Chemical and biochemical control of skin pigmentation with special emphasis on mixed melanogenesis

Abstract

Melanins are widely distributed in animals and plants; in vertebrates, most melanins are present on the body surface. The diversity of pigmentation in vertebrates is mainly attributed to the quantity and ratio of eumelanin and pheomelanin synthesis. Most natural melanin pigments in animals consist of both eumelanin and pheomelanin in varying ratios, and thus, their combined synthesis is called “mixed melanogenesis.” Gene expression is an established mechanism for controlling melanin synthesis; however, there are multiple factors that affect melanin synthesis besides gene expression. Due to the differential sensitivity of the eumelanin and pheomelanin synthetic pathways to pH, melanosomal pH likely plays a major role in mixed melanogenesis. Here, we focused on various factors affecting mixed melanogenesis including (1) chemical regulation of melanin synthesis, (2) melanosomal pH regulation during normal melanogenesis and effect on mixed melanogenesis, and (3) mechanisms of melanosomal pH control (proton pumps, channels, transporters, and signaling pathways).

1 ∣. INTRODUCTION

Melanin pigments, widely distributed in vertebrates, are composed of insoluble brown to black pigments termed eumelanin (EM) and alkali-soluble yellow to reddish-brown pigments termed pheomelanin (PM) (Ito & Wakamatsu, 2003, 2008). Melanin pigments in vertebrates are produced in melanocytes within membrane-bound organelles termed melanosomes; thereafter, the melanosomes in the hair follicle and epidermal melanocytes are transferred to the surrounding keratinocytes leading to a diverse range of hair and skin colors (Del Bino et al., 2015; Delevoye, 2014; Lin & Fisher, 2007; Ohbayashi & Fukuda, 2020; Sturm, 2009). Melanocytes are distributed in the epidermis, hair follicles, choroid, iris, inner ear, and other tissues (Dubey & Roulin, 2014). In the melanosome, l-tyrosine is oxidized by tyrosinase and is converted into melanin through a series of biochemical reactions (d'Ischia et al., 2013; Ito & Wakamatsu, 2008; Sánchez-Ferrer et al., 1995). The light absorption of melanin in skin and hair leads to photoprotection, thermoregulation, camouflage, and display coloring (d'Ischia et al., 2015; Prota, 1992). Melanogenesis is a complex pathway leading to the production of both EM and PM (d'Ischia et al., ,2013, 2015; Ito & Wakamatsu, 2008; Micillo et al., 2016; Prota, 1988, 1992, 1995; Sugumaran & Barek, 2016). The observation that EM production begins with l-tyrosine was established by the pioneering work of Mason (1948) and Raper (1927) and is thus called the “Raper-Mason pathway.” However, most natural melanin pigments are actually copolymers or mixtures of EM and PM in varying ratios (Ito & Wakamatsu, 2003), and thus, melanin production can be thought of as a process of “mixed melanogenesis” (Ito, 2003; Ito & Wakamatsu, 2003, 2008, 2011a; Simon et al., 2009). Melanosomal pH is critical for the ratio of EM and PM production during mixed melanogenesis. This review summarizes how melanogenesis proceeds chemically and biochemically and describes the factors that regulate melanogenesis by controlling melanosomal pH.

2 ∣. CHEMICAL CONTROL OF MELANIN SYNTHESIS

2.1 ∣. Mixed melanogenesis leading to the production of EM and PM

EM and PM are highly oxidized and complex pigments. Both types of pigment are derived from the common precursor dopaquinone (ortho-quinone of L-DOPA, DQ) which is formed via the oxidation of l-tyrosine by the action of the melanogenic enzyme tyrosinase. The catalytic conversion of tyrosine to DQ by tyrosinase is the initial and key step of melanogenesis (Ito & Wakamatsu, 2011a). DQ is a highly reactive ortho-quinone intermediate that reacts extremely rapidly with thiol compounds leading to its pivotal role in the chemical control of melanogenesis (Ito, Sugumaran et al., 2020; Ito & Wakamatsu, 2008). In the absence of thiol compounds, DQ undergoes intramolecular cyclization of its amino group to produce cyclodopa (or leucodopachrome), which is then rapidly oxidized by a redox reaction with DQ to give dopachrome (DC) and L-DOPA (Land et al., 2003) (Figure 1). DC is then gradually and spontaneously converted mostly to form 5,6-dihydroxyindole (DHI) by decarboxylative rearrangement and to a lesser extent into 5,6-dihydroxyindol e-2-carboxylic acid (DHICA) (Ito et al., 2013; Palumbo et al., 1987; Raper, 1927), the ratio of which is determined by a distinct melanogenic enzyme termed DC tautomerase (DCT, tyrosinase-related protein-2) (Kroumpouzos et al., 1994; Pawelek et al., 1980; Tsukamoto et al., 1992). Before the TRP2 locus was characterized, DCT was found as a new enzyme in the pigment pathway (Aroca, García-Borrón, et al., 1990; Aroca, Solano, et al., 1990; Barber et al., 1984; Orlow et al., 1994; Pawelek et al., 1980). It is known that tyrosinase has copper in its active site. On the other hand, Solano et al., (1996) reported that the metal cofactor for the DCT active site is zinc and proposed a reaction mechanism (Aroca et al., 1992; Solano et al., 1994).

FIGURE 1.

Biosynthetic pathways leading to the production of EM and PM. Note that the activities of tyrosinase, Tyrp1 and Dct/Tyrp2, and the precursor tyrosine are involved in the production of EM, while only tyrosinase (and the precursors tyrosine and cysteine) is necessary for the production of PM. Additionally, note that human TYRP1 does not act as a DHICA oxidase (Boissy et al., 1998). MFSD12 is a component of the melanosomal cysteine import system (Adelmann et al., 2020). (Modified from Ito & Wakamatsu, 2011a)

Oxidative polymerization of these indoles leads to the production of EM. Copper ions can catalyze this process (Ito et al., 2013). Oxidative polymerization of DHI and DHICA in various ratios produces black to dark brown EM. Oxidation of DHI is catalyzed directly by tyrosinase or indirectly by DQ, while oxidation of DHICA appears to be catalyzed by tyrosinase-related protein-1 (TYRP1), the brown locus protein, at least in mice (Jiménez-Cervantes et al., 1994; Olivares et al., 2001). However, the human homolog of TYRP1 may not act in the same way as in mice (Boissy et al., 1998), and its precise enzymatic function in humans is not yet clear. In the presence of l-cysteine, DQ preferentially gives rise to cysteinyldopa (CD) isomers.

Mixed melanogenesis is thought to proceed in three distinctive steps: The first step is the production of CDs by the rapid addition of cysteine to DQ, which continues as long as cysteine is present (Ito & Wakamatsu, 2008; Land et al., 2003; Land & Riley, 2000; Thompson et al., 1985). Due to its high reactivity, DQ can easily undergo both Michael 1,6- and 1,4-additions with nucleophiles such as thiols and amines (Ito, Sugumaran, et al., 2020). These reactions lead to the production of melanin pigment found in the skin, hair, and eyes of all animals. Tse et al., (1976) showed that the addition of sulfhydryl compounds proceeds very quickly to generate thiol adducts. 5-S-cysteinyldopa (5SCD) is rapidly formed as the major CD isomer (Ito & Prota, 1977). The second step is the oxidation of CDs to produce PM, which continues as long as CDs are present. Oxidation of CD proceeds by redox exchange with DQ to form the quinone form of CDs. Cyclization of the quinone and rearrangement generates benzothiazine (BT) intermediates that are then oxidized to form PM consisting mainly of the BT moiety (BT-PM), which are gradually converted to PM consisting of the benzothiazole (BZ) moiety (BZ-PM) (Wakamatsu et al., 2009). The last step is the production of EM, which begins only after most CDs are depleted. The creation of a PM internal core covered by EM is known as the “casing” model, which will be described in a later section. Prota and his collaborators have carried out detailed biosynthetic studies to clarify melanogenic intermediates produced beyond DHI and DHICA during eumelanogenesis and produced after the formation of cysteinyldopas during pheomelanogenesis (Di Donato & Napolitano, 2003; Napolitano et al., 1996; Panzella et al., 2007; Prota et al., 1998).

In addition to tyrosinase, two tyrosinase-related proteins modulate eumelanogenesis (Hearing, 1993, 2000). DCT, whose presence was first suggested by Pawelek at al. (1980), catalyzes the tautomerization of DC to DHICA (Tsukamoto et al., 1992) and is encoded by the tyrosinase-related protein-2 (Tyrp2) gene in mice (Jackson et al., 1992). Certain divalent metal ions can also promote tautomerization (i.e., isomerization with a shift of hydrogen atoms) from DC to DHICA (Palumbo et al., 1987, 1991). The oxidative polymerization of DHI can be catalyzed by mammalian tyrosinase (Tripathi et al., 1991); however, a pulse radiolysis study indicates that DHI can also be effectively oxidized by DQ alone (Edge et al., 2006). Although mouse TYRP1 can oxidize DHICA, human TYRP1 cannot; however, human tyrosinase is able to oxidize DHICA, as well as tyrosine, L-DOPA, and DHI. The activities of these proteins greatly affect the quantity and composition (the ratio of DHI to DHICA and the degree of polymerization) of EMs produced.

2.2 ∣. Kinetic analysis of mixed melanogenesis

Melanogenesis is a complex pathway leading to the production of EM and PM. Chemical studies on mixed melanogenesis have been accelerated by the introduction of the pulse radiolysis method, which is a powerful tool for following the rapid reactions promoted by DQ and other related ortho-quinones (Edge et al., 2006; Ito & Wakamatsu, 2008; Land et al., 2001, 2003; Land & Riley, 2000; Thompson et al., 1985). Land et al., (2001) reported rate constants, based on pulse radiolysis, for the four important steps in the early phase of melanogenesis (Figure 2a). Dibromine radical anions Br₂^{. −} are produced within 1 nanosecond by pulse radiolysis of a N₂O-saturated KBr solution. Addition of DOPA to this system produces dopasemiquinone which then disproportionates to generate DQ and DOPA within a millisecond. Then, the fate of DQ can be followed by spectrophotometry in the presence or absence of a target molecule such as cysteine.

FIGURE 2.

Kinetics of early and late stages of mixed melanogenesis. (a) Kinetics of early stages of mixed melanogenesis. Note that the rate constants r1-r4 are controlled by the intrinsic chemical reactivity of DQ. No enzymes other than tyrosinase are necessary to promote these reactions. The intramolecular cyclization of DQ to cyclodopa requires deprotonization of the amino group (actually present in the form of −NH₃⁺). Adapted from Land and Riley (2000), Land et al., (2003), and Thompson et al., (1985). (b) Kinetics of late stages of eumelanogenesis from DHI and DHICA. Note that DQ acts again here as a redox exchanger. Data taken from Edge et al., (2006). (c) Kinetics of late stages of pheomelanogenesis from 5-S-CD quinone. Note that the dihydrobenzothiazine is produced via redox exchange, while benzothiazine intermediates are generated by rearrangement (with/without decarboxylation). Adapted from Edge et al., (2006), Napolitano et al. (1994), and Thompson et al., (1985)

The first step in eumelanogenesis is the intramolecular 1,4-Michael addition of the amino group to DQ giving cyclodopa (r1 = 3.8 s⁻¹). This step is a fairly slow step because this cyclization is a base-catalyzed reaction with rate constants of 0.20 s⁻¹ at pH 5.6 and 7.6 s⁻¹ at pH 7.6 due to a non-protonated −NH₂ on the amino groups (Thompson et al., 1985). However, as soon as cyclodopa is formed, it is rapidly oxidized by DQ to form DC through a redox exchange giving DOPA (r2 = 5.3 × 10⁶ M⁻¹ s⁻¹) (Land et al., 2003). On the other hand, the first step in pheomelanogenesis is a 1,6-Michael addition of cysteine to DC, which proceeds very fast (r3 = 3 × 10⁷ M⁻¹ s⁻¹) (Thompson et al., 1985). The second step in pheomelanogenesis, the redox exchange with DQ giving rise to CD quinone, proceeds at a slower rate (r4 = 8.8 × 10⁵ M⁻¹ s⁻¹) (Land & Riley, 2000). The chemical reactivity of DQ controls these four reactions through an intramolecular addition of an amino group, an intermolecular addition of a sulfhydryl group, and two redox exchange reactions. From these kinetic data, an “Index of Divergence” (D) between eumelanogenesis and pheomelanogenesis can now be derived. By taking DC and CD quinone as representatives of those two pathways, Land et al. (2003) proposed the formula:

$$\mathrm{D}=\mathrm{r}3\times \mathrm{r}4[\text{cysteine}]∕\mathrm{r}1\times \mathrm{r}2=0.78\mu M.$$

From these kinetic data, comparison of the rate constant for the addition of cysteine to DQ (r3) with the rate constant for intramolecular cyclization (r1) shows that the production of CDs is preferred as long as the cysteine concentration is higher than 0.13 μM. Secondly, the redox exchange giving rise to CD quinone (r4) proceeds 30 times slower than the addition of cysteine to DQ (r3). Thus, CDs accumulate in the early phase of pheomelanogenesis. Thirdly, comparison of the rate constant for the redox exchange producing CD quinone from DQ (r4) to the rate constant for intramolecular cyclization (r1) suggests that pheomelanogenesis is preferred over eumelanogenesis as long as the CD concentration is higher than 9 μM.

One of the most intriguing aspects of the reactivity of ortho-quinone with different nucleophiles is the fact that while most nucleophiles including amines exhibit normal Michael 1,4-addition reaction to quinones, addition of thiols with ortho-quinone proceeds mostly at the C5 positions via Michael 1,6-addition (Ito, Sugumaran, et al., 2020, Figure 1). In this regard, Kishida et al., (2021) recently investigated the binding mechanism of l-cysteine to DQ using density functional theory-based calculation. They calculated the binding energies of Cys-S⁻ attacked intermediates and minimum energy pathways for the approach/migration of Cys-S⁻ on the aromatic carbons. Based on the results, they proposed that the binding of Cys-S⁻ to DQ proceeds in the following sequences: (i) coordination of Cys-S⁻ to C3-C4 bridge, (ii) migration of Cys-S⁻ to C5 (or C2), (iii) proton rearrangement from cysteinyl-NH₃⁺ to O3 (O4), and (iv) proton rearrangement from C5 (C2) to O4 (O3). These results were consistent with a previous kinetic study on the thiol binding reactions (Jameson et al., 2004).

In the last stages of eumelanogenesis (Figure 2b), DC accumulates, because DC is fairly stable, with a half-life of about 30 min (first-order rate constant of 4.0 × 10⁻⁴ s⁻¹). It spontaneously decomposes to produce mostly DHI at neutral pH in the absence of DCT or metal ions. The ratio of DHI to DHICA produced under these conditions is 70:1 (Palumbo et al., 1987). However, in the presence of DCT, DC undergoes tautomerization to preferentially produce DHICA (Palumbo et al., 1991). Thus, the ratio of DHICA to DHI in melanin is determined by the activity of DCT. In addition to DCT, a number of biorelevant transition metal ions, such as Cu²⁺ and Zn²⁺, can also direct the rearrangement of DC to DHICA at physiological pH, although DCT seems to be the most effective at catalyzing the tautomerization (Ito et al., 2013; Napolitano et al., 1985; Palumbo et al., 1987, 1988). The redox exchange reaction between DHI and DQ has been studied (Edge et al., 2006) and proceeds with a rate constant of r5 = 1.4 × 10⁶ M⁻¹ s⁻¹, but not quite to completion. This is in the same range as the rate constants for reactions with cyclodopa (r2) and with 5SCD (r4) as shown in Figure 2a. The reaction with DHICA also does not go to completion and proceeds at a much slower rate (r6 = 1.6 × 10⁵ M⁻¹ s⁻¹). During eumelanogenesis, DHI oxidation takes place by redox exchange with DQ, which also occurs with DHICA although less efficiently. Thus, DHICA oxidation to the quinone form may require tyrosinase in humans (Olivares et al., 2001) or TYRP1 in mice (Jiménez-Cervantes et al., 1994; Kobayashi et al., 1994).

In the last stages of pheomelanogenesis (Figure 2c), 5SCD quinone is rapidly cyclized via attack of the cysteinyl side chain amino group by the carbonyl group to produce a cyclic ortho-quinonimine intermediate. The rate (r7) of quinonimine formation was determined by pulse radiolysis to be 10 s⁻¹ (Edge et al., 2006; Napolitano et al., 1994). The rate of 5SCD quinone formation is controlled by the rate of DQ formation as 5SCD is a poorer substrate for tyrosinase than L-DOPA (Agrup et al., 1982). The ortho-quinonimine then undergoes rearrangement to benzothiazine intermediate(s) with (85%) and without (15%) decarboxylation (Napolitano et al., 1994). The rate (r8) of decay (k = 6.0 s⁻¹) of cyclic ortho-quinonimine to benzothiazine was determined by pulse radiolysis (Napolitano et al., 1999). In late stages of pheomelanogenesis, the benzothiazines are produced rapidly from CDs. Benzothiazines are also unstable and decay over a few seconds with a rate constant of 0.5 s⁻¹ (Napolitano et al., 1999). An alternative pathway for the generation of ortho-quinonimine is by redox exchange with 5SCD, which leads to the production of a reduced form of the quinonimine (dihydro-1,4-benzothiazine derivatives) and 5SCD quinone (Napolitano et al., 1994; Wakamatsu et al., 2009). The reactions beyond the benzothiazines, which lead to PM, have been reviewed (Di Donato & Napolitano, 2003).

Glutathione (GSH) is the predominant thiol in cytosol but may not be in melanosomes. Cysteine is transported into melanosomes, but GSH is not transported into melanosomes, nor does it seem likely that simple diffusion or a membrane channel could allow GSH entry into melanosomes (Adelmann et al., 2020; Potterf et al., 1999). Benathan (1996) and Benathan and Labidi (1996) reported several lines of biochemical evidence that 5SCD is the relevant precursor of PM (but not GSH), being formed directly through reaction of DQ with cysteine.

3 ∣. BIOCHEMICAL CONTROL OF MIXED MELANOGENESIS BY TYROSINASE ACTIVITY AND CYSTEINE CONCENTRATION

3.1 ∣. Chemical phenotype

Visual inspection of hair color is limited to subjective assessment (Naysmith et al., 2004). To study mixed melanogenesis objectively, quantities of both EM and PM need to be analyzed. We developed the widely used microanalytical methods of alkaline H₂O₂ oxidation (AHPO) and hydroiodic acid (HI) hydrolysis followed by high-performance liquid chromatography (HPLC). AHPO generates pyrrole-2,3,5-tricarboxylic acid (PTCA), a EM marker, and thiazole-2,4-5-tricarboxylic acid (TTCA), a PM marker (Ito et al., 2011, Ito, Del Bino et al., 2020; Ito & Fujita, 1985). PM can also be analyzed as 4-amino-3-hydroxyphenylalanine (4-AHP) generated by HI hydrolysis (Wakamatsu et al., 2002). The genetic basis for the diversity of human hair color has been extensively studied (for reviews, see Ito & Wakamatsu, 2011b; Pavan & Sturm, 2019; Sturm & Duffy, 2012). Based on the analysis of melanin content in human hair samples of various colors (n = 228), the EM content decreases in human hair of black, dark brown, brown, light brown, blond, and red color, in that order, while the PM content remains low but constant across all hair colors, except for red hair which has elevated PM (Ito et al., 2011; Valenzuela et al., 2010). Thus, human dark hair contains high levels of EM (and trace levels of PM), light hair contains low levels of EM (and trace levels of PM), whereas red hair contains low to medium levels of both EM and PM. This approach to the chemical analyses of melanin has been used on hair from patients with various hypopigmentary disorders including Hermansky–Pudlak syndrome, Menkes disease, proopiomelanocortin deficiency, cystinosis, malnutrition, and trace metal deficiency (Ito & Wakamatsu, 2011a); the chemical phenotype has helped establish the precise melanin defects of each disease.

3.2 ∣. Tyrosinase activity

As mentioned above, mixed melanogenesis proceeds in three distinct stages: (1) the production of CDs; (2) the oxidation of CDs to produce PM; and (3) the production of EM (Ito, 2003). The amount of melanin produced is proportional to DQ formation, which is in turn proportional to tyrosinase activity. Therefore, the ratio of EM to PM is determined by tyrosinase activity and the availability of cysteine in melanosomes (Land et al., 2003).

Several studies have presented evidence demonstrating that tyrosinase activity plays a major role in controlling mixed melanogenesis. Burchill et al., (1986) investigated the effect 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 PM, whereas pubertal mice produce a mixed-type melanin. Injection of α-MSH into pubertal mice resulted in a 1.7-fold increase in tyrosinase activity and more eumelanic hair production with a concomitant increase in total melanin. Injection of bromocriptine (which reduces α-MSH secretion) into these pubertal mice reduced tyrosinase activity to 8% of the untreated control, resulting in pheomelanic hair production and a decrease in total melanin. Injection of α-MSH into newborn yellow mice increased tyrosinase activity 2.2-fold and led to more eumelanic hair (Burchill et al., 1993). These results indicated an important role of tyrosinase activity in controlling mixed melanogenesis in vivo.

The significance of tyrosinase activity in mixed melanogenesis in vitro was also examined (Hunt et al., 1995). Treatment of human melanocytes with synthetic α-MSH analog resulted in an increase in EM content and variable responses in PM content, leading to a significant increase in eumelanin. 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 (Coleman, 1962; Lamoreux et al., 2001). EM content was reduced by 40% in black chinchilla mice without any increase in PM content compared with black mice, whereas PM content was reduced nine-fold in lethal yellow chinchilla mice compared with lethal yellow mice (Lamoreux et al., 2001). In non-agouti black, brown, or slaty mice, chinchilla mutations reduced EM content by only 20%–40%. In lethal yellow or recessive yellow mice, on the other hand, chinchilla mutations reduced PM content by 5 to 10-fold. Agouti mice produce yellow-striped black hair, while agouti chinchilla mice produce white striped gray hair. Similar results were obtained by analyzing the agouti pattern of baboon hairs (Ito et al., 2001). Dark-colored baboons produced high levels of PM in the yellow bands of their agouti hairs, whereas the corresponding bands in light-colored baboons contained little melanin. The above results suggest that pheomelanogenesis is affected more by a decrease in tyrosinase activity than eumelanogenesis (Barsh, 1996; Ollman et al., 1998).

3.3 ∣. Cysteine concentration in melanosomes

Another mechanism that biochemically controls eumelanogenesis and pheomelanogenesis, aside from tyrosinase activity, is the availability of cysteine (or cystine) in melanosomes (Adelmann et al., 2020; Chintala et al., 2005; del Marmol et al., 1996). For example, decreasing the extracellular concentration of cystine in cultures of human melanoma cells led to a eumelanic shift presumably from the dramatic decrease in intracellular cysteine concentration (del Marmol et al., 1996). Smit et al., (1997) found a two-fold increase in melanin production with a decreased ratio of PM to total melanin when cells were cultured at a higher concentration of tyrosine relative to cystine.

The role of cysteine/cystine transport as a mechanism for controlling mixed melanogenesis requires more investigation. Chintala et al., (2005) showed that the murine subtle gray (sut) phenotype occurred because of a mutation in the Slc7a11 gene, which encodes the plasma membrane cystine/glutamate exchanger xCT. The low rate of extracellular cystine transport into sut melanocytes reduced PM synthesis with little to no effect on EM synthesis. The effect of the sut mutation on PM synthesis was enhanced on an A^y/a background leading to a decrease in the levels of PM in the hair to one-sixth of the A^y/a littermates.

Using MelanoIP, a method for rapidly isolating melanosomes and profiling their labile metabolite contents, Adelmann et al., (2020) found that MFSD12 (major facilitator superfamily domain-containing protein 12) is required to maintain normal levels of cystine, the oxidized dimer of cysteine, in melanosomes, and to produce CD (Figure 1). MFSD12 is part of the large major facilitator superfamily of 12-transmembrane domain proteins and causes darker pigmentation in mice and humans when suppressed (Adhikari et al., 2019; Crawford et al., 2017). Isolated melanosomes were known to import cysteine, but the transporter responsible for this activity had not been identified (Potterf et al., 1999); however, cystine was known to efflux out of these organelles through the transporter protein cystinosin (CTNS) (Chiaverini et al., 2012; Gahl et al., 1982; Jonas et al., 1982). Comparing metabolites in melanosomes isolated from wild-type melanocytes and MFSD12-deficient melanocytes, the levels of cystine and CD were significantly reduced in MFSD12-deficient melanosomes. The work by Adelmann et al. (2020) indicates that the pigmentation protein MFSD12 is a component of the melanosomal cysteine import system, and is necessary and probably sufficient for the tranport of cysteine into the melanosome, processes that have been long described but lacked a clear mechanism (Potterf et al., 1999).

3.4 ∣. Casing model of mixed melanogenesis and the factors that impact its chemistry

It is conceived that when tyrosine levels are high or cysteine levels are low, melanogenesis passes rapidly through the early pheomelanogenesis stages (Figure 2a), resulting in more eumelanic cells. In contrast, when tyrosine levels are low or cysteine levels are high, pheomelanogenesis occurs as long as cysteine is present (Ito, 2003; Ito & Wakamatsu, 2008). When tyrosinase activity is significantly reduced, melanogenesis is suppressed to a stage where only the production of CD is apparent and little to no pigment is produced (Lamoreux et al., 2001). Variations in human pigmentation are also affected by homologues of mouse coat color genes. Therefore, a comparison of the pigmentation effects of the human hair color genes with those of the mouse hair color genes can be informative. How two (or even three) coat color genes interact to affect pigmentation can be accurately analyzed in mice. Unfortunately, this level of analysis is not so simple in humans. Human hair color ranges from black, brown, blonde to red, and this diversity results mostly from the quantity and proportion of black-dark brown EM and reddish-brown PM (Ito & Wakamatsu, 2011b; Valenzuela et al., 2010). Thus, human hairs are mostly eumelanic, despite the diversity in color. The fact that PM content remains low but at a constant level fits well with the casing model of mixed melanogenesis (Ito, 2006; Ito & Wakamatsu, 2008; Simon et al., 2009), which was originally suggested by Agrup et al., (1982) based on in vitro biochemical findings (Figure 3a). The casing model of mixed melanogenesis implies that PM is always produced first, and then, EM is deposited on the preformed PM (Ito & Wakamatsu, 2008; Simon & Peles, 2010; Simon et al., 2009). A study by Bush et al., (2006) provided biophysical evidence supporting this model. It is thought that melanosomes are assembled through a casing process in which an initially formed PM core is encapsulated within an EM coating. Evidence for this architecture was provided by studies of iridal melanosomes and neuromelanin, a dark pigment in the substantia nigra and some other regions of the brain (Bush et al., 2006; Peles et al., 2009).

FIGURE 3.

Casing model for mixed melanogenesis. (a) The process of mixed melanogenesis. Note that pheomelanic pigment is produced first, followed by the deposition of eumelanic pigment. In the granule with the EM surface, the side was intentionally cut away to reveal the inner PM core. EM is believed to act as a photoprotective anti-oxidant while PM as a phototoxic pro-oxidant. (b) Proposed mechanism for CD-melanin promoted DOPA polymerization and coating formation. (Adapted from Greco et al., 2011)

In humans, PM and EM are present in various parts of the eye, such as in the iris and choroid-retinal pigment epithelium (Sturm & Frudakis, 2004) (Figure 4a), where they mainly play photoprotective and anti-oxidant roles (Biesemeier et al., 2011; Peles & Simon, 2012; Rózanowska, 2011; Weiter et al., 1986). Iris pigmentation affects the incidence of several important ocular diseases, including uveal melanoma and age-related macular degeneration (Duffy et al., 2004; Sun et al., 2013; Wakamatsu et al., 2008). The amount of EM and PM in cultured human uveal melanocytes was measured by the afore-mentioned chemical degradation—HPLC method (Wakamatsu et al., 2008) (Figure 4b). Uveal melanocytes derived from eyes with dark-colored irides (dark brown and brown irides) contained a significantly greater quantity of EM than uveal melanocytes derived from eyes with light-colored irides (blue, yellow-brown, green, and hazel-colored irides). A similar but less pronounced difference was observed between dark brown and brown irides. The amount of PM in uveal melanocytes derived from eyes with light-colored irides was slightly greater than that from dark-colored irides. The EM/PM ratio in uveal melanocytes reflected iris color: The darker the iris color, the higher the EM/PM ratio. This indicates that darker melanocytes produce not only a higher quantity of melanin but also more eumelanic pigment than lighter melanocytes. Melanocytes derived from dark brown irides contain 93% EM, while those derived from blue irides contain only 44% EM. The total quantity of melanin (either EM + PM or total melanin measured with HPLC or spectrophotometry, respectively) was also correlated with iris color. EM + PM and total melanin values in uveal melanocytes derived from eyes with dark-colored irides were greater than those in uveal melanocytes derived from eyes with light-colored irides. This suggests that uveal melanin granules in eyes with dark-colored irides are eumelanic at the surface and act as an anti-oxidant, while those in eyes with light-colored irides have an exposed PM core and behave as a pro-oxidant.

FIGURE 4.

Eye colors and melanin characterization. (a) Representative eye colors ranging from blue, gray, green, hazel, light brown to dark brown. (Adapted from Sturm & Frudakis 2004). (b) Quantity and ratio of melanin in uveal melanocytes derived from eyes with various colored irides. Data are presented as means ± SEM. *p and **p in melanocytes were compared between dark brown and brown, and between brown and light-colored irides (blue, yellow-brown, green, and hazel combined, as indicated by the horizontal bars), respectively. (Adapted from Wakamatsu et al., 2008)

Greco et al., (2011) reported an oxygen-dependent biomimetic study for L-DOPA, dopamine, and norepinephrine polymerization exploiting the redox properties of 5SCD-melanin polymer. They reported the decay profiles of L-DOPA after multiple additions to fine suspensions of CD-melanin at pH 7.4. The results were consistent with the formation of an insoluble L-DOPA polymer coating and encapsulating the PM redox system. The polymerization was mediated by redox exchange with CD-melanin units (Figure 3b). The candidates are probably 1,4-benzothizine units, for example, in the ortho-quinone-imine form. This compound may oxidize L-DOPA with concomitant reduction to dihydrobenzothiazine units, which would enter a redox cycle sustained by oxygen. Synthetic CD-melanin was able to accelerate the aerial polymerization of L-DOPA leading to eumelanic-like deposits that encapsulate the core. Kinetic and chemical evidence by scanning electron microscopy indicated conversion of L-DOPA into a black insoluble polymer encapsulating the active CD-melanin core.

To study mixed melanogenesis in melanosomes, Gorniak et al., (2014) applied X-ray fluorescence (XRF) analysis with synchrotron radiation to survey the nanoscale distribution of metals within purified iris melanosomes in mice. XRF analysis with synchrotron radiation allows visualization of metals at the nanoscale. From the color distribution, it was apparent that within some melanosomes Ca and Zn were concentrated within their core regions. On the other hand, Cu, which is a trace metal and used as a biomarker for eumelanin pigment in the fossil record (Wogelius et al., 2011), was more likely to be found close to their surface. The observation of some melanosomes with a Cu-rich periphery supports the casing model, whereas the observation of homogenous metal distributions in other melanosomes challenges this model, but also suggests that in at least some tissues and genetic contexts, arrangements of melanin other than casing model may co-exist.

3.5 ∣. Melanosomal ph regulation of mixed melanogenesis

Most natural melanin pigments, including epidermal and hair melanins, are mixtures (or copolymers) of EM and PM (mixed melanogenesis) (Del Bino et al., 2015; Ito et al., 2011; Ito & Wakamatsu, 2003). The role of pH in controlling mixed melanogenesis is receiving more attention (Ancans, Tobin, et al., 2001; Fuller et al., 2001; Ito et al., 2013; Wakamatsu et al., 2017). The catalytic activity of human tyrosinase is suppressed at low pH (activity at pH 5.8 was ~20% of pH 6.8) (Fuller et al., 2001) leading to a more pheomelanic phenotype (Ancans, Tobin, et al., 2001). In contrast, an increase in melanosomal pH by inhibition of V-ATPase proton pumps increases tyrosinase activity (Ancans, Tobin, et al., 2001; Chen et al., 2021). An increase in intramelanosomal pH from 5.0 to 6.8 is needed for the full maturation of melanosomes (Schallreuter et al., 2008). Melanosomes in melanocytes derived from white/fair skin are acidic, while those derived from black/dark skin are near neutral (Smith et al., 2004). Yet, tyrosinase in white or black melanocytes has the same optimum pH of 7.4 and is equally inhibited at acidic pHs (Fuller et al., 2001). Furthermore, the diversity of normal human skin pigmentation appears to stem from polymorphisms in only a few genes, including oculocutaneous albinism 2 (OCA2), membrane-associated transporter protein (MATP), and solute carrier family 24 member 5 (SLC24A5) (Lamason et al., 2005; Lao et al., 2007; Norton et al., 2007). Available evidence suggests that polymorphisms in these genes result in the acidification of melanosomes. However, the significance of pH in controlling mixed melanogenesis has only been addressed directly in a few studies (Ancans, Tobin, et al., 2001; Chen et al., 2021). In both of these studies, intramelanosomal pH was elevated by inhibition of V-ATPase proton pumps in melanocytes leading to increased tyrosinase activity and a preferential increase in the production of EM. Using a newly described method of melanin measurement based on [U-¹³C] tyrosine liquid chromatography–mass spectrometry (LC-MS) fate tracing, Chen et al., (2021) demonstrated that elevation of melanosomal pH leads to an immediate (~1 hr) increase in eumelanin synthesis. Furthermore, the sensitivity of the LC-MS technique allowed Chen et al., (2021) to measure simultaneous changes in eumelanin and pheomelanin synthesis in live cells within hours of melanosomal pH elevation. Thus, alteration of melanosomal pH actively controls mixed melanogenesis.

Ito et al., (2013) examined the effects of pH (5.3–7.3) and Cu²⁺ ions on the conversion of DC to DHI and DHICA (and then to EM). The rate of DC conversion increased gradually with increasing pH values and almost reached a plateau at pH 7.3. The rate at pH 7.3 was 4.6-fold greater than at pH 5.3. The presence of Cu²⁺ (50 μM) accelerated the conversion by ca. two-fold throughout the pH range examined. In the absence of Cu²⁺, however, the DC conversion after 30 min led almost exclusively to DHI. In the presence of Cu²⁺, the DC conversion proceeded much faster leading to DHICA levels comparable to DHI. Melanins produced from DC after a 4 hr reaction at various pH values were analyzed with respect to yield and DHICA content. In the absence of Cu²⁺, the DC conversion at all pHs tested led to melanins consisting mostly of DHI (>90%). In the presence of Cu²⁺, the DC conversion produced mixed-type eumelanin consisting of DHI and DHICA in ca. 2–3:1. Finally, they examined the effects of pH on the production of DHI + DHICA-melanin from an equimolar mixture of DHI and DHICA as a representative example of the last stage of eumelanogenesis. In the absence of Cu²⁺, the oxidation proceeded rather slowly, especially at lower pH values, and the melanins produced were DHI-rich. In the presence of Cu²⁺, however, the oxidation proceeded much faster, leading to the production of melanin with a DHICA content of ca. 50%. These data suggest that an acidic pH greatly suppresses the late stages of eumelanogenesis and that the presence of Cu²⁺ ions accelerates the conversion of DC to DHICA and its subsequent oxidation. As mentioned above, both Cu²⁺ ions and DCT are known to catalyze the rearrangement of DC to give DHICA rather than DHI. However, at these comparable activities, the ratio of DHICA/DHI formation is significantly higher in the enzyme-catalyzed than in the metal-catalyzed reaction. These results provide a clear-cut differentiation between the effects of the two factors when both are present in biological extracts (Palumbo et al., 1991).

Wakamatsu et al., (2017) then extended the study by Ito et al., (2013) by analyzing whether pheomelanin synthesis is chemically promoted at an acidic pH. They found that the tyrosinase-catalyzed production of PM from either L-DOPA or l-tyrosine, in the presence of l-cysteine, was greatest at pH values of 5.8–6.3, while the production of EM was suppressed at pH 5.8. This suggests that mixed melanogenesis is chemically shifted to a more pheomelanic state at a weakly acidic pH (Figure 5a). They also performed a time course study on the oxidation of tyrosine + cysteine catalyzed by mushroom tyrosinase. As shown in Figure 5b,c, concomitant with a decrease of tyrosine, CD increased and then decreased gradually after 60 min. 4-AHP (after HI hydrolysis), a marker of the benzothiazine moiety of PM, appeared at 90 min. Interestingly, benzothiazole amino acid (BZ, after HI hydrolysis), a marker of the benzothiazole moiety of PM (Wakamatsu et al., 2009), increased in parallel with absorbance at 400 nm (A400). These data clearly support the following sequence of reactions: tyrosine → CD → benzothiazine-PM → ben-zothiazole-PM. It was determined that oxidation proceeded at a slower rate at pH 7.3 (Figure 5c) compared with pH 5.8 (Figure 5b), as evidenced by a slower increase in A400 values and the fact that tyrosine consumption and CD production at pH 7.3 was slower (two-fold) than at pH 5.8. The above results indicated that pheomelanogenesis proceeds faster at an acidic pH, while eumelanogenesis proceeds faster at a neutral pH.

FIGURE 5.

Control of eumelanogenesis and pheomelanogenesis at a weakly acidic pH. (a) An acidic pH greatly suppresses the late stages of eumelanogenesis beyond DC, and pheomelanogenesis is promoted at a weakly acidic pH. The up arrow and down arrow indicate acceleration and suppression by a weakly acidic pH, respectively. (b, c) Time course of oxidation of tyrosine + Cys produced by mushroom tyrosinase at pH 5.8 (b) and at pH 7.3 (c). Changes in yields of tyrosine, DOPA, CDs, 4-AHP, A400 value, and BZ, as assessed by HPLC, are shown. A400 and BZ values were multiplied by 500 and 10, respectively. CDQ and QI denote cysteinyldopa quinone and quinone-imine intermediate, respectively. Data are averages from two separate experiments. (Adapted from Wakamatsu et al., 2017)

Since the mechanisms of cyclization are different for DQ vs CD quinones, the effect of pH on the cyclization reaction would be expected to be different. The observed first-order reaction of DQ decay increased two orders of magnitude from 0.2 s⁻¹ to 23 s⁻¹ when the pH was increased from 5.6 to 8.6 (Chedekel et al., 1984). On the other hand, the decay of CD quinone exhibited a different pH dependence than that of DQ. The decay rate for 5SCD quinone increased 5-fold (3.3 to 15 s⁻¹) when the pH was decreased from 7.7 to 5.8 (Thompson et al., 1985). In this regard, Wakamatsu et al., (2009) clearly show that 1,4-benzothiazine-3-carboxylic acid (DHBTCA), produced by the consumption of CD, is oxidized by DQ through redox exchange. DQ is less prone to cyclize and thus more available for redox exchange with CD at an acidic pH. The ortho-quinone-imine (QI) is produced faster from CD quinone at an acidic pH and plays a major role in PM production; QI works together with DQ to accelerate pheomelanogenesis at an acidic pH.

It has become clear that melanosomal pH is a critical mechanism in the control of both eumelanogenesis and pheomelanogenesis and hence drives mixed melanogenesis. Thus, it is important to understand the factors that affect melanosomal pH.

3.6 ∣. Mechamisms of melanosomal ph control

Melanosomal pH is not static and changes during the maturation of the organelle (Ancans, Tobin, et al., 2001). Melanosomes originate from endosomes and proceed through a set of steps involving the delivery of specific ion channels, co-transporters, and proton pumps (Bellono & Oancea, 2014; Raposo et al., 2001). As a lysosome-related organelle, melanosomes begin their maturation at an acidic pH that increases toward a neutral pH when fully mature. In humans, the final melanosomal pH set point varies with light-skinned people having a relatively more acidic set point as compared to melanosomes in those with dark skin (Ancans, Tobin, et al., 2001). Regardless of the final pH set point, melanosomal pH is a balance between increased melanosomal H⁺ concentration from proton pumps and the activity of a variety of ion channels and transporters which control the electrochemical gradient. We will discuss the established proteins which regulate melanosomal pH and the mechanisms that regulate the melanosomal pH set point (Table 1, also reviewed in Wiriyasermkul et al., 2020).

TABLE 1.

Melanosomal proteins: Function, disease, and regulation

Proteins	Ions and pH	Localization	Disease	Regulation
OCA2/P Protein	Cl−, alkalinize	Melanosome II/III	OCA2	Unknown
SLC45A2	H+/sugar co-transporter, alkalinize	Melanosome III/IV	OCA4	Unknown
SLC24A5	Na+/Ca2+ and K+, alkalinize	TGN, Mitochondria	OCA6	Unknown
TPC2	Ca2+ or Na+, acidify	Melanosome, Lysosome	N/A	NAADP, PI(3,5)P2, mTOR, VEGF, p38/JNK, MAPK
V-ATPase	H+, acidify	Melanosome	N/A	cAMP (trafficking)

Note: The recent review by Wiriyasermkul et al., (2020) provides additional detail on these proteins.

3.7 ∣. Proton pumps, channels, and transporters

The vacuolar H⁺-ATPase (V-ATPase) is a proton pump that is expressed on melanosomes during all stages of maturation (Tabata et al., 2008). This pump facilitates the acidification of the premelanosomes and continues to drive H⁺ ions into the melanosomes during maturation. As the premelanosomes mature, other proteins are trafficked to the melanosome to counteract the V-ATPase activity and raise the pH. These proteins consist of ion channels, exchangers, and transporters, which affect V-ATPase activity and pH by affecting H⁺ concentration and the electrochemical gradient in the melanosome (Table 1).

OCA2, which is also known as pink-eyed dilution/P protein (Brilliant et al., 1991), is expressed on melanosomes (Bellono et al., 2014). Recent work suggests that this channel is trafficked to the melanosome early during maturation and plays a role in the initial neutralization of melanosomal pH (Le et al., 2020). This conclusion was based on the limited co-localization of OCA2 at melanosomes with TYRP1, a marker of late-stage (III-IV) melanosomes (Le et al., 2020). Based on patch clamp experiments, OCA2 induces a large outwardly rectifying current with a negative charge of chloride anions (Bellono et al., 2014). The loss of negative charge in the melanosome is thought to prevent the delivery of additional H⁺ ions and lead to the neutralization of the pH (Ancans, Tobin, et al., 2001). Consistent with the role of OCA2 as a neutralizer of melanosomal pH, loss of OCA2 leads to an acidic pH set point and may affect tyrosinase processing (Ancans et al., 2001; Ancans, Tobin, et al., 2001; Bellono et al., 2014; Chen et al., 2002; Manga & Orlow, 2002). Mutation or loss of OCA2 is the cause of oculocutaneous albinism type 2 (Brilliant et al., 1991; Oetting et al., 2005) (Table 1).

Solute carrier family 45 member 2 (SLC45A2), which is also known as membrane-associated transporter protein (MATP)/absent in melanoma 1 (AIM1), is expressed on melanosomes and was thought to be a tumor suppressor in melanoma models (Ray et al., 1997; Trent et al., 1990). Based on sequence homology, it is predicted to function as a H⁺/ sugar co-transporter and expression in yeast appears to confirm this function (Bartölke et al., 2014); however, the source of sugar within the melanosome is not known and this specific function at the melanosome has not been confirmed. SLC45A2 is also known to facilitate tyrosinase trafficking from the Golgi to the melanosome (Costin et al., 2003). This protein is known to affect pigmentation; mutation or loss of this protein is the cause of OCA4 and leads to reduced pigmentation in mice (underwhite) and zebrafish (Fukamachi et al., 2001; Newton et al., 2001). In contrast to OCA2, SLC45A2 is trafficked to melanosomes at late stages of maturation (III/IV) as evidenced by a high level of correlation with TYRP1 expression at the melanosome (Le et al., 2020). This suggests that SLC45A2 may be important for controlling pH in mature melanosomes. Functional studies have revealed that loss of SLC45A2 leads to an acidic melanosomal pH set point and likely explains why loss of this protein leads to reduced pigmentation (Bartölke et al., 2014; Bin et al., 2015; Reinders & Ward, 2015) (Table 1).

Two pore segment channel 2 (TPC2/TPCN2) is an ion transporter that is not exclusively expressed on melanosomes but instead is found on endosomes, lysosomes, and lysosome-like organelles (Ambrosio et al., 2016; Calcraft et al., 2009; Zong et al., 2009). Unlike SLC45A2 and OCA2, which are exclusively expressed in melanocytes, TPC2 is expressed in most tissues (Ambrosio et al., 2016; Calcraft et al., 2009; Zong et al., 2009). TPC2 is distinct from other TPC family members in that TPC2 is a voltage-independent cation channel that responds to ligands instead of changes in membrane potential (She et al., 2019). TPC2 localization on melanosomes has been confirmed (Ambrosio et al., 2016; Sulem et al., 2008). TPC2 function in non-pigmented cells is well established where it is known to function as a cation transporter (Na⁺ or Ca²⁺) (Bellono et al., 2016; Pitt et al., 2016). TPC2 has multiple effects in cells such as cancer cell invasion and autophagy (Alharbi & Parrington, 2019; Nguyen et al., 2017; Sun & Yue, 2018). In melanocytes, TPC2 plays a role in melanosomal pH and size (Ambrosio et al., 2016; Bellono et al., 2016). In contrast to the previously mentioned proteins, TPC2 functions as a negative regulator of melanin synthesis and melanosomal pH neutralization (Ambrosio et al., 2016; Bellono et al., 2016). TPC2 appears to promote an acidic pH since loss of TPC2 leads to an increase in melanosomal pH and pigmentation (Ambrosio et al., 2016). It is suggested that TPC2 generates membrane potential via its positive conductance thereby supporting V-ATPase activity (Bellono et al., 2016). TPC2 mutations or loss is not associated with any pigment diseases; however, polymorphisms in the protein are associated with hair color and UV sensitivity mostly in Icelandic and Dutch individuals (Böck et al., 2021; Chao et al., 2017; Sulem et al., 2008) (Table 1). The proposed interplay between the above proteins in regulating melanosomal pH was recently reviewed in the context of other melanosome proteins highlighting the role of other ions and pathogenic mutations/polymorphisms in pigmentation (Wiriyasermkul et al., 2020).

SLC24A5, also known as sodium/potassium/calcium exchanger (NCKX5), is specifically expressed in pigmented tissues where it is predicted to transport K⁺ and Ca²⁺ ions in exchange for Na⁺ based on homology with NCKX family members. It is this transport of ions that is thought to drive melanogenesis and melanosome biogenesis (Ginger et al., 2008; Zhang et al., 2019). However, its exact function is not known. Interestingly, although fractionation experiments suggest SLC24A5 is present at mature melanosomes (Chi et al., 2006), immunolocalization of SLC25A5 suggests it to not present on melanosomes but instead is localized to mitochondria and the trans-Golgi network (Ginger et al., 2008; Rogasevskaia et al., 2019). One possible mechanism proposed for SLC24A5 in the trans-Golgi network is as a functional coupler between sodium/hydrogen exchanger 7 (NHE7) and V-ATPase leading to pH regulation of the melanosome (Ginger et al., 2008). Even though the exact mechanism of SLC24A5 remains unknown, this protein is clearly important for pigmentation. Loss of SLC24A5 activity leads to reduced pigmentation and an albino phenotype in zebrafish (golden), mice, and humans (OCA6) (Vogel et al., 2008; Wei et al., 2013). In addition, polymorphisms in SLC24A5 are associated with light skin color among Europeans and East Asians (Lamason et al., 2005) (Table 1).

Aside from the above well-characterized channels and exchangers discussed above, there are a few less well-characterized exchangers which may impact melanosomal pH. Na⁺/H⁺ exchangers (NHE) are localized to melanosomes. Specifically, NHE-3 and NHE-7 co-localize with TYPR1 in melanocytes (Smith et al., 2004). However, the role of NHEs in pigmentation is not established. In addition, voltage-gated Cl⁻/H⁺ exchangers (ClCs) have been investigated. Specifically, chloride (Cl⁻) channel, isoform 7 (ClC-7), mediates Cl⁻/H⁺ exchange in lysosomes (Graves et al., 2008). ClC-7 cannot function without another membrane protein called Ostm1, and mice lacking ClC-7 or Ostm1 have coat color and bone abnormalities (gray-lethal mouse) (Chalhoub et al., 2003; Lange et al., 2006; Meadows et al., 2007). It remains unclear whether ClC-7 is simply required for melanocyte survival or is important for melanosomal pH regulation (Stauber & Jentsch, 2012). These two exchangers are known to control H⁺ efflux in lysosomes, and mutations in these proteins can lead to hyperacidification and albinism; however, their localization to melanosomes is not known.

3.8 ∣. Signaling pathways that regulate melanosome development and pH

There are numerous signaling cascades that regulate melanocyte biology; however, few are known to directly regulate melanosome biology or pH. Perhaps the best-characterized pigment-altering signaling cascade is the melanocortin 1 receptor (MC1R) pathway. MC1R is a G protein-coupled receptor expressed on melanocytes and is stimulated by α-MSH. MC1R activation leads to stimulation of heterotrimeric G proteins and transmembrane adenylyl cyclases (tmACs) (Sassone-Corsi, 2012). cAMP generated by tmACs then stimulates protein kinase A (PKA) which activates MITF-dependent gene expression (García-Borrón et al., 2014). MITF-dependent gene expression is important for baseline pigmentation and the ultraviolet (UV) tanning response (Abdel-Malek et al., 2008; Bang & Zippin, 2020; D'Orazio et al., 2006, 2013; García-Borrón et al., 2014; Rouzaud et al., 2006; Walker & Gunn, 2010). However, while MITF regulates the expression of numerous pigment enzymes and melanosome proteins, there is no direct connection between MITF gene expression and melanosomal pH. One study suggested that PKA activity affects melanosomal pH via expression of V-ATPase subunits (Cheli et al., 2009). However, recent studies using methods that measure immediate changes in melanosomal pH did not confirm those results (Zhou et al., 2018). Thus, it appears that MC1R signaling does not directly affect melanosomal pH.

Recently, a second cAMP signaling pathway defined by the soluble adenylyl cyclase (sAC) was found to regulate melanosomal pH (Zhou et al., 2018). In contrast to tmACs, sAC activation does not increase MITF or pigment enzyme expression via PKA (Zhou et al., 2018). Instead, loss of sAC activity leads to an immediate increase in melanosomal pH without requiring new protein synthesis (Zhou et al., 2018). sAC signals through the exchange protein activated by cAMP (EPAC) to regulate melanosomal pH (Zhou et al., 2018). Increased melanosomal pH following sAC inhibition leads to an increase in total melanin levels both in vitro and in mice (Zhou et al., 2018). Consistent with biochemical studies mentioned above, elevation of melanosomal pH following sAC inhibition led to an increase in eumelanin synthesis and a concomitant decrease in pheomelanin synthesis (Zhou et al., 2018). Recent work using LC-MS-based ¹³C-tyrosine tracing confirmed that sAC regulation of melanosomal pH can change mixed melanogenesis within a few hours (Chen et al., 2021). It is presently unknown which melanosomal proteins are affected by sAC activity or whether sAC signaling pathways affect non-melanosomal proteins which then impact melanosomal pH. Interestingly, sAC and EPAC are known to affect trafficking of ion channels to membranes; thus, one possible mechanism of sAC/EPAC control of melanosomal pH is the movement of channels, exchangers, or V-ATPase subunits to the melanosome (Chang & Oude-Elferink, 2014; Lobo et al., 2016; Sheikh et al., 2013; Zhao et al., 2013).

Phospholipid phosphatidylinositol 3,5-bisphosphate [PI(3,5)P₂] is a critical regulator of late endosomes/multivesicular bodies and lysosome membrane trafficking (Hasegawa et al., 2017). The phosphoinositide 5-kinase complex, which is composed of phosphoinositide kinase, FYVE-type zinc finger containing phosphoinositide kinase (PIKfyve), FIG4, and VAC14, synthesizes PI(3,5)P₂ from PI(3)P (Hasegawa et al., 2017). PIKfyve is involved in the fusion of stage 1 melanosomes and lysosomes, and is important for the trafficking of melanosomal cargo and premelanosome protein (PMEL) processing (Bissig et al., 2019; Liggins et al., 2018). Thus, PI(3,5)P₂ is important for the normal development of the melanosome which affects the final pH set point.

Currently, there are no known signaling pathways that directly regulate OCA2, SLC45A2, or SLC24A5; however, multiple signaling pathways are known to affect TPC2 activity. Nicotinic acid adenine dinucleotide phosphate (NAADP) is a naturally produced metabolite generated by CD38 (Guse & Diercks, 2018), a membrane-bound receptor that is thought to respond to many different ligands (Guse & Diercks, 2018; Malavasi et al., 2008). NAADP induces a potent release of Ca²⁺ from the endolysosomal system (Pitt et al., 2016). Although the exact mechanism is not established, evidence suggests that NAADP binds to and activates TPC2 to release Ca²⁺ (Pitt et al., 2016); thus, NAADP may acidify melanosomal pH. Consistent with that hypothesis, Ned-19, an antagonist of NAADP at TPC2, can inhibit TPC2 in melanocytes and raise melanosomal pH and melanin levels (Ambrosio et al., 2016). Another regulator of TPC2 is PI(3,5)P₂. As mentioned above, this lipid is essential for the development of the melanosome. In addition, this lipid signaling molecule appears to directly induce TPC2-dependent Na⁺ conductance; therefore, PI(3,5) P₂ may also induce melanosomal acidification (She et al., 2019; Wang et al., 2012). TPC channels are also activated by decreases in intracellular ATP triggered by depletion of extracellular nutrients via the mechanistic target of rapamycin (mTOR)-dependent pathway (Cang et al., 2013). Although TPC2 is required for mTOR-induced calcium release, the mechanism of mTOR regulation of TPC2 is not established (Ogunbaya et al., 2018). Vascular endothelial growth factor (VEGF) signaling is important for angiogenesis and requires TPC2-induced calcium release. VEGF is thought to activate TPC2 by inducing NAADP production (Favia et al., 2014). TPC2 activity is also regulated by multiple kinase cascades, such as p38/JNK and mitogen-activated protein kinase (MAPK) (Jha et al., 2014). Thus, while the mechanisms are not firmly established it appears that multiple signaling pathways converge on TPC2 with the potential to regulate melanosomal pH and mixed melanogenesis (Table 1).

4 ∣. FINAL THOUGHTS/CONCLUSIONS

Two types of melanins production, eumelanogenesis and pheomelanogenesis, have been identified and characterized. Mechanisms for each pathway at the monomer level have been clarified using biosynthetic methods and pulse radiolysis approaches. The biosynthetic methods have the strength of isolating various monomeric and oligomeric melanin intermediates, while the pulse radiolysis method is able to follow rapid reactions involving DQ. Some unsolved problems in the chemistry of mixed melanogenesis remain: (1) the nature of post-polymerization modifications of EM and PM (such as photo and thermal modifications of EM and PM; Ito et al., 2018; Jarenmark et al., 2021); and (2) the nature of copolymerization of EM and PM. The biological functions of melanin are closely related to their structural features. In this review, we have detailed how the process of mixed melanogenesis can be interpreted and showed some mechanisms for the chemical and biochemical control of skin pigmentation. One of the 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 melanosome. Another unsolved problem includes the role of pH in regulating mixed melanogenesis. New methods, such as LC-MS tyrosine fate tracing, will be helpful in revealing the impact of real-time changes in melanosomal pH on melanin synthesis. Furthermore, it is important to examine how cellular signaling impacts melanosomal pH and establish which melanosomal proteins are required. In conclusion, the great diversity in normal human skin pigmentation appears to stem from a combination of polymorphisms in several melanocyte proteins and a balance of multiple biochemical and cellular signals affecting melanin synthesis.