To identify putative members of the SAGA complex in V. dahliae isolate AT13, we first searched the V. dahliae genome sequence by BlastP using protein sequences of the known S. cerevisiae SAGA complex members as queries. Using this approach, nearly all SAGA homologs in V. dahliae were identified except Sus1, one of the subunits belonging to DUB module. Among the 18 subunits identified, Ada1 (FFUJ_07552) and Sgf11 (FFUJ_06279) were identified as present in F. fujikuroi (Rösler et al., 2016) since there were no homologs in S. cerevisiae (Figure 1A).

Given the importance of Ada1 in SAGA complex, the function of VdAda1 in V. dahliae was further explored. The full VdAda1 (VDAG_08660) gene is 1860 bp in length, encoding a protein with 539 amino acids with a conserved SAGA_Tad domain. The predicted protein sequence of VdAda1 shares 75.3% identity with the sequence of Ada1 from F. fujikuroi (Figure 1A).

Although a large number of studies have shown that the SAGA complex and Ada1 homologs play an important role in multiple physiological processes of fungi, the potential role of VdAda1 in V. dahliae-plant interactions has not been explored. To assess the relationship between VdAda1 expression and virulence in V. dahliae, the expression pattern of VdAda1 was monitored by reverse transcription-quantitative PCR (RT-qPCR) at different times following inoculation. The results indicate that the expression of VdAda1 was induced by cotton, especially at 3 dpi (Figure 1B), suggesting that VdAda1 may play an important role in the virulence of V. dahliae.

To further understand the role of Ada1 subunit during V. dahliae – host interactions, we generated deletion mutants of Ada1 subunit (ΔVdAda1) and complemented transformants (EC_VdAda1) of the ΔVdAda1 mutant (Supplementary Figure 1). Since Ada1 homologs in yeast are involved in vegetative growth (Horiuchi et al., 1997; Sterner et al., 1999), we assessed the growth of ΔVdAda1 strain in comparison with the wild type (AT13) and EC_VdAda1 strains on different carbon sources. The ΔVdAda1 strain grew significantly slower than the wild type strain on medium containing four different carbon sources (sucrose, starch, pectin, and CMC-Na) (Figures 2A,B). In fact, the colony diameter of the ΔVdAda1 strain was less than half of that of AT13, whereas the defect was remedied in the EC_VdAda1 strain (Figures 2A,B). We also examined the sensitivity of ΔVdAda1 strain to cell wall integrity and osmotic stressors including sorbitol, Congo red, and sodium dodecyl sulfate (SDS). The sensitivity of ΔVdAda1 to these stress factors was not different from that of the wild type strain (Figure 2C) nor was there a statistically significant difference between any of the strains (Figure 2D), suggesting that VdAda1 is not involved in the regulation of cell wall integrity and osmotic stress response in V. dahliae.

Analyses of conidia harvested from Czapek plates at 7 days post incubation revealed that the ΔVdAda1 strain did not produce conidia while the EC_VdAda1 strain produced as many conidia as the wild type (Figure 2E). After placing each strain onto hydrophobic cover slips, we observed the hyphae and conidia on cover slips under a fluorescence microscope at 5 dpi and found that only hyphae but no conidia could be observed in the ΔVdAda1 strain. In the wild type and EC_VdAda1 strains, we observed both hyphal growth and production of conidia (Figure 2F). However, the morphology of hyphae in ΔVdAda1 strain was not different from that of wild type and EC_VdAda1 strain (Figure 2F). Collectively, these data demonstrate that VdAda1 is not involved in the regulation of cell wall integrity but is necessary for vegetative growth and conidia production.

Fungal melanins can enhance resistance to environmental stresses (Wang et al., 2018; Eisenman et al., 2020) and may also modulate immune responses of the host (Chai et al., 2010; Volling et al., 2011). Though DHN melanin itself is not a virulence factor in V. dahliae (Wang et al., 2018), production of melanin and microsclerotia are tightly linked in V. dahliae and some of the pathways that regulate melanin biosynthesis also influence virulence in V. dahliae (Fan et al., 2017). To investigate the effect of VdAda1 on melanin synthesis and microsclerotia formation in V. dahliae, the wild type, ΔVdAda1, and EC_VdAda1 strains were incubated on PDA medium and V8 medium at 25°C for 2 weeks. The ΔVdAda1 strain failed to produce melanin on PDA medium or on V8 medium while the wild type strain and EC_VdAda1 strain were clearly melanized (Figure 3A). The conidial suspension of AT13 strain and EC_VdAda1 strain together with protoplasts of ΔVdAda1 strain were adjusted to approximately 1 × 10⁶ propagules ml^–1 and coated on microsclerotia induction medium (BMM medium) overlaid on a cellophane layer. Cultures of these strains at 25°C for 2 weeks followed by microscopy revealed that the AT13 strain formed many melanized microsclerotia, while the ΔVdAda1 strain did not produce melanin or microsclerotia (Figure 3B). Melanized microsclerotial production was restored in EC_VdAda1 strain (Figure 3B). Analyses of gray-scale curves using Image J software verified the defect in melanin synthesis in the ΔVdAda1 strain (Figure 3C).

Numerous genes in V. dahliae have been characterized for their roles in melanin synthesis and microsclerotia formation (Klimes et al., 2015; Zhang et al., 2017; Wang et al., 2018). The expression of melanin biosynthetic genes is tightly correlated with microsclerotia development (Duressa et al., 2013). RT-qPCR expression analyses of five genes involved in melanin synthesis in V. dahliae revealed that the majority of these genes were significantly downregulated in the ΔVdAda1 strain and that their expression in complemented strains was restored to the wild type level (Figure 3D). These results suggested that VdAda1 plays an essential role in melanin synthesis and microsclerotia formation in V. dahliae by regulating the transcription of genes involved in these processes.

To investigate the role of VdAda1 during the initial infection of V. dahliae, we analyzed the penetration abilities of VdAda1 by incubating wild type, ΔVdAda1, and VdAda1 complemented transformants on a cellophane membrane overlaid on minimal medium. At 3 dpi, the cellophane membrane was removed and hyphal penetration into the cellophane membrane was examined at 8 dpi (Figure 4A). The results in Figure 4A suggest that VdAda1 is not involved in penetration. To analyze the role of VdAda1 on virulence, seedlings of cotton (Gossypium hirsutum cv. Junmian 1) and Nicotiana benthamiana were inoculated with hyphae from cultures of the wild type, ΔVdAda1, and EC_VdAda1 strains on PDA. The results revealed that the ΔVdAda1 strain had lost the ability to infect cotton (Figure 4B). There was also a significant reduction in the biomass of the ΔVdAda1 strain in inoculated cotton plants (Figure 4C). The ΔVdAda1 strain was also non-pathogenic on N. benthamiana (Figure 4D). Virulence and fungal biomass measurements were restored in the VdAda1 complemented strains (Figures 4B–E).

The SAGA complex primarily is described as an acetylation modifier of certain lysine residues on histones that regulates gene expression (Fischer et al., 2019). Ada1, as a member of the core module in the SAGA complex, plays an essential role in ensuring the structural integrity of the SAGA complex. In this study, several putatively relevant acetylation marks at histone 3 (H3) were analyzed at a global level. To investigate the involvement of Ada1 in acetylation of H3 in V. dahliae, the wild type, ΔVdAda1, EC_VdAda1 strains were tested in western blot analyses using the anti-H3K9ac, -H3K18ac, -H3K27ac antibodies, which can detect acetyl modifications. H3K27 acetylation was significantly reduced in the deletion mutants compared to wild type strains (Figure 5A). H3K9 and H3K18 acetylation were nearly lost in the ΔVdAda1 strain, while the acetylation levels in the EC_VdAda1 strain was restored to the wild type levels (Figure 5A). We confirmed that VdAda1 deletion resulted in the impaired ability to acetylate histone 3, and likely influenced gene expression.

To further study the mechanisms of how VdAda1 affects histone 3 acetylation, we examined the transcript levels of other SAGA members in ΔVdAda1 strain by RT-qPCR. Interestingly, the expression of almost all the SAGA subunits were induced after deletion of VdAda1 (Figure 5B). The results revealed the important role of VdAda1 in maintaining structural integrity of the SAGA complex, thus ensuring the ability to acetylate histone.

To investigate the regulatory targets of VdAda1 in V. dahliae, transcriptomes of ΔVdAda1 and wild type strain AT13 were compared after 5 days of growth on Czapek medium. For these analyses, differentially expressed genes (DEGs) between the wild type and ΔVdAda1 strains were considered significant with the cutoff log₂ ratio ≥2 and P-value of <0.01. The RNA-seq analysis identified 1671 DEGs in total, including secreted proteins, small cysteine-rich proteins (SCRPs), CAZymes, protein kinase, transcription factors (TFs), and genes responsible for serine hydrolase activity, gene expressed in mitochondria and those involved in responses to stimuli (Figure 6A). GO annotation also indicated that DEGs mainly involved in the signal transduction, cellular response to stimulus, cell communication, response to stimulus, and signaling biological processes (Supplementary Figure 2). Among these DEGs, secreted proteins (secretome) accounted for the largest proportion (DEGs of secretome/1671 DEGs, 14.1%), followed by CAZymes (DEGs of CAZymes/1671 DEGs, 7.1%) and small cysteine-rich proteins (SCRPs, and DEGs of SCRPs/1671 DEGs, 5.4%). Of the 293 SCRPs encoded in AT13 genome, 91 genes (accounting for 31.1% DEGs of SCRPs/293) were differentially expressed, while 29.7% (DEGs of secretome/794 secreted proteins in DK185 genome) of the secreted proteins were differentially expressed, and thus were the two types of genes most significantly affected by the lack of VdAda1 (Figure 6B). Furthermore, the ratio between downregulated and upregulated DEGs was obviously unbalanced (downregulated total DEGs/upregulated total DEGs, 1.3) (Figure 6B), indicating that VdAda1 positively regulated expression of many genes. However, the downregulation of transcription factors in the ΔVdAda1 strain were 2.7 times that of upregulated transcription factors (Figure 6B). The ratio of downregulated and upregulated secreted proteins was 1.9, while the ratio of downregulated and upregulated SCRPs was 2.1 (Figure 6B). The heat map further illustrated that particular transcription factors were significantly up or downregulated by VdAda1, and most of these belonged to the family of Zn₂Cys₆ TFs (Figure 6C). Moreover, the expression levels of the DEGs compared to the total DEGs of ΔVdAda1 strain further confirmed that the expression levels of secreted proteins and SCRPs was different from the total DEGs, and an increased proportion was genes downregulated in the ΔVdAda1 strain versus the wild type (Figure 6D). To confirm the role of VdAda1 in regulating genes encoding secreted proteins and transcription factors, the expression of selected genes was quantified in the ΔVdAda1 strain relative to the wild type by RT-qPCR. The results revealed that VdAda1 showed a significant regulatory effect on the expression of these genes (Figure 6E). In summary, we confirmed that transcription factors and secreted proteins are targets regulated by VdAda1, thereby affecting multiple physiological functions, including virulence.

Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations revealed that DEGs regulated by VdAda1 were enriched in 169 pathways (Supplementary Table 2). Among these, multiple genes involved in carbon metabolism (ko:01200) were downregulated in the ΔVdAda1 strain versus the wild type strain, and this may be related to the growth defect of ΔVdAda1 strain (Figure 2 and Supplementary Figure 3). We also identified DEGs involved in starch and sucrose metabolism (ko:00500), which is speculated to be required for the degradation of plant cell walls (Supplementary Figure 4), suggesting that VdAda1 regulates not only the expression of virulence-related genes but also the genes involved in growth and nutrient utilization.

The V. dahliae isolate in this study (AT13) was isolated from Acer truncatum Bunge in Shandong province, China. This isolate and its transformants were stored at –80°C in 25% glycerol and were cultured on potato dextrose agar (PDA) medium (potato, 200 g/L; glucose, 20 g/L; agar, 15 g/L) at 25°C or in a shaking incubator in complete medium (CM) (yeast extract, 6 g/L; casein acids hydrolyzate, 6 g/ml; sucrose, 10 g/L) at 25°C.

For phenotypic analysis, the equal-sized agar plugs from the wild type, deletion mutant, and the complemented strain colonies were placed in the center of 6-cm-diameter Czapek plates (NaNO₃, 3 g/L; K₂HPO₄, 1 g/L; MgSO₄⋅7H₂O, 0.5 g/L; KCl, 0.5 g/L; FeSO4, 0.01 g/L; sucrose, 30 g/L or starch, 17 g/L or pectin, 10 g/L or CMC-Na, 10 g/L; agar, 18 g/L) containing different carbon sources. The colony diameter was measured after 7 days of incubation at 25°C. To evaluate the conidial production, three agar plugs were collected from the edge of the colonies on Czapek plates after 5 days incubation using a 7-mm-diameter hole puncher, and conidia were counted by a hemocytometer after the agar plugs had been shaken in 3 mL of sterile water.

To examine the sensitivity of ΔVdAda1 to different cell wall stresses, agar plugs of similar sizes from the wild type, deletion mutant, and the complemented strain colonies were placed on PDA medium containing 1 M sorbitol, 200 μg/ml Congo red, and 0.02% SDS. After 7 days of incubation at 25°C, the colony diameters were measured and the inhibition ratio was calculated. Inhibition ratio (%) = (colony diameter on PDA medium – colony diameter on medium containing stressors)/colony diameter on PDA medium × 100%.

To investigate the effect of VdAda1 on melanin synthesis in V. dahliae, agar plugs of similar size from the colonies of the wild type, deletion mutants, and the complemented strains were incubated on PDA and V8 media at 25°C for 2 weeks. In addition, the conidial suspensions from the wild type and complemented strains together with protoplasts from the deletion mutants were adjusted to 1 × 10⁶ propagules/ml and plated on a cellophane membrane (Solarbio, Beijing, China) overlaid on basal agar modified medium (BMM) (Hu et al., 2013). After 2 weeks of incubation at 25°C, microsclerotia on the cellophane layers were photographed.

To compare hyphae and conidia of the wild type, ΔVdAda1, and EC_VdAda1 strains, the strains were plated on PDA medium and hydrophobic cover slips were inserted around the colony. After the cover slips were removed at 5 dpi, the morphology of hyphae and the production of conidia ofΔVdAda1, wild type and EC_VdAda1 strains were observed and photographed under differential interference contrast microscopy (#DM6 B & DMC6200, Leica, Germany).

To assess the ability of ΔVdAda1, wild type, and EC_VdAda1 strains to produce microsclerotia, each were cultured in a shaking incubator in liquid complete medium (CM) at 25°C for 3 days. The protoplasts of deletion mutants were cultured in a shaking incubator in liquid TB3 medium (sucrose, 200 g/L; yeast extract, 10 g/L; casamino acid, 10 g/L; glucose, 10 g/L) at 25°C for 3 days. Fungal protoplast preparation of ΔVdAda1 was conducted according to the method described by Li et al. (2019). The conidial suspensions of wild type and EC_VdAda1 strain together with protoplasts of ΔVdAda1 strain were adjusted to 10⁶ conidia or propagules/ml and coated on a cellophane membrane that was overlaid onto BMM medium in plates. After culturing at 25°C for 2 weeks, the morphology of microsclerotia on the cellophane membrane was observed under a stereomicroscope and photographed (#SMZ18, Nikon, Japan).

To identify putative compositions of the SAGA complex in V. dahliae isolate AT13, the V. dahliae genome sequence was searched with the BlastP algorithm on the JGI website using protein sequences of known members of the SAGA complex in S. cerevisiae S288C as queries. The Sgf11 and Ada1 homologs were identified in F. fujikuroi. The Ada1 homolog was also identified in A. fumigatus and V. nonalfalfae. The Protein domains of all the putative V. dahliae SAGA subunits were examined using the Pfam 35.0 network.

To generate the VdAda1 deletion vector, approximately 1 kb upstream and downstream fragments of VdAda1 were amplified from AT13 genomic DNA. The plasmid pDHt2 (Zhou et al., 2013) was linearized by restriction endonuclease EcoRI or XbaI, and the amplified fragments were attached before and after the hygromycin resistance gene cassette (hyg) by homologous recombination (ClonExpress^® II One Step Cloning Kit, Vazyme, Nanjing, China). After the recombinant plasmids were transferred into Escherichia coli DH5α competent cells, the deletion vectors were transformed into Agrobacterium tumefaciens strain AGL-1, using the A. tumefaciens mediated transformation (ATMT) described by Wang et al. (2016). The transformants were selected on PDA infused with hygromycin at 50 μg/mL. The deletion vector pDHt2 encodes a geneticin (G418) resistance gene, so the randomly inserted transformants can grow on medium containing 50 μg/mL geneticin, thus that can exclude the randomly inserted transformants.

The VdAda1 complemented transformants were produced by polyethylene glycol-mediated protoplast transformation because of the defect of ΔVdAda1 in conidia production. To generate the VdAda1 complementation vector, the coding sequence for VdAda1 with its native promoter and terminator were amplified from AT13 genome and cloned into vector pCOM that carried the geneticin (G418) resistance, and the vector was introduced into the ΔVdAda1 strain by polyethylene glycol-mediated protoplast transformation. Fungal protoplast preparation of ΔVdAda1 was conducted according to the method described by Li et al. (2019). For protoplast transformation, 200 μl of 5 × 10⁶ protoplast/ml suspension was added to 5 ng plasmid mixed with an equal volume of 2 × STC [sucrose, 0.4 g/ml; 1M Tris-Cl (pH 8.0), 0.1 ml/ml; CaCl₂, 0.0147 g] and incubated for 10 min. This protoplast suspension was mixed with 1 × PEG (4 g PEG 4000/1 ml distilled water mixed with an equal volume of 2 × STC) and incubated for 20 min. To this mixture, 3 ml of liquid TB3 medium was added and shaken at 60 rpm for 6 h, 25°C; Cultures were added into 40 ml melted TB3 solid medium (liquid TB3 medium with 8g/L agar) with 50 μg/ml geneticin for selection.

For penetration assays, hyphae of each strain were incubated on minimal medium (MM) overlaid with a sterilized cellophane membrane for 3 days. After removal of the cellophane membrane, the cultures were incubated for another 5 days. The experiments were repeated independently at least three times.

Since ΔVdAda1 is unable to produce conidia, the virulence assay was carried out by inoculating agar plugs from each strain. To prepare for inoculation, the wild type strain AT13, deletion mutants and complementary transformants were cultured on PDA medium for 5 days at 25°C. Cotton (Gossypium hirsutum cv. Junmian 1) and N. benthamiana were grown in a greenhouse (14 h:10 h, day: night photoperiod) at 25°C ahead of the cultures. For inoculation with V. dahliae, a wound at the junction of the root and stem was made using a knife on 2-week-old cotton seedlings or 5-week-old N. benthamiana plants and were inoculated with agar plugs collected from the edge of the colonies from each strain. Uninoculated agar plugs served as negative control. Verticillium wilt symptoms were observed 14 days post inoculation on cotton and 21 days post inoculation on N. benthamiana. Fifteen cotton plants and five N. benthamiana plants were inoculated with each strain. Vascular discoloration of cotton was observed in longitudinal sections of the shoots at 14 dpi. The experiments were repeated independently at least three times.

Biomass of V. dahliae was estimated by qPCR in cotton and N. benthamiana infected with AT13, ΔVdAda1, and complemented strains to investigate the influence of VdAda1 on colonization of plants. DNA was extracted from the junction of the roots and stems of cotton and N. benthamiana inoculated with each strain using DNA secure Plant Kit (TIANGEN, Beijing, China). DNA was quantified by spectrophotometry and 100 ng of DNA of each sample was used for qPCR reaction. A qPCR reaction was performed using genomic DNA as a template and the 2 × TransStart Top Green qPCR SuperMix (TransGen Biotech, Beijing, China) under the following conditions: an initial 95°C denaturation step for 3 min, followed by 40 cycles of 95°C for 15 s, 60°C for 20 s and 72°C for 20 s. V. dahliae elongation factor 1-α (VdEF-1α) was used to quantify fungal colonization. The cotton 18S gene and N. benthamiana EF-1α were used as endogenous plant reference gene. The primers are listed in supporting information (Supplementary Table 1).

To detect the expression level of melanin-associated genes and SAGA subunits, total RNA was isolated from mycelium of AT13, ΔVdAda1, and complemented strains using a total RNA Miniprep kit (Aidlab, Beijing, China). cDNA was reverse transcribed using TransScript II One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). RT-qPCR was performed under following conditions: an initial 95°C denaturation step for 5 min, followed by 40 cycles of 95°C for 30 s, 60°C for 30 s and 72°C for 20 s. The V. dahliae VdEF-1α gene was used as endogenous reference. Relative transcript levels of the above genes in various samples were determined using the 2^–ΔΔCT method (Livak and Schmittgen, 2001). Primers used to quantify each gene in RT-qPCR are listed in supporting information (Supplementary Table 1). The experiment was performed three times, and each treatment had three replicates.

For expression pattern analysis of VdAda1, wild type strain AT13 was cultured in CM liquid medium for 3 days at 25°C. The roots of cotton seedlings (2-weeks-old) were inoculated with AT13 conidial suspensions (1 × 10⁹ cfu/ml) and roots were harvested at 0, 0.5, 1, 2, 3, 5, 7, and 9 dpi. The RNA was isolated and cDNA was reverse transcribed as described above. The primers used to quantify the expression of VdAda1 are listed in Supplementary Table 1. The RT-qPCR procedure was carried out as described above.

Mycelium from AT13 and ΔVdAda1 were plated on Czapek medium overlaid with sterilized cellophane membrane and incubated at 25°C for 5 days. The hyphae on cellophane membrane were collected and total RNA was extracted for RNA-seq using total RNA Miniprep kit (Aidlab, Beijing, China). The AT13 strain cultured on Czapek medium was used as a control.

The SOAP-aligner/SOAP2.0 software package (Li et al., 2009) was used to map the reads to the reference sequence of AT13 genome (unpublished), and less than two mismatches were allowed in the alignment. The uniquely mapped read counts were normalized and the expression levels were calculated using the reads per kilobase per million mapped reads (RPKM) method (Mortazavi et al., 2008). A strict algorithm was used to identify significantly differentially expressed genes between the wild type and ΔVdAda1 strain. The FDR was set to 0.001 to determine the threshold P < 0.001 in multiple tests, and the log₂ ratio ≥2.0 (Audic and Claverie, 1997).

To verify the RNA Seq analysis, the expression level of some of the differentially expressed transcription factors and secreted proteins was examined by RT-qPCR after the total RNA was isolated from mycelium of AT13 and ΔVdAda1 strains. The RT-qPCR procedure was carried out as described previously. The V. dahliae elongation factor 1-α (VdEF-1α) was used as a reference gene for the relative expression analysis. The primers used to quantify the expression of these genes were shown in Supplementary Table 1.

For Western blot analyses, total protein was extracted from lyophilized mycelium of wild type AT13, ΔVdAda1, and complemented strains cultured in CM liquid medium for 3 days. Mycelia (50 mg) were finely ground and suspended in extraction buffer (RIPA lysis buffer: 500 μl; phenylmethylsulfonyl fluoride (PMSF), 1 mM). After homogenization by vortexing, the lysate was centrifuged at 12,000 rpm for 10 min at 4°C. 100 μl supernatant was mixed with an equal volume of 2× loading buffer and boiled for 10 min. Thirty micrograms of protein were separated by SDS PAGE (15%). After transferring to Immobilon-P transfer membranes (Merck Millipore, Darmstadt, Germany), the membranes were probed with anti-H3K9ac (#39138, 1:10000), -H3K18ac (#39756, 1:10000), -H3K27ac (#39136, 1:10000) (Active Motif, La Hulpe, Belgium), anti-H3 (M130918, Hunan Biotechnology Co., Ltd., Hangzhou, Zhejiang, China, 1:5000) primary antibodies and a goat anti-rabbit IgG-HRP secondary antibody (CWBIO, Beijing, China, 1:10000). Chemiluminescence was detected with Immobilon Western Chemoluminescent HRP Substrate (Merck Millipore, Darmstadt, Germany). The experiment was conducted three times independently.