Aspergillus nidulans Cell Wall Composition and Function Change in Response to Hosting Several Aspergillus fumigatus UDP-Galactopyranose Mutase Activity Mutants

Deletion or repression of Aspergillus nidulans ugmA (AnugmA), involved in galactofuranose biosynthesis, impairs growth and increases sensitivity to Caspofungin, a β-1,3-glucan synthesis antagonist. The A. fumigatus UgmA (AfUgmA) crystal structure has been determined. From that study, AfUgmA mutants with altered enzyme activity were transformed into AnugmA▵ to assess their effect on growth and wall composition in A. nidulans. The complemented (AnugmA::wild type AfugmA) strain had wild type phenotype, indicating these genes had functional homology. Consistent with in vitro studies, AfUgmA residues R182 and R327 were important for its function in vivo, with even conservative amino (RK) substitutions producing AnugmA? phenotype strains. Similarly, the conserved AfUgmA loop III histidine (H63) was important for Galf generation: the H63N strain had a partially rescued phenotype compared to AnugmA▵. Collectively, A. nidulans strains that hosted mutated AfUgmA constructs with low enzyme activity showed increased hyphal surface adhesion as assessed by binding fluorescent latex beads. Consistent with previous qPCR results, immunofluorescence and ELISA indicated that AnugmA▵ and AfugmA-mutated A. nidulans strains had increased α-glucan and decreased β-glucan in their cell walls compared to wild type and AfugmA-complemented strains. Like the AnugmA▵ strain, A. nidulans strains containing mutated AfugmA showed increased sensitivity to antifungal drugs, particularly Caspofungin. Reduced β-glucan content was correlated with increased Caspofungin sensitivity. Aspergillus nidulans wall Galf, α-glucan, and β-glucan content was correlated in A. nidulans hyphal walls, suggesting dynamic coordination between cell wall synthesis and cell wall integrity.


Introduction
The cell wall is essential for the survival of fungi in natural environments. In fungi, the cell wall is about one quarter of the average fungal biomass [1], and about a third of the fungal genome (,4000 genes) is involved in cell wall biosynthesis and maintenance [2]. Polysaccharides and glycoconjugates contribute to proper functioning of the cell wall and cell membranes that collectively form the interface between the fungal cell and its environment [3]. Due to the importance of cell walls for fungal growth and host invasion, and the relative lack of conservation with animal cell components, the cell wall is expected to be a good source of potential antifungal targets [4,5].
Aspergillus fumigatus is considered to be the most important airborne, human pathogenic fungus [6]. Aspergillus nidulans is a closely related model species with easy molecular genetic manipulation [7]. Carbohydrate analysis of A. fumigatus and A. nidulans walls shows that both species have similar composition [8], ,40% each of a-glucan and b-glucan [1,9]. Despite decades of research characterizing fungal cell wall composition and structure [2,5], many details remain poorly understood. Notably, recent evidence shows that fungal walls can be rapidly remodeled in response to environmental change [10], adds another layer of complexity.
Galactofuranose (Galf) has an important role in fungal wall maturation [11][12][13][14]. Galf is found in glycoconjugates that are important for survival or virulence of microorganisms including some bacteria, protozoa and fungi [15][16][17][18][19], but it is lacking in vertebrates. Since Galf is absent from higher eukaryotes and is involved in growth or virulence of some bacteria and fungi, the enzymes involved in its biosynthesis are potential drug targets [19,20]. Notably, loss of or reduction in wall Galf following gene deletion or repression in A. nidulans correlated with increased caspofungin sensitivity in vitro [12,13]. In vivo pathogenicity studies related to wall Galf are few, and to date are equivocal [15,18]. We have characterized the Galf biosynthesis pathway in A. nidulans as a model system [11][12][13][14]21] and are extending this to A. fumigatus in preparation for future testing of antifungals.
UDP-galactopyranose mutase is the first committed enzyme in the Galf biosynthesis pathway in both A. fumigatus (AfGlfA [18] also called AfUgmA [15]) and A. nidulans (AnUgmA [11]). Aspergillus nidulans strains deleted for AnugmA have 500-fold reduced hyphal growth and spore production [11]. UgmA is unique in A. nidulans and A. fumigatus, but is not essential in either species.
The crystal structure of AfUgmA has been determined [22]. AfUgmA has conserved arginine residues, R182 and R327 that are important for function in vitro. In addition, H63 and F66 are expected to contribute to AfUgmA activity because the loop containing them changes position depending on the redox state of the FAD cofactor [22]. Assessing the roles of these amino acids for the in vivo Aspergillus phenotype is expected to refine the in vitro characterization of UGM [22], since in vitro enzymatic studies might not be precisely comparable to their in vivo function [23], and because H63N mutant did not express sufficiently for in vitro characterization. Here we examined the in vivo effects of structurally and functionally important amino acid residues of AfUgmA using A. nidulans as a host system, comparing cell wall composition, hyphal surface adhesion and response to antifungal drugs.

Results
We used A. nidulans as a host for wild type (WT), wild type complemented (WC), and single amino acid mutants in AfUgmA to assess their effect on colony growth, wall composition, wall surface adhesiveness, and drug sensitivity. We hypothesized there would be concordance between AfUgmA function in vitro and the in vivo phenotype; in vivo characterization provided additional information.

Fungal phenotype for mutations affecting AfUgmA activity
We generated an [AnugmA:: wild type AfugmA] complemented (WC) strain in A. nidulans to confirm functional homology, and for quantitative characterization that had not been done previously [11]. The WC strain phenocopied wild type A. nidulans hyphal morphometry and colony development ( Figure 1, Table 1, and Figure S1 in File S1). Table 2 presents the relative enzyme activity of wild type and mutated UgmA strains, derived from data in Table 3 of [22]. As expected, the R327K and R327A strains phenocopied the AnugmAD strain for spore formation, colony growth, and hyphal morphometry ( Figure 1, Table 1). Reduced sporulation was due to fewer conidiophores and aberrant conidiophore formation ( Figure S2 in File S1), comparable to the AnugmAD strain. The R182K in vivo phenotype had improved sporulation but not colony growth compared to AnugmAD. There were no significant differences between the R182A and AnugmAD phenotypes ( Figure 1, Table 1). Sporulation in A. nidulans appears to have a less stringent requirement for Galf content than does hyphal growth (Table 2).
H63 is expected to contribute to AfUgmA activity because it is part of the flexible loop (loop III) above the si-face of the isoalloxazine ring that changes position due to the redox state of the FAD cofactor [22]. Structural analysis suggests H63N might keep UgmA loop III in the conformation similar to that seen in prokaryotic UGMs [24][25][26] and the reduced AfUgmA structure [22] by forming an H-bond with Q458. The in vitro H63N enzyme activity was not tested because of low protein expression for this construct. Notably, the in vivo H63N phenotype was comparable to that of R182K (Table 1, Figures 1 and Figure S2 in File S1), implying that the H63N strain likely would show a similar enzyme activity to R182K. This suggests that the flexibility of loop III plays a role in the catalytic activity of AnUGM. As with R182K, the major improvement compared to AnugmAD was in sporulation. This is consistent with Galf having multiple roles in the Aspergillus phenotype, with sporulation being more responsive to low levels of UgmA activity than hyphal morphology.
F66 is at the end of AfUgmA loop III and may control loop III flipping and consequently opening of the mobile loops depending on the redox state of the cofactor [22]. The AfUgmA-F66A in vitro enzyme activity was almost double that of R182K (Table 2). In vivo, the F66A phenotype was relatively comparable to wild type (Table 1, Figure 1) regarding sporulation ( Figure S2 in File S1) and hyphal morphometry, but restricted for colony diameter.
Together, these data show that AfUgmA R327 is critical for function in vitro and in vivo. H63, F66, and R182 each contribute to catalytic efficiency but to a much lesser extent. For all mutants, sporulation was less dependent on UgmA activity than hyphal or colony morphometry, consistent with the fact that ugmAn and R327A strains had limited sporulation. In sum, the in vivo AfUgmA mutant phenotypes correlated well with in vitro UgmA activity, and for H63N the in vivo data extended otherwise unobtainable characterization.

Loss of wall Galf is associated with increased hyphal surface adhesion
We used fluorescent latex beads visualized with confocal microscopy to assess hyphal surface adhesiveness, as adapted from [13]. The WT and WC hyphae had relatively fewer fluorescent beads attached compared to AnugmAn. Bead attachment for the strains with R182K, R182A, R327K, or R327A was similar to AnugmAn ( Figure 2). Figure 2 used bead adhesion by the R327A strain to represent all of these latter mutant strains. For the F66A and H63N strains, bead attachment was slightly higher than wild type or WC strains but much lower than AnugmAn or the other mutant strains (Figure 2). The bead adhesion assay showed major changes in adhesion, but did not discriminate between subtle differences. Atomic force microscope (AFM) force mapping might reveal these changes, but this is currently beyond our scope. However, our previous published paper showed using AFM force mapping that ugmAD strain hyphal wall surfaces had higher adhesion and reduced resilience compared to wild type A. nidulans walls [21]. and AnugmAD strains, grown on complete medium at 306C for 3 d. The colour difference between WT and WC strains was due to slightly different ages of culture. See Figure S1 in File S1 for a direct comparison of the spore colours of these two strains. doi:10.1371/journal.pone.0085735.g001 AfUgmA active site mutations do not affect UgmA-GFP cytoplasmic distribution El-Ganiny et al. [14] showed that AnUgmA-GFP was cytoplasmic and evenly distributed along hyphae, consistent with UgmA lacking a signal peptide (also suggested in [15]). To confirm that expression of wild type and mutated AfUgmA did not affect cytoplasmic distribution in A. nidulans, strains were C-terminal GFP-tagged under the control of the ugmA endogenous promoter ( Figure S3 in File S1). Previously, we had found that N-terminal AnUgmA-GFP abolished wild type UgmA function, since an RFP-UgmA strain had an AnugmAn phenotype (El-Ganiny and Kaminskyj, unpublished). The WC-GFP strain had WT hyphal and colony morphology, and abundant conidiation (data not shown), indicating that AfUgmA-GFP did not interfere with A. nidulans growth. The A. nidulans strains containing GFP-tagged R182A, R327A or H63N had compact colony growth, phenotypically similar to the respective untagged strains. GFP distribution in each strain was cytosolic, excluded from membrane-bounded organelles, and lacked a pronounced longitudinal gradient ( Figure 3A), consistent with wild type distribution of AnUgmA in A. nidulans [14].
To confirm that the even GFP distribution was not due to cleavage of the GFP tag, we extracted total proteins from the wildtype and mutated GFP-tagged strains (WT, R182A-GFP, R327A-GFP, H63N-GFP and AfUgmA-GFP) strains, followed by separation of proteins using 10% SDS-PAGE and western blotting with anti-GFP ( Figure 3B). Notably, the great preponderance of signal was at 87 kDa, consistent with GFP-tagged UgmA, and essentially none was at 27 kDa, which would be consistent with cleaved GFP. The anti-GFP western blotted protein preparations from GFP-tagged mutant strains ( Figure 3B) were consistent with wild type AnUgmA-GFP distribution [14] and with AfUgmA-GFP distribution ( Figure S4 in File S1). This suggested that the phenotype effects of mutated AfUgmA constructs did not relate to alterations in sub-cellular distribution. We do not have an anti-UgmA, so we were unable to perform the complementary experiments to directly measure the intracellular UgmA levels in vivo in order to confirm the stability of AfUgmA.

Alteration in Galf affects wall glucan composition in A. nidulans
Alam et al. 2012 [13] showed using qPCR that gene expression of a-glucan synthase, agsB and b-glucan synthase, fksA were influenced by the level of AnugmA activity. However, synthase gene expression is not necessarily directly related to wall glucan content. We quantified cell wall a-glucan, b-glucan, and Galf using ELISA ( Table 2) and immunofluorescence  in WT, WC, mutated and AnugmAn strains.
Aspergillus nidulans WT and WC strains had strong Galf immunolocalization in hyphal walls ( Figure 4). The F66A strain had the highest level of hyphal wall Galf immunofluorescence of any of the mutants, consistent with ELISA quantification (Table 2) and with its relatively high in vitro UgmA activity ( Table 2). The R182K and H63N strain hyphal wall Galf immunofluorescence quantitation was half that of F66A ( Figure 4), comparable to ELISA results (Table 2) showing Galf content in these strains is 35-45% of F66A. Low hyphal wall Galf was consistent with phenotypic differences, showing concordance between wall composition and hyphal morphology ( Figure 1, Tables 1, 2). WT, WC and F66A cell wall Galf content in (Table 2) was correlated with sporulation (Table 1, Figs. 1, S2 in File S1). The R327K, R327A, R182A strains lacked immunolocalizable Galf and all aspects of their phenotype resembled the AnugmAD strain. Taken together, wall Galf content was consistent with in vitro AfUgmA activity and with other aspects of hyphal and colony development phenotype.
Aspergillus nidulans wild type and WC hyphal walls had relatively low a-glucan content (Table 2, Figure 5) compared to mutant and AnugmAD strains. This is consistent with our previous qPCR studies on gene expression [13], but was more nuanced regarding cellular distribution. Alpha-glucan staining was most pronounced in the older hyphal regions, and was undetectable at wild type and WC hyphal tips. The a-glucan immunofluorescence intensities were stronger in the R327K, R327A, R182K, R182A and AnugmAn strains ( Figure 5), but again were substantially lower at hyphal tips (not shown). ELISA results for a-glucan (Table 2) were comparable to immunofluorescence quantification in older regions of the hyphae. In sum, wall a-glucan content of all the mutants and AnugmAn strains was influenced by alteration in Galf and it was inversely correlated with UgmA enzyme activity levels ( Table  2). Aspergillus nidulans wild type and WC hyphal walls had relatively higher b-glucan levels than the mutants and AnugmAD strains ( Figure 6, Table 2). The quantitative difference in b-glucan level between WT, WC, F66A and all other strains was more pronounced than for a-glucan or Galf. ELISA results for b-glucan were consistent with immunofluorescence quantification, showing a decline in wall b-glucan content that correlated with decrease in UgmA enzyme activity levels ( Table 2).
These data show the relative importance of AfUgm amino acid residues in and near the active site is R327 . R182 , H63 .. F66. Clearly, in vitro enzyme and in vivo phenotype analysis have strong complementarity. We plan to use phenotype analysis as part of our strategy to assess the efficacy of potential UgmA-inhibitor compounds.
Loss of AfUgmA activity leads to increased sensitivity to some antifungal compounds If Galf biosynthesis is to achieve its promise as an antifungal drug development target, most likely as a part of a combination therapy, strains with reduced AfUgmA should be expected to be more sensitive to antifungal compounds. We compared Caspofungin (b-glucan synthesis), Itraconazole (ergosterol synthesis), and Calcofluor White (chitin crystallization) sensitivity on all the strains ( Table 2 and S7 in in File S1). Analysis of the raw drug sensitivity data showed that index values that differed by 0.2 were significantly different, as it had been seen in our previous studies [13]. None of the mutations notably affected sensitivity to Calcofluor White. Strains with substantially reduced Galf (R182 and R327 mutants) were significantly more sensitive to Itraconazole. Furthermore, Caspofungin sensitivity was significantly increased for all of the mutant strains compared to wild type, almost 2-fold for R327K and AnugmAD. Even F66A, which still had a substantially wild type hyphal and colony phenotype, and wall composition was significantly more sensitive to Caspofungin ( Table 2).

Discussion
We used quantitative methods to correlate in vitro UgmA enzymatic function and in vivo fungal cell and developmental phenotype related to site-directed changes in conserved amino acid residues in the AfUgmA catalytic site. We were able to assess residues for which there was structural and in vitro enzymatic function information [22], as well as residues where enzymatic activity had not yet been assessed. These analyses provide a coherent picture of changes in A. nidulans related to the efficacy of an enzyme involved in an early step in Galf biosynthesis in Aspergillus that affects wall composition-structure-function.
Previously we showed qualitatively that A. fumigatus AfugmA restored wild type hyphal morphology in the A. nidulans ugmAD strain [11]. We have extended that preliminary observation with quantitative assessment of colony phenotype and wall composition for A. nidulans whose native AnUgmA had been replaced with wild type AfUgmA or with AfUgmA constructs mutated in or near the enzyme active site. This approach provided a way to correlate structure and function of AfUgmA, regardless of whether gene products could be analyzed in vitro. According to van Straaten et al. [22], for in vitro enzymatic studies where all the mutants (except H63N) were over-expressed, purified and studied, there were no apparent changes in protein stability. There was no evidence of  changes to the overall protein structure since the mutants crystallized under similar conditions and resulted in similar crystallographic structures [22]. In addition, none of the mutations affected the cellular AfUgmA distribution assessed with GFPtagging and anti-GFP western blotting. These results are consistent with the effects we observed being due to changes in AfUgmA activity or other aspects of cellular interaction. Since all the mutants were expressed under endogenous AnugmA promoter to assess phenotype in vivo, it is also possible that these effects might be due to the instability of mutant protein. Although western blot showed the same level of fluorescence intensity for GFP tagged mutants and wild type AfUgmA against anti-GFP antibody, direct intracellular evaluation of these mutants and wild type AfUgmA stability in vivo is beyond our scope at this moment because of lacking an antibody directed against AfUgmA.  Comparing protein structure analysis and in vitro enzymatic activity of wild type AfUgmA with conservative (RK) and nonconservative (RA) mutants showed that R327 was essential to UgmA function [22] in vivo. Structural data in [22] shows that R327 stabilizes the position of the diphosphates of the nucleotide sugar and facilitates positioning of the galactose for catalysis. The R327K mutation impaired enzyme function whereas the R327A mutation produced a strain with no UgmA activity detectable in vitro ( Table 2, and [22]). Consistent with this, A. nidulans strains whose wild type AnUgmA had been replaced with AfUgmA-R327A qualitatively and quantitatively resembled the AnugmAD strain. The role of R182 was slightly less critical than that of R327 (Table 2). R182 is important for sugar orientation related to catalysis, as well as for catalytic efficiency [22]. When AfUgmA-R182K replaced the native AnUgmA, the R182K strain showed significantly better growth and sporulation than AnugmAD, whereas the R182A strain resembled the AnugmAD deletion strain. This is the first time that we have been able to partially rescue the AnugmAD phenotype by genetic means. Previously we had shown that low levels of Calcofluor White, or 1 M sucrose, partially remediated the AnugmAD defects, however this was only for morphogenesis of submerged hyphae, and not for sporulation [11]. Similarly, the R182K strain showed that sporulation was more effectively remediated by low levels of Galf than was hyphal morphogenesis. Galf immunolocalization using EBA2 [11] vs L10 [13] showed substantially different localization patterns. EBA2 preferentially stained conidia, metulae and phialides, whereas L10 stained hyphae. Together these suggest that there are at least two different types of Galf-containing compound in Aspergillus walls.
We also examined the roles of two amino acids that could be functionally important because they are part of loop III, which moves upon redox state of the cofactor [22]. H63 is a highly conserved residue for prokaryotic and eukaryotic UGMs. The H63N construct expressed poorly in vitro so it had not been studied for in vitro enzyme activity. Histidine is positively charged at physiological pH whereas asparagine is polar but not charged. Correlating strain phenotype characteristics with estimates of UgmA activity suggested that the H63N strain UgmA activity would be comparable with the R182K strain. F66 is a highly conserved residue amongst eukaryotic UGMs. The AfUgmA F66A strain showed a 6.3-fold decrease in UgmA activity compared to wild type AfUgmA in vitro ( Table 2). The in vivo F66A phenotype was similar to wild type, indicating that in vivo function of UgmA does not impair colony morphogenesis at this level of decrease in vitro activity. Our results further indicate that changes in AfUgmA loop III affect catalytic activity and that H63 is important but not critical for UGM activity.
Our previous work showed that deletion or down-regulation of any of three sequential genes in A. nidulans Galf biosynthesis was associated with comparable reductions in hyphal growth rate and sporulation [11][12][13][14], with increased sensitivity to fungal walltargeting compounds [12,13], with changes in a-glucan and bglucan synthase gene expression [13], and with changes in wall surface adhesion [21]. Here we have correlated the effect of mutations in AfUgmA (transformed into A. nidulans) with morphometry, with quantitative immunofluorescence, with ELISA quantification of wall carbohydrates, with changes in wall adhesiveness, and with sensitivity to anti-fungal compounds.
Notably, reductions in wall Galf content correlated with increased wall a-glucan content (r 2 = 0.972, P = 0.0001) and in decreased wall b-glucan content (r 2 = 0.980, P = 0.0001) shown in this study were consistent with increased agsB and decreased fksA expression [13]. Not surprisingly, A. nidulans cell wall a-glucan and b-glucan content were also strongly correlated (r 2 = 0.993, P,0.0001). Given that in wild type cells a-glucan and b-glucan are each thought to comprise about 40% of the A. nidulans wall [1,9], this strongly suggests that the cell is metabolically constrained as to where it apportions metabolites for wall synthesis, so that increases in abundance of one major polymer must limit the resources available for synthesis of the others. However, this does not directly explain why engineered changes in Galf content (perhaps 5% of total [15]) are so strongly related to changes in a-glucan and b-glucan content. We suspect these will be coordinated through the cell wall integrity pathway, which is a focus of current research.
Decreased fksA expression and b-glucan content correlated with increased sensitivity to Caspofungin (r 2 = 0.855, P,0.0005), which is mechanistically satisfying. Notably, a ugmA overexpressed strain also showed Caspofungin hypersensitivity [13] suggesting that balanced expression of wall components is important for wild type phenotype and drug resistance.
The mutant (except F66A) and deletion strains also showed smaller but still significant increases in sensitivity to Itraconazole. Previously using AFM force mapping [21] we had found that ugmAD strain wall surfaces were more force-compliant than wild type walls, suggesting that they had reduced resilience. Itraconazole targets ergosterol biosynthesis, and so is expected to affect membrane fluidity. It appears likely that a combination of wall and membrane defects is particularly difficult for a fungal cell to resist. Once UGM inhibitors have been developed, we expect that they will show enhanced efficacy if given in combination with amphotericin B as well as Itraconazole.
Compared to wild type, AnugmAn A. nidulans walls had higher adhesion to AFM Si 3 N 4 probe tips [21]. In addition, AfugmAn strains have been shown to be highly adherent to pulmonary epithelial cells, to glass or plastic surfaces, compared to wild type [15]. Consistent with these previous results, AnumgAn and mutant strains have increased hyphal adherence to latex beads compared to wild type strains (Figure 2). Counter-intuitively, higher adhesion [15] has been correlated with both higher and lower pathogenicity [15,18] but there are few studies of this type and further research is needed. This disparity could be due to lower growth rates in AfUgmA mutants, thus potentially also to lower tissue penetration.
Taken together, we have shown that fungal phenotype can provide useful estimates of functional information on the role of some amino acid residues, even though their mutant gene products were poorly expressed in vitro (i.e. H63N). Our results in this study suggest that UgmA activity correlated with hyphal phenotype: a,7-fold decrease produced a wild type phenotype (wild type AfUgmA, F66A), a ,20-fold decrease resulted in partial function (H63N and R182K), and a $ 70-fold decrease (R182A, R327K, R327A, AnUgmAD) was insufficient to synthesize a functional level of wall Galf.
In conclusion, we have shown that reduced AfUgmA activity due to mutation did impair A. nidulans growth in a manner substantially similar to gene deletion and gene down-regulation. UgmA distribution was not affected by these manipulations. Loss or absence of Galf increased wall a-glucan but reduced wall bglucan (which were strongly correlated), and with increased hyphal surface adhesion and Caspofungin sensitivity. Thus, it seems that A. nidulans needs a balanced expression of UgmA (not high, not low [13]) to generate a resilient wall for normal growth and wall surface integrity. In the future, we anticipate that our approach will also be useful for assessing the effects of UgmA-targeting drugs (under development by the Sanders group).

Strains, plasmids, and culture conditions
Strains, primers and plasmids are listed in Table S1 in File S1. The A. nidulans strains were maintained on complete medium (CM) supplemented for nutritional markers as described in [27]. Bacterial strains were grown on LB with antibiotics as required (Table S1 in File S1). All water was freshly prepared 18 MegOhm ultrapure (Barnstead NanoDiamond).

Site-directed mutagenesis and overexpression of F66A-AfUGM
Site-directed mutagenesis of the F66A AfUGM mutant was performed using the QuikChange TM site-directed mutagenesis kit (Stratagene, Inc.) according to the manufacturer's protocol. Comparable methods were used for creating the other SDM strains [22]. An over-expression vector pET22b harboring the AfUGM gene was used as the template DNA. The PCR mixture contained 50 ng of template DNA and 15 pmole of each primer. PCR amplification was carried out in a GeneAmp PCR PTC100 System. The original methylated plasmid was digested with DpnI, then 2 mL of the reaction was used to transform competent E. coli Dh5a cells (Novagen). Ampicillin-resistant colonies were selected from the LB plates, and the specific mutation was verified by DNA sequencing. The mutant enzyme was overexpressed and purified as previously described [28].

Enzyme kinetics
Kinetic constants for F66A AfUGM mutant were determined as previously described [22]. A fixed concentration of AfUGM mutant protein (500 nM) was chosen so as to have less than 40% conversion to the product UDP-Galp. Reactions were carried out with 0 -300 mM of UDP-Galf in a final volume of 100 mL 50 mM phosphate buffer pH 7.0 and 20 mM freshly prepared sodium dithionite. The incubations were carried out for 1 min at 37uC then quenched with 100 mL n-butanol. The conversion of UDP-Galp to UDP-Galf was monitored at 262 nm using HPLC (Waters). The amount of conversion was determined by integration of the UDP-Galp and UDP-Galf peaks. The initial velocity was calculated from the substrate concentration and percentage UDP-Galp conversion. Kinetic parameters were determined with GraphPad Prism software (GraphPad Software, San Diego, CA) using nonlinear regression analysis.

Mutagenesis and transformation
Strains used in this study are shown in Table S1 in File S1. The wild type (WT) and wild type complemented (WC) strains were compared with AfUgmA mutants R327K, R327A, R182K, R182A, F66A and H63N, and with the AnugmAn strain.
AfugmA constructs for mutations in the AfUgmA active site and in AfUgmA loop III that had been used for structural studies [22] were used to generate AnugmA replacement constructs using fusion PCR ( Figure S3 in File S1) according to [29]. We used an AfpyrG selectable marker controlled by the a-tubulin promoter [tubA(p)-pyrG] to ensure a constitutive level of marker expression, whereas the wild type or mutated version of AfugmA was controlled by their endogenous promoters ( Figure S3 in File S1). Constructs were transformed into wild type protoplasts [29,30]. Confirmation of the correct gene manipulations used genomic DNA (isolated as described in [31]) from putative transformant strains as a template for PCR ( Figure S5 and S6 in File S1) with combinations of primers as shown Table S1 in File S1. Prior to phenotype analysis, genomic DNA from each A. nidulans strain was extracted. The ugmA sequence was amplified from these genomic DNA samples using PCR. Primers that we used to amplify ugmA target 100 bp upstream and 100 bp downstream (inside of tubA promoter) ( Figure S3 in File S1) of ugmA. PCR products were sequenced in order to confirm the expected construct generation and replacement. Fusion construction for replacement cassettes used primers that were ,50 bp inside of upstream and downstream primers from both sides. To confirm the correct insertion site, we used the outside pair of primers.

Colony growth, sporulation and surface adhesion
Colony characters were examined as in [11]. Strains were streaked on CM and incubated for 3 d at 28 uC to give isolated colonies. The diameter of ten colonies/strain was measured to the nearest millimeter using a dissection microscope. The number of spores produced per colony was counted for four colonies/strain.
The hyphal surface adhesion assay was modified from [15]. Fluorescent (excitation maximum 520 nm; emission maximum 540 nm) 0.5 mm diameter polystyrene beads (Sigma: aqueous suspension, 2.5% solids content) were diluted 1:10 in sterile phosphate buffered saline (PBS). One hundred microliters of bead solution was added to CM liquid containing 20,000 spores and incubated for 8 h at 37uC with 150 r.p.m. Images of germlings were collected using a Zeiss META510 confocal microscope with a 63x, 1.2 N.A. multi-immersion objective, a 514 nm excitation from Ar ion laser at 20% power, and a 530-600 nm emission filter.

Cell wall preparation
Cell wall extraction was performed according to [32]. Colonies were grown in shaken liquid at 37uC for 48 h, filtered through Whatman #1 filter paper, washed with ultrapure water, and then with 0.5 M NaCl. Fungal hyphae were broken using 1 mm glass beads in buffer [20 mM Tris, 50 mM EDTA, pH 8.0]. Cell walls were separated from cytoplasmic debris by centrifugation at 3000 x g for 10 min. The pellet containing the cell wall fraction was washed with same buffer with stirring for 4 h at 4uC, followed by a wash with ultrapure water. The pellet was frozen at -80uC, then lyophilized overnight.

Protein extraction and western blot
To extract total protein, A. nidulans conidia were grown in complex media (CM) for 20 h. Mycelia were then collected from liquid cultures by filtration through Miracloth paper, washed with distilled water, dried on paper towels, and ground to a powder with mortar and pestle in liquid nitrogen [33]. The ground mycelium was resuspended and boiled in the urea/SDS buffer [34]: 1% SDS, 9 M urea, 25 mM Tris-HCl (pH 6.8), 1 mM EDTA, and 0.7 M 2 mercaptoethanol). The debris was removed by centrifugation at 13000 x g. Total protein concentration was determined by the Bio-Rad protein assay. UgmA-GFP fusions were detected on Western blots using purified mouse anti-GFP polyclonal antibody (eBioscience). IRDye 800-conjugated antimouse IgG was used as a secondary antibody. Bands were detected using the Odyssey Infrared Imaging System (LI-COR Biosciences). Images were processed by Odyssey 3.0.16 application software (LI-COR Bioscience).

Microscopy
Samples were prepared for confocal microscopy as described in [11]. Galf, a-glucan and b-glucan were immunolocalized using the same primary antibodies as described for ELISA. Primary antibodies were used at full strength for L10, at 1:50 dilutions for a-glucan and b-glucan. GFP was immunolocalized using anti-GFP primary antibody (eBioscience) at a dilution of 1:100. TRITC-conjugated goat-anti-mouse was used as secondary antibody at 1:100 dilution. Immunofluorescence procedures followed [13]. Samples were examined using confocal epifluorescence microscopy. The relative immunofluorescence intensity was quantified using Zeiss image browser software at 25 isolated cellular sites at hyphal wall where we had near-median confocal slices of subapical cells.
Scanning electron microscopy was used to examine conidiation following preparation as in [13]. Strains were grown on dialysis tubing laid on CM for 3 d at 28uC. Colonies were fixed by immersion in 1% glutaraldehyde, dehydrated in acetone, criticalpoint dried (Polaron E3000, Series II), and gold sputter coated (Edwards model S150B). Samples were imaged with a JEOL 840A scanning electron microscope.

Drug sensitivity testing
Sensitivity to Caspofungin, Itraconazole and Calcofluor White of wild type and engineered UgmA strains was measured using a disc-diffusion assay as described in [12,13]. Stock solutions: itraconazole (1.6 mg/mL in DMSO), calcofluor white (10 mg/ mL in 25 mM KOH), and caspofungin (20 mg/mL in sterile water) were stored at -80uC. For this test, 1610 7 spores were mixed into 20 mL of 50uC CM agar and immediately poured in 9 cm Petri plates. Stock solutions were pipetted onto sterile 6 mm paper discs placed on the agar surface: 20 mL of CFW, 10 mL of itraconazole, and 20 mL of caspofungin. Solvent controls, DMSO and 50% ethanol, showed no zone of inhibition. The plates were incubated at 30uC for 48 h, before measuring the zone of inhibition [13].
Drug sensitivity is presented as an index compared to wild type on the same medium, so that values greater than 1.0 indicate hypersensitivity. Statistical analysis of the raw data and comparison to sensitivity indices showed that a difference of .0.2 in index value correlated with statistical significance. Representative plates from the caspofungin study are shown in Figure S7 in File S1. Quantitative results are shown in Table 2.