Biosynthesis of Melanin PrecursorsMelanocytes produce two chemically distinct types of melanin pigments: dark-colored

brown-black, insoluble eumelanin derived from the oxidative polymerization of

dihydroxyindolequinones that can be found in almost every type of human skin

and the light-colored red-yellow, alkali-soluble, sulfur-containing pheomelanin derived

from cysteinyldopa which is abundant mostly in fair-skinned persons with

red hair (15). The chemistry and enzymology of the biosynthetic pathway involved

in the synthesis of the monomers that make up eumelanin and pheomelanin are

now well understood and thought to be synthesized in vitro from tyrosine through

the Raper-Mason enzymatic pathway (1618) (Fig. 3). The conversion of tyrosine

to melanin is a complex series of reactions. Tyrosinase, a key enzyme has at least

two functions. It converts tyrosine (that has been selectively transported into the melanosome by a specific melanosome membrane protein (19) to dopa and then to

dopaquinone. Two soluble forms, an insoluble form (bound to ER and Golgi apparatus),

and a melanosome-bound form of tyrosinase are described as T1, T2, T3, and

T4, respectively. Of these, 90% is the T3 form. These forms differ slightly in molecular

weight possibly because of differences in glycosylation. Presumably, these are

successive forms of each molecule of tyrosinase. Interestingly, tyrosinase has been

used as a model to study protein processing and fate (20).

Tyrosinase, with two copper atoms and oxygen at its active site, binds a monohydric

phenol (e.g., tyrosine) at its active site, which, by successive rearrangements,

is converted to an orthoquinone (dopaquinone) and a deoxy enzyme. Orthoquinones

are known for their ability to react with nucleophiles such as thiol or amino

groups. The presence of an amino group in dopaquinone permits cyclization to

yield cyclodopa, a transient product, due to rapid redox exchange yielding dopachrome

(DC). DC can undergo spontaneous decarboxylation to yield DHI, which

could then lead to generation of indolequinones. DHICA that is generated from DC

by a tautomerase (TRP-2) or by metal ions is less easily oxidizable than DHI and

probably requires the enzyme TRP-1 to be converted to an indolequinone. Uncolored

indolequinones from DHI and DHICA then react with each other to generate

the colored polymer. Eumelanins are derived from DHI and DHICA, whereas

pheomelanins and trichromes are derived from cysteinyldopa/glutathionyldopa

derived benzothiazine units (Fig. 3). Melanogenic inhibitors have been reported but

not well characterized. It is likely that these molecules have fine-tuning effects

on the constitutive and facultative pathways. Some investigators speculate that all

intermediates being reactive could participate in the formation of the melanin polymer.

It is for this very reason that we still do not have a structure for melanineach

time a polymer is made, it is different from the one made before!

Regulation of tyrosine substrate availability in the melanosomes has also

been postulated as a possible mechanism to influence the rate of melanin synthesis.

Indeed, the cloning of a specific tyrosine transporter and its localization to the

melanosome membrane is supportive of this notion (19,21). A cysteine transporter

has also been identified in melanocytes and is thought to influence the switch to

pheomelanin synthesis (22). Peroxidase and -glutamyl transpeptidase have all

been implicated in the control of the type of melanin synthesized and in polymerization

of oligomers (23). Some investigators believe that the activity of tyrosinase

determines whether eumelanin will be formed or pheomelanin will be formed. A

high tyrosinase activity results in eumelanin and a low tyrosinase activity results

in pheomelanin.

Control of Melanin Polymerization

Although the preliminary steps in melanin synthesis are well characterized, the

structure, composition, and polymerization/aggregation of the melanin monomers

that lead to color remain poorly understood. This is partly because of its amorphous

physical naturemature melanin is composed of a combination of DHI, DHICA,

and, possibly, other monomers as well, with great variability in the amount of these

two precursors resulting in a featureless absorption spectrum (24). At the melanosomal

pH of 5 or less, the synthesis of a colored monomer does not take place in

vitro under controlled conditions (see below). A black pigment is easily formed

even without tyrosinase at pH 7 and above. It was clear that a mechanism had to

be defined for the conversation of uncolored monomers to a colored polymer at

pH 5. There is one report in literature that ascribes a DHICA polymerase function

Biology of Skin Pigmentation and Cosmetic Skin Color Control 69

to the product of the gene pmel17 (25). We also knew from literature that CVs (see

Melanosome Biogenesis) were acidic and had an active tyrosinase and melanin

monomers but no polymer. We hypothesized that melanosomal proteins played a

key role in polymerization by providing local alkaline microenvironments in an

overall acidic milieu. We showed that melanosomal proteins nonspecifically promoted

polymerization in an acidic milieu possibly by proton abstraction by side

groups of amino acids (dimerization of monomers: the initiation of polymerization

requiring abstraction of protons) (26). The pH of the melanosome is known to fall

further with progressive polymerization, and this, we believe, is an additional feedback

loop that controls both tyrosinase activity and polymerization. The role of free

radicals in polymerization cannot be excluded acting independently/concurrently

or as part of the above-proposed mechanism.

We have further shown that proteins, depending on the nature of melanin

formed and its interaction with protein, keep melanin in a soluble or insoluble form.

We have thus hypothesized that initiation of polymerization and binding to proteins

are independent but spatially and temporally related steps in polymerization.

Also, the charges on proteins could aid in increasing the concentration of monomers

locally, thus aiding polymerization. Additionally, we have shown that the two forms

of melanin, soluble and insoluble, can be interconverted and do exist in cultured

cells and that soluble melanin is more reactive than the insoluble form (12). Finally,

we have speculated on why the pH of the melanosome is acidic. Based on molecular

weight determinations of melanoproteins by gel permeation chromatography,

we have shown that at an acidic pH, more melanins are protein bound and that

protein-bound melanin is also less reactive (27). We speculate that nature maintains

the pH of the melanosome below 5, at considerable energy cost, because at

this pH, tyrosinase is less active, and melanin polymerization occurs intimately

bound to proteins and not in the cytoplasm. Thus, reactivity of melanin is contained

by this interaction. Containment of reactive species is easier when they

are bound to protein because of diffusional constraints imposed by the high molecular

weight of the melanoprotein complex. Further, because we have found

protein-bound soluble melanin to be more reactive, we speculate that from uncolored

monomers, soluble melanin is first formed, which, after subserving its function,

is converted to an insoluble form and deposited on the melanosomal matrix. It is

still unclear what functions reactive intermediates subserve. Clearly, if proteins play

a role in polymerization, they probably determine the length, nature of branching,

and size of the polymer, which, in turn, is expected to influence the UV-absorptive,

free radical-quenching, and light-scattering properties of melanin and consequently

constitutive and facultative pigmentation. We must question the general view that

melanin, once made, is an unchangeable polymer. Perhaps during polymerization,

there is a window when monomers are exchangeable (just as the two forms of melanin

we have described are interconvertible). We also assume that no proteins other

than those that came with the melanosome when it was transferred to keratinocytes

play a role. Perhaps as evolution progressed, keratinocytes developed a mechanism

to incorporate proteins into melanosomes that could affect the nature of melanin.

Such intrakeratinocyte incorporation of proteins into melanosomes has not been

described.

As alluded to earlier, most investigations study tyrosinase-mediated reactions

at pH 6.8 (at which the enzyme is most active) instead of the melanosomal pH of 5

or below. At pH 6.8, there is a considerable lag phase before reaction products can be

detected. This lag phase is seen only if tyrosine is used as substrate and not if dopa

70 Loy and Govindarajan

is used. In fact, the lag phase can be abolished by the addition of dopa, which thus

raises the question of where does the first molecule of dopa come from? Devi et

al. (28) have argued that a lag phase can actually be indefinite, and thus tyrosinase

would have no activity. They have argued that this lag phase is an artifact of pH 6.8,

but at melanosomal pH 5, the lag phase would not be evident. Recent data suggest

that tyrosinase converts tyrosine directly to DQ, and dopa is generated from this

DQ directly (by action of tyrosinase) or from the redox reaction of cyclodopa to

DC. Although generation of dopa by this mechanism may help to overcome the lag

phase, we propose that all investigations should be carried at the relevant pH (i.e.,

5 or below).

The Molecul