High‐variance phenome database reveals important roles of WD40 proteins in the plant pathogenic fungus Fusarium graminearum

Summary

• WD40 is a highly conserved protein domain in eukaryotes that functions as a versatile platform for protein–protein interactions and participates in diverse biological processes.

• We performed a genome‐wide functional analysis of WD40 domain‐containing proteins in Fusarium graminearum, a phytopathogenic fungus that causes severe yield losses and mycotoxin contamination in major cereal crops.

• Comprehensive phenotypic profiling of 119 WD40 gene deletion mutants across 22 phenotypic traits established a systematic WD40 phenome dataset, revealing the broad functional involvement of WD40 proteins and a strong correlation between sexual reproduction and virulence.

• Protein interaction analyses of selected WD40 proteins revealed diverse WD40‐mediated interaction patterns and provided further insights into WD40‐mediated protein interactions and their roles in protein complex formation. This study provides a foundation for further characterization of WD40 proteins in filamentous fungi.

Functional analysis of predicted interaction‐associated WD40 blades in Fgwd101 and Fgwd048 of Fusarium graminearum.

Introduction

WD40 is a protein domain named after a conserved region containing a tryptophan‐aspartic acid (WD) dipeptide at the C‐terminus of each repeat (Neer et al., ¹⁹⁹⁴; Smith et al., ¹⁹⁹⁹). Each WD40 repeat contributes to the formation of a four‐stranded antiparallel β‐sheet blade, and multiple blades assemble into a characteristic β‐propeller structure that provides a stable platform for protein interactions (Wall et al., ¹⁹⁹⁵; Lambright et al., ¹⁹⁹⁶; Sondek et al., ¹⁹⁹⁶). WD40 is one of the most abundant protein families in eukaryotes (Wang et al., ²⁰¹⁵). There are 743 WD40‐coding genes in wheat (Triticum aestivum L.; Hu et al., ²⁰¹⁸), 200 in rice (Ouyang et al., ²⁰¹²), 237 in Arabidopsis thaliana (van Nocker & Ludwig, ²⁰⁰³), and 262 in humans (Zou et al., ²⁰¹⁶).

The WD40 family is a large and diverse protein family that plays versatile roles in protein–protein interactions (PPIs). WD40 proteins act as scaffolds for protein binding, thereby promoting the formation of protein complexes involved in diverse cellular processes, including signal transduction and gene expression regulation (Xu & Min, 2011; Wu et al., ²⁰¹²; Zhang & Zhang, ²⁰¹⁵). However, despite their functional importance, the characterization of WD40 family proteins remains challenging. WD40 proteins frequently interact with multiple binding partners, and the structural diversity of WD40 domains contributes to the formation of distinct interaction networks. Therefore, defining the biological functions and interaction partners of individual WD40 proteins remains an important challenge.

Fusarium graminearum is an important plant pathogenic fungus that causes Fusarium head blight (FHB) in wheat and barley, as well as ear rot in maize (Windels, 2000; Desjardins, ²⁰⁰⁶). Fusarium infection causes severe loss in yield and contamination with mycotoxins, such as zearalenone (ZEA) and deoxynivalenol (DON), which pose substantial risks to human and animal health (Windels, 2000; Goswami & Kistler, ²⁰⁰⁴; Desjardins & Proctor, ²⁰⁰⁷). To develop effective strategies for controlling FHB, extensive genetic and biochemical studies have been conducted to understand fungal development, stress responses, mycotoxin production, and pathogenicity in F. graminearum.

Genome‐wide deletion mutant collections have served as powerful resources for systematically investigating gene functions and their contributions to biological processes. In F. graminearum, several genome‐wide functional studies targeting specific gene families, including transcription factors (Son et al., ^2011b), kinases (Wang et al., ²⁰¹¹), and phosphatases (Yun et al., ²⁰¹⁵), have provided valuable insights into the genetic regulation of fungal development, stress responses, and pathogenicity. Given that WD40 proteins primarily function as scaffolds for protein interactions, integrating systematic phenotypic analysis with protein interaction studies provides an effective approach for understanding their functional diversity. However, despite their biological importance, a genome‐wide functional investigation of WD40 proteins has not yet been conducted in filamentous fungi, including fungal pathogens.

In this study, we identified 157 genes encoding WD40 domain‐containing proteins and generated a genome‐wide deletion mutant library for non‐essential WD40 genes in F. graminearum. Comprehensive phenotypic profiling of these mutants was performed across 22 phenotypic traits, including vegetative growth, sexual and asexual reproduction, pathogenicity, mycotoxin production, and responses to various stress agents. Based on this genome‐wide phenome dataset, we analyzed relationships among WD40‐associated phenotypes and further characterized selected WD40 proteins through protein interaction analyses. This study provides a comprehensive overview of the functional characteristics of WD40 proteins in F. graminearum, establishing a foundation for further investigation of their molecular functions in filamentous fungi.

Materials and Methods

Fungal strains and culture conditions

F. graminearum Schwabe wild‐type strain Z‐3639 was used as the parental strain throughout this study (Bowden & Leslie, 1999). Fungal strains were stored at −80°C as mycelial suspensions in 20% glycerol solution. Culture media were prepared as described in the Fusarium laboratory manual (Leslie & Summerell, 2006). All fungal strains were grown at 25°C.

Identification of WD40 genes in F. graminearum

Genes encoding WD40 domain‐containing proteins were identified from the FungiDB database (fungidb.org; Stajich et al., ²⁰¹²), with InterPro IDs used as a reference. The FungiDB database was used to analyze the chromosomal location of WD40 genes and obtain sequence information. To confirm the presence of WD40 motifs in the extracted WD40 protein sequences, we used the WD40‐repeat Protein Structure Predictor (www.wdspdb.com/wdsp). To identify orthologs of F. graminearum WD40 proteins in S. cerevisiae, BLASTp analysis was performed using annotated information from the Saccharomyces Genome Database (www.yeastgenome.org). Hits with an E‐value cut‐off of < 0.0001 were retained. The distribution of WD40 genes across chromosomes was mapped with MapGene2Crom (MG2C; Chao et al., ²⁰²¹), and the architecture of WD40 genes was visualized using IBS software (Illustrator for Biological Sequences; Xie et al., ²⁰²²).

Phylogenetic analysis and classification of putative WD40 genes

The amino acid sequences of WD40 proteins were aligned using ClustalW and MUSCLE algorithms (Thompson et al., ¹⁹⁹⁴; Edgar, ²⁰⁰⁴). An unrooted neighbor‐joining phylogenetic tree was constructed using MEGA11 (Molecular Evolutionary Genetics Analysis) software, with 1000 bootstrap replicates for statistical support (Tamura et al., ²⁰²¹). WD40 proteins were classified using the OmicsBox software (www.biobam.com/omicsbox). The finalized phylogenetic tree was visualized using iTOL (itol.embl.de; Letunic & Bork, ²⁰¹⁹).

Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis

GO IDs were identified using OmicsBox with default settings to identify biological functions of WD40 proteins (Götz et al., ²⁰⁰⁸). Functional annotation data of WD40 proteins were submitted to WEGO 2.0 (Web Gene Ontology Annotation Plot, wego.genomics.cn; Ye et al., ²⁰¹⁸) to generate GO classification histograms of WD40 proteins. In addition, KEGG enrichment analysis of WD40 genes was performed using KOBAS v.3.0 (kobas.cbi.pku.edu.cn/) to obtain the KEGG IDs of WD40 proteins. The resulting graphs were generated using the R program.

DNA extraction and Southern blot hybridization

Three‐day‐old mycelia grown in 5 ml of complete medium (CM) broth were lyophilized to extract genomic DNA using the CTAB method (Leslie & Summerell, 2006). Restriction endonuclease digestion and agarose gel electrophoresis were performed as previously described (Sambrook & Russell, 2001). Southern blot analysis was performed using the North2South Biotin Random Prime Labeling Kit and North2South Chemiluminescent Hybridization and Detection Kit (Thermo Scientific, Middlesex County, MA, USA). The ChemiDoc XRS+ Gel Imaging System (Bio‐Rad) was used to detect specific bands.

Generation of WD40 gene deletion mutants and green fluorescent protein (GFP)‐tagged Atg8 strain

WD40 genes were deleted using the split‐marker method based on homologous recombination (Son et al., ^2011a). The geneticin resistance gene cassette (GEN) was amplified from the pII99 plasmid, and 5′ and 3′ flanking regions of the target genes were amplified from wild‐type genomic DNA. Three fragments were fused by double‐joint PCR (Yu et al., ²⁰⁰⁴). PCR products were extracted using the MEGAquick‐spin Plus Total Fragment DNA Purification Kit (Intron Biotech, Seongnam, Republic of Korea). Purified products were transformed into protoplasts of the wild‐type strain.

To generate N‐terminally GFP‐tagged Atg8, the hygromycin resistance‐GFP cassette (HYG‐GFP), including the Neurospora crassa isocitrate lyase promoter, was amplified from the pIGPAPA plasmid (Horwitz et al., ¹⁹⁹⁹) and fused with the corresponding ATG8‐targeting fragments to enable insertion immediately upstream of the ATG8 start codon. The resulting construct was introduced into protoplasts of the wild‐type strain using the transformation procedure described previously.

The transformants were screened using diagnostic PCR, and mutants were confirmed by Southern blot hybridization using either flanking region probes or ORF‐specific probes. At least two independent mutants were obtained for each gene deletion. The oligonucleotides used were synthesized by Bioneer (Daejeon, Republic of Korea; Supporting Information Table S1).

Genetic manipulation for complementation and overexpression strains

For complementation of Δfgwd101 and Δfgwd133, the ORF and native promoter of each gene were amplified from wild‐type genomic DNA using PrimeSTAR (Takara Bio, Otsu, Japan). The resulting constructs and XhoI‐digested pDL2 (Zhou et al., ²⁰¹¹) were co‐transformed into the yeast strain PJ69‐4A using the Alkali‐Cation Yeast Transformation Kit (MP Biomedicals, Irvine, CA, USA). The vectors obtained from yeast transformants were transformed into Escherichia coli DH10B. Plasmid DNA was extracted with the DNA‐spin Plasmid DNA Purification Kit (Intron Biotech) and verified by sequencing (Macrogen, Seoul, Republic of Korea).

To produce large quantities of protein, overexpression strains for FgWD101 or FgWD133 were generated. The ORF of each gene was amplified and co‐transformed into the yeast strain PJ69‐4A with XhoI‐digested pDL2. RP27‐FgWD101‐GFP and RP27‐FgWD133‐GFP fusion vectors were obtained from the yeast transformants. The subsequent steps were performed as described previously. The resulting fusion vectors were used for transforming the corresponding deletion mutants.

Phenotypic analysis of vegetative growth, conidiation, and sexual development

For phenotypic analysis, at least two independent strains were used for each gene. Colony morphology was assayed on CM 4 d after inoculation. Conidiation assays were conducted by inoculating fresh mycelial blocks from each strain into 5 ml carboxymethyl cellulose (CMC) medium and incubating for 5 d on a rotary shaker (200 rpm). The number of conidia was determined using a hemocytometer (Paul Marienfeld, Lauda‐Königshofen, Germany). Each experiment was repeated three times.

For sexual development analyses, each strain was inoculated on carrot agar medium, and after 5 d of incubation, aerial mycelia were removed with 0.4 ml of 2.5% sterile Tween 60 (Leslie & Summerell, 2006). The cultures were incubated under near‐UV light conditions (wavelength: 365 nm; Sankyo Denki, Tokyo, Japan) for 7–9 d. The formation and maturation of perithecia were observed using a Zeiss SteREO Lumar.V12 microscope (Carl Zeiss, Oberkochen, Germany). Ascospore discharge was assessed by observing discharged ascospores on the medium lid. The morphology of rosettes, asci, and ascospores was examined using a Leica DM6 B microscope (Leica Microsystems, Wetzlar, Germany).

For outcrosses, the female strain was fertilized with 1 ml of conidial suspension (10⁶ conidia ml⁻¹) from the male strain. Fertilization was performed 5 d after inoculating the strains on carrot agar. The ∆mat1‐1/pIGPAPA‐GFP strain was used to examine 1 : 1 segregation of ascospores in each ascus (Lee et al., ²⁰⁰³).

Virulence test

The point inoculation method was employed to assess fungal virulence (Son et al., ^2011b). Conidia were collected and resuspended in 0.01% Tween‐20 at a concentration of 10⁵ conidia ml⁻¹. For mutants showing reduced conidiation, larger volumes of CMC cultures were prepared to obtain sufficient numbers of conidia, and the conidial concentration was adjusted to the same level as that used for the wild‐type strain before inoculation. Then, 10 μl of conidial suspension was injected into the central spikelet of a wheat head at the early anthesis stage (cv Eunpamil). For mutants that completely failed to produce conidia, mycelial agar blocks were directly placed onto wheat heads to evaluate virulence. Inoculated wheat heads were covered with plastic bags to maintain humidity for 3 d. The number of spikelets exhibiting disease symptoms was evaluated 14 d after inoculation.

For the coleoptile virulence test, c. 2 mm of the upper part of 3‐d‐old wheat seedlings was trimmed and inoculated with 2 μl of conidial suspension of each strain (10⁶ conidia ml⁻¹). The inoculated seedlings were placed in a humid chamber, and disease symptoms were examined 7 d post‐inoculation (dpi).

For the cellophane penetration assay, autoclaved cellophane sheets were used on CM plates. To determine if a strain penetrated the cellophane membrane, strains were cultured on CM with a cellophane membrane covering the surface. After 2 d of incubation, the cellophane membrane was removed from each plate. The plates were incubated for 2 d and photographed. All experiments were repeated three times independently.

Mycotoxin analysis

To assess mycotoxin production, each strain was cultured on 1.5 g of sterilized rice substrate under dark conditions for 3 wk. The rice culture was harvested and ground. Mycotoxins were extracted by vigorously mixing the ground rice culture with 6 ml of 84% acetonitrile (ACN) for 30 min at 25°C on a rotary shaker (200 rpm), resulting in phase separation. The upper layer was filtered through a 0.45‐μm syringe filter (Merck Millipore, Burlington, MA, USA). Reverse‐phase HPLC analysis was performed using a Prominence HPLC system (Shimadzu, Kyoto, Japan) equipped with a C18 column, with minor modifications (Kim et al., ²⁰⁰⁵; Ok et al., ²⁰¹⁸). Zearalenone (ZEA) was detected using a mobile phase consisting of 70% aqueous methanol, at a flow rate of 1 ml min⁻¹. Deoxynivalenol (DON) was detected with a mobile phase containing 10% aqueous ACN, at a flow rate of 1 ml min⁻¹. The following gradient elution program was used for DON detection: 10% ACN was maintained for 11 min; increased linearly to 30% ACN at 12 min; decreased linearly from 30% ACN at 12 min to 10% ACN at 18 min; maintained for 17 min to re‐equilibrate the column; and the next sample was injected, resulting in a total runtime of 35 min. Both ZEA and DON were detected at a wavelength of 235 nm using the diode‐array detector.

Stress tolerance analysis

Stress responses were examined using 96‐well plates under various stress conditions, with the concentration of each stress agent adjusted to the half‐maximal effective concentration (EC₅₀) of the wild‐type (Choi et al., ²⁰²³). Conidia were collected from CMC cultures, filtered using Miracloth (Merck Millipore), and resuspended in CM broth. All strains were normalized to a concentration of 10⁵ conidia ml⁻¹. In 96‐well plates, 100 μl of conidial suspension was mixed with an equal volume of CM supplemented with various stress agents: oxidative stress (0.31 mM hydrogen peroxide (H₂O₂)), osmotic stress (1 M sorbitol and 1 M NaCl), cell wall stress (0.1 mg ml⁻¹ sodium dodecyl sulfate (SDS)), fungicide stress (0.05 μg ml⁻¹ fludioxonil, 125 μg ml⁻¹ iprodione, and 0.625 μg ml⁻¹ tebuconazole), endoplasmic reticulum stress (2 μg ml⁻¹ tunicamycin and 1.25 mM dithiothreitol (DTT)), and pH stress (pH 2 and 12). After inoculation, the plates were incubated in a humid plastic box for 48 h. We measured OD₆₀₀ using a Synergy HTX microplate reader (BioTek, Winooski, VT, USA). The obtained values were adjusted by subtracting the average absorbance of the medium. The results were normalized to growth in CM, and log₂ fold change was calculated based on the average growth. The heat map was generated using Morpheus software (software.broadinstitute.org/morpheus).

Phenotypic correlation analysis

Phenotypic correlation analysis was performed using quantified phenotypic data from WD40 deletion mutants. For consistency across phenotypic traits, sexual reproduction‐related phenotypes were converted into directional scores. Pairwise correlations between phenotypic traits were calculated using Spearman's rank correlation coefficient. Correlation matrices were generated and visualized as heat maps using Python packages (v.3.11.15), including pandas, scipy, and seaborn.

Western blotting and immunoprecipitation

Fresh mycelia were harvested from 3‐d‐old CM culture, finely ground with liquid nitrogen, and suspended in 1 ml extraction buffer containing protease inhibitors (1 mM phenylmethyl sulfonyl fluoride, 20 μg ml⁻¹ leupeptin, 2 μg ml⁻¹ pepstatin A, and 20 μg ml⁻¹ aprotinin). For analysis of autophagic activity, mycelia were cultured in CM for 24 h and then transferred to nitrogen starvation medium (MM‐N) as previously described (Zheng et al., ²⁰¹⁸) for 2 h before protein extraction. The samples were sonicated using an ultrasonic homogenizer, and cell lysates were collected by centrifugation at 16 000  g for 20 min at 4°C.

For immunoprecipitation (IP), GFP‐tagged proteins were pulled down using anti‐GFP monoclonal antibody‐conjugated magnetic beads (Code No. D153‐11, MBL, Tokyo, Japan). For co‐immunoprecipitation (co‐IP), strains expressing both GFP‐ and FLAG‐tagged proteins were generated, and protein complexes were immunoprecipitated using anti‐GFP magnetic beads. The immunoprecipitated proteins were analyzed by western blotting.

Proteins were separated using SDS‐PAGE and transferred to nitrocellulose membranes using the Trans‐Blot Turbo Transfer System (Bio‐Rad). The membranes were blocked with 7% blotting grade blocker in TBS‐T buffer. Primary antibodies against GFP (Cat# ab290; Abcam, Cambridge, UK), FLAG (Cat# A00170; GenScript, Piscataway, NJ, USA), and GAPDH (Cat# A19056; ABclonal, Woburn, MA, USA) were used for immunoblotting. Goat anti‐rabbit IgG‐HRP antibody (Cat# ab6721, Abcam) was used as the secondary antibody. Signals were detected using Amersham ECL Select Western Blotting Detection Reagent (Cytiva, Marlborough, MA, USA) and photographed using the ChemiDoc XRS+ Gel Imaging System (Bio‐Rad).

Affinity purification‐mass spectrometry (AP‐MS)

Protein samples were separated using 4–20% precast SDS‐PAGE gels and stained with Coomassie Brilliant Blue (Cat# 4561095, Bio‐Rad). The corresponding deletion mutants were used as negative controls to identify non‐specific proteins bound to anti‐GFP mAb‐conjugated magnetic beads. After excluding non‐specific bands detected in the negative controls, bands of interest were excised and digested with trypsin. Tryptic peptides were analyzed using a Q Exactive system (Thermo Scientific; Kim et al., ²⁰¹⁸).

Yeast two‐hybrid assay

Yeast two‐hybrid (Y2H) assay was performed using Matchmaker Gold Yeast Two‐Hybrid System (Takara Bio, CA, USA), according to the manufacturer's instructions. The indicated full‐length or truncated coding sequences of candidate interacting proteins were amplified from cDNA of the wild‐type strain Z‐3639 and cloned into the corresponding bait and prey vectors. WD40 protein‐encoding sequences, including FgWD101, FgWD133, Atg2, and the truncated structure of FgWD101, were cloned into pGADT7 as prey constructs. The sequences encoding their putative interacting partners, including Fgwd072, septin proteins, and FgWD048 variants, were inserted into pGBKT7 as baits. The bait and prey constructs were co‐transformed into S. cerevisiae strain Y2HGold. Transformants were selected on synthetic dropout (SD) medium lacking leucine and tryptophan (SD/‐Leu/‐Trp), and protein–protein interactions were assessed on SD lacking histidine, leucine, and tryptophan (SD/‐His/‐Leu/‐Trp).

Prediction of protein interaction interfaces

Protein complex structures were predicted using AlphaFold2‐Multimer via the ColabFold platform with the amino acid sequences of the corresponding protein pairs as input (Evans et al., ²⁰²²; Mirdita et al., ²⁰²²). Default prediction parameters were used. Putative interface residues between the two proteins were identified using the Interfaces tool in UCSF ChimeraX v.1.9 (Meng et al., ²⁰²³), excluding solvent molecules and using the default interface analysis parameters.

Results

Identification and classification of WD40 genes in F. graminearum

To identify WD40 genes in F. graminearum, we searched for WD40 repeat‐containing domains and WD40 domain‐encoding genes in FungiDB (fungidb.org; Stajich et al., ²⁰¹²) based on the InterPro database. Duplicates of InterPro IDs were removed, and 157 putative WD40‐encoding genes were identified (Table S2). For convenience, we designated the identified genes as FgWD001‐FgWD157 (F. graminearum WD40 gene no.), and genes named in previous studies were referred to by their established names. To verify the gene list, we used WDSP (WD40‐repeat protein Structure Predictor) for search‐based domain analysis. The F. graminearum genome contains 157 putative WD40 genes, which represent 1.1% of the whole F. graminearum protein‐coding genes (King et al., ²⁰¹⁵). Most WD40 genes (152/157) have not been functionally characterized in F. graminearum.

Subsequently, we identified 203 WD40 domains comprising a total of 1254 WD40 repeats. Among the 157 WD40 proteins, 107 contained 6 to 8 repeats, indicating that c. 68% of these proteins maintain the typical structure of the WD40 protein domain. Additionally, 33 WD40 proteins were found to contain more than 10 repeats, suggesting potential structural or functional divergence within this protein family.

To examine the evolutionary relationships of WD40 proteins in F. graminearum, we constructed a phylogenetic tree using the 157 identified WD40 proteins (Fig. 1a). All WD40 protein sequences were subjected to clustering analysis based on sequence similarity. The results showed that in F. graminearum, WD40 proteins are divided into six major clusters containing 9, 9, 7, 28, 53, and 51 proteins, respectively (Table S2), illustrating the complexity and differentiation of the F. graminearum WD40 protein family. Moreover, some chromosomes had a relatively high density of WD40 genes: chromosome 1 contained 61 genes. Chromosomes 1–4 comprised 38.85%, 17.19%, 21.01%, and 22.92% WD40 genes, respectively (Fig. 1b,c).

Fig. 1.

Genome‐wide distribution and classification of WD40 genes in Fusarium graminearum. (a) Phylogenetic classification of 157 WD40 proteins in F. graminearum. WD40 proteins were grouped into six clusters (clusters I–VI), indicated by different colors. Core WD40 proteins, defined in this study as those involved in both pathogenicity and sexual reproduction, are highlighted in red. An unrooted phylogenetic tree was constructed using the neighbor‐joining method in MEGA11 with 1000 bootstrap replicates. Gray circles at the nodes indicate bootstrap support, with circle size proportional to the bootstrap value. (b) Number of WD40 genes on each chromosome of F. graminearum. (c) Chromosomal distribution of WD40 genes in F. graminearum. The chromosome map was generated using MapGene2Chrom (MG2C), with chromosome numbers indicated at the top. Core WD40 genes are highlighted in red. (d) Schematic representation of the domain architecture of F. graminearum WD40 proteins, classified into 16 groups. One representative protein from each group is shown. WD40 domains are shown as black boxes, and additional domains are indicated by different colors and labeled accordingly. The figure was generated using IBS (Illustrator for Biological Sequences).

Based on their domain architectures, the WD40 genes were classified into 16 distinct classes (Fig. 1d). Class 1 comprised 75 WD40 proteins that contained only WD40 domains. The remaining proteins, which harbored additional domains, were grouped into Classes 2 through 16. For example, 16 WD40 proteins with the quinoprotein alcohol dehydrogenase‐like superfamily domain were categorized as Class 2, while 13 proteins containing the NACHT nucleoside triphosphatase domain were placed in Class 4. Domains that occurred only once, with the exception of the NWD2 domain, were not assigned to specific classes, as genes with such unique domain architectures were not grouped into separate classes. The domain‐based classification closely aligned with the clustering of WD40 proteins, reflecting a similar pattern in the organization of classes.

Functional enrichment analysis

To explore the potential function of F. graminearum WD40 proteins, we performed GO distribution analysis. GO analysis classified WD40 genes into three categories: molecular function, cellular component, and biological process (Table S3), and WD40 proteins were extensively distributed across all three categories (Fig. S1a). Particularly, many WD40 proteins were predicted to be involved in binding (molecular function), cell and organelle (cellular component), and cellular and metabolic processes (biological process; Table S3).

The WD40 proteins of F. graminearum were also classified using KEGG pathway enrichment analysis. In total, 18 KEGG pathways were enriched for the selected WD40 proteins (Table S4; Fig. S1b). Ubiquitin‐mediated proteolysis (fgr04120), ribosome biogenesis in eukaryotes (fgr03008), RNA transport (fgr03013), and autophagy (fgr04138) were significantly enriched. The results of functional enrichment analyses predicted that the WD40 proteins of F. graminearum participate in several biological processes and interact with multiple interaction partners. Furthermore, these findings are consistent with the well‐established functions of the WD40 protein family and provide a foundation for further characterization of conserved genes in F. graminearum.

Construction of the WD40 gene deletion mutant library

To establish a resource for systematic functional characterization of WD40 genes, we generated a WD40 gene deletion mutant library. Among the identified 157 WD40 genes, five genes were excluded due to our inability to construct amplicons. In total, we successfully obtained deletion mutants for 119 WD40 genes, with at least two independent strains for each gene (Fig. S2). Despite more than four independent deletion attempts, knockout mutants could not be obtained for 33 WD40 genes, suggesting that these genes are putatively essential for viability or normal growth. Among the 33 genes, 29 were homologous to WD40 genes reported to be essential for growth in S. cerevisiae (Table S2).

Phenotypic profiling of WD40 gene deletion mutants

We analyzed phenotypic differences between the 119 WD40 gene deletion mutants and the wild‐type strain under various developmental and stress conditions. The phenotypic traits examined included vegetative growth, asexual and sexual development, virulence, and production of the mycotoxins deoxynivalenol (DON) and zearalenone (ZEA).

Vegetative growth was examined by cultivating WD40 deletion mutants on CM. Based on their relative growth rates compared with the wild‐type strain, the mutants were classified into six growth groups (Fig. 2a): group 6, increased growth compared with the wild‐type strain; group 5, 81–100%; group 4, 61–80%; group 3, 41–60%; group 2, 21–40%; and group 1, 0–20% of the wild‐type growth rate. A total of 27 mutants exhibited reduced vegetative growth and were classified into groups 1–4, whereas 85 mutants belonged to group 5 and displayed growth comparable to the wild‐type strain. Representative mutants displaying altered vegetative growth phenotypes (groups 1–4 and 6) are shown in Fig. 2(b). In addition, five WD40 deletion mutants exhibited abnormal hyphal morphology, including bulbous and irregular hyphal tips (Fig. 2c).

Fig. 2.

Vegetative growth and conidiation phenotypes of WD40 deletion mutants in Fusarium graminearum. (a) Classification of WD40 gene deletion mutants into six vegetative growth groups based on their relative growth rates on complete medium (CM) compared with the wild‐type strain. (b) Representative WD40 gene deletion mutants exhibiting altered vegetative growth phenotypes (groups 1–4 and 6) on CM. Images of group 1 mutants were captured at 7 d post‐inoculation (dpi), whereas those of groups 2–4 and 6 were captured at 4 dpi. (c) Mycelial morphology of WD40 gene deletion mutants exhibiting abnormal hyphal morphology. Bar, 20 μm. (d) Number of WD40 gene deletion mutants in each conidiation group based on their relative conidial production compared with the wild‐type strain. (e) Conidial morphology of WD40 gene deletion mutants exhibiting altered conidial morphology. Bright‐field (BF) and calcofluor white (CFW)‐stained images were captured after 3 d of incubation in carboxymethyl cellulose medium. Bar, 20 μm.

Conidial production was assessed to investigate the role of WD40 genes in asexual development. Based on relative conidiation levels compared with the wild‐type strain, WD40 deletion mutants were classified into five groups (Fig. 2d). The conidiation groups were defined according to relative conidial production, ranging from group 1 (0–30%) to group 5 (120.1–234%), with intermediate groups representing 30.1–60% (group 2), 60.1–90% (group 3), and 90.1–120% (group 4) of wild‐type conidiation. Based on this classification, 40 mutants were classified into group 4, which included the wild‐type conidiation level range, whereas 39 mutants exhibited increased conidial production and were classified into group 5. The remaining mutants showed various degrees of reduced conidial production, with 20, 9, and 11 mutants classified into groups 3, 2, and 1, respectively. In addition to changes in conidial production, several WD40 deletion mutants displayed altered conidial morphology. The Δfgwd018 and Δfgwd129 mutants produced shorter conidia with fewer septa (2–3 septa), whereas Δfgwd143 produced longer conidia compared with the wild‐type strain, which typically produced conidia with 5–6 septa (Fig. 2e).

Sexual reproduction was evaluated by examining perithecium formation, internal maturation, ascospore morphology, and forcible ascospore discharge. Based on these phenotypes, WD40 deletion mutants were classified into four SD groups (Fig. 3a). A total of 85 mutants showed normal SD (group 1), whereas 34 mutants exhibited defects in sexual reproduction. In SD group 2, seven WD40 mutants produced normal‐shaped perithecia containing asci and ascospores but displayed defects in ascospore morphology and/or forcible ascospore discharge (Figs 3b, S3a). Among them, Δfgwd094 and Δfgwd139 formed normal melanized perithecia with eight ascospores per ascus, but showed impaired ascospore discharge. Δfgwd044 produced abnormal ascospores with fewer septa compared with the wild‐type strain, whereas ΔFgama1 and Δfgwd118 formed irregularly shaped ascospores. In addition, Δfgwd125 and Δfgwd143 produced fewer ascospores per ascus and exhibited defects in ascospore discharge. Five mutants classified into SD group 3 exhibited impaired perithecium maturation and internal development, producing immature perithecia with few or no intact ascospores (Figs 3c, S3b). Twenty‐two WD40 mutants belonging to SD group 4 failed to produce perithecia, resulting in self‐sterility. To determine whether these defects were associated with male or female fertility, self‐sterile mutants were outcrossed as either male or female parents with the ∆mat1‐1/pIGPAPA‐GFP strain (Figs 3d–f, S4). Except for Δfgwd142, which was excluded from the fertility analysis due to its severe growth defect, all tested self‐sterile mutants retained male fertility, whereas their female fertility varied. Among them, seven mutants retained both male and female fertility, four mutants produced immature perithecia when used as female parents, and 10 mutants exhibited complete female sterility with no mature perithecium formation.

Fig. 3.

Sexual development phenotypes of WD40 gene deletion mutants in Fusarium graminearum. (a) Classification of sexual development (SD) groups based on perithecial formation and internal structures. ‘Normal’ indicates a wild‐type phenotype. ‘Abnormal’ indicates reduced numbers or altered morphology of perithecia or internal structures, whereas ‘Defect’ indicates the absence of perithecia or internal structures. (b, c) Self‐fertility of WD40 gene deletion mutants belonging to SD groups 2 and 3. Perithecial formation, rosette asci, and forcible ascospore discharge were examined 7–9 d after sexual induction. Forcible ascospore discharge was assessed at 24 and 48 h. White cloudy material represents discharged ascospores. SD group 2 is shown in (b), and SD group 3 is shown in (c). Bars: (black) 200 μm; (white) 20 μm. (d–f) Representative male and female fertility phenotypes of self‐sterile WD40 gene deletion mutants belonging to SD group 4. Each mutant was crossed as either the male (M) or female (F) parent with the ∆mat1‐1/pIGPAPA‐GFP strain. Male and female fertility was evaluated by examining ascus formation and the expected 1 : 1 segregation of green fluorescent protein (GFP)‐positive and GFP‐negative ascospores. Representative mutants with normal male and female fertility are shown in (d), those exhibiting impaired female fertility are shown in (e), and those exhibiting female‐specific sterility are shown in (f). The remaining mutants are presented in Supporting Information Fig. S4. BF, bright field; N/A, not applicable. Bar, 20 μm.

Virulence of WD40 deletion mutants was assessed by point inoculation of wheat heads. Based on disease progression, the mutants were classified into three virulence groups. Overall, 37 mutants exhibited reduced virulence compared with the wild‐type strain. Twenty‐four mutants in virulence group 1 showed severe defects in disease progression and only colonized the inoculated spikelets, with limited spread to adjacent spikelets (Fig. S5a). In virulence group 2, 13 WD40 mutants were able to spread to adjacent spikelets; however, their virulence was reduced compared with the wild‐type strain (Fig. S5b). The remaining 82 mutants were classified into virulence group 3, which showed wild‐type‐level disease progression. Representative disease symptoms of each virulence group are shown in Fig. 4a. Notably, mutants exhibiting significantly reduced vegetative growth (growth groups 1–3) were enriched in virulence group 1, suggesting a close association between vegetative growth and disease progression in F. graminearum (Fig. 4b). To further investigate the reduced disease progression observed in WD40 mutants belonging to virulence group 1, penetration ability was first examined using a cellophane membrane assay and in planta penetration assays (Fig. 4c). Among these mutants, four WD40 deletion mutants failed to penetrate the cellophane membrane. We further examined whether these penetration defects affected fungal spread during host infection by generating cytosolic GFP‐expressing strains of the wild‐type strain and the four mutants (Fig. 4d). At 6 d after inoculation, GFP signals from the wild‐type strain were observed in adjacent spikelets through rachis nodes, whereas GFP signals from the four mutants were restricted to the inoculated spikelets, indicating impaired fungal spread through rachis nodes in planta.

Fig. 4.

Virulence and mycotoxin production phenotypes of WD40 gene deletion mutants in Fusarium graminearum. (a) Representative virulence phenotypes of WD40 gene deletion mutants belonging to virulence groups 1 and 2. (b) Number of WD40 gene deletion mutants in each virulence group. Colors indicate the vegetative growth groups (groups 1–6) of the corresponding mutants. Vir 1–3 denote virulence groups 1–3. (c) Cellophane penetration assay of the wild‐type strain and the four WD40 gene deletion mutants that failed to penetrate the cellophane membrane. Colonies were grown on complete medium (CM) overlaid with a cellophane membrane for 2 d (Before). After removal of the membrane, plates were incubated for an additional 2 d (After). (d) Longitudinal sections of wheat heads inoculated with cytosolic GFP‐expressing wild‐type and the four WD40 gene deletion mutants exhibiting defective penetration. Wheat heads were harvested at 6 d after inoculation and longitudinally sectioned. GFP fluorescence indicates fungal colonization and spread from the inoculated spikelet. Arrowheads indicate the inoculated spikelets. ‘Reflected’ indicates reflected‐light images. (e) Pie charts showing the distribution of WD40 gene deletion mutants according to their relative zearalenone (ZEA) and deoxynivalenol (DON) production. Numbers indicate the number of mutants in each category.

Production of the mycotoxins ZEA and DON was analyzed in WD40 deletion mutants. Altered ZEA production was observed in 48 mutants, with 19 mutants showing increased production and 29 mutants showing reduced production compared with the wild‐type strain (Fig. 4e). In addition, 38 mutants exhibited altered DON production, including 26 mutants with increased accumulation and 12 mutants with reduced production. All 12 mutants with reduced DON production also belonged to virulence group 1 and SD group 4.

Comprehensive phenotypic summary and phenotypic correlation observed among WD40 proteins

In addition to developmental and pathogenicity‐related phenotypes, we evaluated the stress responses of WD40 deletion mutants under various environmental and chemical stress conditions. Stress tolerance was assessed under conditions affecting multiple cellular processes, including cell wall integrity, osmotic stress response, oxidative stress response, endoplasmic reticulum stress response, and antifungal drug sensitivity.

Together with the stress response data, we generated a comprehensive genome‐wide phenome dataset for 119 non‐essential WD40 deletion mutants, covering vegetative growth, asexual and sexual development, virulence, mycotoxin production, stress tolerance, and antifungal drug responses (Table S5). To provide an overview of the phenotypic contribution of WD40 proteins, all phenotypic data were integrated and visualized as a heatmap (Fig. 5). This integrated phenotypic landscape revealed that several WD40 deletion mutants exhibited defects across multiple biological processes. To further investigate relationships among these phenotypic traits, we performed Spearman correlation analysis using the integrated dataset (Fig. S6). The analysis revealed strong associations between vegetative growth and sexual reproduction, as well as between virulence and sexual reproduction (Spearman correlation coefficient > 0.5), indicating that these phenotypic alterations frequently occurred together among WD40 deletion mutants. Because sexual reproduction and virulence represent distinct biological processes in the fungal life cycle, the strong association between these traits suggests that a subset of WD40 proteins contributes to both processes. Based on this observation, we defined WD40 genes required for both sexual reproduction and virulence as core WD40 genes associated with fungal development and pathogenicity.

Fig. 5.

Phenotypic clustering of non‐essential WD40 gene deletion mutants in Fusarium graminearum. Hierarchical clustering of 119 WD40 gene deletion mutants based on phenotypic profiles was performed using Pearson's correlation in the Morpheus software. Log₂‐transformed fold changes relative to the wild‐type strain were used for hierarchical clustering and heat map generation. Blue dots indicate WD40 genes designated as core genes in this study. In the heat map, blue indicates reduced phenotypes or increased chemical susceptibility, whereas red indicates enhanced phenotypes or increased chemical tolerance. The right panel summarizes the phenotypes associated with sexual development, where white, light green, and dark green indicate wild‐type, abnormal, and defective phenotypes, respectively. The complete phenotypic dataset used for hierarchical clustering and heat map generation is provided in Supporting Information Table S5. Abbreviations used in the heat map: AD, ascospore discharge; AF, ascospore formation; CM, complete medium; CON, conidiation; DON, deoxynivalenol; DTT, 1.25 mM dithiothreitol; F, female fertility; FLU, 0.05 μg ml⁻¹ fludioxonil; H₂O₂, 0.31 mM hydrogen peroxide; IPR, 125 μg ml⁻¹ iprodione; TBZ, 0.625 μg ml⁻¹ tebuconazole; M, male fertility; NaCl, 1 M NaCl; PF, perithecia formation; PM, perithecia maturation; SDS, 0.1 mg ml⁻¹ sodium dodecyl sulfate; SOR, 1 M sorbitol; TM, 2 μg ml⁻¹ tunicamycin; VIR, virulence; ZEA, zearalenone.

Identification of interaction networks associated with core WD40 proteins

To investigate the molecular networks associated with core WD40 proteins, we performed affinity purification‐mass spectrometry (AP‐MS) analysis. Because WD40 proteins primarily function as scaffolds for protein–protein interactions, identification of their interacting partners is essential for understanding their functional roles. Among the core WD40 proteins associated with fungal development and virulence, Fgwd101 and Fgwd133 were selected as representative candidates because their orthologs, Sec13 and EED, respectively, have been extensively characterized in model organisms. Given the limited information available regarding WD40 protein interaction networks in F. graminearum, analyzing conserved WD40 proteins provided a reference framework to validate our interaction‐based approach and to investigate conserved and potentially species‐specific protein associations.

To perform AP‐MS analysis, we generated GFP‐tagged complementation and overexpression strains of FgWD101 and FgWD133. For FgWD101, immunoprecipitation of Fgwd101‐GFP was confirmed by western blot analysis, detecting a protein band corresponding to the expected size of c. 63 kDa (Fig. S7a). AP‐MS analysis identified 33 candidate interacting proteins enriched in the Fgwd101‐GFP samples compared to the control (log₂ fold change > 1; Table S6). Among these candidates, Fgwd072 was detected in both FgWD101 complementation and overexpression strains and showed the highest enrichment frequency. STRING‐based network (Szklarczyk et al., ²⁰²³) predicted a specific interaction between Fgwd101 and Fgwd072 (Fig. 6a), whereas other identified candidates mainly clustered with ribosomal proteins, which were considered background‐associated proteins in the AP‐MS dataset. The interaction between Fgwd101 and Fgwd072 was further confirmed by co‐immunoprecipitation (co‐IP) and yeast two‐hybrid (Y2H) assays (Fig. 6b,c).

Fig. 6.

Validation of protein interaction networks associated with Fgwd101 and Fgwd133 in Fusarium graminearum. (a) STRING network of Fgwd101‐associated proteins identified by affinity purification‐mass spectrometry (AP‐MS). Fgwd072 was selected for interaction validation. (b) Co‐immunoprecipitation (co‐IP) analysis confirming the interaction between Fgwd101 and Fgwd072. Total protein extracts (Input) and protein immunoprecipitated using anti‐GFP magnetic beads (GFP pull‐down) were detected with anti‐GFP or anti‐FLAG antibodies. GAPDH served as a loading control for the input samples. (c) Yeast two‐hybrid (Y2H) assay confirming the interaction between Fgwd101 and Fgwd072. (d) STRING network of Fgwd133‐associated proteins identified by AP‐MS. Septin complex proteins were selected for interaction validation. (e–h) Co‐IP analysis confirming the association between Fgwd133 and septin complex proteins. Total protein extracts (Input) and proteins immunoprecipitated using anti‐GFP magnetic beads (GFP pull‐down) were detected with anti‐GFP or anti‐FLAG antibodies. GAPDH served as a loading control for the input samples. (e) Cdc3; (f) Cdc10; (g) Cdc11; (h) Cdc12. (i) Y2H assay examining the interaction between Fgwd133 and Cdc10. AD, activation domain; BD, DNA‐binding domain; SD, synthetic dropout medium.

Using the same approach, we investigated the protein interaction network associated with Fgwd133. Immunoprecipitation of GFP‐tagged Fgwd133 was confirmed by western blot analysis, detecting a protein band corresponding to the expected size of c. 86 kDa (Fig. S7b). AP‐MS analysis identified candidate Fgwd133‐associated proteins using three independent strains expressing GFP‐tagged Fgwd133 (Table S7). STRING‐based network analysis showed that Fgwd133‐associated proteins were mainly divided into ribosome‐associated proteins and septin complex proteins (Fig. 6d). The septin complex cluster included Cdc3, Cdc10, Cdc11, and Cdc12, which were selected for further validation as potential Fgwd133‐associated proteins. Co‐IP analysis confirmed the association between Fgwd133 and septin proteins (Fig. 6e–h). However, in the Y2H assay, Fgwd133 exhibited only a weak direct interaction with Cdc10, which became detectable after one additional day of incubation (Figs 6i, S7c). No direct interactions were observed with the other septin proteins (Fig. S7d). These results suggest that Fgwd133 weakly and directly interacts with Cdc10, whereas its association with the other septin components is likely mediated through the septin protein complex rather than through individual direct interactions.

WD40 domains are required for the biological functions of core WD40 proteins

We next investigated whether WD40 domain‐mediated protein interactions contribute to the biological functions of core WD40 proteins. To this end, we generated strains expressing proteins lacking specific WD40 blades in the corresponding deletion mutant backgrounds. To select WD40 blades likely to play important roles in WD40‐mediated protein–protein interactions, we used previously identified interacting proteins to guide AlphaFold prediction of interaction interfaces.

For Fgwd101, AlphaFold structural prediction identified blade 6 as part of the predicted interaction interface with the previously identified interacting protein Fgwd072 (Fig. 7a). We therefore generated a strain expressing a blade 6‐deleted version of FgWD101 (FgWD101 ^Δb6) and first examined whether deletion of this blade affected its interaction with Fgwd072. Co‐IP analysis showed that the association between FgWD101 ^Δb6 and Fgwd072 was no longer detectable, and Y2H analysis further demonstrated a marked reduction in the direct interaction between Fgwd101^Δb6 and Fgwd072 compared with the full‐length Fgwd101 protein (Figs 6c, 7b). The ability of Fgwd101^Δb6 to complement the defects of the Δfgwd101 mutant was then compared with that of the full‐length complementation strain (FgWD101c). The FgWD101c strain substantially complemented defects in vegetative growth and virulence and restored sexual reproduction compared with the Δfgwd101 mutant (Fig. 7c,d). By contrast, the FgWD101 ^Δb6 strain failed to complement defects in vegetative growth, sexual reproduction, or virulence and exhibited phenotypes indistinguishable from those of the Δfgwd101 mutant.

Fig. 7.

Functional analysis of predicted interaction‐associated WD40 blades in Fgwd101 and Fgwd048 of Fusarium graminearum. (a) AlphaFold‐predicted interaction interface between Fgwd101 (red) and Fgwd072 (purple). The predicted interaction‐associated WD40 blade (blade 6) of Fgwd101 is highlighted in yellow. (b) Co‐immunoprecipitation (co‐IP) analysis examining the association between FgWD101 ^Δb6 and Fgwd072. Total protein extracts (Input) and proteins immunoprecipitated using anti‐GFP magnetic beads (GFP pull‐down) were detected with anti‐GFP or anti‐FLAG antibodies. GAPDH served as a loading control for the input samples. (c) Vegetative growth and sexual reproduction of the wild‐type strain (Z‐3639), Δfgwd101, FgWD101c, and FgWD101 ^Δb6. CA, carrot agar; CM, complete medium. Bar, 200 μm. (d) Virulence assay of the indicated strains using wheat coleoptiles. (e) AlphaFold‐predicted interaction interface between Fgwd048 (red) and Atg2 (purple). The predicted interaction‐associated WD40 blade (blade 2) of Fgwd048 is highlighted in yellow. (f) Yeast two‐hybrid (Y2H) assay examining the interaction between Fgwd048 and Atg2. AD, activation domain; BD, DNA‐binding domain. (g) Vegetative growth and sexual reproduction of the wild‐type strain, Δfgwd048, FgWD048c, and FgWD048 ^Δb2. CA, carrot agar; CM, complete medium. Bar, 200 μm. (h) Virulence assay of the indicated strains using wheat coleoptiles. (i) Immunoblot analysis of GFP‐Atg8 processing in the indicated strains to evaluate autophagic activity. GFP‐Atg8 and free GFP were detected using an anti‐GFP antibody, and GAPDH served as a loading control. MM‐N, nitrogen‐free minimal medium.

We also attempted to generate a blade‐deleted FgWD133 strain by introducing a construct lacking the WD40 blade predicted to contribute to the interaction interface with Cdc10. However, transformants carrying the blade‐deleted construct could not be obtained. Therefore, we selected Fgwd048, another core WD40 protein previously characterized as an autophagy‐related protein in F. graminearum, to further investigate the contribution of WD40 domains to protein function. Based on the previously reported interaction between Fgwd048 and Atg2 (Obara et al., ²⁰⁰⁸), we selected blade 2 for functional analysis. AlphaFold structural prediction identified blade 2 as part of the predicted interaction interface between Fgwd048 and Atg2 (Fig. 7e). We therefore generated a strain expressing a blade 2‐deleted version of Fgwd048 (FgWD048 ^Δb2). Y2H analysis showed that deletion of blade 2 markedly weakened the interaction between Fgwd048 and Atg2 (Fig. 7f). We next examined whether FgWD048 ^Δb2 retained the biological functions of Fgwd048. Compared with the full‐length complementation strain (FgWD048c), FgWD048 ^Δb2 failed to rescue defects in sexual reproduction and virulence, exhibiting phenotypes comparable to those of the Δfgwd048 mutant (Fig. 7g,h). Since Fgwd048 has previously been shown to function in autophagy through Atg8 processing, we further investigated whether deletion of blade 2 affected this activity. GFP‐Atg8 was expressed in each strain, and Atg8 processing was evaluated by immunoblot analysis following nitrogen starvation. In the full‐length FgWD048‐expressing strains, transfer to MM‐N for 2 h resulted in efficient Atg8 processing, as indicated by the accumulation of free GFP. By contrast, GFP‐Atg8 processing was markedly reduced in FgWD048 ^Δb2, although not completely abolished (Fig. 7i).

Together, these findings indicate that disruption of the predicted interaction blades compromises both WD40‐mediated protein–protein interactions and the biological functions of the examined WD40 proteins.

Discussion

WD40 domain‐containing proteins participate in diverse biological processes through protein–protein interactions, but the functional contribution of individual WD40 proteins remains largely unexplored in filamentous fungi. In this study, we identified 157 putative WD40 genes in F. graminearum and generated a deletion mutant library containing 119 non‐essential WD40 genes. Comprehensive phenotypic profiling under 22 different conditions, including fungal development, pathogenicity, mycotoxin production, and various stress responses, allowed us to establish a genome‐wide WD40 phenome dataset. This systematic resource provides a foundation for understanding the functional diversity of WD40 proteins and exploring WD40‐mediated regulatory networks in F. graminearum.

Among the 33 putative essential WD40 genes identified in this study, 29 had essential orthologs in S. cerevisiae, suggesting that a subset of WD40 proteins retains evolutionarily conserved functions required for fundamental cellular processes. However, this conservation was not observed for all WD40 proteins, as several orthologs of essential WD40 genes in S. cerevisiae were successfully deleted in F. graminearum, with some mutants displaying alterations in specific biological processes. These observations suggest that, although some WD40 proteins maintain conserved roles across fungal species, others may have acquired species‐specific functions. Given that WD40 proteins function as scaffolds for protein–protein interactions, these differences may be associated with changes in their regulatory roles or interaction networks.

In addition to characterizing the contribution of individual WD40 proteins, the genome‐wide phenome dataset allowed us to investigate relationships among phenotypic traits affected by WD40 gene disruption. Spearman correlation analysis of the WD40 deletion mutant phenome revealed a relatively strong positive correlation between sexual reproduction and virulence. Sexual reproduction and pathogenicity represent distinct stages of the F. graminearum life cycle and involve different developmental processes (Guenther & Trail, 2005; Leplat et al., ²⁰¹³). Sexual reproduction occurs primarily on plant residues and involves the differentiation of specialized reproductive structures, including perithecium development and ascospore formation, which contribute to survival and inoculum production. By contrast, pathogenicity occurs during host infection and relies on invasive hyphal growth and colonization within host tissues. Interestingly, a positive correlation between these distinct biological processes has also been observed in a genome‐wide transcription factor mutant library of F. graminearum, in which defects in virulence were strongly correlated with impaired perithecium formation and maturation (Son et al., ^2011b). These observations suggest that the correlation between sexual reproduction and virulence is not specific to WD40 proteins but may reflect the presence of shared genetic components involved in both biological processes. Further investigation of these shared regulators may provide deeper insights into the mechanisms coordinating fungal development and pathogenicity in F. graminearum.

Beyond phenotypic characterization of WD40 proteins, we further investigated selected WD40 domain‐containing proteins, Fgwd101, Fgwd133, and Fgwd048, to explore their roles as protein–protein interaction platforms and understand how WD40‐mediated interaction networks contribute to their biological functions. AP‐MS analysis identified Fgwd072 as a major interacting partner of Fgwd101. Fgwd101 and Fgwd072 are orthologs of S. cerevisiae Sec13 and Sec31, respectively, which form a conserved protein complex through direct interaction as components of the COPII vesicle coat (Novick et al., ¹⁹⁸⁰; Salama et al., ¹⁹⁹⁷; Nie & Ma, ²⁰²⁵). The identification of the Fgwd101–Fgwd072 interaction suggests that this conserved Sec13–Sec31 interaction module is maintained in F. graminearum. Interestingly, although FgWD072 appears to be essential in F. graminearum, as deletion mutants could not be obtained, disruption of the Fgwd101–Fgwd072 interaction resulted in developmental defects rather than loss of viability. This observation suggests that the essential functions of Fgwd072 are not solely dependent on its interaction with Fgwd101. Instead, the Fgwd101–Fgwd072 interaction likely contributes to a subset of COPII‐associated processes required for fungal development and pathogenicity, whereas other interaction partners or functions of Fgwd072 may be sufficient to support essential cellular processes.

In contrast to Fgwd101, whose conserved interaction with its orthologous partner was confirmed in this study, Fgwd133 exhibited a previously unrecognized interaction profile. Fgwd133, an ortholog of the WD40 repeat‐containing Eed protein, has been reported as a component of the Polycomb Repressive Complex 2 (PRC2; Connolly et al., ²⁰¹³). Likewise, Kmt6, Suz12, and Eed have been shown to form a PRC2 complex in F. graminearum (Tang et al., ²⁰²¹). However, our AP‐MS analysis identified components of the septin complex rather than previously reported PRC2‐associated proteins as major interaction partners of Fgwd133. Because the interaction between Fgwd133 and Cdc10 was further supported by both Co‐IP and Y2H analyses, these findings are unlikely to reflect experimental artifacts. Instead, they likely reflect the context‐dependent nature of protein interaction networks, which are influenced by developmental stage, cellular status, and environmental conditions. These observations further suggest that Fgwd133 may participate in distinct protein complexes depending on the biological context.

We further investigated whether the WD40 domain itself contributes to the observed phenotypes and biological functions of WD40 proteins using Fgwd101 and Fgwd048 as representative examples. Based on experimentally identified or previously reported protein interactions, we selected WD40 blades predicted to play important roles in protein–protein interactions. In both proteins, deletion of the selected blade markedly weakened protein interactions and abolished recovery of the corresponding mutant phenotypes, demonstrating that the WD40 domain contributes substantially to the biological functions of these proteins. However, the observed defects cannot be attributed solely to disruption of the representative protein interactions used to identify the candidate blades. Because WD40 proteins function as molecular scaffolds that coordinate interactions with multiple binding partners, disruption of a single interaction‐associated blade is likely to affect additional protein interactions beyond those examined in this study. Although the specific protein interactions underlying the observed phenotypes remain to be identified, our findings support an important role for WD40‐mediated protein interactions in the biological functions of WD40 proteins.

Interestingly, although deletion of the predicted interaction‐associated blades markedly reduced the representative protein interactions in both Fgwd101 and Fgwd048, weak interactions between each WD40 protein and its representative interacting partner remained detectable. Nevertheless, neither blade‐deletion strain restored the biological functions of the corresponding deletion mutant. These observations suggest that individual WD40 blades make major contributions to protein–protein interactions but are unlikely to function as independent interaction modules. Instead, stable protein interactions are likely maintained through the cooperative architecture of the WD40 β‐propeller, in which multiple blades collectively contribute to interaction interfaces.

Collectively, our findings reveal the functional diversity of WD40 proteins in F. graminearum and highlight their potential roles as versatile platforms for protein complex formation. By combining genome‐wide phenotypic profiling with protein interaction analysis and functional validation of WD40 interaction‐associated blades, this study expands our understanding of how WD40 proteins contribute to diverse biological processes in filamentous fungi.

Competing interests

None declared.

Author contributions

SC, NL, Y‐WL, D‐GS and HS conceived the study and designed the experiment; SC, NL, JP, HJ, SK, J‐EK, JS, HM, KM, YC, AH, HK and GJC performed experiments and collected the data; SC, NL, JP and HJ analyzed the data; SC, NL, JP, HJ, D‐GS and HS wrote the manuscript. SC and NL contributed equally to this work.

Disclaimer

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.

Supporting information

Fig. S1 Gene Ontology classification and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis of WD40 proteins in Fusarium graminearum.

Fig. S2 Southern blot validation of WD40 gene deletion mutants and the GFP‐Atg8 strain in Fusarium graminearum.

Fig. S3 Perithecial cirrhus formation on carrot agar medium at 9 d post‐inoculation.

Fig. S4 Assessment of male and female fertility in WD40 gene deletion mutants of Fusarium graminearum.

Fig. S5 Virulence of Fusarium graminearum WD40 deletion mutants on wheat heads.

Fig. S6 Phenotypic correlation analysis of Fusarium graminearum WD40 gene deletion mutants.

Fig. S7 Immunoprecipitation of GFP‐tagged Fgwd101 and Fgwd133 for affinity purification‐mass spectrometry analysis and yeast two‐hybrid assays examining interactions between Fgwd133 and septin complex proteins.

Table S1 List of oligonucleotides used in this study.

Table S2 List of 157 WD40 genes in Fusarium graminearum.

Table S3 Gene Ontology distribution of WD40 genes of Fusarium graminearum.

Table S4 Kyoto Encyclopedia of Genes and Genomes pathway enrichment of WD40 proteins in Fusarium graminearum.

Table S5 Phenotypes of 119 WD40 deletion mutants in Fusarium graminearum.

Table S6 The result of affinity purification‐mass spectrometry for Fgwd101.

Table S7 The result of affinity purification‐mass spectrometry for Fgwd133.

Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

Data availability

All data supporting the findings of this study are included in this article and its Supporting Information (Figs S1–S7; Tables S1–S7).