Functional analysis of the Aspergillus fumigatus kinome identifies a druggable DYRK kinase that regulates septal plugging

Abstract

More than 10 million people suffer from lung diseases caused by the pathogenic fungus Aspergillus fumigatus. Azole antifungals represent first-line therapeutics for most of these infections but resistance is rising, therefore the identification of antifungal targets whose inhibition synergises with the azoles could improve therapeutic outcomes. Here, we generate a library of 111 genetically barcoded null mutants of Aspergillus fumigatus in genes encoding protein kinases, and show that loss of function of kinase YakA results in hypersensitivity to the azoles and reduced pathogenicity. YakA is an orthologue of Candida albicans Yak1, a TOR signalling pathway kinase involved in modulation of stress responsive transcriptional regulators. We show that YakA has been repurposed in A. fumigatus to regulate blocking of the septal pore upon exposure to stress. Loss of YakA function reduces the ability of A. fumigatus to penetrate solid media and to grow in mouse lung tissue. We also show that 1-ethoxycarbonyl-beta-carboline (1-ECBC), a compound previously shown to inhibit C. albicans Yak1, prevents stress-mediated septal spore blocking and synergises with the azoles to inhibit A. fumigatus growth.

Fig. 1. Generation of the kinase knockout collection in A. fumigatus.

A A phylogenetic tree (Maximum-likelihood) of the protein kinases identified in Af293. The protein kinase family as determined using Kinomer v1.0 (first colour ring), the Kinomer score (bar chart with higher values representing higher confidence matches), the occurrence of each kinase within the pangenome of A. fumigatus (circle size with larger circles representing kinases that occur more frequently), if a homokaryotic null was constructed (green squares) and if the protein kinase is fungal specific (purple square) is shown as annotations around the phylogenetic tree. B Four putative essential kinases were validated using a Tet-OFF construct to replace the native promoter. Shutting down expression via addition of doxycycline to the medium resulted in no growth after 72 h at 37 °C, adding further evidence to the essentiality of these kinase for viability.

Fig. 2. Bar-seq of A. fumigatus identifies YakA as a potential antifungal target.

A Heatmap of fitness for each protein kinase null mutant relative to control conditions (n = 5). Relative fitness values was calculated by DESeq2, comparing barcode counts obtained from 20 h fRPMI1640 (pH 7, 37 °C, shaking 130 rpm) cultures with those supplemented with 1.5 mM H₂O₂ (to induce oxidative stress), voriconazole 0.15 mg/l, olorofim 0.002 mg/l, pH adjusted to pH4 or pH8 or with modified culture temperature (48 and 30 °C). and different iron concentrations (Fe−, BPS, 30 μM, 300 μM, 5 mM). The mouse data fitness values was calculated by DESeq2, comparing barcode counts obtained from the input pool with the infected samples. The heatmap was generated with Pheatmap, full linkage clustering on rows and columns (cutree_cols = 2). Source data are provided as a Source Data file. B Dry weight was measured after 16 h culture in HMM, with and without 0.5 mg/l itraconazole, at 37 °C shaking 130 rpm. Data are presented as mean values with SEM. Statistical difference was assessed by using one-way ANOVA (n = 2 or 3 biological replicates). Source data are provided as a Source Data file. C 1000 spores of MFIG001, ∆yakA and the reconstituted isolate (yakA+) were spot inoculated on HMM agar without iron and with 1 µM iron added and incubated for 72 h at 37 °C. Representative pictures are shown. D Survival curves in a Galleria mellonella model of aspergillosis. Statistical difference was assessed by Kaplan–Meier survival analysis and p values between groups are shown on the graph. E Survival in a leukopenic mouse model of aspergillosis (n = 11). Statistical difference was assessed by Kaplan–Meier survival analysis p values between groups are shown on the graph.

Fig. 3. Detailed phenotypic analysis of the ∆yakA isolate.

A 10³ spores of the isotype control MFIG001, ∆yakA, yakA+ were spotted on HMM + 30 µM Fe with increasing concentration of agarose (0.5–6%) and imaged after 72 h incubation at 37 °C. Representative images are shown. B Representative histological sections of lesions found in mouse lungs infected for 3 days with MFIG001, ∆yakA and yakA+. H&E and GMS stains were performed on neighbouring sections. C The hyphal length within histological sections (n = 15 biological replicates) was measured. Data are presented as mean values with SEM. Statistical differences were assessed by one-way ANOVA. Source data are provided as a Source Data file. D Recovery from cytorrhysis after addition of glycerol to liquid cultures. Liquid cultures were grown for 16 h at 37 °C in an Ibidi imaging chamber. 1.8 M glycerol was added and images were taken every 150 s. The % cytoplasm compared to the total cell size was measured for individual septal compartments (n = 10 biological replicates) using Fiji. Data are presented as mean values with SEM. Source data are provided as a Source Data file.

Fig. 4. Phosphoproteomics in response to iron limitation.

A A heatmap showing associations between changes in phosphopeptide abundance between MFIG001 and ∆yakA or upon low iron stress. K-means clustering was performed to identify clusters of phosphopeptides differentially phosphorylated (1 + 2 iron responsive and YakA dependent; 3 + 4 iron responsive YakA independent). B GO-term analysis in Revigo of yakA-dependent phosphopeptides (clusters 1 + 2) in response to iron limitation. Size of circles is the size of each enrichment category. Colour represents the p value. C An analysis of phosphopeptides that are present as components of the septal pore. Only phosphopeptides that are differentially regulated in at least one condition are shown (>2-fold up or down).

Fig. 5. YakA localisation to the septal pore.

A YakA-GFP localisation following overnight growth (16 h) in AMM, and replacement with either AMM (+Fe), iron limiting AMM (−Fe), or AMM supplemented with rapamycin for 1 h. Localisation of the GFP signal can be observed at the centre of septa. Scale bar = 10 µm. B The total fluorescent intensity was measured normalised to the background intensity following overnight growth (16 h) in AMM, and replacement with either AMM (+Fe), iron limiting AMM (−Fe), or AMM supplemented with rapamycin for 1 h. Quantification was performed using Fiji (n = 3 biological replicates). Source data are provided as a Source Data file. C Quantification of signal through a cross-section of septum shows higher fluorescent intensity in the central region of the septum, indicating septal pore localisation of YakA-GFP (n = 7). Source data are provided as a Source Data file.

Fig. 6. Microscopic characterisation of septal pore components; HexA and Lah.

A Microscopy of single hyphae of HexA-GFP and LacC-GFP upon iron deleted conditions show yakA-dependent localisation to the septum or the septal pore. Representative images of HexA-GFP ∆yakA and LahC-GFP ∆yakA showing dysregulation of HexA and Lah upon deletion of yakA. Scale bar = 10 µm. B Quantification of fluorescent intensity measured across the septum for HexA-GFP and LahC-GFP in the parental isolate or upon deletion of yakA (n = 3). Fluorescent intensity was normalised to the total hyphal width to account for variability. Statistical difference was assessed by two-way repeated measures ANOVA. Source data are provided as a Source Data file.

Fig. 7. YakA as a druggable target in A. fumigatus.

A An Alphafold2 model of the YakA protein docked to a library of small molecules using blind docking. Druggable pockets P1–P3 are shown as surface renders. The biggest pocket, P1, binds the largest number of fragments (778/1000), and is at the active site of the kinase. The druggability score from PockDrug for P1 is 0.81. Two smaller pockets, P2 and P3, are found on the C-lobe of the YakA. Druggability scores for P2 and P3 are 0.73 and 0.64 respectively. Pocket P1 is shown in further detail in complex with the 1-ECBC molecule. Charged residues in the pocket are indicated. The ∆G of 1-ECBC at the top scoring binding position is −7.7 kcal/mol. B Microscopic evaluation of YakA-GFP upon iron limitation and addition of 1-ECBC. Hyphae were followed for 1 h upon 1-ECBC addition. ECBC prevents localisation of YakA to the septal pore, but does not induce removal once it is in place. Scale bar = 10 µm. C Checkerboard assay (n = 3) for 1-ECBC and voriconazole to assess synergism of these drugs to inhibit growth of A. fumigatus MFIG001. The FICI (top-left) was calculated and shown synergism (<0.5).