Genetic and structural validation of phosphomannomutase as a cell wall target in Aspergillus fumigatus

Aspergillus fumigatus is an opportunistic mold responsible for severe life‐threatening fungal infections in immunocompromised patients. The cell wall, an essential structure composed of glucan, chitin, and galactomannan, is considered to be a target for the development of antifungal drugs. The nucleotide sugar donor GDP‐mannose (GDP‐Man) is required for the biosynthesis of galactomannan, glycosylphosphatidylinositol (GPI) anchors, glycolipid, and protein glycosylation. Starting from fructose‐6‐phosphate, GDP‐Man is produced by the sequential action of the enzymes phosphomannose isomerase, phosphomannomutase (Pmm), and GDP‐mannose pyrophosphorylase. Here, using heterokaryon rescue and gene knockdown approaches we demonstrate that the phosphomannomutase encoding gene in A. fumigatus (pmmA) is essential for survival. Reduced expression of pmmA is associated with significant morphological defects including retarded germination, growth, reduced conidiation, and abnormal polarity. Moreover, the knockdown strain exhibited an altered cell wall organization and sensitivity toward cell wall perturbing agents. By solving the first crystal structure of A. fumigatus phosphomannomutase (AfPmmA) we identified non‐conservative substitutions near the active site when compared to the human orthologues. Taken together, this work provides a genetic and structural foundation for the exploitation of AfPmmA as a potential antifungal target.

new antifungal targets against A. fumigatus in order to feed into drug discovery pipelines.
The fungal cell wall is a complex and highly dynamic structure that is essential for cellular morphology and protection against environmental stresses and is considered to be a potential drug target. It is primarily composed of the polysaccharides chitin, glucan, and galactomannan (Gow et al., 2017;Latge et al., 2005). In A.
fumigatus, galactomannan is a highly complex structure containing different types of glycosidic linkages, produced by several mannosyltransferases (Lee & Sheppard, 2016). As an active form of mannose, GDP-mannose is not only required for biosynthesis of fungal galactomannan, but also plays an important role in biosynthesis of O-and N-linked glycoproteins, glycosylphosphatidylinositol (GPI) anchors and glycolipids Onoue et al., 2018).
For example, Saccharomyces cerevisiae Pmm was shown to be essential for cell viability under in vitro laboratory conditions (Kepes & Schekman, 1988). Similar findings were observed in Arabidopsis thaliana, where failure of obtaining the pmm deletion and knockdown mutants indicated pmm essentiality. Reduced expression of pmm leads to a decrease in levels of the antioxidant ascorbic acid (AsA) and protein glycosylation (Hoeberichts et al., 2008).
Antifungal drug development has historically been impeded by evolutionary similarities between fungi and their human host.
Thus, characterizing the biological functions and structural properties of potential drug targets is an essential prerequisite for rational design of novel inhibitors (Hu et al., 2007). Structure-based drug discovery has been developed rapidly in the last two decades. Particularly, with the emergence of fragment-based drug discovery (FBDD) (Erlanson et al., 2016), it is possible to identify weakly binding small fragments (usually molecular mass < 300 Da) targeting binding sites away from highly conserved active sites.
These fragments can be converted to potent inhibitors with high selectivity by iterative optimization based on structural and enzymological information (Scott et al., 2012). To date, although the function of phosphomannomutase has been well characterized in many eukaryotes, the physiological function of PmmA in the human opportunistic pathogen A. fumigatus remains unclear.
Moreover, active site sequence conservation of this enzyme family with the human orthologues hampers the development of specific inhibitors. Here, we demonstrate that AfPmmA is indispensable for viability, morphogenesis, and cell wall integrity in A.
fumigatus. Importantly, we reveal potential exploitable differences in AfPmmA structure compared to its human orthologues. This work forms the basis for the initiation of structure-based inhibitor design against AfPmmA.

| AfpmmA is essential for A. fumigatus viability in vitro
To investigate the role of pmmA in A. fumigatus, we attempted to construct a null mutant using the Neurospora crassa pyr-4 selectable marker to replace the open reading frame of AfpmmA by homologous recombination ( Figure 2a). However, we failed to obtain any positive transformants, suggesting that AfpmmA may be essential for growth under in vitro laboratory conditions. To further confirm essentiality of AfpmmA, a heterokaryon rescue technique was employed (Osmani et al., 2006). As shown in Figure 2b, all heterokaryons were not viable on selective media (YAG) but were able to grow on nonselective (YUU) media. Moreover, PCR analysis showed that these heterokaryons contained both wild type and deleted alleles (Figure 2c), indicating that pmmA is an essential gene in A. fumigatus which is consistent with previous studies in S. cerevisiae and Kluyveromyces lactis (Kepes & Schekman, 1988;Staneva et al., 2004). As an alternative strategy, we constructed a conditional inactivation mutant by replacing the native promoter of the pmmA gene with the Aspergillus nidulans alcohol dehydrogenase promoter (P alcA ) that is inducible by ethanol, glycerol, or threonine and repressed by glucose (Romero et al., 2003). Over fifty transformants were obtained and genotyped. The correct F I G U R E 1 Kinetic analysis of recombinant Aspergillus fumigatus phosphomannomutase (AfPmmA). (a) Schematic illustrations of AfPmmA assays. Phosphoglucomutase assay: coupled assay with glucose-6-phosphate dehydrogenase (G6PDH) using Glc-1P as substrate. Phosphomannomutase assay: coupled assay with A. fumigatus phosphomannose isomerase (Pmi) and phosphoglucose isomerase (Pgi) using Man-1P as substrate. (b,c) Kinetic parameters of AfPmmA catalyzed conversion of Glc-1P to Glc-6P (b) or Man-1P to Man-6P (c) in 50 mM HEPES pH 7.1, 5 mM MgCl 2 , 0.25 mM NADP + by monitoring NADPH production. 10 μM Glc-1,6-bisP was used as cofactor. The results are the mean ± SD for three determinations. (d) Effect of Glc-1,6-bisP, EDTA and metal ions on AfPmmA activity. The metal ions were added at 5 mM concentration and the activities were measured using 200 μM Glc-1P as substrate. Control indicates 10 μM Glc-1,6-bisP and 5 mM Mg 2+ were added but without addition of EDTA or other metal ions   gether, these data suggest that pmmA is essential for A. fumigatus viability in vitro.

| pmmA is required for morphogenesis in A. fumigatus
When grown under solid MMTG the radial hyphal growth of the P alcA ::pmmA conditional strain was decreased to approximately 45% of that of the wild type at each time point investigated, indicating in vitro growth was affected by the repression of pmmA ( Figure 4a). Apart from this, the P alcA ::pmmA conditional strain also showed reduced conidiation under repressive conditions ( Figure 4b). We next used differential interference contrast (DIC) microscopy to examine the edges of the colonies under inducing and repressing conditions. An abnormal morphological phenotype of hyper-branching at the hyphal tips was observed when pmmA is repressed ( Figure 4c).
To further explore the effect of pmmA repression in A. fumigatus, the germination rate and pattern were investigated under the repressing condition. The P alcA ::pmmA conditional strain displayed 8 hr delayed germination comparing to the wild type ( Figure 5a).
Moreover, a defective germination pattern was observed for the P alcA ::pmmA conditional strain with 22% morphological abnormalities
CFW and CR interfere with the cell wall structure by inhibiting the enzymes involved in connecting chitin to β-1,3-glucan and β-1,6glucan (Ram & Klis, 2006) whereas Caspofungin inhibits the synthesis of β-1,3-glucan (Kahn et al., 2006). Repression of pmmA caused hypersensitivity of the P alcA ::pmmA conditional mutant to CR and CFW but did not alter sensitivity to Caspofungin (Figure 6a Figure S2). Next, to test whether the phenotypic defects with cell wall perturbing agents were due to changes in the cell wall, we examined hyphal cell wall ultrastructure of wild type and the P alcA ::pmmA mutant by transmission electron microscopy (TEM). As shown in Figure 6c,d, the P alcA ::pmmA conditional mutant showed a much thinner cell wall compared to the wild type under repressing conditions. In contrast, no difference was observed between the wild type and conditional mutant with the induction of pmmA expression (Figure 6c,d). Moreover, we quantified the individual cell wall F I G U R E 3 Growth phenotype of the P alcA ::pmmA conditional mutant under inducing and repressing growth conditions. (a) A series of 10-fold dilutions of the indicated strains were inoculated at 37°C for 48 hr on YEPD, CM, and MM supplemented with 1% and 2% glucose (w/v), 0.1 M ethanol, 0.1 M glycerol, 0.1 M threonine. (b) Relative transcription level of AfpmmA gene under induction (MMT) and partial repression (MMTG, MMT with 1% glucose) conditions. Gene expression levels were normalized to the reference gene tbp. Error bars indicate mean ± SD from three independent experiments. Data were analyzed using an unpaired of t-test, statistical significance is indicated by p values; ns, not significant (b) (a) 10 6 10 5 10 4 10 3 10 6 10 5 10 4 10 3 YEPD CM MMT + 1% Glucose MMT + 2% Glucose 10 6 10 5 10 4 10 3 10 6 10 5 10 4 10 3 10 6 10 5 10 4 10 3 10 6 10 5 10 4 10 3 10 6 10 5 components of the wild type and P alcA ::pmmA mutant strains by highperformance ionic chromatography (Francois, 2006). Repression of pmmA expression led to the reduction of the galactose and mannose content by 56% and 57%, respectively, whereas the amounts of chitin and glucan were increased by 47% and 49%, respectively Conidia (1 × 10 6 ) were incubated in liquid partial repression (MMTG) media for the time indicated. One hundred conidia for each strain were assessed for germination rate and the experiment was performed in triplicate. Values represent the mean ± SD. (b) Germination pattern of the P alcA ::pmmA conditional mutant and the wild-type conidiospores. 1 × 10 6 conidia of the conditional mutant and the wild type were allowed to germinate in liquid partial repression (MMTG) media for 9 hr and 24 hr, respectively. Germination morphology was classified (1-4) as shown in figure. The experiment was performed in triplicate. Values represent the mean ± SD. (c) 1 × 10 6 conidia from each strain were inoculated into 2 ml liquid partial repression (MMTG) medium for the indicated time point at 37°C and examined by differential interference contrast (DIC) using Axio Scope A1 (Zeiss). Scale bar represents 20 μm were similar to those in the wild type upon induction of pmmA expression ( Figure 7a). As GDP-Man is not only important for cell wall integrity but also involved in protein glycosylation, we extracted the total and secreted proteins from the wild type and P alcA ::pmmA conditional mutant to detect protein glycosylation levels by western blot using biotinylated Concanavalin A. As shown in Figure 7b, the repressed P alcA ::pmmA conditional mutant exhibited reduced glycosylation levels in secreted proteins, whereas no significant change was observed for intracellular proteins, suggesting that repression of AfPmmA leads to a reduction of secreted mannosylated proteins.
Collectively, these results show that AfpmmA expression affects A. fumigatus cell wall organization and secreted protein glycosylation.

| The AfPmmA crystal structure reveals exploitable differences
Our data so far suggest that AfPmmA is a genetically validated antifungal target in A. fumigatus. However, the high sequence conservation to the human orthologues suggests that mechanism-inspired inhibitors could elicit toxicity. We next determined the crystal structure of AfPmmA to find potential exploitable differences compared The overall crystal structure of AfPmmA is similar to that of C.
Enzymatic activity was completely abolished for Man-1P, suggesting that these residues are critical for catalytic activity of AfPmmA. As a member of the β-D-phosphohexomutase superfamily a conserved aspartate, corresponding to Asp25 in AfPmmA, was proposed as a nucleophile to form a transient phosphoenzyme intermediate (Jin et al., 2014). However, we did not see crystallographic evidence of phosphorylation of Asp25, which is in agreement with the reports on human and L. mexicana Pmm structures (Kedzierski et al., 2006;Silvaggi et al., 2006). This is likely due to the fact that aspartic acid is only transiently phosphorylated during catalysis.
Structural superposition of AfPmmA with the structure of

| D ISCUSS I ON
In order to circumvent resistance and treat infections of the opportunistic pathogen A. fumigatus, the development of novel antifungals targeting critical cellular components or biological mechanisms is urgently needed (Denning & Bromley, 2015). Enzymes participating in fungal cell wall biogenesis present attractive drug targets due to their essential biological roles. For instance, the antifungal drug class echinocandins target β-1, 3-glucan synthase, a key enzyme essential for cell wall integrity. Several lines of evidence suggest that GDP-mannose, the donor substrate for biosynthesis of cell wall mannan, O/N-glycans, and GPI anchors, plays an important role in fungal growth and development F I G U R E 7 Cell wall monosaccharide contents and glycosylated proteins of the P alcA ::pmmA conditional mutant compared to the wild type. (a) 1 × 10 6 conidia were incubated in 100 ml liquid partial repression (MMTG) (left panel) and induction (MMT) (right panel) media at 37°C for 48 hr. The mycelia were harvested and 10 mg dry mycelium samples were used for analysis as described in Experimental procedures. Measurements were performed in three independent replicates with six technical replicates. Data were analyzed using an unpaired of t-test, statistical significance is indicated by p values; ns, not significant. Several genetic studies have demonstrated that Pmm is indispensable for viability in many eukaryotes (Hoeberichts et al., 2008;Kepes & Schekman, 1988;Staneva et al., 2004). In agreement, heterokaryon rescue and phenotypic analysis of a conditional mutant employing the alc(A) promoter confirmed that pmmA is an essential gene in A. fumigatus (Figures 2 and 3).
A Glc-1,6-bisP cofactor is required for maintaining the active site of this family of enzymes in the phosphorylated state (Knowles, 1980).
In this work, there was no electron density for the phosphorylation of the AfPmmA active site Asp25, in line with the absence of AfPmmA activity in the absence of Glc-1,6-bisP (Figure 1d).

PDB entry 6I5X
Note: Data between brackets represent the highest resolution shell.
Compounds targeting essential genes such as pmmA may have substantial toxicity due to evolutionary conserved enzymatic properties and the high structural similarity with its human orthologues.
To date, information on inhibitors of this class of enzymes is limited to a single report describing the dye Disperse Blue 56 (2-chloro-1,5- Trypanosoma brucei (Fang et al., 2013b;Urbaniak et al., 2013). Similarly, although N-myristoyltransferase (Nmt) from Plasmodium falciparum only displays a single residue difference at the active site from human Nmt, selective inhibitors have been successfully identified and further optimized (Bell et al., 2012;Rackham et al., 2014). Thus, it is possible to develop potent selective inhibitors against AfPmmA based on the structure differences from human Pmms we presented here.
In conclusion, the genetic and structural analysis of PmmA in A.
fumigatus reported here suggests that AfPmmA is a potential target for the development of antifungal drugs. Future efforts are required to discover inhibitors targeting this key enzyme in cell wall biosynthesis that may possess anti-fungal activity.

| Strains and culture conditions
A. fumigatus Ku80ΔpyrG was the recipient parental strain for generating the mutants described in this work (da Silva Ferreira et al., 2006).  Minimal medium (MM) supplemented with 0.1 M glycerol, 0.1 M threonine or 0.1 M ethanol as carbon sources was used to induce gene expression. YEPD (2% w/v yeast extract, 2% w/v glucose and 0.1% w/v peptone) medium and CM (complete medium) were used to completely and partially repress gene expression (Armitt et al., 1976;d'Enfert, 1996). YAG (2% w/v glucose, 0.5% w/v yeast extract, trace elements and 2% w/v agar) and YUU (YAG supplemented with 1.2 g/L each of uracil and uridine) were used for heterokaryon rescue assays. (Todd et al., 2007). A. fumigatus conidia were grown on minimal medium at 37°C for 48 hr and harvested in sterile water supplemented with 0.02% (v/v) Tween 20, and counted in a hemocytometer. Conidia were stored in 20% glycerol stock at −80°C for long-term storage or in sterile water at 4°C for short term storage.

| Heterokaryon rescue
Heterokaryon rescue assays were performed as previously described (Osmani et al., 2006). Conidia from heterokaryotic primary transformants were replica streaked onto selective (YAG) and nonselective (YUU) plates for the pyrG marker of the pmm deletion cassette. Heterokaryons were confirmed by diagnostic PCR using primer P19/P20 for wild-type alleles and primers P21/P22 for deletion alleles.

| Construction of the A. fumigatus pmmA conditional inactivation mutant
The pAL3 plasmid (Romero et al., 2003) containing the alcohol dehydrogenase promoter (P alcA ) and the N. crassa pyr-4 gene as a fungal selectable marker was used to construct a vector allowing the replacement of the native promoter of the pmmA gene by P alcA . The fragment from −60 to +957 of the pmmA genomic DNA sequence was amplified with primers P7 and P8 (Table S1). The PCR-amplified fragment was cloned into the expression vector pAL3 to yield pALP-mmN and confirmed by sequencing using the University of Dundee sequencing service. Generation of protoplasts and polyethylene glycol-mediated transformation were performed as previously described (Tilburn et al., 1983) and positive transformants were se-

| Analysis of the P alcA ::pmmA conditional mutant
To test the sensitivity of the conditional mutant to cell wall perturbing reagents, serial dilutions of P alcA ::pmmA and wild-type conidia from 10 5 to 10 2 were inoculated on MMT or MMTG plates containing 100 μg/ml Calcofluor White, 100 μg/ml Congo Red and 5 μg/ml Caspofungin, respectively. After incubation at 37°C for 48 hr, the plates were photographed.
For quantitative determination of cell wall monosaccharides, conidia were inoculated into 100 ml MMTG liquid medium at a concentration of 10 6 conidia ml -1 and incubated at 37°C with shaking at 200 rpm for 48 hr. The mycelia were harvested, washed with deionized water and stored at −80°C. Fungal cell wall monosaccharides were extracted and quantitatively determined as described previously (Francois, 2006). For the conidial germination assay, 1 × 10 6 conidia ml -1 were inoculated into 20 ml liquid MMTG containing a single coverslip and incubated at 37°C.

| Western blotting
The indicated strains were inoculated into liquid MMTG and cultured at 37°C, 200 rpm for 48 hr. The mycelia were harvested and ground in liquid nitrogen with a mortar and pestle, then resuspended in lysis buffer (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1 mM PMSF, protease inhibitor mixture), and centrifuged at 10,000g for 15 min at 4°C. Supernatants were collected as total proteins. Secreted proteins from culture media were concentrated by acetone precipitation. The total and secreted proteins were separated in 12% SDS-PAGE gels and then transferred into a polyvinylidene difluoride (PVDF) membrane (Millipore). The blots were probed with biotinylated ConA (1:2000, Vector Laboratories, B-1005-5) and subsequently with horseradish peroxidase-conjugated streptavidin (1:1,000, Vector Laboratories, SA-5014-1). Blots were detected by Enhanced ECL luminescence detection kit (Vazyme, E411), and images were acquired with a Tanon 4200 chemiluminescent imaging system (Tanon).

| High-pressure freeze substitution transmission electron microscopy
Strains were grown in liquid MMTG or MMT media at 37°C for 24 hr.
The harvested mycelia were frozen under pressure using a Leica EM AFS2 automatic freeze substitution system and EM FSP freeze substitution processor (Leica Microsystems). The freeze substitution and embedding, sectioning, and staining was performed by the Microscopy and Histology Facility of the Institute of Medical Sciences, University of Aberdeen as described previously (Hall et al., 2013). Samples were examined using a Philips CM10 transmission microscope (FEI UK Ltd., Cambridge, United Kingdom), and images were captured using a Gatan BioScan 792 camera system (Gatan UK, Abingdon, United Kingdom).
The average thicknesses of cell wall layers were calculated from 10 measurements for each strain.

| Cloning of A. fumigatus pmmA
The A. fumigatus pmmA gene (accession no. Q4WNF2) was amplified by PCR from an A. fumigatus cDNA library using primers P1/P2 (Table S1) for cloning into plasmid pGEX-6P1 (GE Healthcare). This generated the final expression plasmid pGEX-AfPmmA 12-269 (amino acids 12-269). This vector encodes a glutathione-S-transferase (GST) tag followed by a PreScission protease cleavage site. Sitedirect mutagenesis of D25N and D27N was performed using pGEX-AfPmmA 12-269 as the template, P15, P16 and P17, P18 (Table S1) as primers following the QuickChange protocol (Stratagene). All plasmids were verified by sequencing using the University of Dundee sequencing service.

| Expression and purification of AfPmmA
The N-terminally truncated pGEX-AfPmmA 12-269 and mutated forms (D25N, D27N) were transformed into E. coli BL21 (DE3) pLysS and cultured in Luria-Bertani (LB) medium supplemented with ampicillin (0.1 mg/ml) at 37°C. Then, 10 ml culture was used to inoculate 1 liter LB medium and grown to an OD 600 of 0.6. Protein expression was induced by 250 μM of IPTG (isopropylβ-D-thiogalactopyranoside) and then incubated at 16°C for 18 hr. The cells were harvested by centrifugation at 3,500 rpm, 4°C for 30 min. The cell pellet was resuspended in 25 ml of ice-cold lysis buffer containing 25 mM HEPES pH 7.5, 150 mM NaCl, 10 mg/ml DNase, 0.5 mg/ml lysozyme and a tablet of protease inhibitor cocktail (Roche) and lysed using a French press at 600 psi. After centrifugation (20,000g, 30 min, 4°C), the supernatant was incubated with pre-washed glutathione Sepharose 4B beads (GE Healthcare) at 4°C on a rotating platform for 2 hr and subsequently the GST tag was cleaved with PreScission protease by incubating at 4°C for 18 hr. The eluted solution was concentrated to 5 ml using a 10 kDa cut-off Vivaspin concentrator (GE Healthcare) and loaded onto a Superdex 200 column (Amersham Bioscience) equilibrated with the same lysis buffer and eluted at a flow rate of 1 ml/min. The fractions were concentrated using a 10 kDa cutoff Vivaspin concentrator (GE Healthcare) and verified by 10% SDS-PAGE.

| Steady-state kinetics
A. fumigatus PmmA activity was determined via a coupled fluorescent assay as reported previously (Pirard, Achouri, et al., 1999) with minor modifications. Briefly, 10 µM glucose-1,6-bisphosphate was used as the co-factor and the reaction mixture was incubated at 30°C for 30 min in a buffer consisting of 50 mM HEPES pH 7.1, 5 mM MgCl 2 , 0.25 mM NADP + and 10 µg/ml glucose-6-phosphate dehydrogenase. Phosphoglucomutase activity was measured in the presence of 0 to 500 µM glucose 1-phosphate, and phosphomannomutase activity was measured in the presence of 0 to 300 µM mannose-1-phosphate using 10 µg/ml of phosphoglucose isomerase (Pgi) and 3.5 µg/ml of phosphomannose isomerase (Pmi), respectively. The production of NADPH was determined using a SpectraMax i3x (Molecular Devices) with emission at 440 nm and excitation at 340 nm. To detect the effects of divalent metal ions on AfPmmA activity, 1 mM EDTA and 5 mM each metal ion (Mg 2+ , Ca 2+ , Mn 2+ , Zn 2+ ) were added and the activities were measured using 200 μM Glc-1P as substrate.  et al., 2006) as the search model. REFMAC (Murshudov et al., 1997) was used for further refinement and iterated with model building using COOT (Emsley & Cowtan, 2004). Figures were produced with PyMol (DeLano, 2004). The atomic co-ordinates and structure factors of AfPmmA were deposited in the Protein Data Bank with accession code 6I5X.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The atomic co-ordinates and structure factors of AfPmmA were deposited in the Protein Data Bank with accession code 6I5X.