Despite considerable advances in the past decade, melanin research still suffers from the lack of universally accepted and shared nomenclature, methodologies, and structural models. This paper stems from the joint efforts of chemists, biochemists, physicists, biologists, and physicians with recognized and consolidated expertise in the field of melanins and melanogenesis, who critically reviewed and experimentally revisited methods, standards, and protocols to provide for the first time a consensus set of recommended procedures to be adopted and shared by researchers involved in pigment cell research. The aim of the paper was to define an unprecedented frame of reference built on cutting-edge knowledge and state-of-the-art methodology, to enable reliable comparison of results among laboratories and new progress in the field based on standardized methods and shared information.
The melanins can be still regarded as the most enigmatic pigments/biopolymers found in nature (Ito et al., 2011a). Unlike the vast majority of natural pigments, the melanins cannot be described in terms of a single well-defined structure and, as a result, there still remains today a lack of general consensus what actually melanin is. A variety of definitions and models are found in the literature, which reflect, however, an arbitrary use of terminology as well as several assumptions and speculations that have never been proven on experimental grounds. Crucial gaps stem from the lack of standardized procedures and methodologies, failure to take in due account melanin properties and the consequences of harsh purification procedures, a widespread tendency to compare materials obtained under different conditions, to extrapolate data referring to natural pigments from studies on synthetic pigments, or to draw conclusions and implications from observations made on unsuitable models.
The aim of this paper was to provide a critical assessment of methodological and practical issues in melanin research concerning (i) isolation and purification of natural melanins; (ii) preparation of synthetic model pigments; (iii) physical, spectral, and chemical characterization; (iv) use of melanins and melanogenic enzymes for biological studies; (v) preparation and assessment of standard compounds for melanin research, including commercially available pigments and enzymes.
Starting from a careful review of current methods and standards, the paper provides a selection of guidelines and procedures that have been verified and optimized, when necessary, through an ad hoc experimental revision. Provided herein as integral part of the paper is also the Appendix S1 section, which contains the first systematic collection of reference data and experimental protocols to be recommended as state-of-the-art for future research in the field.
Definition and classification
The term ‘melanin’ was first coined by Berzelius in 1840 to refer to black animal pigments. Since then, it has been widely used to indicate any black or dark brown organic pigment occurring throughout the phylogenetic scale without any specific structural, biogenetic, or functional implication. Nicolaus (1969) suggested a classification of melanins into three main groups, eumelanins, pheomelanins, and allomelanins, the former two groups comprising animal pigments and the latter encompassing the broad variety of dark non-nitrogenous pigments of plant, fungal, and bacterial origin. Pseudomonas and Aspergillus fumigatus can produce in the presence of tyrosine a eumelanin-like pigment termed pyomelanin via homogentisic acid (Schmaler-Ripcke et al., 2009). Aspergillus fumigatus can also synthesize melanin-type pigments from 1,8-dihydroxynaphthalene, while Serratia marcescens (Trias et al., 1989) or the pathological fungus Cryptococcus neoformans (Casadevall et al., 2000) can produce similar pigments from alternate pathways. Plants also produce dark phenolic pigments that have sometimes been referred to as catechol melanins, although the term conveys no information about the broad diversity and complexity of the polyphenolic precursors. To limit confusion, Prota proposed a more restrictive usage of the term ‘melanin’ to include only those pigments that are formed intracellularly by the oxidation of tyrosine and related metabolites (Prota, 1995), thus taking both the biogenetic origin from tyrosine and the metabolic activity of melanocytes (and, occasionally, related cell systems) as stringent requisites.
Although in the authors’ opinion this restrictive classification of melanins remains valid, it is nonetheless difficult to find alternative definitions for the broad variety of dark phenolic pigments from plants and microorganisms or for the diverse synthetic pigments of biomedical and technological interest, which are driving much of current progress in the field. Accordingly, it may be convenient to maintain a wider, general purpose classification, rooted in the tradition but revisited in light of recent progress, that includes (i) sensu stricto melanins, (ii) the dark phenolic pigments from lower organisms, and (iii) synthetic pigments produced either chemically or enzymatically from natural precursors. The following set of practical definitions is thus proposed and recommended.
Melanins: Pigments of diverse structure and origin derived by the oxidation and polymerization of tyrosine in animals or phenolic compounds in lower organisms.
Eumelanins: Black-to-brown insoluble subgroup of melanin pigments derived at least in part from the oxidative polymerization of l-dopa via 5,6-dihydroxyindole intermediates.
Examples: sepia melanin, black hair melanin.
Pheomelanins: Yellow-to-reddish brown, alkali-soluble, sulfur-containing subgroup of melanin pigments derived from the oxidation of cysteinyldopa precursors via benzothiazine and benzothiazole intermediates.
Examples: red hair melanin, hen feather melanin (it should be noticed, however, that the red hair pigment is rarely pure pheomelanin).
Neuromelanins: Dark pigments produced within neurons by the oxidation of dopamine and other catecholamine precursors.
Example: substantia nigra melanin.
Pyomelanins: Dark pigments produced by microorganisms mainly, but not exclusively, from homogentisate.
Moreover, for all types of natural pigments, the term ‘melanin’ should be preceded by the natural source, for example sepia melanin, hair melanin, while for synthetic pigments the term ‘melanin’ (or eumelanin, pheomelanin whenever appropriate) should be preceded by that of the precursor, for example dopamine melanin, 5,6-dihydroxyindole melanin, cysteinyldopa-melanin, dihydroxynaphthalene melanin.
For the direct investigation of melanins, it is essential to keep in due consideration certain peculiar and critical properties of the pigments (Prota, 1992):
Melanins are not stable indefinitely. They may undergo more or less profound structural degradation on chemical treatment with acids (decarboxylation), alkali (oxidative ring fission), oxygen (oxidation of catechol units), and hydrogen peroxide, even during their (bio)synthesis. They also undergo alteration with physical agents such as heat and light, as well as with aging, even if left dry on a shelf.
Melanins vary their properties with the degree of hydration. Dry melanins (especially eumelanins) are tightly aggregated and may exhibit physical properties that are different from those of freshly collected wet samples.
Consideration of these caveats was central to all experimental protocols reported in this paper for structural and biological studies of melanins.
The impact of the isolation procedure on the structural and physical properties of natural melanins has been addressed earlier (Liu and Simon, 2003). Three important issues emerged, namely: (i) the close association of proteins and other biological components with melanin; (ii) the variety of metal cations present in the natural pigment; and (iii) the influence of the drying method on the physical properties of the pigment, for example aggregation state, surface area-to-mass ratio, and porosity of the material.
The presence of tightly bound cellular components is the major obstacle to the isolation of melanins from natural sources. Harsh hydrolytic treatments with boiling mineral acids or alkali have been abandoned following realization that, in spite of the lack of visual changes, pigment skeleton and functionalities are profoundly affected. Heating of melanins with or without hydrochloric acid at reflux has been shown to lead to extensive decarboxylation (Ito, 1986; Prota, 1992), and care must be taken to avoid extremes in pH and temperature. In general, no attempt should be made to separate melanin from these internal proteins, because such a separation would destroy the granules. Current strategies focus on the attack of the keratin matrix by breaking disulfide bonds and exposure to proteolytic digestion (Novellino et al., 2000).
Physical disaggregation of particulate melanin using a wet milling step was also found to facilitate the removal of significant quantities of adsorbed protein and other molecular structures in various forms. Issues are posed by the large number of free carboxylic acid residues in natural melanins, responsible for the cation-exchange properties.
Methods for purifying melanin in the form of intact melanosomes have been reported (Watabe et al., 2005). These methods utilize highly pigmented cells to maximize recovery and employ sucrose density gradients to separate melanosome fractions based on their density (which is determined in large part by the amount of melanin pigment that they contain). See also Appendix S1 pages S3–S5.
Sepia ink has been widely investigated as one of the most accessible natural sources of eumelanin, and a variety of isolation procedures have been described (Liu and Simon, 2003, 2005; Simon et al., 2008). Iterative washings of the ink with water and centrifugation give a pellet containing largely spherical granules (150 nm in diameter) with a protein content of 6–8% by mass. A very simple isolation procedure that would not affect pigment integrity involves centrifugation of freshly collected ink in 0.01 M HCl in the cold, followed by storage at 4°C without desiccation and in the absence of oxygen (Pezzella et al., 1997). The pigment thus obtained usually shows a protein content of approximately 5%. Sepia melanin has been a proposed standard for natural eumelanin (Chedekel et al., 1992).
Pigments from hair have been widely used as a model for human melanin. Melanin in human hair is deposited in the form of granules in the medulla, in the porous core of the hair fiber, and in the cortex surrounding it. Access to the inside of the hair by a proper swelling of the overlapping layers of the external covering, the cuticle, is therefore necessary for pigment release. By sequential use of proteinase K, papain and protease, and pretreatment of the tissue with dithiothreitol, protein removal can be efficient. The enzymatic extraction preserves the integrity of the melanosome, removes most of the external proteins, and therefore should be the preferred choice for the isolation of melanin from hair samples (Liu et al., 2003; Novellino et al., 2000).
Homogenization of the finely minced hair sample is required prior to exposure to proteolytic agents to favor their action. This can be achieved by the use of a glass/glass potter such as Tenbroeck homogenizer, while other mechanical devices (grinder, ultrasonic disrupters) currently employed for tissue homogenization prove often inappropriate. It is preferable to use freshly collected hair samples, as photoaging induces structural modifications not only in pheomelanins (Greco et al., 2009) but also in eumelanins (Wakamatsu et al., 2012a).
Retinal pigment epithelium (RPE) melanosomes
In some cases, pigmented cells can be lysed and intact melanosomes can be isolated from other cellular constituents, for example RPE. Ultracentrifugation on a discontinuous sucrose gradient yields high-purity samples where the integrity of the melanosomes is preserved. This is exemplified by the atomic force microscopy (AFM) images of bovine RPE melanosomes isolated in this manner (Liu et al., 2005).
Treatment of the iridial tissue with collagenase results in a substantial removal of the protein as the result of a specific attack to collagen, which cannot be achieved with other proteolytic enzymes, such as those used for hair. Non-collagenic protein structures can be specifically attacked by trypsin, while lipids and glycoproteins may be removed by pancreatin. The protocol ensures an efficient protein removal without detectable modification of pigment structure (Novellino et al., 2000; Peles et al., 2009).
B16 mouse melanoma
Intact melanosomes from B16 melanoma can be obtained by sucrose density gradient ultracentrifugation followed by the treatment of the pigment fraction with Triton and SDS in 0.1 M Tris–HCl, pH 7, to remove strongly retained impurities. Extensive washings give a pigment that can be used for structure determination purposes (Jimbow et al., 1984). The purity of melanosomes can be checked under electron microscopy after fixing melanosome pellets with 1% osmium tetroxide. An electron micrograph showing the ultrastructure of the starting material and a purified melanosomal preparation is provided in Appendix S1.
Most isolation protocols involve the use of alkaline conditions (0.1 M NaOH) to disrupt the hard keratin matrix (Prota, 1992) and cause the release of the yellow brown pigment in solution. Actually, the mild repeated enzymatic treatment used for hair samples may be likely extended to these tissues provided that an efficient homogenization of the sample is previously performed.
Human neuromelanin differs from other melanins in that it consists of a pheomelanin core generated from cysteinyldopamine and wrapped by a eumelanin-like dopamine-derived shell (Bush et al., 2006). The isolation and purification of neuromelanin is very complex due to the presence of the tissue matrix that tightly retains the pigment. Human neuromelanin contains as main impurities glycolipids, dolichol and chemically bound metals such as iron and zinc, and proteins (Fedorow et al., 2005; Wakamatsu et al., 1991; Zecca et al., 2001). Accordingly, recommended protocols for the preparation of substantia nigra neuromelanin involve a complex sequence of steps, including homogenization of tissues, washings with phosphate buffer to remove water-soluble components, treatment with SDS and proteinase K to remove proteins, dialysis to remove inorganics and other low molecular weight species, and finally washing with methanol and hexane to remove lipid components (Zecca et al., 2004).
Commercially available melanin precursors include l-tyrosine and l-dopa, as well as dopamine and related catecholamines (for comments on dopamine-derived melanin, see following section). Non-commercial precursors include 5,6-dihydroxyindole (DHI), 5,6-dihydroxyindole-2-carboxylic acid (DHICA), and cysteinyldopas (Figure 1). These may be used to prepare model melanin pigments for structural and functional studies. Although multistep, gram-scale syntheses of 5,6-dihydroxyindoles and cysteinyldopas have been reported, we restrict the present discussion to expedient subgram-scale preparations useful for non-chemists and that do not require organic synthesis skills nor special laboratory equipments. Criteria for purity are provided for each compound (HPLC, elemental analysis, UV, NMR, MS).
The preparation of DHI and DHICA in subgram quantities takes advantage of the chemical behavior of dopachrome, a red-colored, reactive intermediate in eumelanogenesis. When a neutral solution of dopachrome, prepared by ferricyanide oxidation of dl-dopa, is kept at neutral pH, decarboxylation prevails and DHI is by far the main product (Wakamatsu and Ito, 1988; see Appendix S1 page S6). When the pH of the dopachrome solution is raised to 13, DHICA is rapidly produced instead through tautomerization. Both dihydroxyindoles are separated by extraction and purified by crystallization, to give almost colorless crystals in good yields (60–80%). Although these dihydroxyindoles, especially DHI, are extremely labile to auto-oxidation, they are rather stable when kept in a deep freezer (sealed under argon).
Cysteinyldopas and cysteinyldopamines
5-S-Cysteinyldopa (5SCD) is regarded as the main biosynthetic precursor of pheomelanins. 5SCD is obtained as main product by the reaction of dopaquinone, generated by chemical or enzymatic oxidation of dopa, with excess cysteine. Small amounts of other cysteinyldopa isomers, namely 2-S-cysteinyldopa (2SCD), 6-S-cysteinyldopa, and 2,5-S,S-dicysteinyldopa, are also formed (Chioccara and Novellino, 1986; Ito and Prota, 1977).
At present, there are two one-pot procedures to obtain 5SCD from subgram to gram scale, involving the generation of dopaquinone by the oxidation of dopa followed by rapid reaction with cysteine. Procedure (i) involves the use of cerium ammonium nitrate in 2 M sulfuric acid to oxidize l-dopa and the resulting dopaquinone is poured into a solution of l-cysteine in 2 M sulfuric acid under a flux of argon. Biomimetic-type procedure (ii) involves the use of mushroom tyrosinase (and O2) as an oxidant and the resulting dopaquinone is trapped by the cysteine present in the solution (Ito and Prota, 1977) (see also Appendix S1 pages S7–S13). Fractionation of the reaction mixture affords 5SCD in moderate to good yields (40–65%), while 2SCD is obtained in lower yields (approximately 10%). Another gram-scale preparation of 5SCD (and 2SCD) was reported using hydrogen peroxide in the presence of iron–EDTA complex as an oxidizing agent (Ito, 1983).
2SCD, the minor constituent of red hair pheomelanin (Greco et al., 2009; Wakamatsu et al., 2009), can be selectively prepared by a gram-scale multistep synthesis starting from commercially available caffeic acid (Panzella et al., 2007b).
5-S-Cysteinyldopamine and its isomer 2-S-cysteinyldopamine are regarded as minor precursors of neuromelanin (Wakamatsu et al., 2003) and can be prepared by tyrosinase oxidation of dopamine in the presence of cysteine (Ito et al., 1986).
Main types of synthetic melanins include eumelanins, pheomelanins, and neuromelanins, the latter modeled by the oxidation of dopamine in the presence of cysteine (Wakamatsu et al., 2003, 2012b). Their preparation should be based on conditions expected to mimic at best melanogenesis in the biological environment. However, a broad variety of methods and procedures have been described, involving diverse parameters that do not always meet the requisite of biological relevance. Some representative procedures reported in the literature for the preparation of synthetic eumelanins are schematically summarized in the Appendix S1 section. Oxidation of dopamine, originally investigated as a model approach to neuromelanin, has recently become popular as a method for preparing a dark eumelanin-like material known as polydopamine, with extraordinary adhesion and coating properties (Lee et al., 2007). Polydopamine possesses structural and physicochemical properties in common with eumelanins, and thus, it may be also referred to as dopamine melanin. It has been shown to consist of both cyclized (DHI) and uncyclized units in variable proportions depending on the mode of preparation (Della Vecchia et al., 2013) and is therefore different from pure DHI-melanin in that it contains non-cyclized catecholamine units.
When preparing synthetic melanins, three important sets of parameters must be considered:
Post-synthetic procedures, including work-up (precipitation, reduction, centrifugation, lyophilization) and storage (dry, cold, under nitrogen).
Structural features and properties of synthetic melanins depend on all of the above factors. Therefore, it may be misleading to compare pigments prepared under different conditions even if from the same substrate.
Main precursors used to prepare synthetic eumelanins include l-tyrosine, l-dopa, DHI, and DHICA. Tyrosine and dopa-melanins consist largely of DHI-related units with some 10% of DHICA, whereas natural eumelanins contain DHI and DHICA units in approximately 1:1 ratios (Ito, 1986; Pezzella et al., 1997). Commonly used protocols for the preparation of polydopamine or dopamine melanin are based on prolonged aerial oxidation of the catecholamine (10 mM) in Tris or other buffers at pH 8.5 leading to black insoluble materials (Della Vecchia et al., 2013).
A suitable precursor concentration to mimic the in vivo situation is 1 mM (the natural substrate l-tyrosine is not soluble above 2.5 mM). However, for the practical reason of preparing subgram quantities of melanins, 10 mM concentration is preferable. Varying this parameter may affect the degree of polymerization and aggregation. Tyrosinase is the oxidizing system of choice, although caution may be in order against the use of the commercial mushroom enzyme, largely because of its low specificity and different mode of action from the mammalian enzyme. Because tyrosinase is inactivated during the oxidation (Land et al., 2007), its concentration should be such to ensure complete substrate consumption without total inactivation. Peroxidase/hydrogen peroxide provides an alternate system for the preparation of synthetic melanins from DHI and DHICA (d'Ischia et al., 2005). Under these conditions, pigment formation proceeds rapidly and the yields are higher than with tyrosinase. Care should be taken to use the minimal amount of hydrogen peroxide to allow for a complete oxidation of the substrate without concomitant degradation of the pigment as formed. This amount can be fixed in two molar equivalents with respect to the substrate.
Air flux (e.g., simple stirring or shaking) in the place of oxygen is recommended to ensure a more biomimetic and reproducible oxygen tension during polymerization and a lower competition of auto-oxidative and peroxidative processes. Chemical oxidants that can be used include potassium ferricyanide or periodate, but their amounts and general use are subject to caveats depending on the scope of the synthesis and whenever a truly biomimetic procedure is required.
A suitable medium is phosphate buffer at 0.05–0.1 M concentration and at pH around neutrality (6.8, 7.0, or 7.4). Use of alkali (pH > 8) should be avoided whenever a biomimetic material is desired, due to profound structural alterations secondary to quinone ring fission (Prota, 1992).
Work-up is typically based on precipitation with dilute acids, which serves the dual scope of stabilizing catechol components of the pigment to further oxidation and favoring collection. Drying may be carried out either in a desiccator or by lyophilization, with the latter being preferable whenever profound physical alterations caused by extensive desiccation are not desired (see General properties).
Synthetic pheomelanins may be prepared by the oxidation of dopa in the presence of cysteine, or by the oxidation of 5SCD. The former protocol gives a more biomimetic pigment because it may contain all of the cysteinyldopa isomers produced by the dopa-cysteine reaction, while the latter gives a chemically more homogeneous standard that may be useful to investigate specific structural aspects (Wakamatsu et al., 2009).
A suitable concentration for l-dopa and l-cysteine or for 5SCD would be 1 mM. However, for preparative purposes, 10 mM of both l-dopa and l-cysteine ensures high product yields and recovery (Wakamatsu et al., 2009). Use of excess cysteine is no longer recommended because the intermediate cysteinyldopas are oxidized too slowly to pheomelanin unless a trace of dopa acts as a catalyst (Ito, 1989).
It is still unresolved whether there is as yet undiscovered enzymology in the synthesis of pheomelanins. Tyrosinase may be used as an oxidizing system for pheomelanin synthesis (for comments on enzyme concentration, see text above concerning eumelanins), although oxidation of 5SCD requires l-dopa (0.05 eq.) as a catalyst. Again, air flux is recommended instead of oxygen to provide a more biomimetic oxygen tension and to minimize auto-oxidative and peroxidative processes. Oxidation of cysteinyldopas can also be carried out with peroxidase/hydrogen peroxide (Panzella et al., 2010). Addition of Zn2+ ions markedly affects the synthetic process and is recommended whenever more homogeneous preparations with a higher benzothiazine content are desired. The role of Zn2+ ions is to direct the reaction course toward the formation of carboxylated benzothiazine units. Phosphate buffer at 0.05–0.1 M concentration and at pH 6.8–7.4 is a suitable medium.
Typical work-up is based on precipitation at pH 3 with acetic acid. Lower pH values would render pheomelanin soluble.
A few companies produce synthetic eumelanins for research purposes. In 2012, Sigma catalogue offers a product labeled melanin (CAS-No. 8049-97-6, EC-No. 232-473-6), which is reported to be prepared by the oxidation of tyrosine with hydrogen peroxide. The product differs from natural eumelanins in that it is freely soluble in slightly alkaline aqueous medium. Considering the marked difference between the tyrosine–hydrogen peroxide reaction and the biogenetic origin of natural melanin, use of the commercial synthetic melanin should therefore be subject to caution. Comparison of various properties of synthetic eumelanins, commercial melanins, and sepia melanin is reported in Table 1.
Table 1. Preparation and analysis of synthetic and commercial melanins
Under the same CAS number, a melanin from Sepia officinalis is also offered. Different preparations of sepia melanin were found to give similar analytical values (Table 1). This sepia melanin may be used as a reference standard (with lot number) for characterization by chemical degradation (see below) and UV/VIS absorbance.
Several markers are currently used for melanin analysis but only two, pyrrole-2,3,5-tricarboxylic acid (PTCA) and 3-amino-4-hydroxyphenylalanine (3-AHP), are commercially available. Thus, most markers must be prepared by simple chemical processes, as reported below. The purity of the markers described in this section was assessed by HPLC, elemental analysis, UV, NMR, and MS analyses (see Appendix S1).
Typical degradation markers for eumelanin analysis and quantitation are pyrrole-2,3,5-tricarboxylic acid (PTCA) and pyrrole-2,3-dicarboxylic acid (PDCA) (Figure 2), which can be readily obtained in 100 mg quantities by oxidative degradation of commercial 5-hydroxyindole-2-carboxylic acid and 5-hydroxyindole, respectively, followed by extraction and crystallization. (Ito and Wakamatsu, 1998). Recently, pyrrole-2,3,4,5-tetracarboxylic acid (PTeCA) and pyrrole-3,4,5-tricarboxylic acid (isoPTCA) were added to a list of oxidation products from eumelanin (Ito et al., 2013a,b). Those pyrrole carboxylic acids appear to derive from cross-linking of dihydroxyindole moiety that occurs during (photo) aging.
Several pheomelanin degradation products can be used as markers for pigment determination in tissues (Figure 2). 4-Amino-3-hydroxyphenylalanine (4-AHP) is a major degradation product of pheomelanin upon hydroiodic acid (HI) hydrolysis, along with the minor isomer 3-amino-4-hydroxyphenylalanine (3-AHP) (Wakamatsu et al., 2002). 4-AHP can be prepared by HI hydrolysis of 5SCD-melanin (Wakamatsu et al., 2009). However, this amino acid can be more readily obtained in a 100 mg quantity by the nitration of commercially available m-tyrosine followed by HI reduction of the resulting 3-hydroxy-4-nitrophenylalanine (along with other possible isomers). Likewise, 4-amino-3-hydroxyphenylethylamine (4-AHPEA) and 3-amino-4-hydroxyphenylethylamine (3-AHPEA), degradation products of neuromelanin, can be obtained by HI hydrolysis of 5-S-cysteinyldopamine-melanin and 2-S-cysteinyldopamine-melanin, respectively (Wakamatsu et al., 2003).
Alkaline hydrogen peroxide degradation of pheomelanin yields 6-alanyl-2-carboxy-4-hydroxybenzothiazole (BTCA), 7-alanyl-2-carboxy-4-hydroxybenzothiazole (BTCA-2), thiazole-2,4,5-tricarboxylic acid (TTCA), and thiazole-4,5-dicarboxylic acid (TDCA). BTCA and BTCA-2 can be obtained by alkaline hydrogen peroxide oxidation of 5SCD-melanin and 2SCD-melanin, respectively (Greco et al., 2009; Napolitano et al., 1996). BTCA can be easily prepared from dopa and cysteine through a one-pot, 4-step procedure involving the use of ferricyanide and Zn2+ to induce oxidative conversion of intermediate 5SCD into 3-carboxybenzothiazine, and of acidic persulfate to promote ring contraction. Purification of the final mixture by HPLC gives BTCA in around 50% isolated yield. Starting from pure 5SCD, the procedure is reduced to three steps plus chromatographic purification. This latter procedure, applied to 2SCD, allows for the preparation of BTCA-2. 6-Alanyl-4-hydroxybenzothiazole (BT) and 7-alanyl-4-hydroxybenzothiazole (BT-2) may be obtained by the same protocol with a final thermal decarboxylation step preceding HPLC fractionation (Greco et al., 2009). TTCA and TDCA are obtained similarly from 5SCD-melanin (Wakamatsu et al., 2003). TTCA can also be obtained in good yield as the triethyl ester by a simple chemical synthesis from commercially available ethyl thiooxamate and diethyl chlorooxaloacetate (von Erlenmeyer et al., 1948) with some modification. Alkaline hydrolysis of the triethyl ester affords tripotassium salt of TTCA, used as the standard. Decarboxylation of TTCA under controlled acidic conditions yields TDCA in good yield.
Chemical oxidation of eumelanins under various conditions gives PTCA as a major product (Prota, 1992), which serves as a basis for quantitative analysis of eumelanins. PTCA is regarded also as a specific index of DHICA units or 2-substituted DHI units, whereas PDCA levels indicate 2-unsubstituted DHI units. The first microanalytical application of the reaction was based on HPLC analysis of PTCA produced upon permanganate oxidation of eumelanins in 1 M sulfuric acid (Ito and Fujita, 1985), which became a standard of eumelanin assay for some time (Ito and Wakamatsu, 2003). However, a number of disadvantages prompted replacement with an alternative method involving hydrogen peroxide oxidation in 1 M NaOH (Napolitano et al., 1996, 2000) or in 1 M K2CO3 (Ito and Wakamatsu, 1998; Ito et al., 2011b; Pezzella et al., 1997). Advantages of the latter method include (i) omission of an extraction step with ether, (ii) production of PDCA from DHI-derived units, (iii) higher yields of PTCA, and (iv) concomitant production of pheomelanin markers BTCA, BTCA-2, TTCA, and TDCA that can be analyzed simultaneously (Ito et al., 2011b). While the alkaline hydrogen peroxide degradation is definitely the method of choice for eumelanin characterization, which of the two variants based on 1 M NaOH or 1 M K2CO3 as a medium is better remains a matter of convenience. With 1 M NaOH artificially produced interfering peaks are more obvious because of the harsh conditions, while yields of BTCA and BTCA-2 are greater than with 1 M K2CO3 (Ito et al., 2011b). Analytical conditions, particularly the HPLC mobile phase, are also critical for satisfactory elution and reliable identification and quantitation of the melanin markers.
Pheomelanin degradation products used in current analytical methods include 4-AHP, 3-AHP, TTCA, TDCA, BTCA, and BTCA-2 (Greco et al., 2009; Ito and Wakamatsu, 2003; Panzella et al., 2007a; Prota, 1992), while recently identified pyridine-containing fragments await further evaluation (Greco et al., 2011, 2012). Choice of the best methodology depends on several factors, including the type of sample, equipment availability, and the researcher's familiarity with laboratory techniques.
The first microanalytical method was based on HI hydrolysis of pheomelanin-containing tissues followed by HPLC determination of 4-AHP as a marker of 5SCD-derived units (Wakamatsu et al., 2002).
An alternate, more recent method is based on alkaline hydrogen peroxide degradation of the sample followed by HPLC quantitation of BTCA and TTCA (together with PTCA) (Ito et al., 2011b; Panzella et al., 2007a). While TTCA is a specific marker of benzothiazole units, 4-AHP, and possibly in part BTCA, derives from benzothiazine-containing structures (Napolitano et al., 2000; Wakamatsu et al., 2009). With both methods, 3-AHP and BTCA-2 can be analyzed as indices of 2SCD-derived units. The ratios of 4-AHP to 3-AHP (HI hydrolysis) and BTCA to BTCA-2 (alkaline hydrogen peroxide degradation) would thus provide quantitative data about the relative proportions of 5SCD/2SCD units in pheomelanins, although the differential stability of 5SCD versus 2SCD benzothiazine units during photochemical and thermal decomposition must be kept into account (Greco et al., 2009; Ito et al., 2011b; Wakamatsu et al., 2012a).
Although both 4-AHP and BTCA are produced in good yields, 4-AHP is usually revealed with an electrochemical detector, whereas BTCA and TTCA analyses require UV detection. Therefore, the 4-AHP-based procedure results in a much greater sensitivity and would be the method of choice for detecting pheomelanins in human epidermis or cultured melanocytes, which are available only in minute quantities. A disadvantage is that removal of HI may not be easy in a biology laboratory. On the other hand, both BTCA and TTCA can be selected as reliable pheomelanin markers that provide complementary information and are suitable for routine pigment analysis, although their determination is affected by the lower sensitivity, due to instrumental limitations, and, in the case of TTCA, by some artificial production from eumelanic human hairs (Ito et al., 2011b). A careful selection of chromatographic conditions may also be critical. Selection of BTCA or TTCA as marker may thus be dictated by considerations relating to synthetic access, structural specificity, and analytical properties.
A comparative view of the various analytical methods applied to the characterization of synthetic melanins described above is given in Table 1.
Spectral and biophysical studies
Although EPR spectra of melanins are quite characteristic, they are by no means unique to melanin. This is particularly true for eumelanins. However, the unusual physicochemical properties of melanin pigments make their identification based on responses of their EPR signal to selected agents relatively unambiguous (Sealy et al., 1982). Especially relevant is the so-called comproportionation equilibrium, that is, the equilibrium between fully reduced (H2Q), fully oxidized (Q) melanin units, and their semi-reduced (semi-oxidized) forms (SQ) (Felix et al., 1978; Sarna and Plonka, 2005). It can be described by a simple equation:
In eumelanins, the reduced units are most likely of DHI or DHICA nature, while the oxidized units are the corresponding indolequinones. Of course, the semi-reduced (semi-oxidized) units are the so-called extrinsic melanin free radicals. The comproportionation equilibrium in melanins can be modified by many factors including pH, light, redox agents, and diamagnetic multivalent metal ions. The intensity of the melanin EPR signal and its anisotropy increases with pH and doping the melanin with zinc ions. At X-band (microwave frequency approximately 9 GHz), at which standard EPR spectrometers operate, EPR spectra of eumelanin represent a single slightly asymmetric line 0.4–0.6 mT wide with a g-factor close to 2.004. For fully hydrated samples, the anisotropy of the melanin EPR signal and its g-factor are a function of pH. Pheomelanin has a broader signal with the total width approximately 3 mT and g = 2.005. The EPR signal of pheomelanin is reminiscent of immobilized nitroxide EPR spectra with distinct hyperfine splitting, except the latter exhibit larger hyperfine splitting than the former. At comparable concentration, the EPR signal amplitude of pheomelanin is about one order of magnitude lower than that of eumelanin (Figure 3).
The observable EPR signal intensity of melanin can be modulated by an order of magnitude and can also be enhanced by steady-state irradiation with visible and near UV light. The effect is transient, and after termination of the irradiation, the signal intensity slowly returns to its initial level if no photo-oxidation occurred. Removal of oxygen from the samples prior to their irradiation may therefore be appropriate. It is important to stress that comproportionation equilibrium of the melanin subunits, responsible for the observable changes in the melanin EPR signal intensity, only operates in fully hydrated melanin. It has been shown that deep dehydration of melanin irreversibly modifies physicochemical properties of the pigment, including susceptibility of its EPR signal to changes in pH, doping with zinc ions, and irradiation with light. It appears that excessive drying of melanin abates its comproportionation equilibrium (Mostert et al., 2012a; Sealy et al., 1980).
The effect of a significant change in the comproportionation equilibrium of the melanin subunits induced by saturating melanin with zinc ions is shown in Figure 4. It is apparent that the same concentration of eumelanin and pheomelanin in the presence of high concentration of zinc ions exhibits EPR signals with intensities over one order of magnitude larger than those observed in PBS (see Figure 3 for comparison). In addition, complexation of zinc ions makes the pheomelanin component significantly more pronounced both in samples containing synthetic cysteinyldopa-melanin (Figure 4B) and in melanoma cells (Figure 4C). Arrows in Figure 4 indicate the eumelanin and pheomelanin components of the pigment examined.
Because eumelanin is a good chelator of multivalent transition metal ions (Meredith and Sarna, 2006) and metal ions bound to eumelanin can modulate the intensity of the observable EPR signal of melanin radicals (Sarna et al., 1976), quantitative determination of eumelanin by EPR spectroscopy may require preliminary removal of bound metal ions. This can be achieved by prolonged washing of the samples with aqueous solutions of strong chelators such as EDTA, DTPA, and desferrioxamine at high concentration (Enochs et al., 1993; Shima et al., 1997; Zecca et al., 2008). An alternative approach is to acidify melanized samples with strong acids – H2SO4 or HCl – to pH 0–1. It is believed that such a treatment will release most of the bound metals. Of course, the melanin standard should be treated the same way, and the risk of pigment degradation cannot be neglected.
If room temperature EPR examination is preferred or required, aqueous suspension of the melanized material is transferred to standard quartz EPR flat cells (approximately 0.3 mm internal thickness and 8 mm width). Considering that lower amount of fully hydrated material can be examined at room temperature compared to liquid nitrogen temperature, the corresponding lower Q factor of the resonant cavity at room temperature and different microwave power saturability of the melanin EPR signal under both experimental conditions, it is safe to conclude that the room temperature EPR examination of melanin is several times less sensitive than that performed at liquid nitrogen temperature. In addition, the proper position of a flat cell with aqueous sample in the resonant cavity is critical for its correct tuning, making quantitative determination of lossy samples at room temperature more difficult than at liquid nitrogen temperature.
Optical, electrical, and microstructural characterization
A detailed understanding of the optical, electrical, and microstructural properties of melanin not only provides another tool to define and identify melanins, but also opens the intriguing possibility of melanin bioelectronics (Bothma et al., 2008; d'Ischia et al., 2009).
Typically, the optical absorption of a well-dispersed melanin suspension is monotonic, broad, and featureless when Mie scattering is eliminated or accounted for (Riesz et al., 2006). The spectrum has an exponential dependence upon wavelength (or sigmoidal with respect to energy). A typical solution absorption spectrum of synthetic melanin formed from the auto-oxidation of dl-dopa in alkaline aqueous solution is provided as Appendix S1 (Meredith et al., 2006). Solid eumelanin pellets, aggregated solutions, or thin films often show pronounced Mie scattering which broadens, flattens, and extends the absorption into the near infrared. An integrating sphere must be used to collect such spectra in reflectance or transmittance to negate these effects, and recover the ‘real’ spectrum. Solid eumelanin thin films suitably disaggregated have almost identical absorption spectra to well-dispersed solutions (Bothma et al., 2008). This indicates that the fundamental absorption of melanins results from the primary chromophore and not some secondary aggregated state. Such results have also been reproduced with eumelanin thin films synthesized in, and cast from, alternative solvents such as dimethylsulfoxide (DMSO) (Abbas et al., 2011). Polymerization of melanin from DHICA or tyrosine produces ‘peaks’ in the monotonic absorption profile, possibly a signature of residual, low molecular weight precursor in the system, or residual protein in purified natural melanins (Tran et al., 2006).
Although generally considered to be non-fluorescent, eumelanin does emit radiation when photostimulated, albeit with a tiny quantum yield of order 10−4. Furthermore, the emission conforms to that expected from a typical organic chromophore (Appendix S1) and is excitation energy dependent with a complex signature characteristic of an ensemble of multiple chromophores (Meredith et al., 2006) according to the idea of ‘chemical disorder’. Such measurements require extremely sensitive emission detection and careful experimental technique (particularly in relation to accounting for so-called inner filter effects and emission re-absorption) (Riesz et al., 2005).
In the solid state (i.e., thin films or pressed powder pellets), eumelanin conducts electricity. Its electrical properties have a strong dependence upon its state of hydration, the form of which is critically dependent upon how the measurement is made and whether the system is allowed to reach equilibrium with its environment. Measurements of electrical conduction in a so-called sandwich electrode configuration (see Appendix S1) produced erroneous results due to the limited sample surface area available to absorb water (Mostert et al., 2012a). This non-equilibrium behavior delivers a hydration dependence consistent with the so-called modified dielectric theory of Powell and Rosenberg (Powell and Rosenberg, 1970) – a fact used to justify the now debunked model of melanin as a natural amorphous semiconductor (Mostert et al., 2012b). However, it has recently been demonstrated that the hydration dependence of the DC electrical properties of eumelanin has a completely different functional form if careful equilibrium measurements are made with knowledge of the water adsorption isotherms (Mostert et al., 2010, 2012a). This delivered strong evidence that eumelanin is a hybrid ionic-electronic material dominated by the flow of protons as the primary charge carrier. Absorbed water locally titrates the comproportionation equilibrium reaction generating semiquinone species and free protons to be conducted through the hydrated matrix via the Grotthus mechanism (Mostert et al., 2012b).
Photoconductivity measurements support this view of eumelanin electrical conductivity. In the solid state, photocurrent is produced when UV or white light is incident upon eumelanin. This photocurrent is derived from the same mechanism as the DC electrical conductivity (i.e., chemical self-doping). Photoexcitation also drives the formation of semiquinone radicals and water is needed to mediate the process – hence, only wet melanin photoconducts (see Appendix S1) (Mostert et al., 2012b).
Hence, electrical measurements on solid eumelanin require a detailed knowledge and control of the hydration state of the material. Furthermore, the appropriate electrode geometry (surface contacts versus sandwich contacts) must be used with full knowledge of the implications of proton-dominated electrical physics (non-ohmic behavior, AC conductivity, and long RC time constants) (Jastrzebska et al., 1998).
Low-contrast transmission electron microscopy (TEM) combined with radial Fourier analysis has been used to show that synthetic and natural melanins form a stacked secondary structure with characteristic interplane spacing of approximately 3.7 Å. Synthetic melanins curl into stacked ‘onion-like’ structures with relatively small primary sheet (oligomer) dimensions (see Supplementary Information), while natural eumelanins (sepia and bovine epithelium) tend to form much larger sheets, but still retain the same characteristic interplane spacing (Watt et al., 2009). The system can be disaggregated using a mild base and/or a π-stack breaker (Bothma et al., 2008; Watt et al., 2009). If this procedure is performed correctly, then the system does not bleach and the usual melanin monotonic absorption is retained, showing again that the optical properties are derived from the primary structure as indicated previously.
Enzymes for in vitro studies
As opposed to the well-defined and relatively simple reaction media employed in the chemical synthesis of model melanins, the melanosomal lumen where the biosynthesis of melanin pigments takes place is a complex environment. Its dynamic nature is underscored by the pheomelanin switch whereby a given melanocyte switches the pattern of pigment production from eumelanogenesis to pheomelanogenesis (Barsh, 2006; Ito and Wakamatsu, 2003, 2011). Thus, synthesizing models resembling as closely as possible natural melanins requires an understanding of the biochemical and physicochemical variables that may have an impact on the final structure of the melanin polymer. Mammalian melanogenesis involves not only tyrosinase but also at least two other melanogenic enzymes: tyrosinase-related protein 1 (Tyrp1 or gp75) and tyrosinase-related protein 2 (Tyrp2 or dopachrome tautomerase, Dct). In mouse melanocytes, Tyrp2 catalyzes the tautomerization of l-dopachrome to DHICA and Tyrp1 acts as a DHICA oxidase. Accordingly, the combined action of both proteins accounts for the incorporation of DHICA to the growing melanin polymer and for the typically higher contribution of carboxylated indolic units in natural melanins, compared with synthetic pigments (Aroca et al., 1990; Ito, 1986; Jiménez-Cervantes et al., 1994; Kobayashi et al., 1994; Korner and Pawelek, 1980; Palumbo et al., 1991; Pawelek et al., 1980). Of note, the Tyrps are expressed exclusively in animals and are not found in prokaryotic organisms, fungi, or plants. Importantly, the relative activity of the melanogenic enzymes seems to determine the final structure of the pigment, and preliminary in vitro studies indicated that higher levels of Tyrp2/Dct yield smaller polymers of lighter color (Aroca et al., 1992).
The different catalytic properties of the structurally similar tyrosinase family enzymes can be accounted for by the differential binding of specific metal cofactors. Tyrosinase, and probably also Tyrp1, binds copper ions in two conserved metal-binding motifs (Furumura et al., 1998), whereas Dct/Tyrp2 is a zinc protein (Solano et al., 1994, 1996). Because Zn ions cannot undergo redox reactions, Dct/Tyrp2 does not function in oxidation reactions, as opposed to Cu-containing tyrosinase, but instead behaves as a tautomerization catalyst.
Extensive genetic and biochemical evidence has shown the involvement of other melanosomal proteins in mammalian melanogenesis. The melanosomal matrix Pmel17 protein (also known as gp100) is the product of the silver locus. The protein undergoes a complex proteolytic processing within eumelanosomes. Pmel17 is required for the formation of the organelle's lamellar network (Hoashi et al., 2006; Kushimoto et al., 2001), where it acts to facilitate the polymerization and deposit of melanins (McGlinchey et al., 2009; Solano et al., 2000). However, so far, no specific enzymatic activity has been proposed for Pmel17. In addition to Pmel17, other melanosomal proteins are required for the correct maturation of the melanosomes that appears a requisite for efficient melanization. The oculocutaneous albinism type 1 (OA1) protein is a G-protein-coupled receptor associated with the melanosomal membrane, where it interacts with the premelanosomal protein MART-1 (Giordano et al., 2009). Inactivation of either OA1 or MART-1 leads to the formation of structurally aberrant melanosomes with decreased melanin contents. Other melanosomal proteins such as the product of the underwhite locus also contribute to efficient melanin deposition by facilitating correct trafficking and processing of tyrosinase (Costin et al., 2003).
In addition to the melanosomal proteins and low molecular weight thiol compounds, the pH of the melanosomal lumen contributes to the regulation of pigment biosynthesis. Mammalian tyrosinases have an optimal pH near neutrality, and the specific activity of the enzyme can be relatively low at acidic pH (Hearing and Ekel, 1976; Martinez et al., 1985). The melanosomal pH can also modify the stability and the evolution of intermediates of the melanogenic pathway, through changes in protonation of ionizable groups (Cánovas et al., 1982).
Estimation of the melanosomal pH shows that the organelle's lumen is acidic in mouse and human melanoma cells or normal melanocytes (Bhatnagar et al., 1993; Fuller et al., 2001; Puri et al., 2000), consistent with their relationship with the endosomal–lysosomal lineage. However, significant differences in the melanosomal pH have been reported for melanosomes from Black donors, which are more neutral, and the more acidic organelles derived from Caucasian skin (Fuller et al., 2001).
Given the complexity of the melanosomal milieu and the variety of components influencing the relative rates of the melanogenic reactions, the design of biomimetic protocols to synthesize model melanins is not an easy task. Even the use of animal tyrosinases which appears at first sight the simplest option to mimic human melanogenesis is severely limited by the absence of commercial sources of the enzyme. The purification of the mammalian enzyme is technically complex and requires significant amounts of starting tissue, most frequently melanomas grown in mice following subcutaneous injection of a suspension of melanoma cells. A variety of procedures have been described that most often rely on crude preparations of melanosomes obtained by differential centrifugation as the starting material (Aroca et al., 1990; Tomita et al., 1983), or even melanosomes purified by density gradient centrifugation (Kushimoto et al., 2001). Because tyrosinase is an integral membrane protein, the melanosomal preparation must be detergent-solubilized (Jiménez-Cervantes et al., 1995) or a soluble, enzymatically active fragment can be obtained by limited proteolysis (Valverde et al., 1992). The resulting extracts can be fractionated by combinations of techniques such as ammonium sulfate precipitation, gel filtration or ion-exchange chromatography, and preparative electrophoresis to obtain highly purified tyrosinase preparations (Garcia-Borron et al., 1985; Hearing et al., 1978; Ohkura et al., 1984; Tomita et al., 1983). However, the different purification protocols described to date are time-consuming, and their yields are relatively low. Another matter of concern is the possibility of subtle differences in the kinetic behavior of the tyrosinases from different mammalian species. For instance, it has been shown that whereas mouse tyrosinase is devoid of DHICA oxidase activity (Jiménez-Cervantes et al., 1994; Kobayashi et al., 1994), human tyrosinase is able to oxidize DHICA to 5,6-indolequinone-2-carboxylic acid, thus accelerating its incorporation into melanins (Olivares et al., 2001). Overall, these limitations exclude the general use of purified mammalian tyrosinases for the preparation of milligram amounts of enzymatically synthesized pigments.
A possible alternative would be the use of melanosomal extracts containing the complete set of melanogenic proteins. These extracts could be standardized by suitable tyrosinase activity measurements and dialyzed to remove low molecular weight thiolic compounds or metal ions that may interfere with the metabolism of melanogenic intermediates. Also, melanosomal extracts can be rather easily and selectively depleted of Tyrp2/Dct activity by a procedure that takes advantage of the much lower thermal stability of Tyrp2/Dct as compared with tyrosinase (Valverde et al., 1993). This technique might be useful to study the impact of dopachrome tautomerase activity in melanin structure.
Because the problems with mammalian pigmentary enzymes are yet to be resolved, simpler methods of enzymatic preparation of model melanins rely on commercial non-mammalian tyrosinases and either mono- or diphenolic melanogenic substrates. A highly active fungal tyrosinase isolated from Agaricus bisporus has been used extensively. The main advantages of this preparation are its unlimited availability at a reasonable cost and its high specific activity toward a variety of mono- or diphenolic substrates. Suitable resuspensions of the commercial powder can be used for the large-scale oxidation of melanogenic precursors under mild conditions, and an oxidizing environment potentially acting on melanogenic intermediates can thus be avoided. However, these commercial preparations should be used with care due to several complicating factors. The fungal enzymes have relevant kinetic differences compared with mammalian tyrosinases. They are a relatively crude, partially purified material containing significant amounts of other proteins, carbohydrates, and phenolic compounds. The presence of laccase, beta-glucosidase, beta-galactosidase, beta-xylosidase, cellulase, chitinase, xylanase, and mannanase can be demonstrated with suitable enzyme activity assays (Flurkey et al., 2008; Sugumaran and Bolton, 1998). Other protein contaminants were found by protein sequencing. Moreover, the ratios of enzymatic and non-enzymatic contaminants may be variable. Importantly, it has been demonstrated that those contaminants have an impact on the apparent kinetic behavior of tyrosinase (Neeley et al., 2009) and may even lead to the misinterpretation of experimental observations (Sugumaran and Bolton, 1998). In addition, enzymatic preparations from mushrooms or plant sources do not contain Tyrp1 and Tyrp2. These preparations are thus expected to mimic poorly the distal phase of melanogenesis beyond dopachrome and the metabolism of carboxylated indolic intermediates.
Use of melanins and related metabolites in biological research
There are conflicting reports on the biological role of melanin and its intermediates, because they have been shown to exert both toxic and antioxidant/scavenger activities. To deal with these compounds in biological research, it is therefore important to take account of different parameters, that is, the specific experimental conditions employed, the incubation time, the doses, and the way of administration of the compounds in relation to their solubility and stability.
Early studies demonstrated potent toxic effects of melanin intermediates on different cell types. Supplementation of tyrosine in the culture medium resulted selectively toxic to pigmented cells (Pawelek et al., 1973). Further reports extended this evaluation on DHI, which resulted highly cytotoxic on both melanoma cells and fibroblasts when used at 100 μM but with no effect when administered at 10 μM (Pawelek and Lerner, 1978). DHI appeared slightly more toxic than DHICA under experimental conditions employing both long time exposure (at concentrations of 10 and 100 μM up to 6 days) and short time exposure (1 mM for 1 h followed by the maintenance in regular medium). Estimation of the half-life of these melanin precursors in the culture medium by HPLC showed the shorter half-life of DHI (10–15 min) with respect to that of DHICA (30–40 min). The unstable nature of the intermediates in the culture medium leads to their spontaneous auto-oxidation and the following generation of H2O2 and other oxygen radical species, that is, superoxide radicals. The overall cytotoxic properties of these melanin intermediates used in cellular systems seem therefore attributable to the production of reactive oxygen species in the culture medium during their auto-oxidation rather than their intrinsic cytotoxicity (Urabe et al., 1994). The stability of DHICA up to 2 h in the culture medium was demonstrated when the compound was used at the concentration of 10−6 M or higher to stimulate murine macrophage activity. No effects were detected at lower or higher doses, which may be ascribed to faster auto-oxidation of DHICA when supplemented in culture at lower doses and to weak cytotoxic effects when applied close to millimolar doses (D'Acquisto et al., 1995). More recently, the protective/toxic effects of DHI have been analyzed on eye-related cells and no cytotoxicity was observed for the dose of 10 μM, while a significant toxic effect resulted at 100 μM (Heiduschka et al., 2007), in agreement with the earlier studies (Pawelek and Lerner, 1978).
Besides the effects under basal culture conditions, melanin precursors have been employed also in cells subjected to UV irradiation. Pre-incubation of the human retinal pigment epithelial ARPE-19 cell line with 10 μM DHI before UV-A irradiation resulted in a weak, although not significant, protective effect (Heiduschka et al., 2007). On the other hand, different results were obtained using DHICA on UV-A-irradiated human keratinocytes. Supplementation of low concentrations of DHICA (from 0.5 up to 2 μM) during UV-A exposure increased DNA damage as frank single-strand breaks (SSB) in a dose-dependent manner. Moreover, while there was no effect of the melanin precursor on SSB in the absence of UV-A irradiation when used at low doses (0.125–0.25 μM), the doses of 0.5 and 1 μM significantly increased SSB with respect to untreated cells (Kipp and Young, 1999). In our experience on primary cultures of human keratinocytes, the addition of DHICA (5–50 μM) to the culture medium for 24, 48, and 72 h decreases proliferation with no reduction in keratinocyte survival. Moreover, DHICA reduces lipoperoxidation and apoptosis following UV-A exposure. DHICA pre-incubation before UV-A exposure at 10 J/cm2 following a recovery of 24 h increases also cell viability when used at 50 μM. A difference to be mentioned is that in our experimental procedure DHICA was removed before irradiating keratinocytes (Kovacs et al., 2012).
5SCD, the main intermediate in the metabolic pathway of pheomelanin, showed selective toxicity to a variety of tumor cells when added to cell medium at a concentration of 1 mM for 1 h. 5SCD was found to be approximately 10 times more cytotoxic than l-dopa and the suggested mechanism of action involved hydrogen peroxide production (Fujita et al., 1980; Ito et al., 1983).
The melanin precursors 5SCD, DHI, and DHICA together with some related metabolites are diffusible and can be released both in body fluids and locally outside the producing melanocytes. Cutaneous levels of 5SCD, DHI, and DHICA have been judged to increase up to 200 μM under melanocyte-stimulating conditions such as UV irradiation or inflammatory processes (Koch and Chedekel, 1987). However, the precise in vitro reproduction of the local modulation of these melanin intermediates occurring under both physiological and pathological conditions is challenging.
Natural and synthetic melanins
One possible way to evaluate the role of melanin in preserving cell membrane lipids from UV-A-induced peroxidation is the selective modulation of melanin synthesis by tyrosine supplementation. In murine melanocytes with different melanization (e.g., black, brown), melanogenesis was stimulated by increasing tyrosine up to 200 μM. Although melanin is mainly present in melanocytes, it should be noted that it is also present in a diffuse state as ‘melanin dust’, extending its pro-oxidant or antioxidant action also to skin cells of epidermal and/or dermal layers (Haywood et al., 2006).
A direct evaluation of the possible contribution of natural or synthetic eumelanin or pheomelanin to the harmful effects of UV shows several critical aspects due to the complicated procedures for synthesis/isolation, purification, and solubilization of these pigments. Melanins should be suspended in water, stored under N2 in the dark, and diluted in culture medium at 10 times the desired concentrations. A number of studies have suggested that synthetic cysteinyldopa-melanin and dopa-melanin are suitable substitutes of natural pheomelanin and eumelanin. However, several aspects should be considered to evaluate photoprotective or photo-oxidative effects of synthetic melanins on cell cultures. Eumelanin-induced photoprotection, for example, is directly influenced by the effect of radiation absorption, its auto-oxidation tendency with the production of superoxide radicals, and its ability to act as radical scavenger.
It is unclear how these properties determine its behavior in vivo in response to physiological levels of radiation. Moreover, several experiments reporting higher phototoxic effects of pheomelanin compared to eumelanin were performed with high-energy ionizing radiation, and thus, the results cannot be extrapolated to UV-A and UV-B irradiation. An additional question is, ‘In what physical state, or states, do the melanins occur within the melanosomes and in the keratinocytes that form stratum granulosum and stratum corneum?’ It has been hypothesized that melanosomes may contain soluble melanin pigments, typically associated with melanoproteins, whereas it is more likely that eumelanin and pheomelanin may form a melanin dust within keratinocytes. Thus, to evaluate the biological action of intramelanosomal melanins, soluble forms of eumelanin or pheomelanin should be synthesized in the presence of bovine serum albumin.
To evaluate the possible correlation between photoprotective and/or photo-oxidative effects of cysteinyldopa-melanin or dopa-melanin and endogenous antioxidant defense in mediating UV-A and UV-B cell damage, an experimental model was set up including irradiation of human keratinocytes in the presence of synthetic pheomelanin or eumelanin homologues (Briganti et al., 2005). The pigments were suspended in PBS and diluted in cell culture medium (15–100 μg/ml). After testing cytotoxicity, 30 μg/ml was chosen as a safe concentration able to reproduce in vitro the amount of melanins dispersed in keratinocytes. Three experimental settings were considered: (i) short (1 h) or overnight pre-incubation of cells with synthetic melanins and their presence during UV-A irradiation, (ii) cell exposure to UV-A with cysteinyldopa- or dopa-eumelanin diluted in PBS, and (iii) employing Nunc OptiCell Culture Systems with pigment dissolved in PBS and maintained separated from keratinocytes by a gas-permeable polystyrene membrane. These methodologies allowed to compare the biological effect of synthetic pheomelanin or eumelanin as solution and ‘dust’ dispersed in PBS during irradiation to the pro-oxidant action of reactive oxygen species generated by pigment's UV-A exposure (in particular singlet oxygen) and diffused to cell cultures through the gas-permeable membrane. Moreover, to distinguish the pro-oxidant or photoprotective effects of soluble melanins from those induced by pigments in a dispersed physical state, cysteinyldopa- or dopa-melanin was filtered through a 0.22-μm filter before treating the cells. After UV-A exposure, the main parameters examined include: (i) gap junction intercellular communication (GJIC), as index of cell integrity; (ii) cell membrane lipids, as oxidation targets; (iii) activity of antioxidant enzymes (catalase and glutathione peroxidase [GSH-Px]); and amount of small molecular antioxidant molecules (GSH and vitamin E), as index of endogenous antioxidant system effectiveness. Synthetic pheomelanin, acting as a photosensitizer, was found to significantly increase UV-A-induced GJIC inhibition, antioxidant system imbalance, and damage of cell membrane integrity.
To summarize, use of melanin intermediates and synthetic melanins on cell systems requires that several experimental parameters are taken into account, including:
The specific sensitivity of the cell type utilized.
The light and temperature sensitivity along with the stability of the compounds (i.e., DHICA is more stable than DHI and leads therefore to a slower generation of radical species in the culture medium; DHI is sensitive to temperature increases (Maeda and Hatao, 2004; Heiduschka et al., 2007); enzymatically prepared dopa-melanin as well as cysteinyldopa-melanin must either be used immediately or stored below 0°C. Moreover, thawed samples should not be studied after few hours at room temperature).
Possible reactivity in culture media during UV exposure (i.e., the quick darkening of the medium when DHI, DHICA, and related metabolites are present during UV-A exposure due to the irreversible generation of brown pigments following their oxidation. Consequently, keeping the compounds in the irradiation medium may both increase the toxicity due to the generation of radical species and decrease the intensity of UV light reaching the cells (Maeda and Hatao, 2004; Heiduschka et al., 2007); similar cautions have to be considered when using cysteinyldopa-melanin to avoid photo-oxidation).
Conditions of the reaction mixture of synthetic melanins.
Possible formation of aggregates when handling synthetic pheomelanin or eumelanin.
Outlook and perspectives
The present paper includes different sets of recommended standards and procedures that have been selected on the basis of a critical review of the literature and experimental revision of selected methods. The main principles and concepts underlying proposed methods have been illustrated; detailed protocols and spectral data of standard precursors and markers have been provided in Appendix S1. Commercial products widely used by pigment cell researchers have also been analyzed and discussed. It is strongly recommended that the proposed methods are agreed upon, and rapidly adopted by, all laboratories working on melanogenesis and pigment cell biology. This should eventually lead to an unprecedented integration and synergism between research groups from different disciplines and with various backgrounds and expertise in the field.
This paper was carried out in the frame of the EuMelaNet Special Interest Group and was possible thanks to the active support of the European Society for Pigment Cell Research, the International Federation of Pigment Cell Societies, and their officers. The contents of the paper reflect in part presentations and discussions during Concurrent Session 5 at International Pigment Cell Conference (IPCC 2011), Bordeaux, September 20-24, 2011 (see d'Ischia et al. Pigment Cell Melanoma Res. 2011, 24, 782-783). The authors are grateful to Lionel Larue, Alain Taieb, Lluis Montoliu, and Ghanem Ghanem for encouragement and helpful discussions. MdI acknowledges financial support by Italian MIUR, PRIN 2010-2011 (PROxi) project, for the experimental protocols executed and verified at Naples University. TS thanks Andrzej Zadlo for running EPR spectra of melanin samples.