Abstract
Aspergillus fumigatus is a fungal pathogen that causes a disease called invasive aspergillosis in immunocompromised patients, with mortality rates of approximately 50%. The risk of A. fumigatus infection is rising due to the increasing use of immunosuppressants in medical treatments such as organ transplantation. Only limited numbers of compound classes (polyenes, azoles, echinocandins) are available for antifungal treatment with multiple issues including drug resistance, toxicity and undesirable drug-drug interactions. In recent decades, pharmaceutical companies have carried out research for novel antifungal agents with modest progression, which is partially due to the lack of genetically and chemically characterized targets.The A. fumigatus cell wall provides the cell with rigidity against hydrostatic pressure within the cell. Carbohydrate is the main (72% of dry weight) component of the A. fumigatus cell wall. Cell wall carbohydrates include glucan, chitin, galactomannan and galactosaminogalactan (GAG). These carbohydrates contribute to viability and virulence of A. fumigatus, suggesting that interfering with cell wall synthesis would inhibit the growth of A. fumigatus. Biosynthesis of cell wall carbohydrates is carried out by Leloir glycosyltransferases, which are membraneembedded enzymes that convert sugar nucleotides to cell wall carbohydrates. Therefore, targeting Leloir glycosyltransferases may affect cell wall carbohydrates and in turn elicit antifungal activity, i.e., Leloir glycosyltransferases are potential antifungal targets against A. fumigatus. A successful example of this is clinical use of echinocandins, lipopeptides inhibiting the activity of β-1,3 glucan synthase that synthesize β-1,3 glucan from UDP-Glc. Although inhibitors of other glycosyltransferases (e.g. chitin synthases) also exhibit antifungal activity, only echinocandins have been approved for the treatment of aspergillosis. This is due to intrinsic drawbacks of targeting glycosyltransferases. Firstly, the membrane-embedded nature of glycosyltransferases complicates protein expression and purification, which therefore hampers inhibitor identification. In addition, functional redundancy of glycosyltransferases may weaken antifungal potency of inhibitors. Moreover, targeting a single class of glycosyltransferases may only affect one carbohydrate, which may trigger increasing biosynthesis of other carbohydrates, i.e., compensatory biosynthesis of other carbohydrates. For instance, loss of β-1,3 glucan can be compensated by an increase in GAG and chitin. Therefore, cell wall integrity is likely to be maintained through compensatory increases in other carbohydrates, which may weaken antifungal activity of glycosyltransferase inhibitors.
As activity of glycosyltransferases relies on supply of the substrates (sugar nucleotides), targeting sugar nucleotides also affects cell wall carbohydrates and in turn elicits antifungal activity. Sugar nucleotide biosynthetic pathway suggests that cell wall carbohydrates are made from three key sugar nucleotides: UDP-GlcNAc, UDP-Glc and GDP-Man. Targeting UDP-GlcNAc may affect chitin and GAG. Targeting UDP-Glc may affect glucan, galactomannan and GAG. Targeting GDP-Man may affect the mannan backbone of galactomannan. Therefore, targeting a single sugar nucleotide may affect several cell wall carbohydrates, through which the compensatory issue may be circumvented. Taken together, targeting sugar nucleotides may elicit antifungal activity and circumvent drawbacks of glycosyltransferases. In addition to affecting cell wall carbohydrates, targeting sugar nucleotides may also affect other biological processes including N-glycosylation and GPI anchor biosynthesis, which may also contribute to antifungal activity. In the thesis, I will focus on UDP-GlcNAc and UDP-Glc.
Sugar metabolic pathway suggests that UDP-GlcNAc is synthesized from GlcNAc-1P by UDPGlcNAc pyrophosphorylase (UAP1). GlcNAc-1-P is synthesized from GlcNAc-6-P by Phospho N-acetylglucosamine mutase (AGM1). Therefore, targeting AGM1 and/or UAP1 may affect UDP-GlcNAc biosynthesis and in turn elicit antifungal activity. Similarly, UDP-Glc is synthesized from Glc-1-P by UDP-Glc pyrophosphorylase (UGP). Glc-1-P is synthesized from Glc-6-P by phosphoglucomutase (PGM). Targeting UGP and/or PGM may affect UDP-Glc and therefore elicit antifungal activity. Taken together, these four enzymes are likely to be antifungal targets against A. fumigatus. In this thesis, I will focus on three enzymes: AGM1, UGP and PGM. To demonstrate that the three enzymes are indeed antifungal targets, it is required to perform target validation of the three enzymes in A. fumigatus, which includes genetic and chemical validation. Genetic validation demonstrates that a target is essential for expected phenotypes (e.g. viability, cell wall integrity). Chemical validation provides information of chemical probes (e.g. enzyme inhibitors) that can phenocopy the expected phenotypes by probing the target in vivo. Furthermore, human orthologues of the three enzymes are essential for human health. To avoid toxicity, chemical probes should preferentially interact with the three enzymes over their human orthologues. Currently only AGM1 has been genetically validated as an antifungal target against A. fumigatus. No genetic validation of UGP and PGM has been performed for A. fumigatus. Furthermore, no chemical validation of the three enzymes has yet been reported.
In this thesis, I will describe target validation of the three enzymes (AGM1, UGP and PGM)for A. fumigatus. Using a promoter replacement approach, a conditional A. fumigatus mutant of ugp has been constructed, revealing that the ugp gene is required for the growth of A. fumigatus. Using the same approach, I have also demonstrated that pgm is essential for A. fumigatus growth. These results are the first genetic validation of ugp and pgm in A. fumigatus.
As the three enzymes have been demonstrated as essential for A. fumigatus growth, inhibitors of these enzymes may be antifungal agents against A. fumigatus, i.e., chemical probes that phenocopy retarded growth elicited by knockdown of the corresponding target. Currently no inhibitor of A. fumigatus PGM (AfPGM) has been reported. Similarly, no inhibitor of A. fumigatus UGP (AfUGP) has been reported. Although inhibitors of A. fumigatus AGM1 (AfAGM1) have been reported, these inhibitors exhibit severe drawbacks (e.g. protein aggregator) and therefore cannot be used as chemical probes. Lack of inhibitor hampers chemical validation of the three enzymes. In this thesis, using kinetic assays, I have demonstrated that 6-hydroxy-DL-DOPA is an uncompetitive inhibitor of AfAGM1 (Ki 10 μM) with marginal selectivity over the human orthologue (HsAGM1). In addition, using crystallography and fragment screening, I have demonstrated that a resveratrol scaffold can fit into a surface pocket on domain IV of the AfAGM1 protein. The binding mode of resveratrol suggests possibility of developing selective allosteric inhibitors of AfAGM1. In terms of AfUGP, I have identified a non-conserved groove proximal to the active site of AfUGP using crystallography, which suggests the possibility of improving selectivity for mechanism-inspired inhibitors. In addition, I have also identified a non-conserved pocket that suggests selective allosteric inhibitors of AfUGP. To identify starting points for inhibitor identification, fragment screening for AfUGP has been performed. Two fragments that stabilize the AfUGP protein have been identified from a library of fragment compounds although the binding modes of these two fragments are still unknown. With regards to AfPGM, I have identified a thiol reactive compound (ISFP1) that inhibits AfPGM activity (IC50 1 μM). A crystal structure of the PGMISFP1 complex suggests that the compound inhibits the activity of AfPGM by modifying a surface cysteine (C353). ISFP1 also inhibits A. fumigatus growth (MIC 23 mg/L) although it is not clear whether this effect is on-target. Furthermore, ISFP1 exhibits marginal selectivity (seven folds) over the human orthologue (HsPGM). Synthetic exploration of ISFP1 scaffold led to a derivative exhibiting similar potency (IC50 1 μM) and improved selectivity (40 folds). These results suggest that it is possible to develop selective covalent inhibitors against AfPGM.
In summary, I have genetically validated ugp and pgm as antifungal targets against A. fumigatus. I have also identified inhibitors or potential fragment binders of AfAGM1, AfUGP and AfPGM. These results provide guidance for developing chemical probes of the three enzymes (AGM1, UGP and PGM), which facilitate chemical validation of the three enzymes as antifungal targets against A. fumigatus.
| Date of Award | 2021 |
|---|---|
| Original language | English |
| Sponsors | China Scholarship Council |
| Supervisor | Daan van Aalten (Supervisor) |