Fungal Peptidomelanin: A Novel Biopolymer for the Chelation of Heavy Metals

Abstract

Melanin is an amorphous, highly heterogeneous polymer found across all kingdoms of life. Although the properties of melanin can greatly vary, most forms are insoluble and strongly absorb light, appearing dark brown to black. Here, we describe a water-soluble form of melanin (peptidomelanin) secreted by the spores of Aspergillus niger (strain: melanoliber) during germination. Peptidomelanin is composed of an insoluble L-DOPA core polymer that is solubilized via short, copolymerized heterogeneous peptide chains forming a "corona" with a mean amino acid length of 2.6 ± 2.3. Based on in vitro experiments, we propose a biochemical copolymerization mechanism involving the hydroxylation of tyrosynylated peptides. Peptidomelanin is capable of chelating heavy metals such as lead, mercury, and uranium (as uranyl) in large quantities. Preliminary data indicates that peptidomelanin may have applications for the remediation of heavy metals in situ, including in agricultural settings.

Figure 1

Morphology of Aspergillus niger strains used in this study. (A–C) The A. niger type strain (MTCC 281). (A) Dense hyphal mass possessing abundant conidiophores. (B) Conodiophore possessing relatively fewer conidiospores (C) Large conidiospores with echinulations evenly distributed across their surface. Polar regions are not discernible. (D–F) A. niger melanoliber (MTCC 13366). (D) Sparse hyphal mass possessing fewer large conidiophores. (E) Conodiophore possessing a large number of small conidospores. (F) Small conidiospores possessing deep, longitudinal striations originating from discernible polar regions.

Figure 2

Peptidomelanin release assay for Aspergillus spps. spores. (A) A. niger melanoliber (MTCC 13366) and A. niger (MTCC 281) spores were inoculated into Sabouraud broth. A. niger melanoliber rapidly released soluble peptidomelanin into the medium. In contrast, A. niger released no soluble peptidomelanin into the medium. (B) Peptidomelanin retention assay for A. niger melanoliber spores in Sabouraud broth. At predetermined time points, samples were isolated and fixed using 8% glutaraldehyde to halt peptidomelanin release. Samples were treated with 0.1 M NaOH to release retained peptidomelanin. Retained peptidomelanin decreased with time, indicating that new peptidomelanin is not synthesized during germination, and only existing peptidomelanin is released. 0.1 M NaOH did not degrade peptidomelain for the duration of incubation (Figure S4). (C) Melanin release assay for UV-irradiated A. niger melanoliber spores. Irradiated spores did not release peptidomelanin into the Sabouraud broth, demonstrating that only live spores release melanin. (Inset) Petri plate confirming UV irradiated spores do not germinate. All experiments were performed in quadruplicate. Lines represent means. The shaded area represents standard deviation.

Figure 3

Spectral analysis of peptidomelanin. (A) Peptidomelanin appears dark brown in color. Peptidomelanin can be concentrated using a 30 kDa centrifugal concentrator. (B) UV–vis spectral overlay of four different melanin spectra. Peptidomelanin displays an absorbance spectrum characteristic of melanin. It displays the greatest absorbance in the UV region but strongly absorbs light in the visible spectrum as well, up to a wavelength of 650 nm, with absorption decreasing for longer wavelengths. Synthetic L-DOPA melanin displays a similar absorbance spectrum. Acid-hydrolyzed melanin (lacking a peptide component) also displays a similar spectrum. Sepia melanin displayed strong absorbance in the UV region while also displaying greater absorbance across the visual spectrum compared to the other samples. (C) FTIR spectra of peptidomelanin, sepia melanin, and synthetic L-DOPA melanin. All spectra have been vertically offset to ease comparison. All spectra appear somewhat similar. All display a broad dip between 3500 and 3100 cm^–1 assignable to the O–H or N–H stretch vibration. Synthetic melanin displays and a sharp dip between 1800 to 1600 cm^–1 assignable to C–C or C=O stretch vibration, while peptidomelanin and sepia melanin display shallower dips. Synthetic melanin displays strong signals in the 1600–1000 region, whose intensity is not matched in peptidomelanin and sepia melanin. This is possibly attributable to the simpler chemical composition of synthetic melanin.

Figure 4

PAGE gels indicate the biochemical composition of peptidomelanin. (A) Native PAGE of peptidomelanin. A brown smear from ∼20 to ≥250 kDa is observed. Coomassie blue staining darkens the smear, indicating the presence of an amino acid component. (B) SDS PAGE and (C) 8 M Urea PAGE of peptidomelanin. Coomassie blue staining reveals that the amino acid component does not separate from the melanin component, indicating a covalent linkage between the two components (ladder not resolved). The observation of smears rather than discrete bands on gels A–C confirms that peptidomelanin is an amorphous polymer. (D) Native PAGE of resolubilized acid-hydrolyzed melanin displays a smear that runs slower on a gel. The smear was only marginally stained by Coomassie blue. Both these observations indicate the loss of the amino acid component upon acid-hydrolysis. (E) Native PAGE of synthetic L-DOPA melanin. The smear does not uptake Coomassie blue as it lacks an amino acid component.

Figure 5

Dansylation of peptidomelanin and its derivatives. Dansyl chloride is a sulfur-containing fluorophore that can be covalently linked to the N-terminal of amino acids and peptides. (A) SDS PAGE of peptidomelanin and dansylated peptidomelanin. Both substances are stained by Coomassie blue. However, only dansylated peptidomelanin displays fluorescence under UV transillumination. (B) SDS PAGE of resolubilized acid-hydrolyzed melanin and dansylated acid-hydrolyzed melanin. Both substances are poorly stained by Coomassie blue as they lack their amino acid component. Only dansylated acid-hydrolyzed melanin displays fluorescence under UV transillumination. Dansylated acid-hydrolyzed peptidomelanin could not be resolubilized and therefore does not migrate on an SDS PAGE gel.

Figure 6

Dansylation experiments using peptidomelanin and its derivatives. (A) Schematic showing all the substances under study. Dansyl chloride is a sulfur-containing fluorophore that can be covalently linked to the N-terminal of amino acids and peptides. Peptidomelanin will possess an amino acid component, whereas acid-hydrolyzed melanin will not. (B) SEM-EDS was used to estimate the sulfur (weight %) content of all four substances depicted in panel A. Sulfur acts as a proxy for the dansyl group. The increase in sulfur (ΔS₁ = 2.12 ± 2.03%) between peptidomelanin and dansylated peptidomelanin represents sulfur originating only from dansyl linked to N-terminals. Likewise, the increase in sulfur (ΔS₂ = 2.75%) between acid-hydrolyzed melanin and dansylated acid-hydrolyzed melanin represents sulfur originating only from dansyl linked to N-terminals. ΔS₁ is not significantly different from ΔS₂ (p = 0.39, 2-sided p-value), indicating that peptides are linked to the melanin core polymer via a peptide bond (see text). Error bars represent the standard deviation from the mean. (C) The mean peptide chain length can be calculated using data from dansylation experiments (panel B) as well as LC-MS amino acid quantification experiments (Table 1). Placeholder data is used here to explain the rationale. The actual mean peptide chain length was calculated to be 2.6 ± 2.3 amino acids. All calculations are provided in Text S2.

Figure 7

In vitro formation of melanin from dansylated L-DOPA shows that it is possible for L-DOPA containing an occupied N-terminus to be incorporated into melanin. (A) L-DOPA and dansyl chloride were incubated under conditions that allowed for the dansylation of primary amines. These conditions also allowed for the in vitro autoxidation of L-DOPA into melanin. (B) An SDS PAGE gel confirms that the autoxidized L-DOPA melanin is linked to dansyl. (C) Reactions along the melanin biosynthetic pathway that can form eumelanin from L-DOPA via nonenzymatic or autodixation reactions. Tyrosinase can increase reaction rates but is not essential for the in vitro formation of eumelanin. Only L-DOPA and L-DOPAquinone possess dansylatable N-terminals (primary amines). Metabolites from leucoDOPAchrome onward only possess secondary amines and cannot be dansylated under our reaction conditions. Likewise, dansylated L-DOPAchrome, lacking a primary amine group, cannot undergo intramolecular addition (cyclization) into dansylated leucoDOPAchrome. Therefore, L-DOPA and L-DOPAquinone containing dansylated/occupied N-terminals can be incorporated into melanin.

Figure 8

Metal-binding stoichiometric assays used to determine which substance tested has the greatest capacity to chelate metal ions. Four substances were tested: silica, activated charcoal, synthetic L-DOPA melanin, and peptidomelanin. (A) A visual depiction of our protocol. An excess of metal salt is added to all substances tested. The unbound metal salt was then removed via dialysis or centrifuging and washing, depending on the solubility of the metal complex. SEM-EDS was used to estimate the percentage metal in all complexes. Placeholder data is used to describe how the stoichiometric ratio (wt%:100%), stoichiometric equivalent ratio, and molar stoichiometric ratio are calculated (AM: atomic mass). (B) The mean metal-binding stoichiometric ratio of all substances tested across 9 metals, ranging from sodium to uranium (as uranyl). Peptidomelanin can chelate significantly larger ratios of metal compared to silica and activated charcoal. Synthetic melanin does not display a significant difference in metal chelation compared to peptidomelanin. Here, the molar stoichiometric ratio is used to normalize our data across metals with different molecular weights. (C) The ratios of metal chelated by peptidomelanin, synthetic melanin, and activated charcoal is positively correlated with the atomic mass of the metal. Silica displays no strong correlation. Regression lines for all substances tested are shown. (D–L) Metal-binding stoichiometric assays used to determine which substance tested has the greatest capacity to chelate a given metal ion. The means obtained from these experiments were used as data for the summary statistics in panels B and C. The individual metal ions tested were: (D) sodium, from NaCl, (E) chromium, from CrCl₃.6H₂O, (F) nickel, from Ni(NO₃)₂, (G) copper, from CuSO₄.5H₂O, (H) zinc, from ZnSO₄, (I) cadmium, from Cd(NO₃)₂, (J) mercury, from HgCl₂, (K) lead, from Pb(NO₃)₂, and (L) Uranium, from UO₂(CH₃COO^–).2H₂O. For panels D to L, ≥10 SEM-EDS spectra were collected per sample (technical replicates). For panels B and D–L, statistical significance testing was performed using the Welch 2-sample t test. All substances were compared to peptidomelanin. P-values ≤0.05 are shown. Error bars represent the standard deviation from the mean. Raw data is provided in Data set S5.

Figure 9

Wheat germination dose–response experiments show that peptidomelanin can ameliorate soil mercury toxicity in agricultural contexts. (A) Wheat seed mass shows a moderate inverse correlation (log–linear) with mercury concentration (r = −0.40, p = 2 × 10^–4) across a range of 0 ppm to 10,000 ppm mercury. (B) When germinated in substrate containing 100 ppm mercury, wheat seed mass shows a moderate positive correlation (log–linear) with increasing concentrations of peptidomelanin (r = 0.51, p = 7 × 10^–4). Wheat seeds germinated in substrate containing 100 ppm mercury completely neutralized by peptidomelanin (at 1:1 and 1:10 stoichiometric equivalent ratios) display significantly larger seed masses compared to wheat seeds germinated in 100 ppm mercury alone (1:0). (C) When germinated in substrate containing 100 ppm mercury, wheat shoot length shows a moderate inverse correlation (log–log) with increasing concentrations of peptidomelanin (r = −0.62, p = 1 × 10^–9). (D) Shoot lengths of wheat seeds grown in substrate containing 100 ppm mercury completely neutralized by peptidomelanin (at 1:1 and 1:10 stoichiometric equivalent ratios) display a moderate positive correlation (log–log) with increasing concentrations of peptidomelanin (r = 0.58, p = 9 × 10^–5). (E) ICP-OES quantification of the mercury concentration of wheat seed homogenates. When germinated in substrate containing 100 ppm mercury, plant tissue mercury concentrations show a strong inverse correlation (log–log) with increasing concentrations of peptidomelanin (r = −0.91, p = 5 × 10^–5). Wheat seeds germinated in substrate containing 100 ppm mercury either partially or completely neutralized by peptidomelanin (at 1:0.1, 1:1, and 1:10 stoichiometric equivalent ratios) display significantly lower tissue mercury concentrations compared to wheat seeds germinated in 100 ppm mercury alone (1:0). (F) Peptidomelanin reuse experiment. We assayed peptidomelanin’s ability to bind mercury after 5 cycles of binding and removal via EDTA chelation. Peptidomelanin shows good reuse characteristics for 4 cycles. Mercury was quantified via SEM-EDS. Green values indicate the change in mercury stoichiometric ratio (wt%:100%) between mercury binding and removal (mean and standard deviation). For panels B, D, and E, statistical significance testing was performed using the Welch 2-sample t test (p-values in blue). All substances were compared to wheat germinated in 100 ppm mercury in the absence of peptidomelanin. P-values ≤0.05 are shown. For panels A–E, “r” (in red) represents the Pearson’s correlation coefficient. For panels A–E, “p” (in red) represents the p-value of our regression model against the null model (a horizontal line) using an ANOVA. All data points, and not just the means/medians, were used to create regression models. For panels A,B, E and F, error bars represent the standard deviation from the mean. For panels C and D, error bars represent the interquartile range. Here, the median and quartiles were considered as shoot lengths do not follow a Gaussian distribution. Raw data (images) are provided in Figure S12.

Figure 10

Our model of the biochemical composition of peptidomelanin. Claims 1–9 are supported by experimental data as described in the text.