G18670) versus previous annotation (FGSG_05949), indicated by green triangles. The position of the corrected PSC editing site (TAG) is indicated. Predicted protein domains: TP6A_N (PF04406) and TOP6A-Spo11_Toprim (PF21180). Two nonsynonymous editing sites: N32D and *210W. (B) Colony morphology, perithecia, asci, and ascospore production in PH-1 [wild-type (WT)] and Fgspo11 deletion mutant on potato dextrose agar (PDA) (3 days) and mating cultures (8 days postfertilization). White triangles, ascospore cirrhi; red triangles, ascospore cloud mass; blue triangles, asci with wild-type eight ascospores or with few or no ascospores. Scale bars, 0.5 mm (white) and 20 μm (black). Bar graph indicates means ± SD of ascospore discharge (relative concentration) at 7 days postfertilization (n = 3; P value from two-tailed t test). Stacked bar chart indicates the percentage of asci with different ascospore numbers at 8 days postfertilization (n = 3, 30 perithecia per replicate). Red asterisk (*) indicates significant difference (P < 0.05) in normal asci with eight ascospores between strains by Fisher’s exact test. Only normal asci tested to simplify analysis. (C) DAPI (4′,6-diamidino-2-phenylindole)–stained asci at 6 days postfertilization visualized by epifluorescence microscopy. White dashed lines, ascospore boundaries; red triangles, ≥2 nuclei enclosed by forespore membrane. Scale bar, 10 μm. Bar graph indicates the percentage of asci with specific nuclear numbers at 6 days postfertilization (n = 3, 20 perithecia per replicate). Statistical significance versus PH-1 (wild-type) by two-tailed t test (*P < 0.05; **P < 0.01). (D) RNA-seq expression profile of FgSPO11 [transcripts per million (TPM)] during sexual development from 1 to 8 days postfertilization. dpf, days postfertilization; DIC, differential interference contrast. This finding prompted us to examine the FgSPO11 transcripts expressed during sexual stages. We identified a PSC correction editing site that converts the TAG stop codon to a tryptophan (TGG) codon at the 210th amino acid position (*210W). Across biological replicates, this site exhibited an average editing level of 81% at 8 days postfertilization (Fig. 1A and fig. S1). This RNA editing event enables the production of the full-length FgSpo11 protein (414 amino acids) specifically during sexual reproduction. Notably, the corrected protein sequence of full-length FgSpo11 after editing contains 57 amino acid residues (surrounding the edited tryptophan) that differ from previous gene model predictions, due to misannotation from alternative reading frames (fig. S1). In addition, we identified two further A-to-I mRNA editing sites within the coding region of FgSPO11 transcripts (Fig. 1A). One is a missense editing site, with an average editing level of 87%, which converts the 32nd amino acid from asparagine to aspartic acid (N32D). The other is a synonymous editing site. These three editing events—including the PSC editing site—completely disappeared in the ame1 deletion mutant and the T2Ssil-P mutant (where the sexual stage-specific FgTAD2 S-transcript is suppressed) (18) (fig. S1). These findings further confirm that the observed RNA editing modifications in FgSPO11 are genuine and specifically associated with the sexual developmental program. FgSPO11 is critical for ascospore formation To investigate the biological function of FgSPO11 in F. graminearum, we generated FgSPO11 deletion mutants using the split-marker method (39). Three independent Fgspo11 mutants with similar phenotypes were obtained (table S1), and the phenotypes of a representative strain are presented here. The Fgspo11 mutant exhibited normal colony growth rate and morphology, as well as normal pathogenicity and conidiation (Fig. 1B and fig. S2). However, during sexual reproduction, the Fgspo11 mutant formed normal perithecia but displayed reduced cirrhus formation and decreased spore discharge on carrot agar plates at 8 days postfertilization (Fig. 1B). Microscopic examination revealed that while ascus development within perithecia appeared normal, ascospore development within the asci was severely impaired. Most asci produced by the Fgspo11 mutant were completely devoid of ascospores. Fewer than 20% of asci contained ascospores, and the number of ascospores per ascus was markedly reduced (Fig. 1B). These results indicate that FgSPO11 is critical for ascospore formation in F. graminearum. FgSPO11 is required for proper meiosis and postmeiotic mitoses during ascosporogenesis To determine the underlying cause of defects in ascospore formation, we examined nuclear division and ascospore delimitation in asci, as the coordination of these two processes is essential for proper ascosporogenesis. In the wild-type strain, most asci at 6 days postfertilization contained eight nuclei, indicating the successful completion of meiosis, followed by a postmeiotic mitotic division. In contrast, the proportion of asci containing eight nuclei was markedly reduced in the Fgspo11 mutant, suggesting impaired nuclear divisions following the deletion of FgSPO11. In the wild type, a small fraction of developing asci with one, two, or four nuclei were observed (Fig. 1C), representing intermediate stages of ongoing meiosis. However, in the Fgspo11 mutant, the frequency of asci with one or two nuclei was significantly reduced, while the proportion with three nuclei was significantly increased (Fig. 1C). This pattern suggests that FgSPO11 is essential for the proper progression and completion of meiosis in F. graminearum. Deleting FgSPO11 appears to cause an accelerated progression through meiosis I, as indicated by the reduced number of early meiotic stages (one and two nuclei), coupled with impaired meiosis II, which leads to the abnormal presence of three-nuclei asci. At 6 days postfertilization, wild-type asci typically contain eight nuclei, each enclosed by a forespore membrane, leading to the formation of eight defined ascospores. Normally, after this delimitation, the nucleus within each ascospore remains undivided until maturation begins. During maturation, these initially unicellular ascospores undergo further mitosis and septation, developing into multicellular ascospores (40). Consequently, asci with more than eight nuclei are generally not observed in wild-type strains at 6 days postfertilization (Fig. 1C). In contrast, over 64% of Fgspo11 mutant asci at 6 days postfertilization contained nine or more nuclei. In addition, many delimited unicellular ascospores in the mutant harbored two or more newly divided or dividing nuclei (Fig. 1C). This indicates an abnormal activation of postmeiotic mitoses due to the absence of FgSPO11. Further abnormalities in the mutant included fewer than eight delimited ascospores per ascus and some nuclei that were not enclosed by forespore membranes. These findings highlight that FgSPO11 is essential for ascospore delimitation and for coordinating nuclear divisions with septum formation during ascosporogenesis. Consistent with these observations, FgSPO11 expression was elevated during ascosporogenesis (5 to 8 days postfertilization). Specifically, transcript levels were low during meiosis (4 days postfertilization) but progressively increased after meiosis (Fig. 1D). This expression profile is highly reproducible across independent studies (41) (fig. S1), although slight variations in the absolute timing of sexual development may occur depending on laboratory-specific conditions. Together, these findings demonstrate that FgSPO11 is indispensable for both proper meiotic progression and the regulation of postmeiotic nuclear division during ascosporogenesis in F. graminearum. FgSpo11’s catalytic activity differentially regulates meiotic and postmeiotic nuclear divisions Besides initiating DSBs to trigger sexual recombination, Spo11 is also known to facilitate homolog pairing and synapsis during meiosis (Fig. 2A). Previous studies have demonstrated that substituting a conserved catalytic tyrosine (Y) residue with phenylalanine (F) in Spo11 abolishes DSB formation but does not affect homolog pairing that occurs independently of DSBs (42–45). Role of FgSpo11 DSB-forming activity in ascospore formation. (A) Schematic of Spo11 function in meiosis I, showing its roles in DSB formation for homologous recombination and DSB-independent homolog pairing and synapsis. (B) Multiple sequence alignment of Spo11 homologs highlighting the conserved tyrosine residue (Y129 in FgSpo11, marked by a red triangle). Mutation of this residue to phenylalanine (Y-to-F) blocks DSB formation but retains homolog pairing and synapsis functions. Sc, Saccharomyces cerevisiae; Fg, F. graminearum; At, Arabidopsis thaliana; Mus, mouse. (C) Colony morphology and ascospore production in PH-1 (wild-type) and FgSPO11Y129F mutant on PDA (3 days) and mating cultures (8 days postfertilization). Blue triangle, asci with wild-type eight ascospores; red triangle, asci with few or no ascospores. Scale bar, 20 μm. Stacked bar chart indicates the percentage of asci with different ascospore numbers at 8 days postfertilization (n = 3, 30 perithecia per replicate). The red asterisk (*) on the stacked bar chart indicates a statistically significant difference (P < 0.05) in the proportion of normal asci with eight ascospores between the indicated strains, determined by Fisher’s exact test. Because abnormal asci were divided into multiple categories by ascospore number, only normal asci were tested to simplify the analysis. (D) DAPI-stained asci at 6 days postfertilization observed by epifluorescence microscopy. White dashed lines, ascospore boundaries. Scale bar, 10 μm. Bar graph indicates the percentage of asci with specific nuclear numbers at 6 days postfertilization (n = 3, 20 perithecia per replicate). Statistical comparisons were performed separately for each category using one-way analysis of variance (ANOVA), followed by Tukey’s test (P < 0.05). Different letters indicate significant differences among the three strains within each nuclear number category. To pinpoint which specific function of FgSPO11 is responsible for the observed defects during sexual development, we identified the conserved catalytic residue Y129 in FgSpo11 (Fig. 2B) and generated an FgSPO11Y129F mutant strain. This FgSPO11Y129F mutant exhibited severe defects in ascospore formation (Fig. 2C), remarkably similar to those seen in the Fgspo11 deletion mutant. Abnormalities in ascospore delimitation and the abnormal activation of mitotic nuclear division during ascosporogenesis were observed in the FgSPO11Y129F mutant, mirroring the Fgspo11 mutant. However, defects in meiotic progression were not observed in the FgSPO11Y129F mutant (Fig. 2D). This suggests that FgSpo11’s role in ascospore delimitation and suppressing mitotic nuclear division during ascosporogenesis relies on its DSB-forming catalytic activity, whereas its role in meiotic progression does not. Genome-wide analysis reveals FgSpo11-dependent and -independent meiotic recombination pathways To determine the meiotic recombination landscape in F. graminearum, we performed whole-genome sequencing of the wild-type parental strains and 10 randomly selected single-ascospore isolates (intraspecific F1 hybrids) derived from a cross between F. graminearum strains PH-1 and Gz3639 (Fig. 3A). To confirm the role of FgSpo11 in meiotic recombination, we generated an Fgspo11 mutant in the Gz3639 background and conducted a similar analysis for an outcross involving the Fgspo11 mutant. Using the single-nucleotide polymorphism (SNP) loci between PH-1 and Gz3639 as markers, we developed a bioinformatic pipeline to accurately identify crossover events in the F1 hybrids. Critical role of FgSPO11 in meiotic recombination. (A) Schematic of the experimental workflow for generating single-ascospore isolates (intraspecific F1 hybrids) from a cross between F. graminearum strains PH-1 and Gz3639 for DNA sequencing (DNA-seq). CA, carrot agar; d, days. (B) Homologous crossover patterns across the four chromosomes (Chr1 to Chr4) in representative single-ascospore isolates from outcrosses involving wild-type and Fgspo11 mutant strains. Blue and orange indicate chromosomal regions inherited from Gz3639 and PH-1, respectively. (C) Number of crossover events detected in each across. Box plots show the median (center line), 25th and 75th percentiles (box), and 1.5× interquartile range (whiskers). P value is from two-tailed t test. The number of crossover events detected per meiosis on the four chromosomes of the wild-type strain ranged from 18 to 27, with a median of 22.5 (Fig. 3, B and C). In contrast, only a median of three crossover events (ranging from 0 to 6) was detected in the Fgspo11 mutant. The substantial reduction in homologous crossover events in the Fgspo11 mutant underscores the critical role of FgSpo11 in initiating meiotic recombination. However, the presence of residual crossover events in the Fgspo11 mutant suggests the existence of an FgSpo11-independent pathway for the initiation of meiotic recombination. PSC editing is essential for FgSPO11 function in sexual reproduction To elucidate the biological significance of PSC editing, we generated an uneditable mutant FgSPO11TAA expressing only the unedited isoform of FgSpo11, by replacing the PSC TAG with TAA in situ in the PH-1 background (Fig. 4A). For the N32D missense editing site, we similarly created an uneditable mutant FgSPO11N32 by mutating the editing-preferred −1 base T to the unpreferred base G in the PH-1 genome without altering the amino acid sequence (Fig. 4A). Reverse transcription polymerase chain reaction (RT-PCR) product sequencing revealed that postediting “G” peaks were predominantly detected at both the PSC and N32D sites in PH-1 Sanger sequencing traces, whereas only “A” peaks were observed in both uneditable mutants (Fig. 4B). These results confirm that RNA editing at both sites was successfully abolished in the uneditable mutants. Essential role of PSC editing in FgSPO11 function. (A) Schematic representation of wild-type, uneditable (ue), and postediting (pe) alleles at the N32D and PSC (*210W) editing sites. Editing and mutated bases are marked in red and green, respectively. Expected transcript isoforms expressed during sexual and asexual stages by wild-type, uneditable FgSPO11TAA, and postediting FgSPO11TGG mutants are shown. (B) Sequencing traces for the regions flanking the N32D and PSC editing sites amplified from cDNA of 7-days postfertilization perithecia of PH-1 (wild-type), uneditable, and postediting mutants. Red lines and black triangles indicate editing and mutated sites, respectively. (C) Colony morphology and ascospore production in PH-1 (wild-type) and the indicated mutants on PDA (3 days) and mating cultures (8 days postfertilization). White triangle, asci with wild-type eight ascospores; red triangle, asci with few ascospores. Scale bar, 20 μm. Stacked bar chart indicates the percentage of asci with different ascospore numbers at 8 days postfertilization (n = 3, 30 perithecia per replicate). (D) DAPI-stained asci at 6 days postfertilization visualized by epifluorescence microscopy. White dashed lines, ascospore boundaries; red triangles, ≥2 nuclei enclosed by forespore membrane. Scale bar, 10 μm. Stacked bar chart indicates the percentage of asci with specific nuclear numbers at 6 days postfertilization (n = 3, 20 perithecia per replicate). Asterisk (*) on the stacked bar chart indicates a statistically significant difference (P < 0.05) in the proportion of asci with a given number of nuclei compared to wild-type, determined by Fisher’s exact test. Only asci with one to three nuclei were tested to simplify the analysis. (E) Western blots of total proteins extracted from hyphae of each indicated strain, shown for two independent experiments (first and second). Proteins were probed with anti–green fluorescent protein (GFP) and anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies, with GAPDH serving as the loading control. The expected sizes of the GFP-fused full-length and truncated proteins, as well as GFP alone, are indicated. Phenotypic analysis showed that the FgSPO11N32 mutant did not display any defects, while the FgSPO11TAA mutant exhibited ascospore formation defects similar to those observed in the Fgspo11 mutant (Fig. 4C). Furthermore, both the FgSPO11TAA and Fgspo11 mutants exhibited similar defective patterns in meiotic and postmeiotic nuclear divisions within asci (Fig. 4D). Quantitative RT-PCR analysis indicated that the expression level of the FgSPO11TAA allele was comparable to that of the wild-type allele (fig. S3), indicating that the observed defects are not due to altered expression. Collectively, these results demonstrate that PSC editing, but not N32D missense editing, is crucial for the biological function of FgSPO11 in sexual reproduction. To further investigate the functional consequence of PSC editing, we generated a postediting mutant that exclusively expresses the edited isoform of FgSpo11 by directly mutating the edited A to G at the PSC editing site in the genomic DNA (Fig. 4A). The resulting FgSPO11TGG mutant displayed normal ascosporogenesis (Fig. 4C), thereby demonstrating that the edited isoform is necessary for normal ascospore formation. The unedited truncated FgSpo11 isoform is nonfunctional because of protein instability and loss of FgSki8 interaction We performed Western blot analyses using C-terminal green fluorescent protein (GFP)–tagged constructs to track the production of both protein isoforms. While the full-length protein expressed from the postediting allele (FgSPO11TGG-GFP) was successfully detected as a stable band in vegetative hyphae, the truncated allele construct (FgSPO11PSC-GFP) yielded only a free GFP signal (Fig. 4E). The absence of the predicted truncated FgSpo11 protein fragment suggests that the unedited transcript produces an unstable or rapidly degraded protein that is unlikely to persist long enough to exert a regulatory effect. Spo11 is known to directly interact with Ski8, an interaction essential for DSB formation during meiosis (46). AlphaFold-based protein-protein interaction predictions revealed that the C-terminal region of FgSpo11 mediates its interaction with FgSki8 (Fig. 5A). Notably, the unedited isoforms of FgSpo11 lack this C-terminal region and, consequently, are unable to interact with FgSki8. Yeast two-hybrid assays confirmed that only the edited isoform of FgSpo11 interacts with FgSki8, while the unedited isoform does not (Fig. 5B). These findings demonstrate that only the edited FgSpo11 is functional in mediating meiotic recombination. Resolving tradeoffs through PSC editing of FgSPO11. (A) AlphaFold3 structure prediction of the protein-protein interaction between FgSki8 and FgSpo11TGG but not FgSpo11TAA. (B) Yeast two-hybrid assays showing interaction between FgSki8 and FgSpo11TGG or FgSpo11TAA. Plates show growth on SD-His-Leu-Trp medium at indicated concentrations, along with Mel1 α-galactosidase activity. (C) DAPI-stained asci at 6 days postfertilization visualized by epifluorescence microscopy. White dashed lines, ascospore boundaries. Scale bar, 10 μm. Stacked bar chart indicates the percentage of asci with specific nuclear numbers at 6 days postfertilization (n = 3, 20 perithecia per replicate). The asterisk (*) on the stacked bar chart indicates a statistically significant difference (P < 0.05) in the proportion of asci with a given number of nuclei compared to wild-type, determined by Fisher’s exact test. Only asci with one to three nuclei were tested to simplify the analysis. (D) Line chart showing editing levels at the PSC site during sexual development from 1 to 8 days postfertilization. (E) RNA-seq expression levels (TPM) of FgSPO11 during asexual stages, including conidia (Coni), 12-hour conidial germlings (Ger12h), 24-hour vegetative hyphae (Hyp24h), 3-dpi infection hyphae (Inf3d), and 3-day-old DON-producing hyphae in trichothecene biosynthesis induction (TBI) culture (DON3d). (F) DAPI-stained hyphal tips of 4-hour conidial germlings visualized by epifluorescence microscopy. White triangle, mitotic nucleus; red triangle, hyphal septum. Bar graph indicates the number of nuclei per tip cell (n = 3, 30 tip cells per replicate). (G) Colony growth of indicated strains on PDA plates with or without camptothecin (CPT; 25 nM) after 3 days of incubation. Bar graph indicates the growth inhibition rate (%) by CPT (n = 3). (H) Growth length of indicated strains in race tubes with PDA cultures over 20 days (n = 3). For (F) to (H), different letters indicate significant differences (one-way ANOVA, Tukey’s test, P < 0.05). dpf, days postfertilization. Furthermore, we investigated whether the unedited isoform exerts any dominant-negative effects. We generated a strain, d-FgSPO11TAA, by inserting an additional copy of the preediting FgSPO11TAA allele into the FgSPO11TAA background. Quantitative RT-PCR analysis confirmed that FgSPO11 expression levels in the d-FgSPO11TAA strain were twofold higher than those in both the wild-type and FgSPO11TAA strains (fig. S3). The d-FgSPO11TAA strain exhibited phenotypes identical to the single-copy FgSPO11TAA mutant (equivalent to the Fgspo11 deletion) (fig. S3), indicating that the unedited isoform lacks dominant-negative activity and likely represents a null allele due to protein instability. PSC editing regulates FgSpo11 abundance to optimize the pace of meiosis I These results raise the intriguing question of why PSC editing is necessary to produce the full-length, functional protein, rather than encoding it directly in the genome. During sexual development, we found that the PSC editing level is highly dynamic. While editing levels reach 25 to 81% during late ascosporogenesis (5 to 8 days postfertilization), editing was undetectable at 1 to 4 days postfertilization in our initial RNA-seq dataset, corresponding to the period before or at the onset of meiosis (Fig. 5D). However, our reanalysis of published RNA-seq data from an independent study (41) revealed editing levels of 44 and 51% at four- and five-stage, respectively (fig. S1). This discrepancy likely stems from variations in the timing of sexual development under different laboratory conditions. To achieve more precise quantification than standard RNA-seq, we performed targeted amplicon sequencing (Hi-TOM) from 4 to 7 days postfertilization (fig. S1). This analysis yielded a similar editing profile and confirmed that PSC editing occurs at 4 days postfertilization—coinciding with meiosis I—albeit at a relatively low level (∼5%). Because only a small fraction of mRNA is edited during meiosis I, the total abundance of full-length FgSpo11 protein in the wild type remains strictly limited. In contrast, in the FgSPO11TGG mutant, 100% of the transcripts inherently carry the sense codon, leading to a substantially higher abundance of full-length FgSpo11 protein compared to the wild type. Notably, the FgSPO11TGG mutant exhibited an increased frequency of asci containing one, two, or three nuclei, despite maintaining normal postmeiotic nuclear divisions (Fig. 5C). These results suggest that excessive FgSpo11 delays progression through meiosis I, resulting in an accumulation of one- and two-nuclei asci. This is notably consistent with the accelerated progression through meiosis I observed in the Fgspo11 deletion mutant. It appears that any deviation from the optimal pace—whether accelerated or delayed—disrupts meiosis II, ultimately leading to the abnormal formation of three-nuclei asci. To further substantiate the regulatory role of FgSpo11 in meiosis I progression, we performed a dosage-dependent experiment. We generated a double-copy strain, d-FgSPO11TGG, containing two copies of the postediting FgSPO11TGG allele. Quantitative RT-PCR analysis confirmed that FgSPO11 expression levels in the d-FgSPO11TGG strain were twofold higher than those in both the wild-type and FgSPO11TGG strains (fig. S3). While the Fgspo11 deletion mutant exhibits accelerated progression, this overexpressing strain showed a significant increase in the frequency of one- and three-nuclei asci relative to the FgSPO11TGG single-copy strain (fig. S3). The further accumulation of one-nucleus asci indicates that increased FgSpo11 dosage imposes a more stringent delay on the entry into or progression through meiosis I. Furthermore, the rise in three-nuclei asci—coupled with the scarcity of the two-nuclei stage—suggests that the few cells successfully completing meiosis I undergo highly asynchronous or defective divisions during meiosis II. This dose-response relationship strongly supports the hypothesis that FgSpo11 plays a functional role in restraining the pace of meiosis I. Therefore, by strictly regulating the timing and abundance of FgSpo11 production, PSC editing ensures proper progression through meiosis I while supporting regular postmeiotic nuclear divisions during ascosporogenesis. PSC editing restricts FgSpo11 activity to prevent vegetative cytotoxicity Although SPO11 expression is generally considered meiosis specific, we detected low but consistent expression levels during nonsexual stages in both F. graminearum and N. crassa (Fig. 5E and fig. S4). We hypothesized that the production of full-length FgSpo11 protein during vegetative growth could be detrimental to mitosis, thereby reducing overall fitness. To test this, we examined mitotic nuclear divisions in the postediting mutant FgSPO11TGG, comparing it with the wild type and the uneditable mutant FgSPO11TAA. We found that the number of nuclei at the hyphal tips was significantly reduced in the FgSPO11TGG mutant (Fig. 5F). This provides direct evidence that the presence of full-length FgSpo11 is cytotoxic at the cellular level. To determine whether this toxicity is dependent on DSB formation, we generated and analyzed a postediting FgSPO11Y129F-TGG catalytic-dead mutant (fig. S4). The absence of nuclear reduction in this mutant confirms that the observed toxicity stems directly from the protein’s DNA-cleaving activity. Since the FgSPO11TGG mutant exhibited no significant defects in vegetative growth, plant infection, conidiation, or stress sensitivity under standard conditions (Fig. 5G, and figs. S2 and S5), we performed competition assays to detect more subtle fitness costs. After mixing conidia of the wild type and the FgSPO11TGG mutant at a 1:1 ratio and passaging them for five generations in 1/2 minimal medium (MM), the proportion of the FgSPO11TGG strain was recorded at average 49.2% (fig. S5). While this value is marginally lower than the initial 50%, the difference did not reach statistical significance (P = 0.6234, one-sample t test). We propose that although this fitness cost appears negligible under stable and nutrient-controlled environments, the cumulative effect of these minor postediting-related cytotoxicities could manifest as a disadvantage over longer evolutionary timescales or under the fluctuating and challenging conditions of natural habitats. We speculate that most DSBs generated by the FgSPO11TGG allele during asexual stages are efficiently repaired via homologous recombination, resulting in only minor fitness costs that require long-term selection to become apparent. To test this, we deleted the key recombinase gene FgRAD51 in both postediting and uneditable backgrounds. While the Fgrad51 and FgSPO11TAA rad51 mutants exhibited normal growth, the FgSPO11TGG rad51 mutant showed significantly reduced growth rates in race tube assays and increased sensitivity to genotoxic stress induced by camptothecin (CPT) (Fig. 5, G and H, and fig. S6). This “synthetic” phenotype confirms that the cellular-level cytotoxicity of FgSpo11 requires an active recombination repair pathway to prevent macroscale growth inhibition. Thus, postediting appears to confer an evolutionary advantage by restricting the production of functional FgSpo11 to sexual stages, thereby shielding the genome from harmful DNA cleavage during mitosis. Evolutionary dynamics and multiple independent origins of PSC editing in fungal SPO11 SPO11 orthologs are generally conserved across ascomycetes, with the vast majority of species harboring a single copy of the gene, except for a few species with multiple copies (Fig. 6A and data S1). Notably, SPO11 homologs were not detected in several species belonging to Saccharomycotina and Eurotiomycetes. Evolutionary dynamics of the PSC editing site in fungal SPO11. (A) Dendrogram of various fungal taxa showing the distribution of SPO11 genes in Ascomycota. Each pie chart indicates the proportion of species with one (gray), two or more (blue), or no (orange) copies of SPO11 orthologs. The number of species analyzed in each taxon is shown in parentheses. (B and C) Sequence logos, predicted RNA secondary structure percentages, and amino acid states at the PSC editing site in SPO11 orthologs across different Fusarium species complexes (B) and Sordariomycetes taxa (C). Each pie chart shows the proportion of species with a PSC (TAG, *) or a postediting residue (W) at the PSC site. Red triangles indicate the PSC editing site. The number of species examined in each group is indicated in parentheses. NA, not available. Evolutionary analysis revealed that the preediting TAG codon is conserved within the Fusarium sambucinum species complex (47), to which F. graminearum belongs (Fig. 6B). This TAG codon is also conserved in the three most closely related species complexes—Chlamydosporum, Incarnatum-equiseti, and Tricinctum—but not in Heterosporum, where the postediting TGG codon is conserved (Fig. 6B). The presence of the TGG (W) codon at the PSC editing site in most Fusarium species complexes suggests that the postediting codon represents the ancestral state, while the preediting TAG codon likely arose in the common ancestor of these five species complexes and was subsequently lost in Heterosporum. Notably, the preediting TAG codon was also identified in two distantly related Fusarium species complexes (Fig. 6B), indicating independent origins of the PSC TAG in different Fusarium lineages. Given that A-to-I mRNA editing is a conserved feature in Sordariomycetes, we analyzed the evolutionary conservation of the PSC editing site in SPO11 orthologs. The postediting TGG (W) codon is commonly conserved across different lineages of Sordariomycetes (Fig. 6C), supporting its status as the ancestral state. Beyond Fusarium, the preediting TAG codon was also detected in species outside Fusarium within Nectriaceae, as well as in individual species from Stachybotryaceae (the sister group of Nectriaceae) and in two distantly related orders, Microascales and Glomerellales (Fig. 6C). This pattern suggests independent origins of the preediting TAG codon in multiple Sordariomycetes lineages. Because RNA stem-loop structures and the sequence context from positions −2 to +4 (especially at the −1 position) are critical for A-to-I editing (23), we analyzed the RNA secondary structure and sequence conservation surrounding the PSC editing site in SPO11 orthologs. While the sequence motifs required for A-to-I editing are relatively conserved in Sordariomycetes, a higher proportion of species with predicted stem-loop structures was observed only in specific lineages, including Nectriaceae, Stachybotryaceae, and Glomerellales (Fig. 6C). This suggests independent origins of the stem-loop structure in these lineages. Notably, stem-loop structures were predicted in almost all species harboring the preediting TAG codon (Fig. 7), implying the potential for PSC editing in these species. In two fungal beetle symbionts, Ambrosiella beaveri and Meredithiella fracta (Microascales) (48, 49), the TAG codon was not predicted to be within a stem-loop structure at 25°C, the optimal temperature for most fungi. However, at elevated temperatures (29°C for A. beaveri and 36°C for M. fracta), the TAG codon was predicted to reside within a stem-loop structure (Fig. 7), suggesting that PSC editing of the SPO11 ortholog in these beetle symbionts may be temperature dependent. Predicted RNA secondary structures at the PSC editing site in SPO11 orthologs of representative fungal species. The edited adenosine (A) and its flanking 30 nucleotides upstream and downstream from various fungal taxa were used for RNA secondary structure prediction under different temperature parameters. Each nucleotide is rainbow color coded by base-pairing probability, with red indicating the highest probability and violet indicating the lowest. Red triangles mark the PSC editing site. In Fusarium, many species have a stem-loop structure at the PSC editing site but have the postediting TGG (W) codon encoded in the genome (Fig. 6). According to the phylogenetic tree of Fusarium species complexes, the most parsimonious explanation is that the stem-loop structure—and thus the capacity for PSC editing—originated in the Fusarium ancestor. Subsequently, the postediting TGG codon was frequently fixed in the genome, while the preediting TAG codon was reacquired during later evolutionary events. Since the efficiency of PSC editing is not 100%, the balance between the preediting TAG codon and the postediting TGG codon across different evolutionary times and lineages likely reflects the context-dependent adaptive advantage of PSC editing, shaped by trade-offs between vegetative growth and sexual reproduction. Comparative analysis of species both with and without PSC editing (i.e., those naturally carrying the TGG codon) confirms that transcriptional up-regulation during the sexual stage is a conserved ancestral feature. However, transcriptional regulation is rarely “absolute,” and trace levels of mRNA can often be detected during vegetative growth (fig. S7). Our findings support the model that RNA editing does not replace transcriptional control but rather functions as an additional layer of regulation. Given that even trace amounts of FgSpo11 are cytotoxic—as evidenced by the impaired mitotic divisions at hyphal tips—PSC editing represents an elegant “add-on” to the ancestral transcriptional program. We propose that this postediting mechanism likely evolved as a secondary fail-safe to eliminate the burden of residual protein, ensuring genomic integrity during nonsexual stages. DISCUSSION Meiosis is essential for sexual reproduction and must be precisely coordinated to balance genetic diversification with genome integrity. In the cereal pathogen F. graminearum, our work reveals that this balance is achieved not only through canonical control of meiotic recombination but also through an RNA editing–dependent gate that constrains the activity of the conserved DSB initiator Spo11 to the correct developmental window. By integrating corrected gene annotation, allele-specific genetics, cytology, and genome-wide crossover mapping, we identify FgSpo11 as (i) a dosage-sensitive, DSB-independent regulator of meiosis I pace and (ii) a DSB-dependent suppressor of supernumerary postmeiotic mitoses during ascosporogenesis, while demonstrating that PSC editing is indispensable for producing functional FgSpo11 at the sexual stage (Fig. 8). These findings connect stage-specific RNA editing to nuclear division fidelity and illuminate how epitranscriptomic regulation can resolve life cycle trade-offs in filamentous fungi. Proposed models for FgSPO11 function and the adaptive significance of its PSC editing. (A) Model summarizing FgSPO11 roles and PSC editing during vegetative growth and sexual development. During sexual development, FgSPO11 performs its canonical meiotic recombination function, delays meiosis I progression, and suppresses excessive nuclear divisions in postmeiotic mitosis during ascosporogenesis. During vegetative growth, constitutive full-length FgSpo11 reduces mitotic nuclear divisions. Thus, FgSPO11 imposes a dose-dependent trade-off among vegetative mitosis, meiosis I progression, and postmeiotic mitoses. A genomically encoded G at the PSC site (full-length protein) benefits ascosporogenesis but is deleterious during vegetative growth and meiosis I due to excessive FgSpo11 activity. PSC editing enables spatiotemporal control of functional FgSpo11 dosage, thereby minimizing harmful effects during vegetative mitosis and ensuring proper timing of meiosis I. Asci with defective nuclear divisions and abnormal ascospores are shown in blue. (B) Model illustrating how PSC editing balances FgSpo11 function across vegetative mitosis, meiosis I, and postmeiotic mitosis. FgSpo11-mediated DSB formation is required for meiotic recombination and for proper regulation of postmeiotic mitoses during ascosporogenesis, whereas leaky DSB activity during vegetative growth can compromise genome stability. In addition, an appropriate DSB-independent FgSpo11 level is needed to pace meiosis I. The evolution of an editable PSC (TAG), together with sexual stage–specific PSC editing and dynamically regulated editing levels, allows fungi to (i) support postmeiotic mitoses during ascosporogenesis, (ii) prevent harmful DSB formation during vegetative mitosis, and (iii) fine-tune functional FgSpo11 dosage throughout meiosis. Together, these models support a unifying view that FgSpo11 functions primarily as a stage-recruited developmental brake: At the molecular level, its functions manifest as inhibitory control, deployed in different phases to satisfy distinct developmental demands—pacing meiosis I and suppressing supernumerary postmeiotic nuclear divisions—while PSC editing tunes the timing and dosage of functional FgSpo11. Spo11 is widely conserved and best known for catalyzing programmed meiotic DSBs to initiate homologous recombination and is essential for meiosis in plants, animals, and many fungi (42–44). Consistent with this conserved role, genome-wide analysis of meiotic recombination in F. graminearum showed a marked decrease in crossovers in the Fgspo11 mutant, confirming that FgSpo11 is the primary driver of homologous crossover formation in this species. At the same time, residual crossovers persisted in the absence of FgSpo11, suggesting that F. graminearum retains a Spo11-independent mechanism capable of initiating a limited amount of recombination, as has been reported in other eukaryotes (50–52). Functionally, however, this residual activity is insufficient to sustain normal ascosporogenesis, underscoring that robust FgSpo11-dependent recombination is tightly coupled to successful sexual development in this pathogen. A key advance in this study is that it rests on an accurate representation of the FgSPO11 transcript. Full-length transcript evidence corrected misspliced introns in prior computational gene models and uncovered a genomically encoded PSC in exon 5. This unexpected feature led us to identify a high-efficiency PSC editing event that converts TAG to TGG during sexual development, thereby restoring a full-length FgSpo11 protein containing the C-terminal TOP6A-related region required for catalytic function and partner interactions. This emphasizes that developmental-stage transcript structures can mask essential regulatory logic when gene models are inferred solely from prediction pipelines and that accurate annotation can be prerequisite for discovering biologically central RNA-level control (38). Beyond its canonical recombination role, our results separate two distinct functions of FgSpo11 across sexual development. First, FgSpo11 restrains the pace of meiosis I independently of its catalytic DSB-forming activity. The Fgspo11 deletion mutant displayed an apparent acceleration through early meiotic stages, accompanied by characteristic downstream defects during meiosis II manifested as abnormal three-nuclei asci. In contrast, the catalytic-dead FgSPO11Y129F mutant progressed through meiosis without the meiosis I acceleration phenotype, indicating that the “brake” function does not require Spo11-catalyzed DSBs, consistent with prior observations that Spo11 can influence early meiotic progression independently of its DSB activity in budding yeast (44). This timing control is quantitatively tuned by the amount of full-length FgSpo11 available. Forcing constitutive restoration of the edited codon (FgSPO11TGG) increased full-length FgSpo11 abundance and delayed meiosis I, and increasing dosage further strengthened this delay, supporting a model in which FgSpo11 abundance sets the timing constraints of meiosis I. Notably, both extremes—too little Spo11 (deletion) and too much full-length Spo11 (constitutive restoration and overdosage)—converged on a common outcome of meiosis II failure, suggesting that the system requires an optimal Spo11 “set point” for orderly progression through the meiotic divisions. Second, we uncovered an unexpected DSB-dependent role of FgSpo11 in suppressing supernumerary postmeiotic mitoses during ascosporogenesis and in coordinating nuclear division with ascospore delimitation. In wild type, asci at 6 days postfertilization predominantly contained eight nuclei consistent with completion of meiosis, followed by a single postmeiotic mitosis, and nuclei remained undivided within newly delimited unicellular ascospores until later maturation. In the Fgspo11 mutant, however, most asci contained nine or more nuclei at 6 days postfertilization, and delimited ascospores frequently harbored multiple newly divided or dividing nuclei, indicating ectopic activation of postmeiotic mitoses. Crucially, these postmeiotic defects were recapitulated by FgSPO11Y129F, demonstrating that suppression of extra postmeiotic nuclear divisions and proper delimitation require FgSpo11 catalytic activity, unlike meiosis I pacing control. Consistent with these stage-specific demands, FgSPO11 transcript abundance rose after meiosis and peaked during ascosporogenesis (5 to 8 days postfertilization), and a similar late-stage expression peak has been observed for NcSPO11 in N. crassa (53). Together, these results expand Spo11’s functional repertoire and underscore its contribution to safeguarding reproductive fidelity in filamentous ascomycetes. The PSC editing event in FgSPO11 is central to both functions because it is required to produce stable, full-length FgSpo11. Editing at the TAG site was abolished in editing-defective backgrounds, consistent with dependence on the sexual-stage Tad2–Tad3–Ame1 machinery (18, 19), and functional tests established its necessity and sufficiency. The uneditable FgSPO11TAA mutant phenocopied Fgspo11 deletion for meiotic/postmeiotic nuclear defects and ascospore formation failure, whereas the postediting FgSPO11TGG allele supported normal ascosporogenesis, demonstrating that the edited isoform is the functional form required in vivo. In contrast, abolishing the missense N32D editing event had no detectable phenotype, sharpening the conclusion that PSC editing is the critical functional editing event at this locus. Mechanistically, the unedited isoform behaves as a null allele. Western blotting using C-terminal GFP fusions detected stable full-length FgSpo11 from the postediting allele but yielded only free GFP from the preediting construct, indicating that the truncated product is unstable or rapidly degraded and thus unlikely to accumulate as a functional or regulatory protein. Consistent with this, only the edited isoform interacted with FgSki8 in yeast two-hybrid assays, matching structural predictions and the conserved requirement of Ski8 for Spo11-dependent DSB formation (46). Moreover, increasing the dosage of the uneditable preediting allele did not elicit dominant-negative effects, reinforcing that the unedited transcript does not produce a stable inhibitory protein and that PSC editing functions primarily as an on-demand production mechanism for the active Spo11 isoform. These results also provide a mechanistic answer to a long-standing question in RNA biology: Why evolution may favor RNA editing over direct DNA changes (27, 28). If the genome encoded the fully functional Spo11 constitutively (TGG), then low basal transcription outside sexual development could translate into ectopic DSB activity during vegetative growth. Despite FgSPO11 being strongly up-regulated in sexual tissues, we detected low but consistent expression during nonsexual stages, consistent with incomplete transcriptional silencing. In this context, PSC editing supplies a second regulatory layer that prevents “promoter leakage” from becoming protein leakage (54–56). Direct evidence for this protective role comes from the postediting FgSPO11TGG mutant, in which full-length FgSpo11 production reduced nuclear number at hyphal tips, and this cytotoxicity was abolished by the catalytic-dead Y129F substitution, proving that ectopic toxicity depends on Spo11-mediated DNA cleavage. While macroscopic vegetative defects of FgSPO11TGG were modest under standard laboratory conditions, genetic interaction tests exposed a strong dependence on homologous recombination repair: Deleting FgRAD51 in the FgSPO11TGG background caused marked growth defects and CPT hypersensitivity, whereas FgRAD51 loss in the uneditable background did not. Thus, PSC editing restricts functional FgSpo11 to the sexual program and buffers vegetative cells from genotoxic DSB formation, consistent with the broader principle that misexpression of meiotic machinery in mitotic contexts promotes genome instability (57–59). In addition, the highly dynamic editing profile provides a mechanism to tune full-length FgSpo11 abundance during sexual development itself, keeping full-length levels low at meiosis I, when only a small fraction of transcripts are edited while allowing high levels later in ascosporogenesis, when editing efficiency increases. This temporal logic matches the dual roles uncovered here: limiting FgSpo11 dosage to set the correct pace of meiosis I while enabling sufficient catalytic activity later to suppress inappropriate postmeiotic mitoses and coordinate delimitation. Our phylogenetic analyses further support an adaptive role for PSC editing as a tunable regulatory module rather than a passive correction. Across Sordariomycetes, the conserved tryptophan position is most often genomically encoded as TGG, consistent with an ancestral functional codon, yet a PSC-editable TAG appears repeatedly in multiple lineages, including several Fusarium species complexes and more distantly related taxa. The enrichment of predicted local RNA stem-loop structures in lineages harboring the TAG codon supports independent acquisition of editing competence, consistent with the known importance of local sequence/structure context for fungal A-to-I editing (23, 24). In certain beetle-associated fungi, predicted temperature-dependent formation of the relevant stem-loop raises the possibility that environmental conditions may modulate editing efficiency (48, 49). Together, these patterns are consistent with recurrent gain and loss of an editable stop codon under context-dependent selection, where layered regulation—sexual-stage transcriptional up-regulation plus editing-based gating—can be favored when vegetative phases are prolonged and the cost of ectopic Spo11 activity is high. This view is also consistent with broader discussions of how restorative A-to-I editing can arise and persist under selection (25, 60, 61). Because F. graminearum relies on sexual reproduction and airborne ascospores for disease initiation and because recombination fuels population diversity and facilitates the emergence of fungicide resistance, mechanisms that couple recombination benefits to stringent genome-protection constraints are likely to shape pathogen evolution (8–14). PSC editing provides a means to retain the adaptive value of high meiotic recombination while limiting genome destabilization during clonal expansion, suggesting that perturbing this gate could, in principle, alter evolutionary trajectories by reducing recombination-driven diversification or by imposing vegetative fitness costs when active FgSpo11 is ectopically produced. More broadly, our findings refine the functional scope of “restorative” A-to-I editing by demonstrating that, at least in this case, it operates as a developmental gate that determines when a potentially harmful yet indispensable activity can be produced and as a quantitative lever that tunes protein dosage to match stage-specific requirements. This conclusion is consistent with prior observations that PSC editing is widespread in F. graminearum and can be required for sexual development while imposing costs when full-length isoforms are constitutively expressed (25), and it complements emerging examples of PSC editing implicated in fungal reproductive conflicts such as meiotic drive (62, 63). In sum, FgSPO11 PSC editing is essential for sexual reproduction in F. graminearum because it enables production of stable, full-length FgSpo11 while restricting its catalytic potential to the appropriate developmental context and dosage. By revealing separable DSB-independent and -dependent roles of FgSpo11 in meiotic timing and postmeiotic nuclear control, respectively, this work expands Spo11 biology and establishes a clear mechanistic and evolutionary rationale for how an organism can maintain a genomically “compromised” allele that is conditionally rescued at the RNA level to resolve the trade-off between meiotic necessity and vegetative cytotoxicity. MATERIALS AND METHODS Strain culture conditions We routinely cultured the wild-type strains of F. graminearum PH-1 and Gz3639 (5), along with their derived mutants and transformants, on potato dextrose agar (PDA) plates containing 20% potato, 2% dextrose, and 1.5% agar at 25°C. Colony growth rate and morphology were assessed on PDA plates after 3 days of incubation. For more detailed growth rate measurements, we monitored fungal growth in race tubes over a 20-day period. To test strain sensitivity to various stresses, we supplemented PDA plates with one of the following agents and incubated the strains for 3 days: 0.7 M NaCl, 0.01% SDS, Congo Red (150 μg/ml), 1 M sorbitol, 0.05% methyl methanesulfonate, 25 or 50 nM CPT, 0.02% hydrogen peroxide (H2O2), or 1 mM hydroxyurea. We harvested conidia from 5-day-old liquid cultures grown in carboxymethyl cellulose (CMC) medium (1.5% carboxymethylcellulose, 0.1% NH4NO3, 0.1% KH2PO4, 0.05% MgSO4·7H2O, and 0.1% yeast extract). To obtain apical hyphal cells, we cultured conidia in liquid YEPD medium (0.3% yeast extract, 1% peptone, and 2% dextrose) for 4 hours. For sexual reproduction assays, we gently pressed down the aerial mycelia of 7-day-old cultures grown on carrot agar plates [1.5% agar and carrot (250 g/liter)] using 1 ml of sterile 0.1% Tween 20 and then incubated the plates under black light at 25°C. We examined perithecium formation, cirrhi production, and spore discharge using an Olympus SZX16 stereomicroscope. The morphology of conidia, apical hyphal cells, asci, and ascospores was observed with an Olympus BX-51 microscope. Gene deletion and site-directed mutagenesis We performed gene deletion using the split-marker approach (39). Approximately 1.0-kb fragments upstream and downstream of the target gene were amplified and fused to the N and C termini of a recyclable marker module conferring resistance to hygromycin and sensitivity to the nucleoside analog 5-fluoro-2′-deoxyuridine (23) via overlapping PCR. Following transformation of PH-1 or Gz3639 mat1-2-1 protoplasts, we selected transformants using hygromycin B (H005, MDbio, China) at final concentrations of 300 μg/ml. For site-directed mutagenesis, we constructed allelic fragments containing the desired mutations by overlapping PCR. Each allelic fragment included ∼1.0 kb upstream and downstream homologous arms and was transformed into the corresponding gene deletion mutant protoplasts. Transformants resistant to floxuridine (25 μg/ml; HY-B0097, MedChemExpress, USA) were screened by PCR, and desired mutations were confirmed by DNA sequencing (DNA-seq). We used quantitative PCR to verify the absence of unintended integration of allelic fragments in the transformants. For each mutation, we identified at least two independent mutants exhibiting similar phenotypes. All generated transformants and primers used are listed in tables S1 and S2, respectively. Reverse transcription polymerase chain reaction Total RNA was extracted from 7-days postfertilization perithecia of each strain. cDNA synthesis was performed using the RevertAid Master Mix (Thermo Fisher Scientific) according to the manufacturer’s instructions. RT-PCR products were gel purified and directly sequenced. Sanger sequencing traces were visualized using SnapGene Viewer 4.3 (www.snapgene.com/snapgene-viewer/) as previously described (26). We quantified relative expression levels by real-time RT-PCR using the 2−ΔΔCT method (64), with the GzUBH gene serving as an internal control. The primers used are listed in table S2. Construction of double-copy d-FgSPO11TGG and d-FgSPO11TAA strains To achieve ectopic overexpression of the FgSPO11TGG and FgSPO11TAA alleles at a defined genomic locus, we used the FG1G36140 locus, as its deletion does not alter the wild-type phenotype (18). In the FgSPO11TGG and FgSPO11TAA mutant backgrounds, we first replaced the FG1G36140 gene with a recyclable marker module, generating recipient strains such as FgSPO11TGG ∆FG1G36140. Next, we amplified ∼1.0-kb fragments upstream and downstream of the FG1G36140 gene and fused them to the FgSPO11TGG or FgSPO11TAA alleles (including their native promoters) via overlapping PCR. We transformed these overexpression constructs into the protoplasts of the respective recipient strains. We then screened for transformants resistant to floxuridine using PCR and performed RT-PCR to confirm that these strains overexpressed the FgSPO11 alleles in 7-days postfertilization perithecia. Generation of the catalytic-dead FgSPO11Y129F-TGG mutant To determine whether the reduction in hyphal tip nuclei caused by ectopic expression of FgSPO11 depends on DSB formation, we generated the FgSPO11Y129F-TGG mutant. Using the site-directed mutagenesis approach described above, we fused 1.0-kb upstream and downstream fragments of the FgSPO11 locus to an allele carrying both Y129F and TAG to TGG (PSC site) point mutations via overlapping PCR. We transformed the resulting construct into protoplasts of the Fgspo11 deletion mutant and confirmed the transformants by DNA-seq. Generation of FgRAD51 deletion strains in uneditable and postediting backgrounds To test whether homologous recombination efficiently repairs DSBs generated by the FgSPO11TGG allele during asexual development, we deleted the key recombinase gene FgRAD51 in both FgSPO11TGG and FgSPO11TAA backgrounds. We performed gene deletions using the same homologous recombination method described previously. For each mutation or deletion, we identified at least two independent mutants exhibiting consistent phenotypes. All generated transformants and the primers used in this study are listed in tables S1 and S2, respectively. Competition assays To detect subtle fitness differences between the wild type and the FgSPO11TGG mutant, we conducted competition assays. Conidia were harvested from 4-day-old liquid cultures grown in CMC medium. The conidial suspensions were adjusted to a final concentration of 2.0 × 105 conidia/ml. Equal volumes of wild-type and FgSPO11TGG conidial suspensions were mixed at a 1:1 ratio, and 10 μl of the mixture was inoculated onto the center of a 1/2 MM plate [containing 0.05% glucose, 30 mM NaNO3, 5 mM KH2PO4, 4 mM KCl, 1 mM MgSO4·7H2O, and agar (15 g/liter)]. To amplify potential fitness differences, hyphae from the colony margin of the mixed population were collected using a sterile toothpick and transferred to a fresh 1/2 MM plate. This transfer process was repeated sequentially up to the F5 generation. At the F5 generation, hyphae from the colony margin were harvested for genomic DNA extraction. A DNA fragment spanning the PSC site was amplified using primers flanking the target region and subjected to Sanger sequencing. The relative proportions of wild-type and FgSPO11TGG strains in the mixed population were estimated by quantifying the A/G peak ratio at the PSC site, as previously described (23). Chromatograms were visualized using SnapGene v8.0.1 (www.snapgene.com). Nuclear staining To stain nuclei in sexual tissues, we mounted samples with DAPI (4′,6-diamidino-2-phenylindole; 20 μg/ml) during slide preparation and incubated them in the dark at room temperature for 10 min. To stain nuclei in hyphal tips, we mounted the samples with DAPI (20 μg/ml) and incubated them in the dark at room temperature for 5 min. We captured fluorescence images using an Olympus BX53 microscope equipped with a DP80 dual CCD camera. Plant infection assays We resuspended conidia harvested from 4-day-old CMC cultures in sterile distilled water to 2 × 105 conidia/ml. We point-inoculated wheat heads (cultivar Xiaoyan 22) at anthesis by applying 10 μl of the conidial suspension to the upper spikelet. We assessed disease severity 14 days postinoculation by counting symptomatic spikelets per head. Yeast two-hybrid assays We used yeast two-hybrid assays to detect protein-protein interactions. The open reading frames of FgSKI8 and FgSPO11 and mutant alleles of FgSPO11 were amplified and cloned into either the pGADT7 (prey) or pGBKT7 (bait) vectors (TaKaRa Bio, Tokyo, Japan). The resulting constructs were cotransformed pairwise into yeast strain AH109 and assayed for growth on SD-His-Leu-Trp plates as previously described (18). Activity of Mel1 α-galactosidase was evaluated on SD-His-Leu-Trp plates containing X-α-gal. Western blotting To facilitate protein detection, we generated FgSPO11TGG-GFP and FgSPO11PSC-GFP constructs to express C-terminal GFP-tagged versions of either the full-length FgSpo11 or the truncated isoform. We then transformed these constructs into PH-1 protoplasts and selected transformants on TOP plates supplemented with G418. PCR analysis confirmed the successful genomic integration of the constructs. We extracted total proteins from vegetative hyphae and separated them via 10% SDS–polyacrylamide gel electrophoresis before transferring them onto nitrocellulose membranes. To prevent nonspecific antibody binding, we blocked the membranes with 5% skim milk for 1 hour at room temperature. Subsequently, we incubated the membranes overnight at 4°C with primary antibodies: anti-GFP (11814460001, Merck, Germany). After washing the membranes three times with 1× TBST [100 mM tris-HCl (pH 7.5), 136 mM NaCl, and 0.1% Tween 20), we incubated them with the corresponding secondary antibodies for 1 hour at room temperature. Following final washes with 1× TBST, we applied a chemiluminescent substrate for 5 min and captured the signals using a chemiluminescence imaging system. For loading control detection, membranes were stripped using stripping buffer (PS107S, Epizyme Biotech, China) for 10 min and washed with 1× TBST for 5 min before incubation overnight at 4°C with anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (Sangon Biotech, China). Secondary antibody incubation and signal detection were performed as described above. We independently repeated all experiments twice. Single ascospore isolation To facilitate outcrosses, we knocked out the key MAT gene MAT1-2-1 in the Gz3639 strain using a GFP-G418 resistance marker, generating a Gz3639 mat1-2-1 strain with GFP fluorescence. We also constructed the Fgspo11 deletion mutant in the Gz3639 mat1-2-1 background. For outcrosses, we fertilized Gz3639 mat1-2-1 or Gz3639 mat1-2-1 Fgspo11 strains grown on carrot agar plates with 1 ml of conidial suspension (1 × 106 conidia/ml) derived from the PH-1 or PH-1-background Fgspo11 strains, respectively. Fifty perithecia from each cross were collected near a flame under a dissecting microscope, crushed with extra fine forceps, and transferred into 100 μl of sterile water in a centrifuge tube. The tube was immediately sealed and vortexed to obtain an ascospore suspension. We transferred 10 μl of the suspension onto a 2% agar plate. By adjusting the focal plane of the dissecting microscope to the medium surface, we located individual ascospores and transferred a single ascospore to the plate edge using a capillary tube. The ascospore position was marked on the plate bottom. Plates were sealed with parafilm and incubated at 25°C for 8 hours. We then excised a small piece of medium containing the germinating ascospore with a sterile syringe and transferred it to a PDA plate, where single ascospore isolates were obtained after further incubation at 25°C. DNA-seq and identification of meiotic crossover events For each wild-type and Fgspo11 outcross, we randomly selected 10 single ascospore isolates for whole-genome DNA-seq, along with the Gz3639 mat1-2-1 strain. DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method. Sequencing libraries were prepared and sequenced on an Illumina NovaSeq 6000 system (2 × 150 bp paired-end) at Novogene Bioinformatics Institute (Tianjin, China). Low-quality reads and adapter sequences were removed using Trimmomatic (65) with default settings. Details of the DNA-seq data generated are provided in table S3. One single ascospore isolate was excluded from further analysis due to contamination. We downloaded the latest genome assembly of F. graminearum PH-1 (38) derived from our laboratory stock (YL1) from FgBase (http://fgbase.wheatscab.com/) as the reference genome. DNA-seq reads from each strain were mapped to the PH-1 reference genome using Bowtie 2 v2.5.1 (66) with default parameters. SAM files were converted to BAM format and indexed using SAMtools v1.17 (67). Duplicated reads were removed using the MarkDuplicates tool from the Picard package v3.1.0 (https://broadinstitute.github.io/picard/) to avoid biases in SNP calling caused by amplification artifacts. SNPs were identified using BCFtools v1.17 (67), and variant calls were filtered based on sequencing depth and quality scores. We used 76,117 high-quality SNP sites identified between Gz3639 mat1-2-1 and PH-1 as genetic markers to detect crossover events in each single ascospore isolate using custom scripts. Crossover points were defined as the midpoint between two adjacent markers flanking a crossover event. Hi-TOM amplicon sequencing We obtained perithecia samples from wild-type strains at 4 to 7 days postfertilization under the sexual reproduction conditions described above. We extracted total RNA and reverse transcribed it to generate cDNA as previously described. Using cDNA from each developmental stage as a template, we amplified a 189-bp fragment spanning the PSC editing site (81 bp upstream and 107 bp downstream) by PCR. We then sent equal amounts of the PCR products for next-generation sequencing using the Hi-TOM platform (China National Rice Research Institute, Chinese Academy of Agricultural Sciences, Hangzhou). We detected RNA editing events by mapping the Hi-TOM reads to the FgSPO11 gene sequence. RNA editing identification, homology searching, protein modeling, and RNA secondary structure prediction We downloaded genome sequences of 577 Ascomycota species—including 339 Sordariomycetes and 154 Fusarium species—from the National Center for Biotechnology Information (NCBI) Genome database (data S1). We also retrieved RNA-seq datasets covering both vegetative and sexual stages for a subset of species from the NCBI Sequence Read Archive (SRA) (detailed accessions are provided in table S3). We mapped the RNA-seq reads to their respective genomes and identified RNA editing sites following previously described protocols (68). Gene-level read counts were quantified with featureCounts v2.0.6 (69) and normalized to transcripts per million (TPM) in R v4.1.2 (www.r-project.org/). Orthologs of SPO11 and the corresponding amino acid states at the PSC editing site were identified using tBLASTn, considering only hits with e values of <1 × 10−5 as homologs. To prevent missing homologs due to sequence variation, we used the identified homologous sequences in each fungal clade as seeds to re-search all genomes within the clade. The phylogenetic tree of the genus Fusarium was based on our previous study (25), and the evolutionary relationships among major groups of Sordariomycetes and Ascomycota were inferred from the MycoCosm tree (https://mycocosm.jgi.doe.gov/mycocosm/home). Sequence logos around the editing sites were visualized using ggseqlogo v0.1 (70). RNA secondary structures of 30 nucleotides upstream and downstream of the PSC editing site were predicted with RNAFold (71). Secondary structure type distributions were analyzed using forgi v2.0.3 (72), and structures were visualized using RNAstructure v6.5 (73). We predicted the three-dimensional structures of protein-protein interactions between FgMms4 and FgSpo11TAA or FgSpo11TGG using AlphaFold3 Server (https://alphafoldserver.com/) and visualized them with ChimeraX v1.4 (74). Statistical significance tests were performed in R.