A. niger strains used in this study are summarized in Table 1. Spore suspensions were prepared from 3 days (for N402 and MA93.1) colonies grown on complete media (CM) agar [55 mM glucose, 11 mM KH₂PO₄, 7 mM KCl, 178 nM H₃BO₃, 2 mM MgSO₄, 76 nM ZnSO4, 70 mM NaNO₃, 6.2 nM Na₂MoO₄, 18 nM FeSO₄, 7.1 nM CoCl₂, 6.4 nM CuSO₄, 25 nM MnCl₂, 174 nM EDTA, 15 g/L of agar supplemented with 0.5 % (w/v) yeast extract, and 0.1 % (w/v) casamino acids]. The spores were gently harvested with a sterile cotton swab and suspended in 0.9 % sodium chloride (NaCl). The resulting spore suspensions were filtered through sterile miracloth to remove hyphal fragments. Fresh spore suspensions, <2 weeks old, were used for all inoculations. Liquid shake flask cultivations were performed in CM medium for 6 days at 30°C at 200 rpm, after inoculation with 1 mL of 10⁶ sps/mL in 200 mL flasks.

All molecular techniques were performed according to standard procedures (Green and Sambrook, 2012). Protoplast-mediated transformation of A. niger, genomic DNA extraction, diagnostic PCR, and Southern blot were performed as described earlier (Fiedler et al., 2017). Quantitative PCR was performed according to Polli et al. (2016) using a 10 μl reaction with 2x Blue SYBR Green Mastermix (Biozyme) on an AriaMX Real-time PCR System (Agilent Technologies). A. niger strains MF41.3 and N402 were used as negative and positive controls, respectively.

All plasmids used in this study were constructed using DNA parts available in the Fungal MoClo Toolkit (Addgene Kit # 1000000191; Mózsik et al., 2021). A measure of 0.5 μl of the required type II restriction enzyme (BpiI, BsaI, or Esp3I, Thermo Scientific) was used per 15 μl reaction and 1 μl of T4 DNA ligase (Thermo Scientific). Golden Gate reactions were performed for 25 cycles (5 min at 37°C, 5 min at 16°C) and a final inactivation step of 10 min at 70°C to denature enzymes. A measure of 7.5 μl of the reaction mixture was transformed into 50 μl of TOP10 chemically competent Escherichia coli cells (NEB) using a heat shock at 42°C for 30 s. Two protospacers for Cas9 within the coding region of An11g02180 (hmgA) and An11g02200 (hppD) were identified using CCTop (Stemmer et al., 2015), and the corresponding sgRNA expression plasmids were constructed using pFTK086 to express the sgRNA under control of the tRNA promoter An08e08800. Sanger sequencing (LGC Genomics) was performed to validate the correct insertion of the protospacer into the sgRNA cassette. For the preparation of donor DNA, KAPA HiFi Polymerase was used (Roche). All primers used in this study are listed in Supplementary material 1. Protoplasts of strain MF41.3 were transformed using a CRISPR approach that harnessed Cas9-aided short-flank (75 bp) donor DNA for integration (Pohl et al., 2016).

For pyomelanin production, 200 mL of CM was supplemented with 1 mM uridine and 2.2 mM L-tyrosine (hereafter abbreviated CM + Uri + Tyr.) in a 500-mL Erlenmeyer flask with 1 × 10⁸ spores/mL of strain OS4.3 and incubated for 6 days at 150 rpm and 30°C. After cultivation, the cultural broth was centrifuged at 4,500 rpm for 20 min. This process was repeated three times. After centrifugation, the collected supernatant was filtered through a sterile miracloth filter to remove any leftover hyphal fragments and was kept in Falcon tubes for further investigation. For the purification of the Pyo_Fun, a modified protocol by Schmaler-Ripcke et al. (2009) was followed. In brief, the supernatant containing pyomelanin was adjusted to pH 2 using 1 M HCl. The samples were precipitated at room temperature in the dark for 15 h. After precipitation, the tubes were centrifuged at 4,500 rpm for 20 min, the supernatant was discarded, and the Pyo_Fun pellet was saved for lyophilization in an Eppendorf Concentrator plus complete system with an integrated diaphragm vacuum pump with rotor F-45-48-11, 230V/50–60Hz. Before lyophilization, the pellet was rinsed with ddH₂O three times, and during washing, several melanin pellets can be concentrated into the same tube(s) which lowers the general time needed for lyophilization. Aliquots of Pyo_Fun were placed lyophilized and concentrated using the V-Aq program for 9 h at 4°C. The full process was performed 10 times (n = 10), and the yield was determined by weighing the lyophilized Pyo_Fun using a precision scale (Sartorius) which was referred to as the cultivation volume (200 mL). Additionally, an in vitro synthesized pyomelanin (Pyo_Syn) was used as the control for all the following experiments. The synthetic pyomelanin was produced in accordance with the protocol described by Schmaler-Ripcke et al. (2009). In brief, 10 mM HGA (Sigma Aldrich) autoxidizes at pH 10 (adjusted with 1M NaOH) under constant stirring for 3 days in the dark at room temperature (RT). The lyophilization and quantification were carried out in the same way as described for Pyo_Fun.

FTIR was used to characterize both lyophilized fungal and synthetic pyomelanin using a Spectrum One^TM from Perkin-Elmer Inv. #429 with the following specifications: Analysis mode: attenuated total reflection (ATR) on diamond crystal and Measuring range: wavenumber 4,000 to 650 cm⁻¹. For FTIR measurements (n = 3 for every measurement), approximately 1 mg of fungal pyomelanin (Pyo_Fun), synthetic pyomelanin (Pyo_Syn), and corresponding precursors for pyomelanin production, L-tyrosine (Sigma Aldrich), and homogentisic acid (Sigma Aldrich) were transferred to the sample plate and pressed onto the ATR diamond window using a stamp.

Raman spectra for Pyo_Fun, Pyo_Syn, L-tyrosine, and homogentisic acid were determined. For the spectral acquisition, a small amount of respective powder was transferred onto the CaF2 slide, and spectra were acquired directly. All the samples were analyzed using the commercial Renishaw Raman spectrometer (Renishaw inVia Raman Spectrometer, Renishaw plc., Wotton-under-Edge, UK) with a 785-nm single-mode diode laser as an excitation source. The laser beam was focused onto a sample using a microscope objective (Leica, Wetzlar, Germany, N PLAN EPI, magnification 50 × , working distance 0.5 mm, and numerical aperture 0.75). The dimensions of a laser spot shape typical for the Renishaw inVia instrument were ~2 μm × 10 μm, with a full axial depth of the excitation region at 8 μm (Mlynáriková et al., 2015; Rebrošová et al., 2017). Before each spectral acquisition, the laser beam was refocused onto the individual fungal spore or the powder to stay within the focal depth of the laser beam excitation and the imaging optics. Considering the possible variability of individual fungal spores, 10 measurements were taken per strain, each for a different spore. Similarly, 10 measurements were taken per powder sample. Five-second acquisition time and 100% laser power (~140 mW on a sample) were used for all powders. Due to a significant autofluorescence of fungal spores in the range 614–1,724 cm⁻¹, the acquisition settings had to be adjusted. To acquire each individual spectrum, accumulations of 10 × 1 s and 10% laser power were used (~14 mW).

The acquired Raman spectra were analyzed using an in-house written program based on MATLAB software (MathWorks, Natick, MA, USA). First, high-frequency noise was removed using Savitzky–Golay filtering (4th order, width 15 points). Then, the spectra were treated with rolling circle filtering (50 passes, 500 points circle radius) to suppress a fluorescence background. Finally, the area of spectra was normalized to 1, and a threshold of 0.001 was applied. A comparison of spectra from individual fungal spores belonging to different A. niger strains was made using centered principal component analysis (PCA). The groups were marked by ellipsoids with a Mahalanobis distance of 2.15, corresponding to the 90% confidence interval (De Maesschalck et al., 2000; Rebrošová et al., 2017).

To characterize the solubility properties of Pyo_Fun and Pyo_Syn, 0.1–0.5 mg/mL of the respective lyophilized pyomelanin was dissolved in 1 mL of EtOH (100%), 0.1 M KOH, DMSO (100%), and ddH₂O and vortexed for 10 s each. Solubility was assessed visually. Additionally, to test for decolorization with oxidizing agents, 100 μL of 30% H₂O₂ was added to 900 μL of dissolved Pyo_Fun, and oxidation was stopped after 10 min or 20 min using catalase (0.1 mg/mL, from bovine liver, Sigma Aldrich). The bleaching process was then evaluated via UV-Vis spectroscopy using a multidetection microplate reader (Infinite M200 PRO, Tecan).

A 2,2,1-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay described earlier (Chen et al., 2013) was used with some modifications. In brief, DPPH is a free radical with a dark violet color that provides hydrogen acceptor capability to antioxidants. Its absorption maximum occurs at 515 nm. When antioxidants react to DPPH by providing an electron or hydrogen atom, the free radical DPPH is reduced to 2,2-diphenyl-1-hydrazine (DPPH-H), and its dark violet color changes toward a colorless or pale-yellow/orange color, which can be detected using a multidetection microplate reader (Infinite M200 PRO, Tecan).

To evaluate the antioxidant activities of Pyo_Fun, a DPPH stock solution was prepared by mixing DPPH free radical (C₁₈H₁₂N₅O₆, Sigma Aldrich) with 100% EtOH to create a stock solution (0.002%) and left to incubate for 2h at 4°C. This stock solution was then vigorously mixed with each sample and incubated in the dark at room temperature. Two comparative samples were taken along, one with known high reactive oxygen species (ROS) scavenging activity (ascorbic acid 5 mg/mL of stock solution, diluted 1:50) and one with low ROS scavenging activity (DMSO 100%, diluted 1:50). The two samples of Pyo_Fun (pure culture supernatant, diluted 1:50 and purified pyomelanin powder dissolved in H₂O, 5 mg/mL, diluted 1:50) and one negative control (the solvent DMSO and DPPH) were measured. The discoloration was measured at 517 nm for 30, 40, and 60 min, respectively, after the reaction was initiated. All tests were performed in technical triplicates (n = 3). The percentage of antioxidative effect was calculated using the following equation:

To characterize the physical and biological shielding properties of Pyo_Fun and Pyo_Syn, 0.5 mg/mL of lyophilized pyomelanin was dissolved in 1 mL of 0.1 M KOH under sterile conditions and pipetted in quartz glass cuvettes (QS 0.5, Hellma). A measure of 1 mL of 0.1 M KOH was additionally pipetted in a cuvette to be used as a negative control (solvent blank). These quartz glass cuvettes are used as “melanin filters” and contain a homogenous mixture of solved pigment. As UV-C radiation can transmit through the quartz glass cuvettes, one can evaluate only the shielding capacities of our dissolved pigments. To evaluate the physical shielding of UV-C radiation, the cuvettes were placed individually on top of the dosimetry sensor of a UVX Radiometer (Analytic Jena) to shield the monochromatic wavelength of 254 nm originating from UV-Lamp (VL-215-LC, Vilber Lourmat, SN.: 14 100595). The corresponding dose rate reaching the dosimetry sensor was recorded. This was performed six times for each cuvette filled with pyomelanin or the solvent control for each of the two radiation forms (see Supplementary Figure SI 4).

Fungal spores of the ΔfwnA were tested against the UV-C dose that is required to eliminate 90% of the wild-type spores (LD₉₀; 1,038 J/m², Cortesão et al., 2020b). Spores of A. niger were exposed to UV-C (254 nm) in a 96-well plate with an initial concentration of 10⁶ spores/ml in 200 μL of saline solution (0.9% (w/v) NaCl). The concentration of 10⁶ spores/ml ensures a spore monolayer and prevents survival due to shielding by the spores themselves. The radiation dose was adjusted through exposure time since the intensity (distance to object) of the UV lamp was kept constant. UV-C exposure time was calculated using the following formula (Cortesão et al., 2020b):

where t = time (in seconds), R = desired radiation dose (in J/m²), and d = dosimeter value for UV fluence (in μW/cm²).

After irradiation, 30 μL of the sample were taken in triplicates from the spore suspension within each well, and viability was calculated by the ability to form colonies (colony forming units, CFU). Radiation exposure included at least three biological replicates per strain and was performed two independent times (n = 6). Samples were serially diluted up to 10⁻⁸ using a 96-well plate, with each well having a total volume of 300 μl. To count the CFUs, 20 μl of each dilution was plated out in triplicate on one-eighth of a Petri dish with MM agar which contained Triton X-100 (0.05%) to facilitate counting. The plates were incubated for 2 days at 30°C before the colonies were counted.

The survival of fungal spores after irradiation was additionally tested using an oCelloScope™ (BioSense Solutions ApS, Farum, Denmark) at 22°C for 48 h. The oCelloScope™ is a digital time-lapse microscopy technology that generates high-resolution micrographs of microorganisms while growing. This is achieved by scanning through a growth chamber, creating a series of images on a z-axis. The oCelloScope™ quantifies fungal mass per well on the best focus level of the z-axis images per timepoint by quantifying the number of pixels belonging to the fungus and to the background (Winters et al., 2022). Germination of fungal spores and formation of hyphae were micrographed in 30 min intervals after samples were diluted to 2.3 × 10⁵ spores/mL and allowed to rest for 10 min before image acquisition. The oCelloScope™ then analyzed corresponding changing fungal mass over time per well (n = 6) using the build-in SESA fungi algorithm (Aunsbjerg et al., 2015).

As several biosynthetic routes are followed in fungi to produce L-DOPA melanin (= eumelanin), DHN-melanin (allomelanin), or pyomelanin, we used orthology screening to identify A. niger genes that encode enzymes predicted to be involved in the degradation of L-phenylalanine/L-tyrosine and thus pyomelanin formation. We focused on the well-described pyomelanin pathway in A. fumigatus (Schmaler-Ripcke et al., 2009) and could indeed identify gene candidates based on MultiGeneBlast analyses (Blin et al., 2013). The pathway involves six genes, which physically colocalize within a gene cluster. All pairwise gene identities between A. fumigatus and A. niger are given in Figure 2A, Supplementary Tables SI 1, SI 2, and Supplementary Figures SI 1, SI 2. We decided to delete two central genes of this pathway in A. niger–hppD, which catalyzes the precursor of pyomelanin, homogentisic acid (HGA), and hmgA, which degrades HGA using CRISPR-Cas9 technology (see Materials and methods section). We received several A. niger transformants for each deletion approach and verified successful deletion of hppD and hmgA, respectively, via diagnostic and quantitative PCR (Supplementary material SI 3). A comparative phenotypic analysis clearly demonstrated that the deletion of hppD resulted in strains that nearly lost the ability to produce any dark pigment (e.g., OS3.1) when cultured in the presence of L-tyrosine, whereas the deletion of the hmgA gene resulted in strains that secreted much more of a dark brown pigment, supposedly pyomelanin, into the medium agar (e.g., strain OS4.3).

To confirm that the dark brown pigment secreted by strain OS4.3 is indeed pyomelanin (hereafter abbreviated as Pyo_Fun), we first established a pyomelanin production and purification protocol for submerged A. niger cultures based on the study published earlier for A. fumigatus (Perez-Cuesta et al., 2020). As shown in Figure 2 and described in detail in the Materials and Methods section, the production of substantial amounts of Pyo_Fun can be induced in strain OS4.3 by adding L-tyrosine to the culture medium. When strain OS4.3 was cultivated for 6 days in liquid CM medium (Figures 3A–C), we could isolate approximately 1.5 g Pyo_Fun out of 1 L of culture supernatant. Remarkably, Pyo_Fun seems to be mostly present extracellularly, from where we were able to easily purify it via acidification (pH 2) (Figures 3D, E) and lyophilization into a powder that appeared dark brown to black in color (Figure 3F).

In order to verify that the dark brown/black powder is indeed Pyo_Fun, it was subjected to different spectroscopic analyses including UV-Vis, FTIR, and Raman spectroscopy. In this study, a synthetically produced pyomelanin (Pyo_Syn) was taken along, which was proven earlier to show the same spectral characteristics as fungal pyomelanin (Lorquin et al., 2021). Pyo_Syn is thus an excellent positive control and can be easily obtained through the autooxidation of HGA when subjected to alkaline conditions (see Materials and Methods section).

As shown in Figures 4A, B, both Pyo_Fun and Pyo_Syn exhibit similar UV-Vis spectra and have their absorbance maximum between 280 and 335 nm, which is characteristic for most types of melanins (Pralea et al., 2019). We used different concentrations of both compounds and detected an exponential increase of absorption in wavelength under 500 nm and a gradual decrease of absorbance in higher wavelengths for both, Pyo_Fun and Pyo_Syn. The maximum absorbance peak of Pyo_Fun (2.4 [a.u]) is slightly lower than that of Pyo_Syn (2.6 [a.u.]). In both trajectories, the two highest concentrations (1 mg/mL and 0.5 mg/mL) depicted a double peak between 280 and 335 nm. Such behavior can be expressed mathematically when plotting the logarithms of each absorbance curve against the wavelength. The absorption regression lines of the original curves are linear with negative slopes (Figure 4C), which agrees with melanin data from the literature (Lorquin et al., 2021).

Data from FTIR and Raman spectroscopy also confirmed that both Pyo_Fun and Pyo_Syn are very similar compounds and display spectra which are characteristic of melanins (Figure 5). For these analyses, we took the precursors L-tyrosine and HGA along. FTIR frequencies of Pyo_Fun and Pyo_Syn show melanin characteristic bands and transmittance peaks covering regions between 3,600–2,900 cm⁻¹, 1,600–1,500 cm⁻¹, and 1,400–1,300 cm⁻¹ (Figure 5A; note that the Pyo_Syn FTIR spectra are better resolved). The corresponding assignment of the prominent bands most likely indicates symmetric carboxylate stretching vibrations (COO–) (1,580 cm⁻¹) and polymeric O-H groups (3,400 cm⁻¹). Within the fingerprint region (400–1,400 cm⁻¹), there is one distinct peak at 1,390 cm⁻¹ for Pyo_Syn which could indicate a C-H bending. All peaks within the transmittance curve seen in Pyo_Syn are more intense than for Pyo_Fun but they show a similar trajectory pattern. The prominent peaks in the Pyo_Syn spectrum (at 1,375, 1,575, and 3,287 cm⁻¹) are generally approximately 30% lower in transmittance (%) than Pyo_Fun. In comparison to the pyomelanin samples, the FTIR transmittance spectra of the two precursors L-tyrosine and HGA show clear differences toward each other and to the pigments. They exhibit very sharp bands between regions 1,000 cm⁻¹ indicating a different molecular structure. HGA additionally shows defined peaks at approximately 3,333.4 cm⁻¹ and 3,479.2 cm⁻¹ and L-tyrosine at 3,195.4 cm⁻¹. The Raman spectra shown in Figure 5B provide complementary information. The relative intensity of the all Raman bands in both the precursors differs from the bands visible in both pyomelanin. HGA and L-tyrosine exhibit high-frequency vibrations between 800 and 850 cm⁻¹. HGA shows another prominent band at 1,290–1,300 cm⁻¹ in contrast to L-tyrosine which exhibits several medium strong vibrations between 1,190 and 1,350 cm⁻¹. Significant similarities can be observed in the vibration frequencies of Pyo_Fun and Pyo_Syn. Both pigments show medium strong bands between 1,290 and 1,390 cm⁻¹ (stretching of the C-C bond) and a higher frequency vibration between 1,550 and 1,621 cm⁻¹, which is indicative of aromatic or heterocyclic ring systems.

The different spectral analyses demonstrated that Pyo_Fun and Pyo_syn are identical molecules and display spectral characteristics of melanins published in the literature. We next analyzed physicochemical properties of Pyo_Fun with a focus on solubility and reactivity since different kinds of melanin display distinctive solubility and reactivity properties (Pralea et al., 2019). We tested the solubility of Pyo_Fun in water (aqueous solvent), ethanol (100 % EtOH, polar solvent), dimethylsulfoxide (100 % DMSO, polar aprotic solvent), and potassium hydroxide (0.1 M KOH, aqueous alkaline solvent). Figure 6A shows that 10 mg/mL of purified and lyophilized Pyo_Fun is poorly soluble in water, well soluble in dimethylsulfoxide and potassium hydroxide, and insoluble in ethanol.

Melanins are known to be bleachable when treated with different oxidizing agents including hydrogen peroxide (Pralea et al., 2019). Bleaching is thus a common method to qualitatively determine any antioxidant activity of a compound of interest. We treated Pyo_Fun with 30 % H₂O₂ solution and stopped the reaction after 10 and 20 min, respectively, via the addition of 0.1 mg/mL of catalase. UV-Vis spectral analyses showed the loss of absorbance capacity of Pyo_Fun between 300 and 500 nm (Figure 6B), which parallels the loss of dark colorization. Indeed, the bleaching result was already visually assessable by eyes (Figure 6C), demonstrating that Pyo_Fun exerts antioxidant activities.

To quantitatively assess the antioxidant capacity of Pyo_Fun, we performed a ROS scavenging assay with culture supernatant from strain OS4.3, lyophilized Pyo_Fun pigment (5 mg/mL), ascorbic acid (positive control), and DMSO (negative control; see Materials and Methods section for details), and measured ROS scavenging after 30, 40, and 60 min, respectively. Figure 7 illustrates that the ROS scavenging capacity of lyophilized Pyo_Fun was found to be approximately 80% of the capacity of ascorbic acid. Moreover, the unconcentrated culture supernatant of strain OS4.3 obtained through filtration also showed approximately 60% of the radical scavenging capacity of ascorbic acid.

The higher order structures of the different kinds of melanin form the basis for their broad optical absorption and potent antioxidant capacities but also for their ability to protect against ionizing radiation such as UV and X-ray (Cordero and Casadevall, 2017). Radioprotection by melanins is thought to be mediated by three main mechanisms: (i) Melanins are able to absorb radiation energy while dissipating the energy through heat. (ii) Melanins trap and neutralize free radicals and ROS which become released by cellular molecules (DNA, proteins, and lipids) as a consequence of radiation (biological shielding; Esbelin et al., 2013; Cordero and Casadevall, 2017). (iii) Melanins scatter radiation, causing it to be redirected in different directions and reducing its intensity upon contact. The physical shielding capacities of 0.5 mg/mL of solutions of Pyo_Fun and Pyo_Syn against UV-C radiation were determined to be 81.13 ± 0.27 J/m² (n = 6), whereby the solvent itself (0.1 M KOH) was responsible for 19.40 ± 0.31 J/m² (n = 6).

To determine the biological shielding capacity of Pyo_Fun and Pyo_Syn against UV-C radiation, we subjected spores of an A. niger wild-type strain (N402) and spores of a strain deficient for fwnA and thus DHN-melanin formation (MA93.1), (Jørgensen et al., 2011) to a UV-C dose rate of 1,038 J/m². We choose this dose rate because it corresponds to the LD₉₀ for strain N402 (Cortesão et al., 2020a). Importantly, before we treated the spores with UV-C, each spore monolayer was protected with a corresponding quartz glass cuvette filled with 0.5 mg/ml Pyo_Fun, Pyo_Syn dissolved in 0.1 M KOH. As a reference, we used a quartz glass cuvette filled with only 0.1 M of KOH was used. Spore survival was followed by a live microscopic recording of spore germination and mycelial formation 48 h post-radiation (Figure 8), and survival rates were also calculated by determining colony forming units 48 h post-radiation (Table 2).

The growth curves shown in Figure 8 demonstrated that both strains were efficiently protected by Pyo_Fun and Pyo_Syn. Although, the trajectories of the growth measure curves in Figure 8 show that the DHN-melanin-deficient strain had a longer recovery phase after irradiation before the onset of germination (30 h of incubation) than the wild-type strain (10 h of incubation).

Comparing the survival rates of the unprotected wild-type strain N402, the DHN-melanin-deficient strain MA93.1 survived significantly less radiation (p = 0.012) (Table 2), as expected, and reported earlier (Cortesão et al., 2020a). However, we see that fungal pyomelanin contributed to the radioresistance of A. niger, as also reported earlier (Cortesão et al., 2020a). Especially in the DHN-melanin-deficient strain MA93.1, the radioprotection effect of Pyo_Fun was significantly higher compared with the non-shielded control (p = 0.007) and the KOH solvent control (p = 0.021). Interestingly, the radioresistance effect of Pyo_Fun was higher for the mutant strain than for the wild type. In this study, the shielding of Pyo_Fun was not statistically significant (p = 0.280) compared with the non-shielded control. In contrast, Pyo_Syn did not show a significant protective effect in any of the exposed strains.