We used the amino acid sequences of Saccharomyces cerevisiae Stu2 (Gene bank ID: 856736) and S. pombe Dis1 (Gene bank ID: 2539505) to find the Dis1/Stu2/XMAP215 homologs in F. graminearum (FgStu2 gene bank ID: 23557431). The protein blast was carried out by online blast services of the National Center for Biotechnology information (NCBI¹) and the Ensembl Fungi². The presumed FgNdc80 gene (FGSG_09262, gene bank ID: 23556224) was also identified by online blast services of the NCBI. The phylogenetic analysis (phylogenetic tree) of Dis1/Stu2/XMAP215 homologs of Aspergillus nidulans (Protein ID: XP_663125.1), F. graminearum (Protein ID: XP_011319522.1), Neurospora crassa (Protein ID: XP_956946.3), S. cerevisiae (Protein ID: NP_013146.1) and S. pombe (Protein ID: NP_587785.1) was finished by the Molecular Evolutionary Genetics Analysis software (MEGA, version: 7.0) (Supplementary Figure 1). The domain prediction was conducted by the online service of Simple Modular Architecture Research Tool (SMART³) (Letunic et al., 2014). Multiple sequences alignments of tumor-overexpressed gene (TOG) domains and terminal areas in Dis/Stu2/XMAP215 homologs of A. nidulans (Protein ID: XP_663125.1), F. graminearum (Protein ID: XP_011319522.1), N. crassa (Protein ID: XP_956946.3), S. cerevisiae (Protein ID: NP_013146.1), and S. pombe (Protein ID: NP_587785.1) were performed using the online service⁴ (Figures 1B,C and Supplementary Figure 2). The amino acids sequences alignment of presumed FgNdc80 (protein ID: XP_011328506.1) and budding yeast Ndc80 (protein ID: NP_012122.3) was performed using the same online service (Supplementary Figure 3). The alignment images were generated by the online service of ESPript 3.0⁵ (Robert and Gouet, 2014).

All strains used in this study are listed in Table 1. The PH-1 strain is the standard strain of F. graminearum which was completely sequenced in Cuomo et al. (2007). In the recent years, PH-1 strain has become a model strain in plant pathology and fungal genetics (Liu et al., 2017; Chen et al., 2019; Jiang et al., 2020). More importantly, the genetic transformation and fluorescence labeling technologies of PH-1 are now perfect and convenient. Taken together, we would like to use PH-1 strain to study the functional role of Dis1/Stu2/XMAP215 homolog in plant pathogenic fungi and the results would be more representative than the other wild-type strain. Conidia and mycelia were produced in mung bean liquid (MBL) medium and yeast extract peptone dextrose (YEPD) medium at 25°C respectively. For colonial morphology, strains were grown on complete medium (CM) and minimal medium (MM) plants at 25°C (Gu et al., 2015). For sensitivity assays, all strains were grown on potato dextrose agar (PDA) containing carbendazim (MBC) at 25°C. DON mycotoxin was produced in trichothecene biosynthesis induction (TBI) medium (Gardiner et al., 2009). MBC were dissolved in 0.1 M HCl at 10 mg/ml as stock solutions (Chen et al., 2009).

The Fgβ₁-tubulin-EGFP, Fgβ₂-tubulin-EGFP single labeling strains (Table 1) were stored in our lab. The FgStu2-3 × FLAG strain was constructed using homologous recombination strategy (Supplementary Figure 4), first a 1.0 kb fragment of FgStu2 3′-terminus coding sequence was amplified from genomic DNA of PH-1 by primer containing 1 × FLAG-tag sequence. This fragment was used as a template for another two rounds of PCR to add 2 × FLAG-tag. Second, a 1.2 kb geneticin (G418) fragment was amplified from pNEO. Third, a 1.0 kb FgStu2 downstream fragment was amplified from genomic DNA of PH-1. Finally, the three fragments were fused by DJ-PCR and used as a template for amplify the FgStu2-3 × FLAG vector. Transformates were identified by PCR, sequencing and western blot (Figure 2 and Supplementary Figure 4). For constructing the FgStu2-Si-β1EGFP and FgStu2-Si-β2EGFP strains, we cloned the Fgβ1-tubulin and Fgβ2-tubulin coding sequences (CDS) into PYF11 plasmid and then transformed these plasmids into protoplasts of FgStu2-Si mutant to generate the FgStu2-Si-β1EGFP and FgStu2-Si-β2EGFP strains (Bruno et al., 2004). For constructing the Fgβ2EGFP-Stu2mRFP strain, we cloned RFP-Stu2 DNA fragment into PYF11 plasmid and then transformed this plasmid into protoplasts of Fgβ₂-tubulin-EGFP strain to generate the Fgβ2EGFP-Stu2mRFP strain. The Fgγ-tubulin-EGFP and FgNdc80-EGFP strains were constructed using homologous recombination strategy. First, a 1.0 kb 3′-terminal CDS of Fgγ-tubulin (Gene bank ID: 23556916) and presumed FgNdc80 gene (Gene bank ID: 23556224) was amplify from the genomic DNA of wild-type strain PH-1; a 0.7 kb EGFP CDS and 1.7 kb hygromycin resistance gene (HPH) fragment were amplified from PYF11 plasmid and pNEO plasmid respectively, and then fused the two fragments by double-joint PCR to generate a 2.4 kb GFP-HPH fragment; a 1.0 kb downstream fragment of Fgγ-tubulin and presumed FgNdc80 gene was amplified from the genomic DNA of PH-1. Finally, fused the three fragments by double-joint PCR to generate the Fgγ-tubulin-EGFP and FgNdc80-EGFP vectors and then transferred the vectors into protoplasts of PH-1. The transformates were identified by PCR, sequencing, and western blot (Figure 3 and Supplementary Figure 5). Protoplasts transformations were performed as previously described (Zheng et al., 2014).

Strains containing EGFP or 3 × FLAG tag were inocubated into MBL medium for 3 days. After that, fresh spores were inoculated into 100 ml YEPD medium for 24 h at 25°C with agitation (175 rpm) and then fresh mycelia were collected and ready for use. For total protein extraction, fresh dried pressed mycelia were quickly frozen in liquid nitrogen and ground into powder. For each strain, 0.15 g frozen mycelial powder was suspended in 1.5 mL of extraction buffer (20 mM Tris, 200 mM NaCl, 0.1% Triton X-100, 1 × protease inhibitor cocktail [20124ES10, YiSheng Biotechnology Inc., Shanghai, China], PH = 7.5). After incubation on ice for 30 min, cell debris was spun down at 12000g for 20 min at 4°C, and the clear supernatant was transferred to a new tube. The Co-IP assay was conducted follow the standard immunoprecipitation protocol of protein G magnetic beads (SureBeads^TM, Bio-Rad Laboratories, lnc).

Protein samples was added 5 × Loading buffer and boiled for 5 min before loaded into Sodium dodecyl sulfate-polyacrylamide gel. Protein separated by Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was transferred to polyvinylidene fluoride membrane (PVDF) with a Bio-Rad electroblotting device. After that, the PVDF membrane was incubated with 5% defatted milk (5 g non-fat dry milk diluted into 100 ml TBST buffer [10 mM Tris, 100 mM NaCl, 0.1% Tween 20, PH = 7.4]) and blocking for 1 h at 25°C. After Blocking, antibody (polyclonal Anti-EGFP Rabbit antibody, monoclonal Anti-FLAG mouse antibody and monoclonal Anti-Actin antibody were purchased form Zhengneng Biotechnology, Inc., Chengdu, China) was added into the defatted milk. After incubation at 25°C for 2 h, the PVDF membrane was washed with TBST buffer for 5 min with 3 buffer changes and then incubated for 1 h using horse radish peroxidase (HRP)-conjugated secondary antibody diluted into 5% defatted milk. Finally, the PVDF membrane was washed three times by TBST and ready for blotting.

To study the functional role of FgStu2 gene, we substituted original FgStu2 promoter with an inducible promoter Pzear (Promoter of the FGSG_04581.3 gene) to generate FgStu2 silent mutant (FgStu-Si) as previously described (Lee et al., 2010). The schematic of promoter replacement was shown in Figure 4. First, a 1.2 kb geneticin resistance gene (G418) and a 0.8 kb Pzear fragment were amplified from plasmid pNEO and genomic DNA of PH-1 respectively and fused the two fragments to obtain the G418-Pzear fragment by double joint polymerase chain reaction (DJ-PCR) as previously described (Yu et al., 2004; Ren et al., 2014). Second, a 1.0 kb upstream region of FgStu2 original promoter (Pstu2) and a 1.0 kb FgStu2 coding sequence were amplified from genomic DNA of PH-1. Finally, the three fragments (upstream region of Pstu2, G418-Pzear and 1.0 kb FgStu2 coding sequence) were fused by DJ-PCR to obtain the Pstu2 substitution vector which was used for protoplast transformation. To induce Pzear replacement, 30 μM zearalenone (ZEA, Sigma Aldrich, St. Louis, MO, United States) was added to the medium during the regeneration (Lee et al., 2010).

To test the expression level of FgStu2 gene, the mycelium was harvested after cultivation for 36 h in YEPD medium. This time is sufficient for gaining enough mycelium and distinguishing the expression level of FgStu2 gene in wild-type strain and FgStu2-Si mutants. To test the expression level of FgTri5 and FgTri6 genes, the mycelium was harvested after three-days cultivation in TBI medium at 28°C in the dark (Gardiner et al., 2009). The reason why we chose three days is that the level of DON increases obviously between 48 and 72 h and then increase slowly after 72 h and reaches the stationary phase (Tang et al., 2018). Total RNA extraction was follow the protocol of Total RNA Extraction Kit (Tiangen, China, DP419). First cDNA was synthesized by HiScript^® II Reverse Transcriptase (Vazyme, Nanjing, China, R223-01). The ChamQTM SYBR^® qPCR Master Mix (Vazyme, Nanjing, China, Q311-02/03) was used for Quantitative real-time PCR (qRT-PCR), which was conducted at CFX Connect Real-Time System (Bio-Rad) with the following procedure: 95°C 30 s; following 40 cycles of 95°C 10 s and 60°C 30 s; and a melt curve step of 95°C for 15 s and 60°C for 60 s, followed by 71 cycles of gradual increase in temperature from 60 to 95°C for 13 s/cycle. The primers used for qRT-PCR analysis are listed in Supplementary Table 1. The endogenous housekeeping gene Fgactin was used for normalization.

For photograph the colonial morphology, 5-mm mycelial plugs from the margin of a 3-day-old colony (on PDA plates) of wild-type strain PH-1 and FgStu2 silent mutant were transferred on PDA, MM, and CM plates with or without 30 mM ZEA and keep it at 25°C for three days. After that, the colonial morphology was photographed. For hyphal morphology, wild-type strain PH-1 and FgStu2 silent mutants were inoculated on PDA plates with or without 30 μM ZEA for 24 h. After that, the margin of the small colony was photographed using an Olympus IX-71 inverted microscope.

To observe the localizations of FgStu2 and Fgβ1-tubulin, Fgβ2-tubulin, the fresh spores of Fgβ2EGFP-Stu2mRFP were inoculated into 20 ml YEPD medium with or without 30 μM ZEA to germinate at 25°C for 10-12 h and then pipette 10 μl medium onto the slide for micro examination. This experiment was conducted by a LEICA TCS SP8 confocal laser-scanning microscope (LEICA).

The growth rate was tested as previously described (Zhu et al., 2018). In brief, 5 mm mycelial plugs taken from the margin of a 3-day-old colony of wild-type strain PH1 and FgStu2 silent mutants were transferred into PDA plates with or without 30 μM ZEA and incubated at 25°C. After three days cultivation, the colony diameters of wild-type strain and FgStu2-Si mutants were significantly different. Then the diameters of each strain were measured to calculate the mean growth rates. This experiment was conducted three independent times.

Radial growth was used to test the sensitivity of the wild-type strain PH-1 and FgStu2 mutants to MBC as previously described (Zhu et al., 2018). Colony diameter was measured after 3 days for PH1; after 7 days for FgStu2-Si mutants. The MBC was added to PDA plates at 0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 μg/ml for PH-1 and FgStu2-Si mutant; at 5, 10, 15, 20, and 25 μg/ml for Fgβ2F167Y and F167Y-FgStu2-Si mutant.

In order to assess the ability of asexual reproduction of FgStu2-Si mutants, we observed the conidial morphology and tested the number of conidial production of FgStu2-Si mutant as previously described (Zhu et al., 2018). For observe the septa and nuclei, fresh spores were stained with calcofluor white (CFW) and 4′,6-diamidino-2-phenylindole (DAPI) and examined with a LEICA TCS SP8 confocal laser-scanning microscope (LEICA). The numbers of conidia produced by the wild-type strain PH-1 and FgStu2-Si mutants were calculated using hemocytometer as previously described (Zheng et al., 2014). This experiment was conducted three independent times.

Sexual reproduction test was performed as previous described (Bui et al., 2016). Briefly, inoculate fresh mycelia of each stain to carrot medium plates with or without 30 μM ZEA and keep it at 25°C incubator until the plates are full of mycelia. Then clean out the mycelia and add 2.5% Tween 60 solution to the plates. Keep the plates into 25°C incubator with near-UV light for about one or two weeks until the black perithecia are appeared and then photograph the plates by stereomicroscope. This experiment was conducted three independent times.

In order to compare the pathogenicity of the wild-type strain PH-1 and FgStu2-Si mutants we preformed wheat coleoptile infection assay as previous described (Zhang et al., 2012). In brief, Huaimai 33 wheat seeds were placed in culture dishes and placed it into a 25°C chamber with light. After 3 days, remove the 2-3 mm tip of all coleoptiles and the wound was inoculated with 2.5 μl (1 × 10⁶ spores/ml) fresh conidia, which was collected from MBL medium with or without 30 μM ZEA. Coleoptiles inoculated with MBL medium with or without 30 μM ZEA were used as control. After 7 to 10 days cultivation, the infected coleoptiles were photographed and the length of lesion was measured. At least 20 wheat seedlings were measured for each strain. This experiment was conducted three independent times.

Deoxynivalenol production of FgStu2-Si mutants were determined by enzyme linked immunosorbent assay (ELISA) followed the instruction of DON detection kit (Weisai Technology, Jiangsu, China). Briefly, two fresh 5 mm mycelial plugs taken from the margin of a 3-day-old colony of wild-type strain PH-1 and FgStu2-Si mutants were transferred into 30 ml TBI medium with or without 30 μM ZEA and cultured at 28°C, 175 rpm in the dark (3 repeats were set for each strain). After 7 days cultivation, the medium and mycelium was collected. The medium was used for DON determination, the mycelium was dried in the dryer for 48h. After that the dry weight of the mycelium was used for calculating the DON production (μg DON/g mycelium). The DON production of wild-type strain PH-1 was used as reference to generate relative DON production of FgStu2-Si mutants. The experiment was repeated three independent times.

For visualize the localization FgTri1p (also known as “toxisomes”) in FgStu2-Si mutant and wild-type strain PH-1, we cloned FgTri1 gene into PDL2 plasmid as previous described (Liu et al., 2017). The recombinant plasmid was identified by PCR, and sequencing. Then we transformed the recombinant plasmid into protoplasm of FgStu2-Si mutant and wild-type strain PH-1 to generate FgStu2-Si-Tri1-EGFP and PH-1-Tri1-EGFP strains. The transformants were identified by PCR, sequencing and fluorescence microscope. The identified FgStu2-Si-Tri1-EGFP and PH-1-Tri1-EGFP strains were inoculated into TBI medium and incubated for 2 days at 28°C, 175 rpm in the dark, the fresh mycelium were used for fluorescence observation. This experiment was conducted by a LEICA TCS SP8 confocal laser-scanning microscope (LEICA). This experiment was repeated three independent times.

The protein blast results of NCBI and Ensembl Fungi website suggested that the FGSG_10528 (gene bank ID: 23557431) is the Dis1/Stu2/XMAP215 family homolog gene in F. graminearum. The FgStu2 gene is predicted to encode an 888 amino acids protein (FgStu2p) which is similar with the lengths of S. cerevisiae Stu2 and S. pombe Dis1 (Figure 1A). Sequence analysis showed that FgStu2p contains two TOG domains at the amino terminal, TOG1 domain (Position 1 to 233) and TOG2 domain (position 273 to 505) (Figure 1A). The FgStu2-TOG1 domain shares only 32% identity with that of S. cerevisiae Stu2 (ScStu2) and S. pombe Dis1 (SpDis1). The FgStu2-TOG2 domain shares only 26 and 33% identity with that of S. cerevisiae Stu2 and S. pombe Dis1. However, the TOG1 domain of FgStu2p share 63 and 71% identify with that of A. nidulans alpA and Dis1/Stu2/XMAP215 homolog of N. crassa respectively (Figure 1B). The TOG2 domain of FgStu2 share 68 and 74% identity with that of A. nidulans alpA and Dis1/Stu2/Xmap215 homolog of N. crassa respectively, which are two folds higher than that of S. cerevisiae Stu2 and S. pombe Dis1 (Figure 1C). Moreover, The carboxyl terminal part of FgStu2p (position 506 to 877) shares 46 and 52% identity with that of A. nidulans alpA (position 510 to 891) and N. crassa stu2 homolog (position 509 to 941) which are nearly three times higher than the 18% identity with that of S. cerevisiae Stu2 (position 567 to 888) and S. pombe Dis1 (position 539 to 882) (Supplementary Figure 2).

The Fgβ-tubulin-GFP and RFP-FgStu2 double labeling experiment indicated that most of the FgStu2p was localized at cytoplasm and some of them were co-localized with free tubulin heterodimers which were not polymerized into microtubules. Another part of FgStu2p was co-localized with microtubules which contain Fgβ1-tubulin or Fgβ2-tubulin (Figure 2A). However, we could not identify the specific location of FgStu2p in the microtubule such as the microtubule plus end. The result of co-immunoprecipitation (Co-IP) assay demonstrated that the FgStu2p binds to tubulin heterodimers which contain Fgβ1-tubulin and Fgβ2-tubulin (Figure 2B). In addition, our Co-IP assay also suggested that FgStu2p binds to Fgγ-tubulin and FgNdc80p, component of kinetochore, which was previously reported in fission yeast (Figure 3; Hsu and Toda, 2011).

After attempted for several times, we failed to obtain FgStu2 gene deletion mutant by homologous recombination method (data not shown). We therefore used promoter replacement strategy to study the genetic function of FgStu2 as previously described (Figure 4A). The transformants were identified by PCR and southern blot. The results indicate that the promoter of FgStu2 gene has been replaced by G418 resistance gene and zear promoter (Figures 4B,C). The expression level of FgStu2 promoter replacement mutants (FgStu2-Si mutants) was significantly decreased compared with those of wild-type strain PH-1 and Fgβ2F167Y strain. Addition of 30 μM ZEA partially restored the FgStu2 gene expression level (Figure 4D). These results indicate that we have successfully obtained the FgStu2-Si mutants.

The FgStu2-Si mutants produced twisted hyphae compared with that of the wild-type strain PH-1. Addition of 30 μM ZEA restored the hyphal morphology (Figure 5A). Moreover, the FgStu2-Si mutants grew significantly slower than the wild-type strain PH-1 on PDA, MM, and CM plates. Addition of 30 μM ZEA partially restored the growth rate (Figure 5B and Table 2). Besides, microtubule network was sparse in FgStu2-Si mutants, which was recovered by addition of 30 μM ZEA (Figure 6).

We tested MBC sensitivity of FgStu2 promoter replacement mutants. The effect concentration against 50% mycelia growth (EC₅₀ value) of FgStu2-Si mutants was slightly decreased compared with those of wild-type strain PH-1 (Table 2). However, the FgStu2-Si mutants cannot grow on PDA containing 0.8 μg/ml MBC. When the PDA plates was added 30 μM ZEA, the EC₅₀ values FgStu2-Si mutants were restored and even a bit higher than that of wild-type strain PH-1 (Supplementary Figure 6A). The same phenomenon was observed in FgStu2 promoter replacement mutant generated from Fgβ2F167Y (Fgβ2F167Y-Stu2-Si), an MBC resistant strain reported previously. The EC₅₀ value of Fgβ2F167Y-Stu2-Si (4.28 ± 0.32 μg/ml) was decreased about 30% compared with that of parental strain Fgβ2F167Y (6.48 ± 0.27 μg/ml) (Table 3). Moreover, the Fgβ2F167Y-FgStu2-Si cannot grow on PDA plates containing 20 μg/ml MBC (Supplementary Figure 6B).

The FgStu2 gene promoter replacement mutant showed curved conidia and decreased number of conidial production compared with those of the wild-type strain PH-1. While the conidial morphology and number of conidial production were restored after addition of 30 μM ZEA (Figures 7A,C). Moreover, the FgStu2-Si mutants lost ability to produce perithecia. Addition of 30 μM ZEA restored the ability of perithecia production (Figure 7B).

We performed the coleoptile infection assay to assess the rule of FgStu2 gene in pathogenicity of F. graminearum. The lesion length of coleoptiles inoculated with conidia of FgStu2-Si mutants was decreased about 35% compared with those of the wild-type strain PH-1 (Figure 8A). However, the conidia of FgStu2-Si mutants collected from MBL medium containing 30 μM ZEA did not cause the same size of lesion (Figures 8A,B).

The DON production was significantly decreased in FgStu2-Si mutants compared with that of wild-type strain PH-1 (Figure 9A). However, addition of ZEA didn’t restore the DON production of FgStu2-Si mutants and decreased the DON production of PH-1. Moreover, expression levels of trichothecene biosynthesis pathway genes (Tri gene) including FgTri5 and FgTri6 were significantly decreased in FgStu2-Si mutants (Figure 9B). The localization of Tri1-EGFP, also known as “toxisome,” in FgStu2-Si mutant was similar with that of wild-type strain PH-1 (Figure 9C).