Identification of the UDP-glucose-4-epimerase required for galactofuranose biosynthesis and galactose metabolism in A. niger

Background Galactofuranose (Galf)-containing glycoconjugates are important to secure the integrity of the cell wall of filamentous fungi. Mutations that prevent the biosynthesis of Galf-containing molecules compromise cell wall integrity. In response to cell wall weakening, the cell wall integrity (CWI)-pathway is activated to reinforce the strength of the cell wall. Activation of CWI-pathway in Aspergillus niger is characterized by the specific induction of the agsA gene, which encodes a cell wall α-glucan synthase. Results In this study, we screened a collection of cell wall mutants with an induced expression of agsA for defects in Galf biosynthesis using a with anti-Galf antibody (L10). From this collection of mutants, we previously identified mutants in the UDP-galactopyranose mutase encoding gene (ugmA). Here, we have identified six additional UDP-galactopyranose mutase (ugmA) mutants and one mutant (named mutant #41) in an additional complementation group that displayed strongly reduced Galf-levels in the cell wall. By using a whole genome sequencing approach, 21 SNPs in coding regions were identified between mutant #41 and its parental strain which changed the amino acid sequence of the encoded proteins. One of these mutations was in gene An14g03820, which codes for a putative UDP-glucose-4-epimerase (UgeA). The A to G mutation in this gene causes an amino acid change of Asn to Asp at position 191 in the UgeA protein. Targeted deletion of ugeA resulted in an even more severe reduction of Galf in N-linked glucans, indicating that the UgeA protein in mutant #41 is partially active. The ugeA gene is also required for growth on galactose despite the presence of two UgeA homologs in the A. niger genome. Conclusion By using a classical mutant screen and whole genome sequencing of a new Galf-deficient mutant, the UDP-glucose-4-epimerase gene (ugeA) has been identified. UgeA is required for the biosynthesis of Galf as well as for galactose metabolism in Aspergillus niger. Electronic supplementary material The online version of this article (doi:10.1186/s40694-014-0006-7) contains supplementary material, which is available to authorized users.


Background
The cell wall is an essential component of the fungal cell. Cells can survive the enzymatic removal of the cell wall but the resulting protoplasts need to be stabilized in an environment with high osmolarity to withstand internal turgor pressure. The cell wall of filamentous fungi consists mainly (90%) of polysaccharide material, including polymers of glucose (β-1,3-and β-1,6-glucans, α-1,3-glucan), N-acetylglucosamine (chitin), mannose and galactofuranose (galactomannan), galactoaminogalactan and of about 10% of cell wall glycoproteins (galactomanno-proteins) [1][2][3]. The different glycoconjugates are either synthesized at the plasma membrane by specific cell membrane-localized synthases (e.g. chitin, β-1,3and α-1,3-glucan) or preassembled in the secretory pathway (galactomannan and galactomannoproteins). During or after transport over the membrane, polymers are cross-linked with each other via covalent or hydrogen bonds to create a sturdy cell wall.
We and others have previously shown that galactofuranose (Galf) is an important component of the cell wall in Aspergillus species. It is found in several glycoconjugates including galactomannan, secreted and cell wall proteins via Nand O-linked chains, and glycosphingolipids (see for review [4]). A key enzyme in Galf biosynthesis is UDP-Galp mutase (UgmA), which converts the pyranose form of UDP-galactose (UDP-Galp) into UDP-Galf. Only UDP-Galf can be transported into the Golgi where Galf is used as a donor sugar for the synthesis of Galf-containing structures [5][6][7]. Gene disruption approaches of the ugmA gene in Aspergillus niger, A. fumigatus, A. nidulans have shown that Galf biosynthesis is required for hyphal morphogenesis and cell wall architecture [8][9][10], whereas disruption of the ugmA homolog in Cryptococcus neoformans did not have an apparent growth phenotype nor did it affect capsule formation [11]. Inactivation of the ugmA genes in Aspergilli results in an increased sensitivity towards cell wall assembly interfering drugs such as calcofluor white (CFW), indicating that galactofuranosecontaining glycoconjugates are necessary for maintaining the integrity of the cell wall [8].
In A. niger the ugmA gene was identified in a screen for mutants in which the cell wall stress reporter gene agsA was constitutively induced [8]. Besides UgmA, proteins required for the biosynthesis of Galf-containing glycoconjugates have been identified in both A. nidulans and A. fumigatus. These proteins include UgeA/Uge5, encoding the UDP-glucose-4-epimerase necessary for the synthesis of UDP-Galp [12,13], UgtA/GlfB encoding a Golgi localized UDP-Galf transporter protein [5,6] and GfsA encoding a Galf transferase [7].
In this study, we have screened a collection of 240 cell wall mutants with induced agsA expression [8] for mutants that do not secrete Galf containing glycoconjugates into the growth medium. In addition to a large complementation group of 9 ugmA mutants, one additional Galf-low mutant (#41), belonging to a different complementation group, was identified. Whole genome sequencing of this mutant revealed that the newly identified Galf-mutant contains a mutation in gene An14g03820. This gene is predicted to encode a putative UDPglucose-4-epimerase gene (ugeA), required for the biosynthesis of Galf as well as for Galp metabolism in Aspergillus niger.

Results
Screening of Galf-deficient mutants within 240 A. niger cell wall mutants To identify additional genes involved in the biosynthesis of cell wall galactomannan in A. niger, we screened a collection of 240 cell wall mutants for Galf-deficient strains. Galf containing structures such as galactomannans and Nor O-glycosylated proteins are secreted in the medium and therefore analysis of the medium for the presence of Galf using an antibody can identify Galf minus mutants. From the collection of cell wall mutants, we previously identified the A. niger UDP-galactopyranose mutase (UgmA) as an essential protein for the formation of Galf-containing cell wall glycoconjugates ( [8], see below). The selection of the three ugmA mutants in that study was not based on their Galf phenotype, but on their Calcofluor white-and SDShypersensitive phenotype. Here, all 240 mutants were grown in liquid medium and 2 μl of medium was used in a dot blot analysis using the anti-Galf antibody L10 [14] as described in Materials and Methods ( Figure 1). The screening confirmed the absence of Galf in the three ugmA mutants already identified (6.13#44, 15.4#17, and 6.13#50) [8], and identified six additional Galf-minus mutants (15.4#30, 15.4#18, 15.4#5, 15.4#50, 15.4#57 and 6.47#41). A heterokaryon complementation test using the ΔugmA strain was performed to determine whether the newly identified mutants were also mutated in ugmA. In this test, spores of mutants of interest are inoculated close to each other. Outgrowth of the spores will result in the formation of heterokaryotic mycelium in the contact zones. The phenotype of all Galf-deficient mutants identified was characterized by a reduced sporulation ( Figure 2A) and this phenotype was used to assay complementation. As shown in Figure 2B only mutant #41 was complemented by the ΔugmA mutant, which is visible by a well sporulating zone of heterokaryotic mycelium in the contact zones of the two colonies. The other mutants were not complemented by the ΔugmA mutant (data not shown) and therefore considered to be mutant alleles of ugmA. In contrast, mutant #41 showed a clear zone of sporulation in the contact zones with the previously identified ΔugmA mutant ( Figure 2B) as well as with the ΔugmA mutants identified in this study ( Figure 2C), indicating that the ΔugmA mutant and #41 mutant are in two different complementation groups. To confirm that the #41 mutant was not mutated in the ugmA gene, the ugmA locus of the #41 mutant was amplified by PCR and sequenced. No mutations in the ugmA gene were found.

Mutant #41 has reduced levels of Galf
To address the Galf-deficient phenotype of the #41 mutant further, medium samples of mutant #41 were analysed using the Platelia assay. This quantitative assay for Galf detection uses the monoclonal EB-A2 antibody which recognizes Galf-moieties on galactomannoproteins [15]. As shown in Figure 3A, titration of the medium samples revealed an approximately 10-fold reduced amount of Galf-reactive epitopes produced by the #41 mutant compared to that of its parental strain RD6.47.
As controls, high reactivity with medium from wild type N402 and no reactivity with medium from the ΔugmA strain were measured in this assay. Western blot analysis of extracellular medium proteins using the L10 antibody also detected less Galf epitopes on secreted glycoproteins in contrast to the wild type strains N402 and RD6.47, but clearly more than in the medium of the ΔugmA strain ( Figure 3B). Detection of N-glycans with the lectin ConcavalinA (ConA) was performed as a control.
To obtain more insight into the presence of Galf in the cell wall of the #41 mutant, galactomannan from two Galf mutants (ΔugmA and #41), and the wild-type strain (N402) was isolated essentially as described previously by Bardalaye and Nordin, 1977 [16] (see Methods for details). A titration of the purified galactomannan fraction was applied to the Platelia assay, revealing again a lowered reactivity of the Galf-antibody with the galactomannan derived from #41 compared to the wild-type strain ( Figure 3C), whereas the polysaccharide isolated from ΔugmA was not reactive at all. This indicated that there are less Galf-moieties present on the galactomannan from #41 compared to wild type galactomannan. To confirm this result, the monosaccharide composition of the galactomannan fraction of these strains was analysed by HPAEC analysis (DIONEX system) after hydrolysis ( Figure 3D). As expected, no Galf was detected in the hydrolyzed polysaccharide isolated from the ΔugmA mutant. The ratio Galf /Man of the wild-type strains was found to be well above 2 (2.2-2.8), while this ratio was only 0.5 for the #41 mutant. These results demonstrate that the amount of Galf is reduced more than 4-fold in the #41 mutant compared to its parent strain. This lowered amount of Galf apparently has escaped detection by the L10 antibody in our initial screening by dot blot, leading to its discovery ( Figure 1), but clearly is detected by the Platelia assay ( Figure 3A and 3C), using the EB-A2 antibody, as well as by Western blot analysis using the L10 antibody ( Figure 3B).

Genomic characterization of mutant #41 by whole genome sequencing
An A. niger genomic cosmid library [17] was used in several unsuccessful attempts to complement the reduced sporulating growth phenotype of mutant #41. As an alternative approach, the genome of both the parental strain (RD6.47) and the #41 mutant were sequenced by pair-end Illumina sequencing. A single nucleotide polymorphism (SNP) analysis between the two strains was performed and 78 SNPs were identified (Additional file 1: Table S1). These SNPs were analysed further by determining whether a particular SNP was located in a predicted ORF and whether this SNP affected the amino acid sequence. We identified 21 SNPs in genes in the #41 mutant that caused changes in the amino acid sequence (Additional file 2: Table S2). Most of these SNPs seemed unrelated to Galf-biosynthesis, but the mutation in gene An14g03280 encoding a putative UDP-glucose-4-epimerase seemed important as, it was recently published that the UDP-glucose-4-epimerase (UgeA) is needed for Galf biosynthesis in A. nidulans [13]. The mutation in the ugeA (An14g03280) gene of mutant #41 (A to G) caused the change of a codon from AAC to GAC which consequently resulted in the change of Asn to Asp at position 191 in the UgeA protein. The mutation identified by the Illumina genome sequencing approach was confirmed by PCR amplification of the ugeA locus of the #41 strain followed by direct sequencing.

Disruption of the A. niger ugeA gene and phenotypic analysis
To confirm that the mutation in An14g03820 (ugeA) caused the Galf-low phenotype in mutant #41, a pyrG based gene deletion cassette was made. The deletion cassette (pΔugeA::pyrG) was transformed to MA169.4 and uridine prototrophic transformants were purified. During purification of the primary transformants it was found that most of the transformants displayed a reduced growth and a reduced sporulation phenotype, which resembled the phenotype of the #41 mutant. Subsequent diagnostic PCR and Southern blot analysis proved that the transformants with the growth phenotype contained an ugeA deletion (data not shown). Further phenotypic characterization of the ΔugeA mutant showed similar morphological alterations (increased branching and irregular length of hyphal compartments) as found for the #41 mutant (data not shown).
As demonstrated above, very low levels of Galf were observed in the dot blot ( Figure 1) and Platelia assay ( Figure 3A) for the #41 mutant. Interestingly, the amount of Galf in the ΔugeA strain was strongly reduced compared to the #41 mutant, indicated by the complete lack of signal on the dot blot ( Figure 4) and a further 10-fold reduction of the Galf-signal in the Platelia assay ( Figure 3A). These results suggest that the UgeA protein in #41 is still slightly active and able to convert some UDP-Glcp to UDP-Galp, which is subsequently utilized for Galf-biosynthesis. Despite the observation that more Galf is present in the #41 mutant compared to the ΔugeA mutant, both mutants display similar phenotypes, although some subtle differences could be seen between the #41 and the ΔugeA mutants. The #41 mutant seems to sporulate slightly more intense on the MM plate and grows somewhat better on the 0.005% SDS plate ( Figure 5).
To further confirm that the mutation in ORF An14g03820 (ugeA) in the #41 mutant and the deletion of ugeA are the cause of the observed phenotypes, a complementation analysis was performed. Therefore, pyrG − derivatives were obtained from the #41 and ΔugeA mutants by selecting on 5′FOA plates. The ugeA gene was PCR amplified and cloned into an autonomously replicating vector (pAMA-pyrG, [18]) and used for complementation. Transformation of the pAMA-ugeA plasmid to the #41 and ΔugeA mutants complemented the reduced growth and sporulation phenotype ( Figure 6A and B) as well as the Galf-negative phenotype in the dot blot analysis (data not shown). Control transformations with the empty plasmid (pAMA-pyrG) or the pAMA-ugmA-pyrG gene were also performed. In general, the transformation of these control plasmids did not complement the sporulation defect of the #41 and ΔugeA mutants. Unexpectedly, we occasionally (with a frequency of 0.2%-1%) obtained transformants with the empty or ugmA plasmid that sporulated well and grew like the complemented strain. From the original transformation plates of the ΔugeA mutant with the pAMA-pyrG or the pAMA-ugmA-pyrG plasmids, two transformants were purified with the improved sporulation phenotype. The pAMA-pyrG or pAMA-ugmA-pyrG transformant that sporulated well were also Galf-positive in the dot blot analysis (data not shown).
The reversion of the phenotype might be caused by second site suppressor mutations or by increased expression of redundant genes (see Discussion).

UgeA is required for growth on galactose in A. niger
Based on sequence comparisons, UgeA (An14g03820) is predicted to encode a UDP glucose-4-epimerase and the result above clearly indicate that this gene is most likely responsible for the synthesis of UDP-Galp, which is further converted to UDP-Galf by the UgmA enzyme. A. niger contains in addition to UgeA (An14g03820) two other homologous genes predicted to encode UDP glucose-4epimerases (An12g10410 and An02g11320). The presence of three putative UDP glucose-4-epimerases in A. niger is similar to A. fumigatus in which the function of the three genes (uge3, uge5 and uge4) was recently analysed [12]. Based on homology, the A. niger UgeA is most similar to Figure 4 Dot blot assay to detect the presence of Galf residues on secreted glycoconjugates of different A. niger strains. Strains were grown to early stationary phase and cell-free medium was spotted in a dilution series as indicated on nitrocellulose filter paper on two blots in parallel. The left blot was incubated with the anti-Galf antibody (L10) the right blot was incubated with the ConA lectin. Figure 5 Susceptibility of A. niger Galf mutants towards SDS and CFW. Ten-fold serial dilutions of spores derived from different Galf mutants were spotted on minimal medium plates containing 0.005% SDS or 100 μg/ml CFW. Pictures were taken after three days of growth at 30°C. the A. fumigatus Uge5 protein and the UgeA protein of A. nidulans (See Table 1 for amino acid identities of the Uge homologs between the three species). To test whether the A. niger ugeA gene is required for growth on galactose, mutant #41, ΔugeA and their respective parental wild type strains were inoculated on plates containing 50 mM galactose and 0.3 mM arabinose. The supplementation of a second carbon source in addition to galactose is included because A. niger spores are unable to germinate on galactose alone [19][20][21]. As shown in Figure 6C, the #41 and ΔugeA mutants do not grow on galactose medium indicating that ugeA is required for growth on galactose. The inability to grow on galactose of the ugeA strain is not a consequence of its inability to produce Galf as the ugmA deletion strain grows normally on galactose ( Figure 6C).

Discussion
Screening of a collection of A. niger cell wall mutants, which were previously isolated by induced expression of the cell wall stress reporter gene (agsA), identified in  total 9 mutants which did not secrete detectable levels of Galf-containing glycoconjugates into the medium. Three mutants were identified previously [8] and six new Galfdeficient mutants were additionally found in this screen. Interestingly, the mutants fell into two unevenly distributed complementation groups. A large group (represented by eight alleles) consisted of mutants in the ugmA gene and the other complementation group only contained a single representative with a mutation in the ugeA gene. The reason for this uneven distribution among the two complementation groups is not clear as both the ugmA and the ugeA mutants have very similar phenotypes when inactivated (Figures 4 and 5). Besides UgeA and UgmA, which are required for the synthesis of the UDP-Galf, proteins involved in subsequent steps of the synthesis of Galfcontaining glycoconjugates have been recently identified and include a specific UDP-Galf transporter (GlfB in A. fumigatus [5] and UgtA in A. nidulans [6] and a Galftransferase [7]. Mutants in the homologs of those genes in A. niger were not identified in our cell wall mutant collection. Analysis of the A. niger genome showed the presence of two homologs of GlfB/UgtA and three homologs of the GfsA protein (data not shown). The inability to detect representative mutants in the cell wall mutant collection suggests possible redundant functions of the pairs of genes in A. niger. Indeed, the two A. niger Galf transporters only display a Galf minus phenotype when both genes were deleted (Park and Ram, unpublished data). This study represents the first example of the power of genome sequencing for characterizing classically obtained mutants in A. niger. The mutant #41 strain was created by UV mutagenesis which is known to preferably induce transitions (purine to purine (A-G) or pyrimidine to pyrimidine (T-C) mutations. From the 78 SNPs a majority of 49 (63%) SNPs were transitions and 29 (27%) were found to be transversions (purine to pyrimidine or reverse) (Additional file 1: Table S1). Of the 29 transversions, a high number of A to T transversions were found. The total number of mutations (21) in non-exonic regions was relative low taking into account the relative stringent dosis of UV treatment (66% survival). This number of SNPs makes it feasible to identify responsible mutants in the case that analysis of the predicted function of the mutated genes does not give a clear clue.
Interestingly, the Asn to Asp mutation at position 191 in UgeA demonstrated in mutant #41 seems to inhibit enzyme function greatly, but not completely. We aligned the amino acid sequences of UgeA from A. niger with that from A. nidulans, human and Escherichia coli (Additional file 3: Figure S1A). This shows that the Asn at position 191 is conserved between the Aspergilli and human, but not in E. coli. Recently, the complex crystal structure of UgeA from A. nidulans together with its substrates NAD + and UDP-Glcp was elucidated [22], which showed a similar orientation as in the human enzyme (GALE). This structure enabled us to visualize the location of Asn191 at the start of β-strand 6 (in red, Additional file 3: Figure S1B). Mutation of this residue to an Asp will result in the disruption of the ideal orientation of β-strands 6 and 7, and thereby of the substrateorientation YFN-domain (residues 197-199). Apparently, this disruption is enough to diminish the UgeA enzyme activity greatly, resulting in lower amounts of available Galp to flow into the Galf-biosynthesis pathway.
This study shows that the ugeA gene of A. niger is required for UDP-Galf biosynthesis and for growth on galactose. This is identical to the situation in A. nidulans in which it has been shown that the ΔugeA mutant also lacks Galf and is unable to grow on galactose [13]. Sequence comparison showed that the A. nidulans UgeA protein is most similar to the UgeA protein of A. niger (Table 1). In A. fumigatus, the uge5 mutant (Uge5 is the paralog of both A. niger and A. nidulans UgeA protein) can grow on galactose medium. It has recently been shown that the A. fumigatus Uge3 protein is responsible for the growth on galactose as disruption of both uge3 and uge5 results in a galactose auxotrofic strain [12]. The galactose auxotrophy of the A. niger ΔugeA and A. nidulans ΔugeA mutants either indicated that the Uge3 homologs which are present in the genome are not expressed or have a more specific activity. The Uge3 protein of A. fumigatus has been shown to be involved in galactosaminogalactan synthesis and can mediate both the production of UDP-galactose and UDP-N-acetyl-galactosamine. Further enzymatic and molecular genetic analysis of the Uge homologs in A. niger and A. nidulans is required to understand their exact function and their possible role in galactosaminogalactan biosynthesis. In this respect, it is interesting that we observed high frequency of suppression of the ΔugeA phenotype after transformation. Although the phenotype of the #41 and ΔugeA mutants were stable during normal propagation, transformation of both strains yielded regularly suppressor mutations that were restored in their ability to produce Galf-decorated galactomannoproteins and were also restored in their ability to grow on galactose (data not shown). A possible mechanism to explain the development of suppressors include activation of expression of one of the uge genes, or occurrence of a mutation in one of these genes that changes substrate specificity.

Conclusion
By screening a collection of UV-generated cell wall mutants, a new Galf-deficient mutant was identified. We used whole genome sequencing to identify the mutation responsible for this mutant phenotype. Complementation and targeted deletion studies confirmed that the UDP-glucose-4-epimerase gene (ugeA) is required for the biosynthesis of Galf as well as for galactose metabolism in Aspergillus niger.

Strains and culture conditions
The Aspergillus niger strains used in this study are listed in Table 2. Strains were grown on minimal medium (MM) [23] containing 1% (wv-1) glucose as carbon source or complete medium (CM) containing 0.5% (wv-1) yeast extract, 0.1% (wv-1) casamino acids in addition to MM. When required, plates were supplemented with 10 mM uridine. 5′FOA selection for the selection of pyrG − strains was performed as described previously [24]. For the plate growth assays, strains were grown on MM plates containing 25 mM glucose, 25 mM galactose, or 25 mM galactose and 3 mM arabinose. For the heterokaryon assay, MMplates were supplemented with 0.005% SDS.
Tween-20. Blots with the L10 antibody were subsequently incubated with rabbit-a-mouse-HRP (PO161, DAKO) as a secondary antibody in TSMT with 5% BSA. For detection of the secondary antibody or the ConA-PO conjugate, a chemiluminescence kit (Thermo Scientific Pierce) was used.

Isolation of cell wall galactomannan
Cell walls were isolated from A. niger wild type strain N402 and the #41 mutant strain after growth in CM for 24 h at 30°C at 250 rpm. Spores (1 × 10 9 ) were inoculated in 1 L of CM in a 2 L Erlenmeyer. The mycelium of both strains was isolated by filtering over myracloth and grinded in liquid nitrogen using a pestle and mortar and the broken mycelia washed 3 times with 1 M NaCl and 3 times with MilliQ at 4°C by centrifugation (3600 rpm, 10 min). Successful breakage of the mycelia was confirmed by microscopy. Isolated cell walls were lyophilized for 24 h using 1.0 gram of cell walls (dry weight) for both the N402 and the #41 strain. The galactomannan fraction was isolated according to Bardalaye and Nordin [16]. The yield of purified galactomannan was 21.2 mg and 10.4 mg from 1 gram of freeze dried cell walls for the N402 and #41 strain, respectively.

Platelia assay
Microtiter plate wells containing coated antibody EB-A2, which recognizes Galf-moieties on galactomannan [14] were filled with 50 μl EB-A2 conjugated to HRP (all from Platelia Aspergillus EIA kit, Bio-Rad). Supernatant of A. niger strains or purified galactomannan was added to the wells and incubated for 90 min at 37°C. After washing 5 times, 200 μl 3,3′,5,5′-tetramethylbenzidine (TMB) detection mixture was added and samples were incubated for 30 min at ambient temperature in the dark. After stopping the coloring reaction by addition of 100 μl 1.5 N H 2 SO 4 , relative amounts of Galf on galactomannan were determined by measuring the optical density (OD) by spectrophotometry at 450 nm.

Monosaccharide analysis by HPAEC
Galactomannan isolated from A. niger was analyzed for its monosaccharide composition by High-Performance Anion Exchange Chromatography (HPAEC), using the method described by [26] with minor modifications. In summary, chromatography of the samples was performed in a Dionex, Bio-LC system, using a CarboPac PA 10 column (250 ×2 mm) in combination with a CarboPac guard column (2×50mm), Dionex Corp. Around 2 mg galactomannan was dissolved in 2 ml water and incubated with 360 μl of TFA for 5 h at 100°C. Hereafter, 3 ml MilliQ water was added and samples were freeze dried overnight. The pellets were dissolved in 100 μl water. After diluting 1:100, 20 μl (equivalent to 4 μg polysaccharide) was injected onto the column. Monosaccharides were eluted isocratically using H 2 O and 1 M NaOH in a ratio of 80:20 for 30 min. The column was washed with water for 10 min and equilibrated with 250 mM NaOH for 10 min before every injection.

Western blot analysis
Medium samples of the various strains were obtained as described for the dot-blot analysis and 20 ul of culture filtrate was used for Western blot analysis. SDS-PAGE and blotting were carried out as described [27]. For labeling with the anti-Galf antibody (L10), the membrane was blocked with 5% low-fat milk in TTBS (Tris-buffered saline, 0.05% Tween-20), and Galf was detected using the L10 antibody (1/5,000) overnight, followed by a goat anti-rabbit horseradish peroxidase secondary antibody (1/20,000) for 1 h. Detection was performed using a chemiluminescence kit (Thermo Scientific Pierce), according to manufacturer's instructions.

Genetic methods
Fungal transformations were done according the protoplast method described by [24]. Complementation tests were done using the heterokaryon test. 2 ul of spores (1 × 10 6 spores/ml) of strains to be tested were inoculated on MM plates containing 0.005% SDS approximately 1 cm apart from each other. Plates were incubated for 3 days at 30°C and analyzed for a zone of conidiation of heterokaryotic mycelium.

Plasmid construction
For the construction of the ugeA deletion cassette the Multisite GatewayR Three-Fragment Vector Construction kit was used. As a marker for deleting the ugeA gene, the pyrG marker of A. oryzae was used. To facilitate removal of the AopyrG marker, A. nidulans tTrpC repeats were cloned around the pyrG gene [28]. The TrpC-pyrG-TrpC fragment was PCR amplified using primer listed in Table 3 and cloned in pDonor221. The 5′ and 3′ flanking regions of ugeA were PCR amplified using primers with appropriate attB sites (Table 3) and the 799 bp fragments were cloned into pDONR P4-P1R and pDONR P2R-P3 respectively. The subsequent LR reaction was performed using pDONR_ugeA5, pDONR_ugeA3 and pDONR_TrpC_pyrG_TrpC and pDEST R4-R3 Vector 2 to create the ugeA deletion plasmid. The final construct was verified by restriction analysis and sequencing. For the complementation analysis the autonomously replicating vector pMA172 was used [18]. The ugeA or ugmA genes including promoter and terminator regions were PCR amplified using primers listed in Table 3, cloned into pJet2.1 and verified by sequencing. The ugeA and ugmA genes were reisolated as SmaI (ugeA) or NotI (ugmA) fragments and cloned into the unique SmaI or