The Aspergillus strains used in this study are listed in Supplementary Table S1. The strains were grown on minimal medium (MM) containing 1% w/v glucose, 10 mM sodium glutamate, 0.052% w/v KCl, 0.052% w/v MgSO₄·7H₂O, and 0.152% w/v KH₂PO₄, plus Hunter’s trace elements (pH 6.5). To cultivate A. oryzae NSPlD1 (Maruyama and Kitamoto, 2008), 1.5 g/l methionine, 1.22 g/l uracil, and 1.21 g/l uridine were added to MM. To cultivate the Aspergillus nidulans strains, 1 mg/l biotin was added to MM.

ugmA was disrupted in A. oryzae NSPlD1 by inserting AnpyrG. A gene replacement cassette encompassing the homology arms at the 5′ and 3′ ends of ugmA was amplified by recombinant polymerase chain reaction (PCR) using A. oryzae RIB40 genomic DNA as the template and the primer pairs ugmA-1/ugmA-2 and ugmA-3/ugmA-4, respectively (Supplementary Table S2). The A. nidulans pyrG (AnpyrG) marker was amplified by recombinant PCR using pHSG396-AnpyrG (Kadooka et al., 2022b) as the template and the primer pair pHSG396-F/pHSG396-R. The resultant DNA fragment, amplified using the primers ugmA-1 and ugmA-4, was used to transform A. oryzae NSPlD1, yielding the ΔugmA strain. MM agar plates without uracil and uridine were used to select the transformants. The introduction of AnpyrG into each gene locus was confirmed by PCR using the primer pair ugmA-F/ugmA-R (Supplementary Figure S1).

ugmA that included 1.5 kbp upstream of ugmA was amplified by PCR using A. oryzae RIB40 genomic DNA as the template and the primer pair pPTR-II-ugmA-IF-F/pPTR-II-ugmA-IF-R. The amplified fragment was inserted into the SmaI site of pPTR-II using the In-Fusion HD Cloning Kit (Takara, Kusatsu, Shiga, Japan) to yield pPTR-II-ugmA.

Total GM (FTGM and galactomannoproteins) from A. oryzae was prepared as previously described (Katafuchi et al., 2017). Briefly, the hot-water-soluble extract from cells was fractionated using cetyl trimethyl ammonium bromide. The resultant fraction was precipitated at pH 9.0 with NaOH in the presence of borate and resolved in distilled water as the total GM fraction. A β-elimination reaction was performed to remove O-glycans from the galactomannoproteins under reducing alkali conditions (500 mM NaBH₄ /100 mM NaOH, 10 ml, at 25°C for 24 h). After neutralization with 50% acetic acid, the samples were dialyzed overnight against distilled water. The purified samples were then lyophilized, resuspended in distilled water, and clarified using 0.45-μm-pore filters. The resultant samples were prepared as GM fractions.

Samples for ¹H-NMR were exchanged twice in D₂O with intervening lyophilization and then dissolved in D₂O (99.97% atom ²H). The ¹H-NMR spectra were recorded using a JNM-LA600 spectrometer (JEOL, Akishima, Tokyo, Japan) at 45°C. The proton chemical shifts were referenced relative to internal acetone at δ 2.225.

Immunoblotting was performed as previously described (Komachi et al., 2013). The EB-A2 antibody of the Platelia Aspergillus enzyme immunoassay (Bio-Rad Laboratories, Hercules, CA, United States) was used at a dilution of 1:10 to detect β-Galf.

Hyphal surface adhesion assay was performed as previously described with slight modifications (Lamarre et al., 2009; Alam et al., 2014). Briefly, 0.5-μm-diameter polystyrene beads (Sigma) were diluted to 1:100 in sterile phosphate-buffered saline (PBS). Mycelia were grown for 18 h at 30°C with shaking at 127 rpm in liquid potato dextrose medium, harvested into PBS-containing polystyrene beads for 1 h, and then washed five times with PBS. Mycelium images were acquired using a microscope equipped with a digital camera.

To investigate the presence of Galf-containing sugar chains in A. oryzae, Aspergillus luchuensis (Aspergillus awamori var. kawachi) and A. luchuensis mut. Kawachii (A. kawachii), galactomannoproteins were extracted from mycelia and subjected to immunoblotting to detect Galf-containing glycoproteins using the anti-Galf antibody EB-A2 (Oka et al., 2005). Smeared bands indicated the presence of Galf-containing glycoproteins (Figure 1), which were thought to be mainly attributable to the O-glycans among the glycoproteins (Komachi et al., 2013). Densitometric quantification of the immunoblot bands was performed using ImageJ software (Schneider et al., 2012). The signal intensity ratios for EB-A2 were then calculated, and the ratio for A. fumigatus was normalized to 1.0. The ratio of EB-A2 intensity to Galf-containing glycoprotein in A. oryzae was 0.6-fold less than that in A. fumigatus (Figure 1), suggesting the presence of few Galf-containing glycoproteins in A. oryzae (Figure 1). Interestingly, the intensities of A. luchuensis and A. kawachii were more than four times higher than that of A. fumigatus (Figure 1). These data indicate that the quantities of Galf-containing glycoproteins differ among different species of koji fungi.

To investigate whether the genes encoding UDP-galactopyranose mutase are present in the genome of koji fungi, we performed an NCBI protein BLAST¹ search using the amino acid sequence of A. nidulans UgmA (AN3112) as a query sequence. AO090012000855 of A. oryzae RIB40, RIB2604_01707610 of A. luchuensis RIB2604, and AKAW2_10730A of A. kawachii IFO4308 exhibited strong homology to A. nidulans UgmA (91.75, 91.18, and 91.35% identities, respectively).

To clarify the diversity of GlfA/UgmA in detail, we performed a phylogenetic analysis of UDP-galactopyranose mutase conserved in fungi (Figure 2). The dataset for analysis was obtained by an NCBI protein BLAST search using the amino acid sequence of A. nidulans UgmA as a query sequence. The BLAST search identified the UDP-galactopyranose mutase of A. oryzae, which was named AoUgmA by Damveld et al. (2008). In the present study, AoUgmA is referred to as UgmA. Damveld et al. (2008) also revealed the presence of a putative UDP-galactopyranose mutase named AoUgmB in A. oryzae. However, as AoUgmB encodes only 168 amino acids, it probably does not function as a UDP-galactopyranose mutase because its protein size is too small. Therefore, AoUgmB was excluded from the present analysis. Phylogenetic analysis revealed that UgmA is widely distributed in the subphylum Pezizomycotina and the phylum Basidiomycota. In the phylum Ascomycota, Saccharomycotina and Taphrinomycotina do not carry UgmA. These results are consistent with previously reported findings and with the fact that galactofuranosyl chains are not present in Saccharomycotina and Taphrinomycotina (Tefsen et al., 2012). UgmA is present in Agaricomycotina, Ustilaginomycotina, and Pucciniomycotina in the phylum Basidiomycota, whereas it is absent in most Agaricomycotina in the phylum Basidiomycota. Functional analysis of UgmA in basidiomycetes is currently limited, and further functional analysis is warranted in future studies.

To investigate the physiological roles of ugmA in A. oryzae, we constructed a ugmA gene disruptant and observed the colonial morphology following culture at 30°C, 37°C, and 42°C for 5 days (Figure 3A). ΔugmA had a smaller colony size than the parental strain, indicating the important role of UgmA in normal mycelial formation in A. oryzae. The growth of the parental and ∆ugmA strains was delayed at 42°C compared with that at 30°C on MM. The colony diameters of the parental strain were 34% smaller after culture for 4 days at 42°C compared with that at 30°C, while those of the ∆ugmA strain were 23% smaller after culture for 4 days at 42°C compared with that at 30°C. These results indicate that the ∆ugmA strain exhibits a temperature-sensitive phenotype (Figure 3A; Supplementary Figure S2). This temperature-sensitive phenotype tended to improve under high osmotic support conditions (Figure 3A), suggesting that the ΔugmA strain exhibits temperature sensitivity attributable to the absence of some cell wall components. To confirm that the ΔugmA phenotype was caused by the disruption of ugmA (Figure 3B), we constructed a ugmA complementation strain (ΔugmA + A. oryzae ugmA) and observed the colonies (Figure 3B). The ΔugmA + A. oryzae ugmA strain had a similar colony size and morphology as the parental strain on MM (Figure 3B). This result indicated that the abnormal phenotypes of ΔugmA were truly attributable to ugmA disruption.

In addition, we investigated the drug sensitivity of ΔugmA. In A. nidulans and A. niger, ugmA-disrupted strains were shown to be sensitive to higher concentrations of calcofluor white (CFW), a chitin-binding reagent (Damveld et al., 2008; El-Ganiny et al., 2008). Therefore, we investigated whether ugmA disruption also affected CFW resistance in A. oryzae. The growth of ΔugmA was delayed on MM supplemented with 30 μg/ml CFW compared with that of the parental strain (Figure 3C), suggesting that ugmA disruption changes the balance of cell wall components, such as chitin and glucan, in A. oryzae.

It was previously reported that ΔglfA and ∆gfsABC strains of A. fumigatus exhibited increased hyphae branching (Lamarre et al., 2009; Chihara et al., 2020). Therefore, we examined whether abnormal mycelial branching also occurred in the ΔugmA strain of A. oryzae. Branching structures at the hyphal tips were observed at a high frequency in the ΔugmA strain (Figure 4A), indicating that the loss of Galf-containing oligosaccharides increases abnormal mycelial branching. It has also been reported that the deletion of Galf-containing oligosaccharides from cells increases cell surface hydrophobicity (Lamarre et al., 2009; Chihara et al., 2020). To confirm the increase in cell surface hydrophobicity of the ΔugmA strain, we examined the level of attachment of latex beads to the mycelium in the ΔugmA strain and found that the level of attachment was clearly increased (Figure 4B), suggesting that Galf-containing oligosaccharides are involved in the cell surface hydrophobicity of A. oryzae.

To determine whether A. oryzae ugmA can complement the growth defect and Galf antigen-lacking phenotype in A. nidulans ΔAnugmA, ugmA was expressed in A. nidulans ΔAnugmA. First, we evaluated whether the expression of A. oryzae ugmA could reverse the growth defects and aberrant conidial formation of A. nidulans ΔugmA. Each conidium was inoculated on MM agar and cultivated at 37°C for 3 days. Expression of A. oryzae ugmA recovered the phenotype of A. nidulans ΔAnugmA, suggesting that A. oryzae ugmA complements the function of A. nidulans ΔAnugmA (Figure 5A). Next, we analyzed the presence of the Galf antigen on the glycoprotein using EB-A2 antibody. Disruption of AnugmA results in the loss of Galf antigens in glycoproteins in A. nidulans (Komachi et al., 2013), but expression of A. oryzae ugmA restored the Galf antigen (Figure 5B). These results indicated that A. oryzae ugmA encodes a UDP-galactopyranose mutase.

To determine whether the galactofuranose-containing sugar chain is deficient in the ∆ugmA strain of A. oryzae, the galactofuranosyl residues were detected by ¹H-NMR (Figure 6). The chemical shift signals at 5.195 and 5.05 ppm of the ¹H-NMR spectra represented the H-1 signal of the underlined Galf residue in the β-Galf-(1,5)-β-Galf-(1,5)-β-Gal and β-Galf-(1,5)-β-Galf-(1,6)-β-Galf structures, respectively, according to previous studies on A. fumigatus (Shibata et al., 2009; Kudoh et al., 2015). Chemical shift signals at 5.195 and 5.05 ppm were detected in the GM fraction of the parental strain, consistent with the chemical shift signals of the β-(1,5)−/β-(1,6)-galactofuran side chain of FTGM in A. fumigatus (Figure 6). This result indicated the presence of the β-(1,5)−/β-(1,6)-galactofuran side chain structure in A. oryzae. In contrast, these signals were absent in the GM fraction of the ∆ugmA strain, indicating that the galactofuranose-containing sugar chain is deficient in the ∆ugmA strain of A. oryzae (Figure 6). In addition, four unique chemical shifts at 5.0–5.2 ppm (signals A–D), indicating the presence of core mannan backbone structures, emerged in the GM fraction of ΔugmA, indicating the presence of a core mannan chain in FTGM (Kudoh et al., 2015; Onoue et al., 2018; Kadooka et al., 2022b). In the ΔugmA + ugmA strain, the chemical shift indicating the presence of the α-core-mannan backbone was masked, and the chemical shifts indicating the galactofuran chain reappeared (Figure 6). These results clearly indicate that a structure consistent with the FTGM of A. fumigatus is also present in A. oryzae.