Melanins can be classified into two major groups—insoluble brown to black pigments termed eumelanin and alkali-soluble yellow to reddish-brown pigments termed pheomelanin. Both types of pigment derive from the common precursor dopaquinone (ortho-quinone of 3,4-dihydroxyphenylalanine) which is formed via the oxidation of l-tyrosine by the melanogenic enzyme tyrosinase. Dopaquinone is a highly reactive ortho-quinone that plays pivotal roles in the chemical control of melanogenesis. In the absence of sulfhydryl compounds, dopaquinone undergoes intramolecular cyclization to form cyclodopa, which is then rapidly oxidized by a redox reaction with dopaquinone to give dopachrome (and dopa). Dopachrome then gradually and spontaneously rearranges to form 5,6-dihydroxyindole and to a lesser extent 5,6-dihydroxyindole-2-carboxylic acid, the ratio of which is determined by a distinct melanogenic enzyme termed dopachrome tautomerase (tyrosinase-related protein-2). Oxidation and subsequent polymerization of these dihydroxyindoles leads to the production of eumelanin. However, when cysteine is present, this process gives rise preferentially to the production of cysteinyldopa isomers. Cysteinyldopas are subsequently oxidized through redox reaction with dopaquinone to form cysteinyldopaquinones that eventually lead to the production of pheomelanin. Pulse radiolysis studies of early stages of melanogenesis (involving dopaquinone and cysteine) indicate that mixed melanogenesis proceeds in three distinct stages—the initial production of cysteinyldopas, followed by their oxidation to produce pheomelanin, followed finally by the production of eumelanin. Based on these data, a casing model of mixed melanogenesis is proposed in which a preformed pheomelanic core is covered by a eumelanic surface.
Pigmentation of hair, skin and eyes in animals mainly depends on the quantity, quality and distribution of melanin. Melanocytes are responsible for the synthesis of melanin within membrane-bound organelles termed melanosomes, and the subsequent transfer of those melanosomes to surrounding epidermal cells, termed keratinocytes. Melanocytes in mammals and birds produce two chemically distinct types of melanin, black to brown eumelanin and yellow to reddish-brown pheomelanin (1–6). Interest in these two types of melanin pigments has been increasing in recent years, mostly because of their apparently opposite responses to UV radiation. Thus, it is generally accepted that eumelanin acts as a photoprotective antioxidant while pheomelanin is a phototoxic pro-oxidant (7–10).
However, in reality most natural melanin pigments consist of both eumelanin and pheomelanin in varying ratios (11–13) and thus should be considered as “mixed melanin.” In this review, we summarize recent advances in the chemistry of “mixed melanogenesis” with special emphasis on the pivotal roles of dopaquinone in chemically controlling the course of melanogenesis. We use the term “mixed melanogenesis” to make it clear that most, if not all, of melanogenesis leads to the production of copolymers (or mixtures) of eumelanin and pheomelanin. Polymerization of melanin monomers and degradation of melanin pigments are not covered in this review. Further information on those topics can be found in Ito and Wakamatsu (6), Prota (1) and in several reviews (2,3,5).
Chemical studies on mixed melanogenesis have been accelerated by the introduction of the pulse radiolysis method, which has been a powerful tool to follow the rapid reactions promoted by dopaquinone (ortho-quinone of 3,4-dihydroxyphenylalanine) and other related ortho-quinones (14–18). It should be added that, in order to facilitate studies on mixed melanogenesis, we independently developed quantitative methods to analyze eumelanin and pheomelanin (19,20). Those methods are based on the production of specific degradation products—pyrrole-2,3,5-tricarboxylic acid produced by the permanganate oxidation of eumelanin and 4-amino-3-hydroxyphenylalanine produced by the hydriodic acid hydrolysis of pheomelanin (21,22).
Current Concepts of Melanogenesis
Structure of eumelanin and pheomelanin
Extensive early studies by Nicolaus’ group in Naples (23) and by Swan’s group in Newcastle (24,25) suggested that eumelanin is a very heterogeneous polymer consisting of different oxidative states of 5,6-dihydroxyindole (DHI) and 5,6-dihydroxyindole-2-carboxylic acid (DHICA) units, and pyrrole units derived from their peroxidative cleavage (Fig. 1). Interestingly, a more recent study reported a high proportion of pyrrole units in Sepia melanin (26), although how much those pyrrole units contribute to the structure of natural eumelanin in mammals remains to be determined.
Most current knowledge of pheomelanin chemistry results from studies in the late 1960s conducted by Prota, Nicolaus and collaborators (27,28). Those early results, combined with more recent findings, led us to propose a representative structure (Fig. 1) for pheomelanin that consists mostly of benzothiazine units with minor contributions from benzothiazole and isoquinoline units. Some of these monomer units may be connected by ether bonds (29,30). However, the contribution of ether bonds should be minimal as monomer units connected by ether bonds should be colorless, which is not consistent with the yellow-red color of natural pheomelanin. The isoquinoline units in the structure may result from postpolymerization modifications of freshly formed pheomelanin as isoquinoline units are not produced during earlier stages of pheomelanogenesis. Modification of alanyl side chains to an aromatic group is consistent with biosynthetic studies using radioisotopes and 13C-NMR spectroscopy (31,32).
Another class of pheomelanic pigments, termed trichochromes, is also produced in vivo. Trichochromes are the only melanin-like pigments with structures that have been fully characterized as having a bi(1,4-benzothiazine) chromophore (1,6,27,28). The close similarity in structural features of trichochromes and pheomelanin, and their coexistence in pigmented tissues, suggests that they are formed oxidatively from the same monomer units and differ only in their mode of polymerization. Therefore, trichochromes are regarded as a variety of pheomelanic pigment.
Prota and his successors have continued biosynthetic studies to clarify the melanogenic intermediates beyond DHI and DHICA that are involved in eumelanogenesis and those produced after the formation of cysteinyldopas involved during pheomelanogenesis. Although considerable advances have been made (3,29,30,33), they do not change our basic views about the structures of eumelanin and pheomelanin, as depicted in Fig. 1.
The Raper–Mason–Prota pathway of melanogenesis
Eumelanin and pheomelanin both are derived from the common precursor dopaquinone formed by oxidation of the common amino acid l-tyrosine by tyrosinase (Fig. 2). It had been generally believed that 3,4-dihydroxyphenylalanine (dopa) is formed first on the way to dopaquinone. However, Cooksey et al. (34) showed that ortho-quinones, such as dopaquinone, are formed directly during the initial stage of melanogenesis. Dopaquinone is highly reactive and in the absence of sulfhydryl compounds it undergoes the intramolecular addition of the amino group to produce cyclodopa (also termed leucodopachrome). A redox exchange between cyclodopa and dopaquinone then gives rise to dopachrome, an orange-red intermediate (35,36) and dopa. This latter reaction is considered the source of dopa formed during melanogenesis (Fig. 2). Dopachrome then gradually rearranges to generate mostly DHI and to a lesser extent DHICA (35,37). Those two dihydroxyindoles are then further oxidized and polymerized to produce eumelanin.
Regarding the production of pheomelanin, the intervention of sulfhydryl compounds (such as cysteine) gives rise exclusively to thiol adducts of dopa (cysteinyldopas) among which 5-S-cysteinyldopa (5-S-CD) is the major isomer (38). Further oxidation of the thiol adducts leads to the formation of pheomelanin via benzothiazine intermediates.
In addition to tyrosinase, two tyrosinase-related proteins have been shown to modulate eumelanogenesis (39,40). Dopachrome tautomerase (Dct), suggested some years earlier (41), catalyzes the tautomerization of dopachrome to DHICA (42) and that catalytic activity is associated with tyrosinase-related protein-2 (Tyrp2) (43). Certain divalent metal ions can also promote tautomerization (i.e. isomerization with a shift of hydrogen atom) of dopachrome to DHICA (37,44). The oxidative polymerization of DHI can be catalyzed by mammalian tyrosinase (45). A recent pulse radiolysis study, however, indicates that DHI can be effectively oxidized by dopaquinone (18); this result is incorporated in Fig. 2. In mice, another tyrosinase-related protein, Tyrp1, can oxidize DHICA (46,47), although human TYRP1 is unable to catalyze the same reaction (48). Instead, human tyrosinase is able to oxidize DHICA, as well as tyrosine, dopa and DHI. The activities of these tyrosinase-related proteins greatly affect the quantity and quality (the ratio of DHI to DHICA and the degree of polymerization) of eumelanins produced.
The intrinsic reactivity of ortho-quinones
Dopaquinone is an ortho-quinone and the chemistry of ortho-quinones is closely involved with melanogenesis. Figure 3 summarizes some important reactions of ortho-quinones, which are extremely reactive compounds. Tse et al. (49) showed that the addition of sulfhydryl compounds proceeds very quickly to generate thiol adducts. Pulse radiolysis studies on the reactivities of 4-substituted ortho-quinones with cysteine and glutathione (50) showed that the rate constants of thiol addition range from 4 × 105 to 3 × 107m−1 s−1 (in the case of cysteine at neutral pH) depending on the nature of the substituents. Reduction to the parent catechols through redox exchange proceeds as quickly as the thiol addition (49), and therefore, these two reactions are competitive. Reactions with amine compounds do not proceed as quickly. However, in cases when the amino group is present within the same molecule, the amino group rapidly undergoes either an addition reaction to give an aminochrome (such as dopachrome) or a condensation reaction to give an ortho-quinonimine. The latter type of reaction occurs during the cyclization of cysteinyldopaquinones (see Fig. 2). It should be stressed that all of these reactions are controlled by the intrinsic chemical reactivity of ortho-quinones.
Pivotal roles of dopaquinone in controlling mixed melanogenesis
Rate constants (r1–r4) for all four important steps in the early phase of melanogenesis were reported based on pulse radiolysis studies (Fig. 4) (15–17). Pulse radiolysis is a powerful tool to study the fates of highly reactive melanin precursors. The technique depends on the production of dibromide radical anions Br2− due to pulse radiolysis of an N2O-saturated aqueous buffer containing KBr. The dibromide radical anions thus formed oxidize dopa to dopasemiquinone, which then disproportionates to generate dopaquinone and dopa. This entire process proceeds within 2–3 ms so that the fate of dopaquinone can be followed by spectrophotometry in the presence (or absence) of a targeted molecule. Some cautions should be drawn to the limitations that dopaquinone generated by pulse radiolysis might behave differently from that generated by tyrosinase and that pulse radiolysis technique is based on the generation of exceedingly low levels of the oxidized species, e.g. dopaquinone, in the presence of large amounts of unreacted substrate, e.g. dopa.
The first step (r1 = 3.8 s−1) in eumelanogenesis is a fairly slow step involving the intramolecular addition of the amino group to produce cyclodopa (15). However, as cyclodopa is formed, it is rapidly oxidized to dopachrome through a redox exchange (r2 = 5.3 × 106m−1 s−1) (17). In contrast, the first step in pheomelanogenesis (r3 = 3 × 107m−1 s−1) is the addition of cysteine, which proceeds very quickly (14). The second step in pheomelanogenesis is the redox exchange which generates cysteinyldopaquinone, and which proceeds at a slower rate (r4 = 8.8 × 105m−1 s−1) (15). From these kinetic data, several important conclusions can be drawn, as follows:
1 Comparing the rate constants for the addition of cysteine to dopaquinone (r3) and for the intramolecular cyclization (r1) reveals that cysteinyldopa formation is preferred over cyclodopa formation as long as the cysteine concentration is above 0.13 μm. This value may be, however, pH-dependent because the rate of dopaquinone cyclization (r1) varies with pH, becoming 100-fold slower when the pH is reduced from 8.6 to 5.6 (14).
2 The redox exchange giving cysteinyldopaquinone (r4) proceeds 30 times slower than the addition of cysteine (r3). Thus, cysteinyldopas accumulate during the early phase of pheomelanogenesis.
3Comparing the rate constants for the redox exchange generating dopachrome from dopaquinone (r2) and for the intramolecular cyclization (r1) reveals that the production of dopachrome is faster than the production of cyclodopa when the cyclodopa concentration is above 0.7 μm. Thus cyclodopa does not accumulate during the early phase of eumelanogenesis.
4Comparing the rate constants for the redox exchange producing cysteinyldopaquinone (r4) and for the formation of dopachrome (2 × r1; ) reveals that pheomelanogenesis occurs preferentially over eumelanogenesis as long as the cysteinyldopa concentration is above 9 μm.
5An “Index of Divergence” (D) between eumelanogenesis and pheomelanogenesis can now be derived. By taking dopachrome and cysteinyldopaquinone as representatives of those two divergent pathways, Land et al. (17) proposed that:
This suggests a “crossover value” (i.e. for D =1) for switching between eumelanogenesis and pheomelanogenesis when the cysteine concentration reaches 0.8 μm.
The considerations above are useful in interpreting the early phase of pheomelanogenesis. Tyrosinase oxidation of dopa in the presence of excess cysteine generates a high yield of 5-S-CD (74%) and 2-S-CD (14%) together with minor amounts of the 6-S-CD (1%) and a di-adduct, 2,5-S,S′-dicysteinyldopa (5%) (38). The high yield of mono-adducts of cysteine confirms the hypothesis that mono-cysteinyldopa formation is preferred as long as cysteine is present. It should also be pointed out that the ratio of these cysteinyldopa isomers is determined by the intrinsic chemical reactivity of dopaquinone (and cysteine). An interesting aspect of the cysteine addition is the marked preference of the addition that occurs at the C5 and C2 positions (adjacent to the carbonyl group) over the C6 position (distal to the carbonyl group). A recent study suggests the participation of the amino group (in the protonated form) in cysteine which leads to this preference (51). Other thiols such as 2-thiouracil preferentially add to the C6 position of dopaquinone (52).
The ratio of r3 to r4 is ca 30, thus the formation of 5-S-cysteinyldopaquinone becomes predominant only after the concentration of cysteine decreases to 30 times lower than that of 5-S-CD. The facts that mono-cysteinyldopas accumulate in the early phase of pheomelanogenesis and that the formation of the diadduct 2,5-S,S′-dicysteinyldopa is only a minor pathway support this interpretation. This explains why high levels of 5-S-CD are produced in melanoma tissues and are secreted into the blood of melanoma patients, making 5-S-CD a useful biochemical marker of melanoma progression (53).
The pheomelanogenesis pathway proposed was supported by the identification of 5-S-CD (54) and other isomers, along with the diadduct, in the urine of melanoma patients (55). The ratios between various isomers of cysteinyldopa found in melanoma urine and tissues (56) are similar to those obtained by incubation of dopa with tyrosinase in vitro in the presence of cysteine (38), which suggests that cysteinyldopas originate in vivo by a similar route involving the addition of cysteine to dopaquinone. As this process is the initial event during pigment synthesis, it is likely that some cysteinyldopas formed are secreted into body fluids, regardless of the type of melanin that is eventually formed. A recent study goes even further to suggest that the secretion of high levels of 5-S-CD (and a eumelanin-related metabolite, 6-hydroxy-5-methoxyindole-2-carboxylic acid) may indicate a deficiency in incorporating melanin precursors into nascent melanins within melanosomes (57).
The proposed pathway for mixed melanogenesis
The studies discussed above suggest a three-step pathway for mixed melanogenesis, as shown in Fig. 5 (4,5). As indicated in that scheme, the amount of melanin produced is proportional to dopaquinone production, which is in turn proportional to tyrosinase activity. In fact, melanogenesis proceeds in three distinct stages. The initial stage is the production of cysteinyldopas, which continues as long as the cysteine concentration is above 0.13 μm. The second stage is the oxidation of cysteinyldopas to produce pheomelanin, which continues as long as cysteinyldopas are present at concentrations above 9 μm. The last stage is the production of eumelanin, which begins only after most cysteinyldopas and cysteine are depleted. Therefore, the ratio of eumelanin to pheomelanin is determined by tyrosinase activity and the availability of tyrosine and cysteine in melanosomes (17).
The time course of mixed melanogenesis is schematically represented by the casing model (Fig. 6), which was originally suggested by Agrup et al. (58) based on biochemical findings. Recently, Bush et al. (59) provided direct, biophysical evidence on neuromelanin granules that supports this model (60). That study employed photoelectron emission microscopy coupled to a free-electron laser to demonstrate that neuromelanin is composed of granules with ca 30 nm diameters consisting of pheomelanin at the core and eumelanin at the surface.
Although direct proof for the casing model has so far been only for neuromelanin, there is some evidence that human epidermal and uveal melanocytes in culture produce pheomelanin at rather constant levels regardless of the degree of pigmentation while they produce eumelanin at levels proportional to pigmentation, as shown in Fig. 7 (61,62). If this casing model holds true for melanin in skin and eye tissues, it should be envisaged that melanin granules produced in dark-colored melanocytes (from the skin or eye) may be covered with eumelanin at the surface while those in light-colored melanocytes may have an exposed pheomelanic core at the surface due to the lower ratio of eumelanin to pheomelanin. Thus, even if the pheomelanin content does not differ significantly between light-colored and dark-colored melanocytes, melanin in light-colored melanocytes, but not in dark-colored melanocytes, would be expected to behave as pheomelanin, i.e. a pro-oxidant. This, coupled with the lower content of melanin, could be the reason why skin cancers (such as melanoma) and eye diseases (such as uveal melanoma and age-related macular degeneration) are more prevalent in light-colored subjects (63–65).
Photoprotective roles of eumelanin, as opposed to phototoxic roles of pheomelanin, have drawn great attention from a diverse field of researchers (for a recent review, see ref. 66). As an example, D’Orazio et al. (67) have shown that chemically induced tanning with eumelanin is protective against UV-induced DNA damage and tumorigenesis in mice. Many studies have compared the chemical reactivity of eumelanin and pheomelanin and found that both melanins act as a free-radical scavenger and inhibit UV-induced liposomal lipid peroxidation. However, pheomelanin, when complexed with Fe3+, stimulates UV-induced lipid peroxidation, whereas eumelanin does not (68). The antioxidant properties of melanin are related to the type of melanin; the greater the ratio of eumelanin to pheomelanin, the more antioxidant the pigment (7,69). Cultured melanocytes with high level of melanin, especially eumelanin, show a better survival after UVB irradiation (8). UV irradiation of melanin also generates reactive oxygen species and this photosenstization is greater for pheomelanin than for eumelanin (9). Furthermore, pheomelanosomes are more pro-oxidant than eumelanosomes (10). Also, synthetic pheomelanin produces superoxide anion even in the absence of UV light (70). Thus, due to the significance of the casing model for mixed melanogenesis in understanding photoprotective/toxic properties of melanin, further biochemical and biophysical studies to confirm this model in pigmentation of melanocytes are warranted.
Biochemical evidence for the proposed scheme of mixed melanogenesis
Some evidence supporting the three-step pathway (and the casing model) of mixed melanogenesis can be summarized as follows.
Ozeki et al. (71) examined the oxidation of tyrosine by tyrosinase in the presence of cysteine and showed that the three steps proceed in sequence as expected (Fig. 8). Thus, tyrosine levels decreased gradually over the reaction time while cysteine was consumed much faster (because of its partial oxidation to cystine). At 1 h, the production of cysteinyldopas reached a maximum, after which cysteinyldopas were oxidized to give pheomelanin by 2 h. After that time, eumelanin was deposited on the preformed pheomelanin, as evidenced by an increase in total melanin.
The effects of varying the ratio of tyrosine to cysteine have been examined by several groups. For example, by decreasing the extracellular concentration of cystine in cultures of human melanoma cells, del Marmol et al. (72) demonstrated a shift to more eumelanic cells resulting from the dramatic decrease in intracellular cysteine concentration. By changing the concentrations of tyrosine and cystine in cultures of human melanocytes, Smit et al. (73) found a two-fold increase in melanin production with a decreased ratio of pheomelanin to total melanin when cells were cultured at a higher concentration of tyrosine.
The proposal that tyrosinase activity plays a major role in controlling mixed melanogenesis is also supported by several studies. Burchill et al. (74) examined the effects of α-melanocyte-stimulating hormone (α-MSH) on mixed melanogenesis in viable yellow mice. This mutation at the agouti locus is unique in that neonatal and adult mice produce almost pure pheomelanin while pubertal mice produce a mixed-type melanin. When pubertal mice were injected with α-MSH, tyrosinase activity increased 1.7-fold and more eumelanic hair was produced with a concomitant increase in total melanin. When these pubertal mice were injected with bromocriptine (which reduces α-MSH secretion), tyrosinase activity was reduced to 8% of the untreated control and pheomelanic hair was produced along with a decrease in total melanin. The injection of pheomelanic neonatal mice with α-MSH resulted in an increase in tyrosinase activity by 2.2-fold with a switch to more eumelanic hair (75). These results illustrate the important role of tyrosinase activity to control mixed melanogenesis in vivo.
The significance of tyrosinase activity in mixed melanogenesis was also studied in vitro. Hunt et al. (76) examined the effects of a superpotent synthetic α-MSH analog on melanogenesis in human melanocytes. Treatment of melanocytes with the synthetic α-MSH resulted in an increase in eumelanin content in all six cell lines examined from different ethnic origins. On the other hand, pheomelanin showed variable responses, and this led to a clear, significant shift to more eumelanic cells.
However, a change in tyrosinase activity itself is not sufficient for the switch of melanogenesis as seen in chinchilla mice where tyrosinase activity is decreased to one-third that of wild-type mice (77,78). The eumelanin content is reduced by 40% in black chinchilla mice without any increase in pheomelanin content compared with black mice, whereas the pheomelanin content is reduced nine-fold in lethal yellow chinchilla mice compared with lethal yellow mice (78). Similar results were obtained by analyzing the agouti pattern of baboon hairs (79). Dark-colored baboons produced high levels of pheomelanin in the yellow bands of their agouti hairs, whereas the corresponding bands in light-colored baboons contained little melanin. The scheme summarized in Fig. 5 fits well to these results in that pheomelanogenesis is more strongly affected by a decrease in tyrosinase activity than is eumelanogenesis. Barsh (80) has put forth a similar proposal.
Until very recently, it remained unsolved how the availability of cysteine is controlled in melanosomes, as pointed out by Ito (5). A cysteine (or cystine) transporter that affects pheomelanogenesis had been sought for years, and was finally identified (81). That study demonstrated that the subtle gray (sut) pigment mutation in mice arose due to a mutation in the Slc7a11 gene which encodes the plasma membrane cystine/glutamate exchanger xCT. The resulting low rate of extracellular cystine transport into sut melanocytes reduces pheomelanin production with minimal or no effect on eumelanin production. In fact, the effect of the Slc7a11 sut mutation on pheomelanin production was markedly accentuated on the Ay/a background, reducing pheomelanin levels in hair to one-sixth of the control level, with a four-fold increase in the low level of eumelanin.
Finally, the significance of pH in controlling mixed melanogenesis was addressed by Ancans et al. (82). By neutralizing intramelanosomal pH in cultured human melanocytes and melanoma cells, they found that cells with increased tyrosinase activity after neutralization had a preferential increase in the production of eumelanin, which resulted in an increase in the eumelanin to pheomelanin ratio. The effects of more acidic pH on mixed melanogenesis in melanosomes are two-fold—a lower activity of tyrosinase and a slower rate of dopaquinone cyclization (r1). Both of those effects favor pheomelanogenesis in acidic melanosomes.
Late stages of eumelanogenesis
Dopachrome accumulates during early stages of eumelanogenesis, and stages after the formation of dopachrome are discussed in this section, mostly with respect to the additional roles of dopaquinone. The orange-red pigment dopachrome is fairly stable having a half-life of about 30 min (first-order rate constant of 4.0 × 10−4 s−1). It spontaneously decomposes to give mostly DHI by decarboxylation at neutral pH values in the absence of Dct or metal cations. The ratio of DHI to DHICA produced under these conditions is 70:1 (37). However, in the presence of Dct (Tyrp2), dopachrome undergoes tautomerization to preferentially produce DHICA (44). The ratio of DHICA to DHI in melanins is thus determined by the activity of Dct. Metal cations, especially Cu2+, accelerate the dopachrome rearrangement and also affect the DHICA/DHI ratio, but Dct seems to be more effective in catalyzing the tautomerization (37,44).
As discussed above, the pulse radiolysis technique is very useful in following the fate of short-lived ortho-quinone intermediates. Using that technique, Lambert et al. (83) proposed that 5,6-indolequinone tautomerizes to its quinone-imine and quinone-methide tautomers which then undergo the nucleophilic addition of water, thus giving trihydroxyindole species.
Recently, the redox exchange reaction between DHI and dopaquinone has been studied (18) (Fig. 9). Redox exchange with DHI does proceed but not quite to completion with a rate constant of r5 = 1.4 × 106m−1 s−1. This is in the same range as the rate constants for reactions with cyclodopa (r2) and with 5-S-CD (r4). The reaction with DHICA more obviously does not go to completion and is much slower (r6 = 1.6 × 105m−1 s−1). It was concluded that, during eumelanogenesis, DHI oxidation takes place by redox exchange with dopaquinone, although such a reaction is likely to be less efficient for DHICA. Thus, DHICA may require its oxidation to the quinone form by tyrosinase in humans (84) or by Tyrp1 in mice (46,47).
Extensive biosynthetic studies have been performed to clarify the stages beyond 5,6-indolequinones. Although beyond the scope of this review, recent studies include the first isolation of a tetramer of DHI by oxidation of its 2,4′-dimer (33).
Late stages of pheomelanogenesis
As discussed above, the early stages of pheomelanogenesis up to the formation of cysteinyldopaquinones are well characterized. In the later stages of pheomelanogenesis, once 5-S-cysteinyldopaquinone is formed, it then rapidly cyclizes via attack of the cysteinyl side chain amino group on the carbonyl group to produce a cyclic ortho-quinonimine intermediate (Fig. 10; ). The rate (r7) of quinonimine formation was determined by pulse radiolysis to be 10 s−1 (18,86). Again, it should be stressed that the rate of 5-S-cysteinyldopaquinone formation is controlled by the rate of dopaquinone formation as 5-S-CD is a much poorer substrate for tyrosinase than is dopa (58).
An alternative pathway for the metabolism of 5-S-cysteinyldopaquinone is possible; this ortho-quinone also undergoes the addition of cysteine with a rate constant of 1 × 104m−1 s−1 to give the diadduct 2,5-S,S′-dicysteinyldopa (14). Thus, unless the cysteine concentration is above 1 mm (a rather high concentration for in vivo), the quinonimine formation predominates.
The ortho-quinonimine then undergoes rearrangement to benzothiazine intermediate(s) with (85%) and without (15%) decarboxylation (85). The rate (r8) of decay (k =6.0 s−1) of cyclic ortho-quinonimine to benzothiazine was determined by pulse radiolysis (86). Thus, in late stages of pheomelanogenesis, the benzothiazines are produced rather rapidly from cysteinyldopas. Benzothiazines are also unstable and decay over a few seconds with a rate constant of 0.5 s−1 (86).
As shown in Fig. 10, an alternative pathway for the generation of ortho-quinonimine is also possible by redox exchange with 5-S-CD which leads to the production of a reduced form of the quinonimine (a 3,4-dihydro-1,4-benzothiazine derivative) and 5-S-cysteinyldopaquinone (87). Whether the redox exchange or the rearrangement prevails is strongly influenced by many factors, including the nature of the oxidant, the presence of metal ions, and the concentration of the precursor 5-S-CD (88,89). It remains to be seen therefore whether this redox change is a significant pathway in vivo.
Reactions beyond the benzothiazines which lead to pheomelanin are rather complex, but nevertheless have also been extensively studied by Prota and his associates (88,89). Some of those results have been reviewed in Di Donato and Napolitano (30).
Two types of melanin production, eumelanogenesis and pheomelanogenesis, have been identified and characterized. Most of those pathways at the monomer level have been clarified using biosynthetic and pulse radiolysis approaches. Both techniques have provided valuable information and have complemented each other. The former method has the strength of isolating various monomeric and oligomeric melanin intermediates, while the latter method is able to follow rapid reactions involving dopaquinone that could not be studied otherwise.
Some unsolved problems in the chemistry of mixed melanogenesis include: (1) the nature of postpolymerization modifications of eumelanin and pheomelanin, (2) the nature of copolymerization of eumelanin and pheomelanin. The first of these problems has been addressed only sporadically (32,71,90), and is also relevant to clarifying the biodegradation of melanin and melanosomes (91,92). An interesting observation related to this is the marked difference in color and intensity among natural pheomelanic pigments, such as seen in red hair color in humans and yellow hair color in mice (93). Regarding the second problem, no study to date has examined in depth the mode of copolymerization of the two types of melanin pigments.
The biologic functions of melanin pigments are closely related to their structural features. In this review, we have shown how the process of mixed melanogenesis can be interpreted in terms of chemistry. However, several important issues remain to be solved through multidisciplinary approaches. One of those important unsolved problems is the precise biochemical mechanism of switching between eumelanogenesis and pheomelanogenesis in vivo, which appears to depend not only on tyrosinase activity but also on the availability of cysteine (or cystine) in melanosomes. In this connection, the role of the recently identified cystine transporter Slc7a11 (81) in regulating mixed melanogenesis deserves more attention.
Another unsolved problem includes the role of pH in regulating mixed melanogenesis. This aspect of melanogenesis is now receiving much attention because it was recently found that melanosomes in melanocytes from White/fair skin are acidic while those from Black/dark skin are near neutral (94). Furthermore, the great diversity in normal human skin pigmentation appears to stem from mutations in only several genes, including P, MATP and SLC24A5, the latter being identified only recently (95–97). Available evidence suggests that mutations in those genes may result in changing the acidic environment of melanosomes (96). Nevertheless, only one study has so far been published that examined the effects of pH on mixed melanogenesis in cultured melanocytes, as mentioned earlier (82). Most pulse radiolysis studies have been carried out only at neutral pH, except for a study by Thompson et al. (14). Thus, the effect of pH on the rate of the addition reaction of cysteine to dopaquinone (r3) is not known.
Finally, neuromelanin, a dark pigment present in the substantia nigra and some other regions of the brain, was found to consist of dopamine with some incorporation of cysteinyldopamine (98,99). Thus, during neuromelanin formation, which is another example of mixed melanogenesis, the chemistry of dopaminequinone appears to play an even more important role compared with dopaquinone because of the slower cyclization of the former. The rate of cyclization of dopaminequinone to cyclodopamine is reported to be 0.37 s−1 (51,100), which is 10-fold slower than the corresponding reaction of dopaquinone. Thus, mixed melanogenesis in neuromelanin formation also needs to be studied further.
Acknowledgement— This work was partially supported by grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (18591262).