Inhibitors of mammalian melanocyte tyrosinase: In vitro comparisons of alkyl esters of gentisic acid with other putative inhibitors

Article abstract of DOI:10.1016/S0006-2952(98)00340-2

To discover safe and effective topical skin-lightening agents, we have evaluated alkyl esters of the natural product gentisic acid (GA), which is related to our lead compound methyl gentisate (MG), and four putative tyrosinase inhibitors, utilizing mammalian melanocyte cell cultures and cell-free extracts. Desirable characteristics include the ability to inhibit melanogenesis in cells (ic₅₀ < 100 μg/mL) without cytotoxicity, preferably due to tyrosinase inhibition. Of the six esters synthesized, the smaller esters (e.g. methyl and ethyl) were more effective enzyme inhibitors (ic₅₀ ~ 11 and 20 μg/mL, respectively). For comparison, hydroquinone (HQ), a commercial skin 'bleaching' agent, was a less effective enzyme inhibitor (ic₅₀ ~ 72 μg/mL), and was highly cytotoxic to melanocytes in vitro at concentrations substantially lower than the ic₅₀ for enzymatic inhibition. Kojic acid was a potent inhibitor of the mammalian enzyme (ic₅₀ ~ 6 μg/mL), but did not reduce pigmentation in cells. Both arbutin and magnesium ascorbyl phosphate were ineffective in the cell-free and cell-based assays. MG at 100 μg/mL exhibited a minimal inhibitory effect on DHICA oxidase (TRP-1) and no effect on DOPAchrome tautomerase (TRP-2), suggesting that MG inhibits melanogenesis primarily via tyrosinase inhibition. MG and GA were non-mutagenic at the hprt locus in V79 Chinese hamster cells, whereas HQ was highly mutagenic and cytotoxic. The properties of MG in vitro, including (1) pigmentation inhibition in melanocytes, (2) tyrosinase inhibition and selectivity, (3) reduced cytotoxicity relative to HQ, and (4) lack of mutagenic potential in mammalian cells, establish MG as a superior candidate skin-lightening agent. Copyright (C) 1999 Elsevier Science Inc.

Full text of DOI:10.1016/S0006-2952(98)00340-2

Biochemical Pharmacology, Vol. 57, pp. 663–672, 1999.
© 1999 Elsevier Science Inc. All rights reserved.
ISSN 0006-2952/99/$–see front matter
PII S006-2952(98)00340-2
Inhibitors of Mammalian Melanocyte Tyrosinase:
In Vitro Comparisons of Alkyl Esters of Gentisic Acid
with Other Putative Inhibitors
Ernest V. Curto,* Cecil Kwong,* Heino Hermersdo¨rfer,† Hansruedi Glatt,†
Chie Santis,‡ Victoria Virador,‡ Vincent J. Hearing, Jr‡ and Thomas P. Dooley*§
*SOUTHERN RESEARCH INSTITUTE, BIRMINGHAM, AL, U.S.A.; †GERMAN INSTITUTE OF HUMAN NUTRITION,
¨
POTSDAM-REHBRUCKE, GERMANY; AND ‡NATIONAL CANCER INSTITUTE, NATIONAL INSTITUTES OF HEALTH,
BETHESDA, MD, U.S.A.
ABSTRACT. To discover safe and effective topical skin-lightening agents, we have evaluated alkyl esters of the
natural product gentisic acid (GA), which is related to our lead compound methyl gentisate (MG), and four
putative tyrosinase inhibitors, utilizing mammalian melanocyte cell cultures and cell-free extracts. Desirable
characteristics include the ability to inhibit melanogenesis in cells (IC Ͻ 100 ␮g/mL) without cytotoxicity,
50
preferably due to tyrosinase inhibition. Of the six esters synthesized, the smaller esters (e.g. methyl and ethyl)
were more effective enzyme inhibitors (IC ϳ 11 and 20 ␮g/mL, respectively). For comparison, hydroquinone
50
(HQ), a commercial skin “bleaching” agent, was a less effective enzyme inhibitor (IC ϳ 72 ␮g/mL), and was
50
highly cytotoxic to melanocytes in vitro at concentrations substantially lower than the IC for enzymatic
50
inhibition. Kojic acid was a potent inhibitor of the mammalian enzyme (IC ϳ 6 ␮g/mL), but did not reduce
50
pigmentation in cells. Both arbutin and magnesium ascorbyl phosphate were ineffective in the cell-free and
cell-based assays. MG at 100 ␮g/mL exhibited a minimal inhibitory effect on DHICA oxidase (TRP-1) and no
effect on DOPAchrome tautomerase (TRP-2), suggesting that MG inhibits melanogenesis primarily via
tyrosinase inhibition. MG and GA were non-mutagenic at the hprt locus in V79 Chinese hamster cells, whereas
HQ was highly mutagenic and cytotoxic. The properties of MG in vitro, including (1) pigmentation inhibition
in melanocytes, (2) tyrosinase inhibition and selectivity, (3) reduced cytotoxicity relative to HQ, and (4) lack
of mutagenic potential in mammalian cells, establish MG as a superior candidate skin-lightening agent.
BIOCHEM PHARMACOL 57;6:663–672, 1999. © 1999 Elsevier Science Inc.
KEY WORDS. melanogenesis; TRP-1; TRP-2; hydroquinone; arbutin; kojic acid; magnesium ascorbylphos-
phate; cosmeceutical; methyl gentisate; pigmentation
Melanogenesis is the process of production (and subsequent
distribution) of melanin by melanocytes within the skin
and hair follicles [1, 2]. Melanocytes have specialized
lysosome-like organelles, termed melanosomes, which con-
tain several enzymes that mediate the production of mela-
nin. The copper-containing enzyme tyrosinase (a tyrosine
hydroxylase; EC 1.14.18.1) catalyzes the oxidation of the
amino acid tyrosine into DOPAʈ and subsequently DOPA-
quinone. At least two additional melanosomal enzymes are
involved in the eumelanogenesis pathway that produces
brown and black pigments, including TRP-1 (DHICA
oxidase), and TRP-2 (DOPAchrome tautomerase). De-
pending on the incorporation of a sulfur-containing reac-
tant (e.g. cysteine or glutathione) into the products, the
melanogenesis pathway diverges, producing eumelanins or
pheomelanins (amber and red pigments) or both.
Various dermatologic disorders result in the accumula-
tion of excessive levels of epidermal pigmentation. These
hyperpigmented lentigenes include melasma, age spots or
liver spots, and sites of actinic damage (i.e. due to solar
ultraviolet irradiation; [3]). Current therapies are inade-
quate for these conditions. Vitiligo is a hyperpigmentary
condition of unknown etiology, in which melanocyte de-
struction results in regional losses of pigmentation. Poten-
tially, it can be treated by pharmacologic depigmentation of
the surrounding unaffected area of skin. Furthermore, a
global market demand has developed recently for skin-
lightening agents as “vanity” cosmeceutical products, be-
cause lighter skin color is preferred by some dark-skinned
individuals in many countries and races [4, 5]. Unfortu-
nately, several purportedly active agents (e.g. arbutin and
kojic acid, among others) have not been demonstrated yet
to be clinically efficacious when critically analyzed in
carefully controlled studies. The U.S. FDA-approved phar-
§ Corresponding author: Thomas P. Dooley, Ph.D., Southern Research
Institute, 2000 Ninth Ave. S., Birmingham, AL 35205. Tel. (205)
222-6145; FAX (205) 581-2877; E-mail: dooley@sri.org
ʈ Abbreviations: DHICA, 5,6-dihydroxyindole-2-carboxylic acid; DOPA,
3,4-dihydroxyphenylalanine; GA, gentisic acid; HQ, hydroquinone; MG,
methyl gentisate; OTC, over-the-counter; TRP-1, tyrosinase-related pro-
tein-1; TRP-2, tyrosinase-related protein-2; FAB, fast atom bombardment;
and MAP, magnesium L-ascorbyl-2-phosphate
Received 26 May 1998; accepted 4 September 1998.

664
E. V. Curto et al.
maceutical products containing 2–4% HQ are moderately
efficacious, but HQ is considered to be cytotoxic to mela-
nocytes and potentially mutagenic to mammalian cells
[3–5; this work].
(FAB), and elemental analysis. For comparison, MG was
also obtained commercially from the Apin Chemicals Co.
Compound Testing in Cultured MelAb Cell Assays
We previously discovered that MG, the methyl ester of
GA (2,5-dihydroxybenzoic acid), is an inhibitor of melanin
accumulation in a murine melanocyte cell culture primary
screen [6]. GA is a natural product from the root of the
genus Gentiana, named after Gentius, an Illyrian (Greco-
Roman) king of the 2nd century B.C., said to have first
discovered the medicinal properties of the plant [7]. The
sodium salt of GA is thought to be an analgesic and an
anti-inflammatory agent. GA is a ubiquitous metabolite,
produced not only by plants, but also by Penicillium patulum
and Polyporus tumulosus, and is excreted into the urine of
mammals following ingestion of salicylates [8, 9]. We have
not determined yet whether the methyl ester is a natural
product, although we anticipate that it is likely to occur.
MG and GA are structurally similar to HQ (see Fig. 2).
HQ has been reported to induce mutations in Salmonella
and at the hprt locus of treated Chinese hamster V79 cells
[10, 11]. HQ appears to be an important intermediate in the
bioactivation of the carcinogen benzene [12]. The muta-
genic potential of MG and GA is examined here for the
first time, using the same conventional mammalian cell
culture system in which HQ showed mutagenic effects.
A general strategy was described recently by one of us for
the discovery and development of novel topical skin-
lightening products [5]. Desirable skin-lightening agents
inhibit the synthesis of melanin in melanosomes by acting
specifically to reduce the synthesis or activity of tyrosinase,
exhibit low cytotoxicity, and are non-mutagenic. A refined
triple-endpoint methodology is presented here for the
discovery and assessment of such agents. Newly synthesized
alkyl esters of GA and four putative inhibitors of tyrosinase
were screened and compared.
Compound testing in vitro was performed using the protocol
of Dooley and coworkers [6], with minor modifications. The
cellular-based assays were performed in triplicate at least
twice. On day 1, a total of 105 MelAb cells were added to
each well of the 24-well plates, and were incubated at 37°
in 5% CO2. On day 2, concentrated stock solutions were
prepared of test compounds at 20 mg/mL in “vehicle” (50%
propylene glycol, 30% ethanol, 20% water), followed by
incubation at 55° for 30 min to dissolve the compounds.
Serial dilutions of each stock solution were prepared at
concentrations ranging from 2000 to 0.64 ␮g/mL, in 10%
vehicle/90% PBS, by attenuating the concentrated stocks
10-fold in PBS solution. Solutions were exposed to cells on
day 2 in triplicate wells at final concentrations ranging from
200 to 0.064 ␮g/mL (1% vehicle, 9% PBS, 90% medium
final concentration). On day 4, the cells were examined
using an inverted phase microscope for viability. On day 6
(following 4 days of agent exposure), the assays were
terminated by replacing the medium with PBS.
Determination of Melanin Content
Melanin content was determined using the published pro-
tocol [6] with minor modifications. The cytotoxicity assays
were performed in triplicate at least twice. After washing
the cells with PBS, we decanted the wash by inversion and
emptied the wells by tapping the inverted plates on paper
towels. The cells were lysed with 1 mL of 1 N NaOH and
pipetted repeatedly to homogenize. For analysis, 200 ␮L of
each crude cell extract were transferred into 96-well plates.
Relative melanin content was determined by optical den-
sity at 405 nm using a SpectramaxPlus spectrophotometer.
MATERIALS AND METHODS
Synthesis of the Alkyl Gentisate Series
Quantitation of Adherent Cell Number by Crystal
Violet Staining
Methyl, ethyl, n-propyl, i-propyl, n-butyl, and n-pentyl
esters of GA were prepared by refluxing GA (Aldrich
Chemical Co.) in an excess of the corresponding alcohol
for 2–7 days, using p-toluenesulfonic acid as a catalyst. The
2-methoxyethanol ester of GA was prepared by refluxing
stoichiometric amounts of the methoxyethanol ester and
GA in dioxane for 14 days, again using p-toluenesulfonic
acid as a catalyst. The courses of the reactions were
followed by TLC, quenching the reactions when they
appeared to be stalled. Work-up included evaporation of
the solvent followed by extraction with ethyl acetate,
water, and small aliquots of aqueous sodium bicarbonate
until no GA was visible in the organic phase by TLC. The
desired esters were obtained by drying the ethyl acetate
solutions over sodium sulfate followed by evaporation to
Quantitation of cell number (to determine whether com-
pounds are cytotoxic/cytostatic) was obtained following the
protocol of Dooley et al. [6], except for minor modifications.
After washing the cells with PBS and emptying the wells by
inversion and tapping the inverted plates against paper
towels, 0.5 mL of 0.1% crystal violet in 10% ethanol was
added to each well. After staining the wells for 5 min at
room temperature, the stain was decanted by inversion, and
the plates were rinsed four times in beakers of water. After
rinsing, the plates were emptied once again as described. To
extract crystal violet from the cells, 1 mL of 95% ethanol
was added to each well, and the plates were placed on a
rotating platform for 20ϩ min. For analysis, 100 ␮L of
solution was transferred from each well into appropriate
wells in 96-well plates. To determine the relative number of
cells in each well, we measured the absorption by crystal
1
dryness. Product structures were verified by H NMR, MS

Tyrosinase Inhibitors
665
violet in the solutions at 540 nm using a SpectramaxPlus
spectrophotometer.
Determinations of IC Values
50
A first order exponential equation was fit to the data to
obtain the 50% inhibition concentration (IC50) values for
each assay. The “Physica” program package [13] was run on
a Silicon Graphics work station. By fitting the equation I ϭ
I0 ϩ I1exp(-C/Ci) to each of the data sets, the IC50 values of
each titration series were obtained. By definition: I ϭ
intensity of measured variable at compound concentration
Tyrosinase Enzyme Assay
To screen compounds for potential to inhibit mammalian
tyrosinase, cell lysates were prepared and used according to
the protocol by Hearing [1] with modifications. The en-
zyme activity determinations were performed in duplicate
at least twice. Approximately 106 MelAb (immortalized
mouse melanocyte) cells were detached from a 10-cm
culture plate using a cell scraper, and then were centrifuged
and resuspended in 6 mL PBS. The cells were lysed by
freeze–thaw cycling the tube five times, by transferring it
between a dry-ice ethanol bath and a water bath at room
temperature. After the final thawing, the cell suspension
was disrupted using a glass homogenizer, and lysis was
confirmed using a microscope. Samples of about 1.5 mL
were aliquoted and stored at Ϫ80° until required.
C; I0 ϭ estimated background intensity constant; I1
ϭ
estimated total intensity change in measured variable; I0 ϩ
I1 ϭ maximum total variable intensity (the negative
control intensity); C ϭ compound concentration; and Ci ϭ
estimated compound inhibition constant, such that Ci
ln(2) ϭ IC of the compound. Mean deviations of each
50
data point were calculated along with standard deviations
of the various IC values.
50
Melanogenesis Enzymatic Profile
One day prior to compound testing, we prepared 1 L of
“Wash” solution (0.1 N HCl), 1 L of “Stop” solution (1 mM
L-tyrosine in 0.1 N HCl), and 100 mL of vehicle solution
(50% propylene glycol, 30% ethanol, 20% water). Serial
dilutions of test compounds were prepared at 5ϫ stock
concentrations ranging from 1000 to 1.6 ␮g/mL (in 5%
vehicle, 95% PBS), and stored at -20° overnight. On the
day of the assay, we prepared 100 mL of “Activator”
solution (0.5 mM l-DOPA in PBS) and 5 mL “Substrate”
solution (125 ␮L of 0.5 mM [U-14C]L-tyrosine at 50
␮Ci/mL in 0.1 N HCl added to 4.875 mL of 0.5 mM
L-tyrosine in PBS). A single 10-cm culture plate of MelAb
cells provides sufficient cell-free “tyrosinase” extract to Test
16 compounds in two 96-well plates.
Tyrosinase assays were performed in 96-well plates, with
10 ␮L of “Inhibitor” per well. For the negative control, 30
␮L of PBS was added, and for the positive control, 10 ␮L of
5% vehicle/95% PBS. After blending 20 ␮L of “Tyrosinase”
solution into all wells (except for the PBS control) by
repeated pipetting, these mixtures were preincubated for 30
min at room temperature. Enzymatic reactions were started
by adding 10 ␮L of “Substrate” solution and 10 ␮L of
“Activator” solution into all wells, yielding final concen-
trations ranging from 200 to 0.32 ␮g/mL (in 1% vehicle/
99% PBS). The plates were incubated at 37° for 2 hr, and
terminated by adding 100 ␮L “Stop” solution to each well.
The reactions were transferred to a Zeta Probe nylon
blotting membrane (Bio-Rad), in a 96-well Schleicher &
Schuell dot blotter, reinforced below with 2 sheets of
Whatman 3MM filter paper. Solutions were allowed to
bind for 15 min at room temperature, before vacuum
filtration through the membrane until dry. Wells were
washed three times with 250 ␮L of “Stop” solution. The
membranes were washed within a plastic box, in 100 mL of
“Wash” solution 3 ϫ 20 min, rotating gently on an orbit
shaker, and were air-dried overnight. Then, the 14C decay
was monitored using an AMBIS plate reader for 2 hr, and
the results were quantified using AMBIS software.
To assay the mechanism of action of MG against the three
major melanogenesis enzymatic activities (i.e. tyrosinase,
DOPAchrome tautomerase, and DHICA oxidase), melano-
cyte cell extracts were prepared from mouse Melan-a cells
(from which the MelAb line was derived), provided by Dr.
D. Bennett [14]. The cellular lysates were prepared using
NP-40 and SDS, and all enzymatic assays were performed at
pH 6.8, 37o for 60 min. The test compounds (MG and GA)
were solubilized in ethanol. Various melanogenesis enzy-
matic assays were performed (listed below). The multi-
enzyme profile characterization of MG was performed at
100 ␮g/mL (595 ␮M) for each assay, a concentration that
is significantly higher than the known IC for depigment-
50
ing activity on MelAb cells [6], equivalent to our arbitrary
threshold for activity in cultured cells [6], and similar to the
reported apparent Km(600 ␮M) for tyrosine as the substrate
on mouse tyrosinase [15].
(A) TYROSINASE. Tyrosine hydroxylase activity was de-
termined by two methods. The first monitored the produc-
tion of tritiated H2O from the hydroxylation of L-[3,5-
3H]tyrosine (Dupont-NEN) to DOPA [1]. This is referred
to as the “direct” method. A second and “indirect” method
to determine tyrosinase activity utilized the [14C]tyrosine
(Dupont-NEN) incorporation assay [1], described above for
compound screening.
(B) DOPACHROME TAUTOMERASE (TRP-2). Inhibition
of enzyme activity by 100 ␮g/mL of MG was assessed using
DOPAchrome as the substrate. The disappearance of the
substrate and the production of DHICA were monitored by
HPLC [16]. DOPAchrome was prepared using the silver
oxide method [17], and DHICA as a standard was provided
by Professor S. Ito.
(C) DHICA OXIDASE (TRP-1). Inhibition of enzyme
activity was determined using the method of Kobayashi and

666
E. V. Curto et al.
coworkers [18], with immunopurified DHICA oxidase and
100 ␮g/mL of MG.
Mammalian V79 Mutagenicity Assays
Chinese hamster V79 cells (subclone V79-MZ) were main-
tained in Dulbecco’s modified Eagle’s medium (DMEM)
supplemented with 5% fetal bovine serum, penicillin (100
U/mL), and streptomycin (100 ␮g/mL), at 37° in a humid-
ified atmosphere containing 5% CO2. The protocol used for
determination of mutagenicity was identical to that used
routinely in previous studies [10, 11, 19–21]. A total of
1.5 ϫ 106 cells and 30 mL medium were added to each
15-cm Petri dish. After 18 hr, the test compound was added
using 60 ␮L DMSO as the solvent. The exposure was
terminated by changing the medium 24 hr later. After an
expression period of 6 days with one subculture, cells were
replated at a density of 1 ϫ 106 per 15-cm Petri dish in
medium containing 6-thioguanine (7 ␮g/mL) for the selec-
tion of mutants (six dishes), or at a density of 100
cells/9-cm Petri dish in medium without 6-thioguanine for
the determination of the cloning efficiency (three dishes).
The cultures were fixed and stained, and the colonies were
counted after about 10 days. Mutant frequencies were
calculated (6-thioguanine-resistant colonies per 1 ϫ 106
plated viable cells). For the evaluation of the results,
empiric criteria were used. A compound was considered to
be negative in the test if the mutant frequencies of all
treatment groups were less than the mean value of the
solvent control cultures plus 10 ϫ 10Ϫ6. It was considered
positive if the mutant frequency was at least three times
above the value of the solvent control cultures and at least
FIG. 1. Mammalian V79 cell mutagenicity results. The frequen-
cies of mutations (number of colonies ؋ 10؊6) at the hprt locus
in Chinese hamster V79 cells are plotted versus the ␮g/mL
concentrations of the test agents (HQ, ●; MG, ; and GA, E).
Concentrations were titrated in cultures of 1 ؋ 106 cells to the
point of cytotoxicity. Frequencies above 20 ؋ 10؊6 are consid-
ered mutagenic. This titration is representative of three exper-
iments.
required to inhibit melanogenesis in cultured cells [6]. This
concentration is equivalent to the apparent Km for tyrosine
as a substrate on mouse tyrosinase (600 ␮M) [15]. Tyrosi-
nase activity was determined by two different radiochemi-
cal methods [1], either “directly” using [3H]tyrosine in a
tyrosine hydroxylase assay, or “indirectly” using the [14C]ty-
rosine incorporation assay. In repeat experiments, this
enzyme from mouse melanocyte extracts (purified enzyme)
was inhibited effectively by 100 ␮g/mL of MG, as expected,
whereas at the same high dose DOPAchrome tautomerase
was not affected, and DHICA oxidase was inhibited only
minimally (retained 76% of control). At this concentra-
tion, tyrosinase activity was reduced to less than 25% of
vehicle control in numerous experiments (data not shown).
Unlike the activity observed for MG, GA was ineffective
20 ϫ 10Ϫ6
.
The cytotoxicity in the mutagenicity test was determined
by counting the cells harvested at the subcultivation during
the expression period [20]. When the cell density was
optically indistinguishable from that of the solvent control,
the cells were not always counted. From numerous other
experiments we estimate that the cell numbers are at least
85% of the control value. The cells also were not counted
when their density was so low that an appropriate determi-
nation of the mutant frequency was precluded. This corre-
sponds to a cell number of less than 2–5% of the control
value.
(IC Ͼ 600 ␮g/mL) as a tyrosinase enzyme inhibitor (data
50
not shown).
Mutagenicity Assays
HQ, MG, and GA were tested in mammalian V79 cell
mutagenesis assays, and the results are presented in Fig. 1.
The standard positive control compound, anti-chrysene-
1,2-dihydrodiol-3,4-oxide, showed strong mutagenic ef-
fects. This potent control produced a mutant frequency of
825 ϫ 10Ϫ6 at a concentration of 3.6 ␮M, which was
similar to the effects observed previously [19]. An unam-
biguous mutagenic effect was detected with HQ as ex-
pected, although the effect was somewhat weaker and
required higher concentrations than those in the previous
studies [10, 11]. Under the same conditions, the mutant
frequencies in the MG- and GA-treated cultures (2–13 ϫ
10Ϫ6) remained within the range observed in the control
cultures (2–15 ϫ 10Ϫ6). In addition, both compounds were
RESULTS
Enzymatic Specificity of MG
To demonstrate that MG is a tyrosinase inhibitor and to
discern whether any other of the key melanosomal enzymes
implicated in the production of melanin were the molecular
targets of MG action, biochemical assays were performed on
each of the key enzymes either from cellular extracts or
from purified proteins. Three enzymes (tyrosinase,
DOPAchrome tautomerase, and DHICA oxidase) were
profiled for inhibition by MG at a high concentration of
100 ␮g/mL (595 ␮M), at levels substantially higher than

Tyrosinase Inhibitors
667
approximately forty times more toxic than MG in these
mammalian cells, in addition to being highly mutagenic,
whereas MG was not mutagenic at any concentration.
Purported Tyrosinase Inhibitors
We tested four commercially available, putative skin-
lightening agents and compared them with MG in our
three primary screens. Structures of these compounds are
shown in Fig. 2. The third column of Table 1 presents
tyrosinase enzyme IC values for these compounds. Kojic
50
acid and MG were the most potent “free” enzyme inhibi-
tors, with IC values of ca. 6 and 11 ␮g/mL, respectively.
50
HQ and arbutin were relatively poor enzyme inhibitors,
with IC values of about 72 and 149 ␮g/mL, respectively.
50
Magnesium ascorbylphosphate (MAP) did not inhibit ty-
rosinase enzyme in these assays, even at 200 ␮g/mL.
Column 4 of Table 1 presents IC50 values for inhibition of
cultured MelAb melanocyte pigmentation (due to both the
amounts of melanin synthesis and total cell number) for
these same compounds. HQ was a very effective agent with
an IC of about 1 ␮g/mL, albeit due to cytotoxicity. Here
50
MG was an effective pigmentation inhibitor, with an IC50 of
about 31 ␮g/mL, falling within the range of numerous prior
similar experiments (range of ca. 15–60 ␮g/mL). The other
compounds in this series did not inhibit pigmentation in
intact melanocytes at concentrations up to 200 ␮g/mL.
Column 5 of Table 1 shows MelAb cell cytotoxicity IC
50
values for these compounds. MG was relatively non-toxic,
FIG. 2. Putative topical skin-lightening agents: Chemical struc-
tures of putative, topical skin-lightening agents. Analogs of
gentisic acid (GA), R ؍ H, include the alkyl ester series: R ؍
CH3, methyl; CH2CH3, ethyl; (CH2)2CH3, n-propyl;
CH(CH3)2, i-propyl; (CH2)3CH3, n-butyl; (CH2)4CH3, n-
pentyl; (CH2)OCH3, methoxy-ethyl. Hydroquinone (HQ) is a
skin bleaching agent, and arbutin is a glycosylated “prodrug”
form of hydroquinone. Kojic acid is a copper-chelating agent.
Magnesium L-ascorbyl-2-phosphate is an antioxidant.
with an IC of cytotoxicity/cytostasis of about 119 ␮g/mL
50
on MelAb cells. HQ was extremely toxic; it had an IC
50
value of about 3 ␮g/mL in this assay. The other commercial
compounds were non-toxic to MelAb cells up to 200
␮g/mL.
Based upon our experience with over 150 compounds in
this assay system [5, 6; this work; and unpublished results],
we arbitrarily define a potentially efficacious skin depig-
mentation candidate agent as one that inhibits tyrosinase
substantially less cytotoxic than HQ. The cytotoxic LC
50
enzyme with an IC of Ͻ25 ␮g/mL, inhibits melanocyte
values of HQ, MG, and GA in V79 cells were approxi-
mately 2.2, 134, and 771 ␮g/mL (or 20, 800, and 5000 ␮M,
respectively). Concentrations that produced Ͻ5% cell
viability were not included in the analysis. Thus, HQ was
50
cell pigmentation with an IC Ͻ 100 ␮g/mL, and is
50
non-cytotoxic to cells with an IC Ͼ 100 ␮g/mL. The
50
threshold criteria are based upon: (1) our published stan-
TABLE 1. In vitro assessments of putative skin-lightening agents
Relative mass
[MG]
Cell-free tyrosinase
Melanocyte pigmentation
Melanocyte cytotoxicity
Compound
IC50 (␮g/mL)
IC50 (␮g/mL)
IC50 (␮g/mL)
Assessment
Desired agent
HQ
<25
72 Ϯ 25
11.2 ؎ 4
6.2 ؎ 2
149 Ϯ 41
Ͼ200
<100
1.1 ؎ 0.2
30.9 ؎ 5
Ͼ200
>100
3.1 Ϯ 0.7
118.7 ؎ 12
>200
Effective
Toxic
Effective
Ineffective
Ineffective
Ineffective
0.655
1.000
0.845
1.62
MG
Kojic acid
Arbutin
MAP
Ͼ200
>200
1.68
Ͼ200
>200
Experimental in vitro evaluations of putative skin-lightening agents, using tyrosinase enzyme, melanogenesis, and cytotoxicity assays, are displayed along with an overall assessment.
The relative mass of each compound is denoted with respect to MG. Based on the characteristics of MG and our experience with Ͼ150 compounds, we designate a desired
candidate as one that inhibits by 50% (IC50) mammalian tyrosinase enzyme at concentrations Ͻ 25 ␮g/mL, inhibits pigmentation in mammalian melanocytes Ͻ 100 ␮g/mL, and
exhibits low cytotoxicity at concentrations Ͼ 100 ␮g/mL. Magnesium ascorbyl phosphate is denoted as MAP.

668
E. V. Curto et al.
FIG. 3. Cell-free tyrosinase assay: A
representative titration assay of the
inhibition of murine tyrosinase in cell-
free lysates, by alkyl esters of GA.
Tyrosinase enzyme converts [14C]L-
tyrosine into [14C]L-DOPA and sub-
sequently into [14C]L-DOPAquinone.
The labeled reaction products bind to
a nylon filter (in a 96-well dot blotter
apparatus), while unconverted ty-
rosine is washed away. Relative prod-
uct (bound
[
14C]L-DOPAquinone)
was measured using an AMBIS radio-
metric plate reader, and results are
displayed as percent of vehicle control
(lacking testing agent) as a function of
inhibitor concentration (100% repre-
sents ca. 2400 counts in 2 hr per
well). The results of duplicate samples
are expressed as the average ؎ range,
plotted using SigmaPlot. These exper-
iments were repeated twice.
dard (i.e. 100 ␮g/mL) for cell-based assays [6]; (2) a more
stringent standard for free enzyme (i.e. 25 ␮g/mL), because
the agents must pass through two cellular membranes to
inactivate the enzyme within melanosomes in cultured
cells, and optimal uptake is not necessarily assured; and (3)
our general experience in screening many phenolic com-
pounds in the various assays. Of the compounds tested in
this current study, only MG satisfied all of these require-
ments. Based on these multiple in vitro endpoints, it is the
best candidate skin-lightening agent that we are aware of
from the literature (and this work), and its behavior in
these assays establishes the standards for further rational
drug design to identify additional comparable or better
active agents. HQ is highly toxic and mutagenic [this work;
6, 10, 11, 22]. Other purportedly active compounds tested
in this series were non-efficacious. They did not act on
MelAb cells or the enzyme, or both. Our assessments of
these compounds are denoted in the last column of Table 1.
Inhibition by Alkyl Esters of GA
Having established that MG is the superior candidate
among the purported skin-lightening agents, we sought to
expand upon our understanding of its mechanism by per-
forming a limited structure–activity relationship (SAR)
study. Results for a representative cell-free screen of tyrosi-
nase inhibition by our alkyl ester series are presented in Fig.
3. The figure shows a systematic diminution in the ability to
inhibit tyrosinase by gentisate compounds with increasing
size of the alkyl ester moiety.
Table 2 presents results for the alkyl esters in all three
assays. The values given are averages for two separate
titration assays. Column 3 lists tyrosinase IC values of
these compounds, obtained using our cell-free enzymatic
assay. Gentisate compounds with smaller ester moieties
inhibited the enzyme more effectively than those with
larger ester moieties. MG was the most effective inhibitor of
50
TABLE 2. In vitro assessments of alkyl esters of GA
Relative
mass
Cell-free tyrosinase
Melanocyte pigmentation
Melanocyte cytotoxicity
R group
IC50 (␮g/mL)
IC50 (␮g/mL)
IC50 (␮g/mL)
Methyl
1.00
1.08
1.17
1.17
1.25
1.33
1.26
11.2 ؎ 4
19.6 ؎ 7
28.5 Ϯ 12
57.5 Ϯ 21
128 Ϯ 53
Ͼ200
30.9 ؎ 5
28.7 ؎ 5
15.6 ؎ 2
27.8 ؎ 8
12.4 ؎ 4
15.6 ؎ 2
43.7 ؎ 8
118.7 ؎ 12
65.9 Ϯ 20
56.5 Ϯ 10
62.7 Ϯ 9
60.1 Ϯ 10
56.5 Ϯ 10
186.5 ؎ 8
Ethyl
n-Propyl
i-Propyl
n-Butyl
n-Pentyl
Methoxy-ethyl
100 Ϯ 29
Experimental in vitro evaluations of a series of alkyl esters of gentisic acid as potential skin-lightening agents using tyrosinase enzyme, melanogenesis, and cytotoxicity assays. The
relative mass of each compound is denoted with respect to MG.

Tyrosinase Inhibitors
669
the series, with an IC50 of about 11 ␮g/mL. The ethyl ester
of gentisate also met our criteria for an effective enzyme
inhibitor. The IC50 of the n-propyl ester was 29 ␮g/mL. This
are concomitantly assessed for inhibition of mammalian
tyrosinase [1] using cell-free extracts of MelAb cells. If
candidate compounds prove efficacious and non-toxic by
these criteria, tests can be performed to determine if they
are mutagenic, using a mammalian V79 cell assay, described
here. Further testing using, for example, animal models can
then be performed as warranted. The strategy described
here is the minimum criterion for the discovery of candi-
date skin-lightening agents using in vitro methods.
is half the IC for the branched i-propyl ester, 58 ␮g/mL,
50
suggesting that the active site prefers the unbranched
n-propyl ester shape. The IC of the n-butyl ester was 128
50
␮g/mL, or more than four times the value of the n-propyl
ester, and suggests that the binding site might accommo-
date up to three carbon atoms beyond the ester. Methoxy-
ethyl ester had an IC of 100 ␮g/mL. This is smaller than
We believe that candidate skin-lightening agents should
satisfy at least the following criteria before further testing.
First, candidate compounds ought to be enzyme inhibitors,
preferably with the capacity to coordinate with the coppers
in the active site of tyrosinase. This property is likely to
confer some specificity to tyrosinase, since two-copper
complexes are rare, and copper coordination can also
increase binding and thus improve inhibition. Second,
candidate compounds must exhibit low toxicity. Unfortu-
nately, some compounds with the capacity to coordinate
metals can be very toxic, and must be ruled out. Some
compounds are non-toxic to non-melanocyte cells (e.g.
keratinocytes), but become toxic when acted upon by
tyrosinase and converted into quinones within melano-
cytes. HQ, for example, is more toxic to melanocytes than
non-melanocytes [22], but is also generally toxic and
mutagenic [6, 10, 11; this work]. Arbutin is a prodrug form
of HQ [23], and is much less toxic, but does not sufficiently
inhibit cell pigmentation or enzyme activity by the criteria
presented here. Candidate compounds must actively or
passively traverse two intact membranes at moderate con-
centrations without metabolic bioinactivation in order to
act on living melanocytes. This requirement likely elimi-
nates a large number of otherwise potentially effective
enzymatic inhibitors. For instance, we speculate that kojic
acid [24], a copper chelator with enzymatic inhibitory
activity against the free enzyme, fails to enter the cells.
Some compounds, such as magnesium L-ascorbyl-2-phos-
phate [25], are non-toxic and might work at very high
concentrations, but are not effective candidates by our
criteria. Lastly, visual microscopic inspection of compound
effects on cell morphology is an essential part of the
evaluation of candidate skin-lightening agents, as an indi-
rect index of cytotoxicity. To our knowledge, only MG
(and perhaps ethyl gentisate) satisfies all of these candidacy
criteria for a topical skin-lightening agent.
50
the value for the n-butyl ester, although the two compounds
are essentially the same size. The n-pentyl ester is the
largest compound that we tested in the ester series. It was
not effective as a tyrosinase inhibitor (IC Ͼ 200 ␮g/mL),
50
suggesting a size limit to that part of the active-site cavity.
The differences in mass among this series had little effect
on activity (i.e. when expressed in molar terms). The
largest of these compounds (n-pentyl; 224 Da) is about
one-third more massive than the smallest (methyl; 168 Da),
yet its activity was over 20-fold less.
Column 4 of Table 2 displays MelAb cell pigmentation
IC
values for these gentisate esters. Each compound
50
inhibited melanocyte pigmentation (both synthesis and
total cell number) within the criterion we set for efficacy.
Interestingly, the n-propyl, n-butyl, and n-pentyl esters,
which are less effective tyrosinase inhibitors, were the most
effective agents of the series on cultured cells, with IC
50
values of near 15 ␮g/mL, albeit due to increased cytotox-
icity characteristics. The next most effective depigmenta-
tion agents of the series were the methyl, ethyl, and
isopropyl esters, with IC50 values of about 30 ␮g/mL against
intact melanocytes. Interpretation of the activities of some
of the larger esters is difficult, since many factors likely
contribute to the overall “depigmentation” of melanocyte
cultures (i.e. combination of inhibition of tyrosinase ex-
pression or activity and/or other components in the mela-
nogenesis pathway, cytostasis, and cytotoxicity).
Column 5 of Table 2 presents melanocyte cytotoxicity
results for the seven alkyl esters. Most of the compounds
were cytotoxic, and only MG, with an IC of about 119
50
␮g/mL, and methoxy-ethyl-ester gentisate, with an IC of
50
about 187 ␮g/mL, met our arbitrary criterion for candidacy.
It is interesting that the methoxy-ethyl-ester, which did not
inhibit the tyrosinase enzyme, and was non-cytotoxic by
our criteria, is a depigmentation agent. However, it is not
likely via action on tyrosinase enzyme, and the reason for
this behavior is not known.
Having analyzed more than 150 compounds in detail [5,
6; this work; and unpublished results], it is our general
impression that testing at only a single concentration point,
40 ␮g/mL, is sufficient during the primary cell-based screen.
Compounds that produce a lightening effect at this con-
centration are usually worthy of further study. Agents that
render cells with normal morphology and pigment levels at
this concentration are ineffective, whereas compounds that
are cytotoxic at this concentration are excludable. Except
for MG, which permits cell viability at this concentration,
all other alkyl esters yielded cells with some evidence of
toxicity.
DISCUSSION
In this work we presented a multi-endpoint screening
strategy for the discovery of safe and effective active
ingredients for topical skin-lightening or depigmentation
products. Candidate compounds are assessed in MelAb cell
cultures [6] for their ability to inhibit pigmentation of
melanocytes and to determine their toxicity. Compounds
that inhibit melanocyte pigmentation with low cytotoxicity

670
E. V. Curto et al.
In the present study, we demonstrated that MG can
rodent gavage studies, and has demonstrated some evidence of
carcinogenic activity [26]. The demonstrated toxicity, muta-
genic activity, and poor enzymatic properties of HQ in
mammalian cells in vitro raise questions of concern about its
continued use in topical products, and provide the opportunity
for discovery and development of new safer replacements, of
which MG serves as a candidate for further testing.
How might MG inhibit tyrosinase? Historically, Kubo-
witz [27] first remarked upon a strong similarity between
tyrosinase and hemocyanin. Since that time a number of
studies have been reported regarding a variety of tyrosi-
nases, their mode of action, substrates, and inhibitors of the
enzyme (see, for example, Yasunobu [28] for a classic
review). However, the structure of mammalian tyrosinase is
still unknown, so one must use indirect methods to address
mechanistic issues. After noting that six histidines in
tyrosinase are homologous to six histidines in hemocyanin
(that coordinate two coppers in its putative active site),
Lerch [29] proposed a mechanism of action for tyrosinase.
Maddaluno and Faull [30] also suggested that the two
coppers will chelate competitive inhibitors. Mutagenesis
experiments by Spritz et al. [31] generally support elements
of Lerch’s active site model, with specific differences noted
in the histidines that coordinate the coppers.
inhibit mammalian tyrosinase in pure form, in crude cell-
free extract, in addition to inhibition of melanin formation
in cultured melanocytes. These findings indicate that MG
can enter viable melanocytes at sufficient concentrations to
block tyrosinase activity and thus de novo melanin biosyn-
thesis. MG does not appear to inhibit DOPAchrome
tautomerase (TRP-2) or DHICA oxidase (TRP-1) at con-
centrations higher than needed to inhibit tyrosinase [1, 2].
A limited GA structure–activity relationship study has
been performed, and only MG demonstrated the capability
of blocking pigmentation in viable melanocytes at subtoxic
doses [6]. These properties appear to be unique to the
methyl ester of the natural product GA. Various related
compounds were screened and did not have these charac-
teristics, including GA, HQ, the 3,4-dihydroxy isomer of
MG, and other simple phenolic derivatives [6; and this
work]. The other compounds were either inactive or toxic
in the melanocyte cell-based assays. Furthermore, measure-
ment of the IC of GA in cell-free tyrosinase assays
50
revealed that it was significantly less active as an enzyme
inhibitor compared with MG (data not shown), and we
speculate that the acid form might be less likely to enter
cells than the “blocked” methyl ester derivative.
We confirmed the previously reported mutagenicity and
cytotoxicity of HQ. However, the mutagenic activity was
somewhat weaker than that observed in previous studies
[10, 11]. We suspect that auto-oxidation products might
have caused the effects. Whereas HQ was mutagenic and
cytotoxic against mammalian V79 cells, no mutagenic
effects were detected with either MG or GA. Perhaps the
presence of the double bond of the carbonyl moiety
adjacent to the benzene ring prevents auto-oxidation and
conversion into the respective semiquinones and quinones
(likely to be more toxic and mutagenic), whereas auto-
oxidation occurs more readily with HQ. These mutagenesis
results (a reconfirmation of published results) are signifi-
cant, as HQ has been an approved ingredient for topical
depigmentation products for many years [3]. In view of
these findings (i.e. mutagenesis, cytotoxicity, and poor
tyrosinase enzymatic inhibition by HQ), we speculate that
HQ may have a difficult time receiving pre-market ap-
proval or continued approval during post-market safety
evaluation, if it were being reviewed today by the U.S. FDA
or equivalent regulatory agencies worldwide. Regulatory
authorities seldom approve agents with known deficiencies,
such as mutagenic potential, except for use in life-threat-
ening or acute disease indications. Although topical HQ is
moderately effective and typically does not demonstrate
overt signs of toxicity clinically, it is certainly debatable
whether topical HQ acts as a mutagen on human skin in
vivo, and it is not desirable to intentionally expose viable
tissue to known mutagens if a non-mutagenic alternative is
available. This is especially true for individuals who use
either high dose formulations (Ͼ4% HQ) or expose large
surface areas to topical treatment. Furthermore, the Na-
tional Toxicology Program has examined HQ in long-term
Based on the evidence gathered here, we suspect that
short chain esters of gentisate bind tyrosinase by coordinat-
ing the two coppers in the active site, possibly in the
presence of oxygen, as depicted in Fig. 4. We show the
coppers in octahedral coordination, although the exact
nature of the active-site complex remains to be determined.
The ester subgroup R2 is depicted in steric conflict with
unknown residues in the active site (which we represent
with an arc). This notion is supported by evidence that the
branched i-propyl-ester gentisate is a lesser inhibitor than
unbranched n-propyl-ester gentisate. Size limitations on
the cavity for even the linear ester moieties place some
constraints on R1. This is consistent with our data that
smaller linear esters of GA are better inhibitors than larger.
We are currently examining in detail the biochemical
mechanism of inhibition of tyrosinase by MG (Curto EV
and Dooley TP, unpublished results). Furthermore, it will
also be of interest to determine the extent of metabolism of
MG in cultured cells or viable skin. For instance, it is not
yet known whether MG is bioinactivated by esterases,
sulfotransferases, or glucuronosyl transferases, for which this
compound might serve as a substrate.
Our goal is to obtain additional active and safe topical
agents, preferably with higher potency than MG, and then
demonstrate clinical safety and efficacy in animal models.
However, the dermatology/cosmeceutical field is in need of
relevant and predictive mammalian in vivo efficacy testing
models for skin-lightening/depigmentation formulations
[5]. New animal models could help to verify that MG or
other new compounds can be effective topical skin depig-
mentation agents, prior to clinical testing on human skin.
Of the more than 150 compounds that we have tested thus
far, MG appears to be the best candidate active ingredient.

Tyrosinase Inhibitors
671
cologic action and bioavailability of the agent are limited to
the viable epidermis to reduce the potential of toxicologic
risks. In theory, cosmeceuticals closely resemble OTC
products, with the possible exception of limiting their use
only for vanity purposes (i.e. enhancement of appearance)
as opposed to treatment of disease, as mandated by govern-
mental regulatory agencies. We and others in this field are
advocating for the establishment of either recommended or
mandatory safety assessment guidelines for cosmeceuticals (i.e.
intended for “vanity” uses only), however, without requiring
demonstrations of efficacy (which are necessary for OTC or
prescription drugs) [4]. Inhibitors of tyrosinase, such as MG,
serve as a new paradigm in dermatologic drug or cosmeceutical
agent discovery research, in view of their recently developed
potential and recognition by world markets.
We thank Jutta Schwenk for excellent technical assistance with the
mutagenesis studies, and Robert Reynolds for helpful discussions
concerning medicinal chemistry. This work was supported financially by
the Alabama Power Endowed Chair fund (T.P.D), BMBF Project
22P1434 (H.G.), and the National Institutes of Health (V.J.H.).
References
1. Hearing VJ Jr, Monophenol monooxygenase (tyrosinase):
Purification, properties, and reactions catalyzed. Methods En-
zymol 142: 154–165, 1987.
2. Spritz RA and Hearing VJ, Genetic-disorders of pigmenta-
tion, Adv Hum Genet 22: 1–45, 1994.
3. Frenk E, Treatment of melasma with depigmenting agents.
Melasma: New Approaches to Treatment, pp. 9–15. Martin
Dunitz Ltd., London, 1995.
FIG. 4. Conceptual model of gentisate ester binding to the active
site of tyrosinase: Hypothetical two-dimensional model of initial
docking of gentisate ester in the putative active site of mamma-
lian tyrosinase. The two coppers are assumed in octahedral
coordination with a gentisate ester and oxygen.
It inhibits melanocyte pigmentation at moderate concen-
trations, with minimal cytotoxicity, and specifically inhib-
its tyrosinase enzyme. Furthermore, it is non-mutagenic,
unlike HQ. No other compound that we are aware of in the
literature has all of these desired characteristics, thus
establishing MG as the superior candidate.
4. Dooley TP, Is there room for a moderate level of regulatory
oversight? In: Drug Discovery Approaches for Developing Cos-
meceuticals: Advanced Skin Care and Cosmetic Products (Ed.
Hori W), Chap. 1.4. International Business Communications,
Southborough, MA, 1997.
5. Dooley TP, Topical skin depigmentation agents: Current
products and discovery of novel inhibitors of melanogenesis. J
Dermatol Treat 8: 275–279, 1997.
Once MG-containing topical formulations are developed
and tested in vivo in animal models and/or clinically, what
types of products could be marketed? Pre-market approval
and regulatory compliance issues will have significant
effects on the potential commercial availability of MG-
containing products in many countries. If the product’s
marketing literature or labeling asserts that it alters the
structure or function of skin or makes a specific “therapeu-
tic” claim (e.g. treatment of a dermatologic disorder), this
compound would likely be formulated for use as an OTC or
prescription drug product. However, in selected global
markets, MG-containing products could be sold as cosmetic
or cosmeceutical agents [4, 5]. Cosmeceuticals represent an
emerging concept in dermatologic agent development and
marketing, ideologically somewhere between OTC drugs
and cosmetic products, yet vaguely defined and not formally
recognized by the U.S. FDA [4]. Cosmeceutical products,
unlike cosmetics per se, should contain ingredients that
have been demonstrated in the laboratory to be pharmaco-
logically active. In addition, it is preferable if the pharma-
6. Dooley TP, Gadwood RC, Kilgore K and Thomasco LM,
Development of an in vitro primary screen for skin depigmen-
tation and antimelanoma agents. Skin Pharmacol 7: 188–200,
1994.
7. Morse JL (Ed.), An Abridgment of The New Funk & Wagnalls
Encyclopedia, The Universal Standard Encyclopedia, Vol, 10,
pp. 3662–3663. Unicorn, New York, 1955.
8. Budavari S (Ed.), Gentisic acid, Merck Index, 11th Edn,
Abstract No. 4290, p. 688. Merck & Co., Rahway, NJ, 1989.
9. J-Hua L and Smith PC, Direct analysis of salicylic acid, salicyl
acyl glucuronide, salicyluric acid and gentisic acid in human
plasma and urine by high-performance liquid chromatogra-
phy. J Chromatogr [B] 675: 61–70, 1996.
10. Glatt HR, Padykula R, Berchtold GA, Ludewig G, Platt K-L,
Klein J and Oesch F, Multiple activation pathways of benzene
leading to products with varying genotoxic characteristics.
Environ Health Perspect 82: 81–89, 1989.
11. Glatt HR, Endogenous mutagens derived from amino acids.
Mutat Res 238: 235–243, 1990.
12. La Du BN, Alcaptonuria and ochronotic arthritis. Mol Biol
Med 8: 31–38, 1991.
13. Physica, © 1993–1997, Chuma JL, TRIUMF, 4004 Wesbrook
Mall, Vancouver, BC, Canada V6T 2A3.

672
E. V. Curto et al.
14. Bennett DC, Cooper PJ and Hart IR, A line of non-
tumorigenic mouse melanocytes, syngeneic with B16 mela-
noma and requiring a tumor promoter for growth. Int J Cancer
39: 414–418, 1987.
15. Adachi K and Halprin KM, A sensitive fluorometric assay
method for mammalian tyrosinase. Biochem Biophys Res Com-
mun 26: 241–246, 1967.
16. Tsukamoto K, Jackson IJ, Urabe K, Montague PM and
Hearing VJ, A second tyrosinase-related protein, TRP2, is a
melanogenic enzyme termed DOPAchrome tautomerase.
EMBO J 11: 519–526, 1992.
17. Korner AM and Pawelek JM, DOPAchrome conversion, a
possible control point in melanin biosynthesis. J Invest Der-
matol 75: 192–195, 1980.
18. Kobayashi T, Urabe K, Winder AJ, Jimenez-Cervantes C,
Imokawa G, Brewington T, Solano F, Garcia-Borron JC and
Hearing VJ, Tyrosinase related protein 1 (TRP1) functions as
a DHICA oxidase in melanin biosynthesis. EMBO J 13:
5818–5825, 1994.
19. Glatt H, Pie´e A, Pauly K, Steinbrecher T, Schrode R, Oesch
F and Seidel A, Fjord- and bay-region diol-epoxides investi-
gated for stability, SOS induction in Escherichia coli, and
mutagenicity in Salmonella typhimurium and mammalian cells.
Cancer Res 51: 1659–1667, 1991.
20. Glatt HR and Oesch F, Activation of N-acetoxy- and
N-hydroxy-2-acetylaminofluorene to mutagenic and cyto-
toxic metabolites by V79 Chinese hamster cells. Mutat Res
149: 265–269, 1985.
tyrosinase activity but independent of melanin content.
Pigment Cell Res 1: 386–389, 1988.
23. Maeda K and Fukuda M, Arbutin: Mechanism of its depig-
mentation actions in human melanocyte culture. J Pharmacol
Exp Ther 267: 765–769, 1995.
24. Cabanes J, Chazarra S and Garcia-Carmona F, Kojic acid, a
cosmetic skin whitening agent, is a slow-binding inhibitor of
catecholase activity of tyrosinase. J Pharm Pharmacol 46:
982–985, 1994.
25. Kameyama K, Sakai C, Kondoh S, Yonemoto K, Nishiyam S,
Tagawa M, Murata T, Ohnuma T, Quigley J, Dorsky A, Bucks
D and Blanock K, Inhibitory effect of magnesium L-ascorbyl-
2-phosphate (VC-PMG) on melanogenesis in vitro and in vivo.
J Am Acad Dermatol 34: 29–33, 1996.
26. National Toxicology Program Technical Report Series No.
366, Toxicology and Carcinogenesis Studies of Hydroquinone
(CAS No. 123–31-9) in F344/N Rats and B6C3F1 Mice.
National Technical Information Service, U.S. Department of
Commerce, Springfield, VA, 1989.
27. Kubowitz F, Cleavage and resynthesis of polyphenoloxidase
and hemocyanin. Biochem Z 299: 32–57, 1938.
28. Yasunobu KT, Mode of action of tyrosinase. In: Pigment Cell
Biology (Ed. Gordon M), pp. 583–608. Academic Press, New
York, 1959.
29. Lerch K, Protein and active-site structure of tyrosinase. In:
Advances in Pigment Cell Research (Ed. Bagnara JT), pp. 85–98.
Alan R. Liss, New York, 1988.
30. Maddaluno JF and Faull KF, Inhibition of mushroom tyrosi-
nase by 3-amino-L-tyrosine: Molecular probing of the active
site of the enzyme. Experientia 44: 885–887, 1988.
31. Spritz RA, Ho L, Furumura M and Hearing VJ, Mutational
analysis of copper binding by human tyrosinase. J Invest
Dermatol 109: 207–212, 1997.
21. Hermersdo¨rfer H, Ozierenski B, Schmalix WA, Doehmer J
and Glatt HR, Cell culture system for the controlled intra-
cellular generation of reactive oxygen species. Toxicol In Vitro
11: 569–572, 1997.
22. Smith CJ, O’Hare KB and Allen JC, Selective cytotoxicity of
hydroquinone for melanocyte-derived cells is mediated by