MGRN1 depletion promotes intercellular adhesion in melanoma by upregulation of E-cadherin and inhibition of CDC42.

Mahogunin Ring Finger 1 is an E3-ubiquitin ligase encoded by the color gene MGRN1. Our previous in vitro and in vivo studies demonstrated that Mgrn1 deletion in mouse melanoma cells induced cell differentiation and adhesion, and decreased cell motility and invasion on collagen I, and lung colonization in an in vivo model. Here, we investigated the role of MGRN1 on human melanoma cell morphology, adhesion and expression of genes/proteins involved in an EMT-like transition. We demonstrated that wild-type BRAF human melanoma cells adopted a clustering-like morphology on collagen I, with permanent MGRN1 abrogation resulting in bigger cell clusters. Enhanced intercellular adhesion was mostly mediated by induction of E-cadherin and higher co-localization with ß-catenin. Transcriptional upregulation of E-cadherin likely occurred through downregulation of the ZEB1 repressor. Finally, pulldown assays showed reduced activation of CDC42 in the absence of MGRN1, which was reverted after E-cadherin silencing. Overall, these findings highlight a new MGRN1-dependent pathway regulating melanoma cell shape, motility, and invasion potential.

Biosynthesis of hamster zona pellucida is restricted to the oocyte.

The zona pellucida (ZP) is an extracellular coat that surrounds the mammalian oocyte and the early embryo until implantation. This coat mediates several critical aspects of fertilization, including species-selective sperm recognition, the blocking of polyspermy and protection of the oocyte and the preimplantation embryo. Depending on the species, the ZP is composed of three to four different glycoproteins encoded by three or four genes. These genes have been cloned and sequenced for different species. However, controversy exists about the cell type specificity of the ZP glycoproteins, for which several models have been proposed. Different groups have reported that ZP is produced only by the oocytes, by the granulosa cells or by both cell types, depending on the species under study. We recently described the expression of four ZP proteins in the hamster ovary. By means of the complete set of the hamster ZP cDNAs, we undertook the study of the origin and expression pattern of the four ZP genes. In the present work, the expression of ZP1, ZP2, ZP3 and ZP4 is carefully analyzed by in situ hybridization (ISH) in hamster ovaries. Our data suggest that the four hamster ZP genes are expressed in a coordinate and oocyte-specific manner during folliculogenesis. Furthermore, this expression is maximal during the first stages of the oocyte development and declines in oocytes from later development stages, particularly within large antral follicles.

Hamster zona pellucida is formed by four glycoproteins: ZP1, ZP2, ZP3, and ZP4.

The zona pellucida (ZP) is an extracellular glycoprotein matrix that surrounds all mammalian oocytes. Recent data have shown the presence of four glycoproteins (ZP1, ZP2, ZP3, and ZP4) in the ZP of human and rat rather than the three glycoproteins proposed in the mouse model. In the hamster (Mesocricetus auratus), it was previously described that ZP was composed of three different glycoproteins, called ZP1, ZP2, and ZP3, even though only ZP2 and ZP3 have been cloned thus far. The aim of the study was to determine whether hamster might also express four, rather than three, ZP proteins. The full-length cDNAs encoding hamster ZP glycoproteins 1 and 4 were isolated using rapid amplification cDNA ends (RACE). The cDNA of ZP1 contains an open reading frame of 1851 nucleotides encoding a polypeptide of 616 amino acid residues. The amino acid sequence of ZP1 revealed a high homology with other mammalian species like human (66%), rat (80%), and mouse (80%). The cDNA of ZP4 contains an open reading frame of 1632 nucleotides encoding a polypeptide of 543 amino acid residues. The deduced amino acid sequence of ZP4 revealed high overall homology with rat (82%) and human (78%). Subsequent mass spectrometric analysis of the hamster ZP allowed identification of peptides from all four glycoproteins. The data presented in this study provide evidence, for the first time, that the hamster ZP matrix is composed of four glycoproteins.

Variant amino acids in different domains of the human melanocortin 1 receptor impair cell surface expression.

The alpha melanocyte-stimulating hormone receptor (MC1R) is a heptahelical G protein-coupled receptor (GPCR) found in the plasma membrane of melanocytes. By mediating the melanogenic response to melanocortins, MC1R is a major determinant of mammalian pigmentation. The human MC1R gene is unusually polymorphic. Many loss-of-function alleles have been described, but the molecular basis for their functional impairment remains most often unknown. Here we report a study of two natural MC1R loss-of-function variants, Leu93Arg and Arg162Pro, and two artificial mutants, Cys35Ala and a deleted form missing the last five amino acids in the carboxyl tail. When expressed in HEK 293T cells, those mutants neither bound an iodinated hormone analogue nor elicited cAMP increases in response to saturating doses of a superpotent agonist. Cell surface expression of mutant receptors was dramatically decreased respect to the wild type form, in spite of smaller changes in total protein abundances and intracellular stability. Accordingly, aberrant processing with intracellular retention is the most likely cause of loss-of-function for those mutants. Therefore, mutations in virtually any region of the heptahelical protein, including its extracellular N terminus, a transmembrane fragment, intracellular loops or carboxyl terminal cytosolic extension, seem to compromise normal MC1R processing.

The melanocortin-1 receptor carboxyl terminal pentapeptide is essential for MC1R function and expression on the cell surface.

The pigmentary actions of the melanocortins are mediated by the melanocortin-1 receptor (MC1R), a seven transmembrane domains receptor positively coupled to Gs and the cAMP cascade. In order to define the structure-function relationships of potentially relevant domains in MC1R, particularly its C-terminal cytosolic tail, we generated and analyzed several variants with C-terminal deletions, as well as point mutants in selected residues of the human MC1R. We show that the MC1R C-terminal pentapeptide is essential for proper receptor expression on the plasma membrane, and that residues Thr314, Cys315 and Trp317 are at least partially responsible for this effect.

Anti-inflammatory and anti-invasive effects of alpha-melanocyte-stimulating hormone in human melanoma cells.

Alpha-melanocyte stimulating hormone (alpha-MSH) is known to have pleiotrophic functions including pigmentary, anti-inflammatory, antipyretic and immunoregulatory roles in the mammalian body. It is also reported to influence melanoma invasion with levels of alpha-, beta- and gamma-MSH correlated clinically with malignant melanoma development, but other studies suggest alpha-MSH acts to retard invasion. In the present study, we investigated the action of alpha-MSH on three human melanoma cell lines (HBL, A375-SM and C8161) differing in metastatic potential. alpha-melanocyte-simulating hormone reduced invasion through fibronectin and also through a human reconstructed skin composite model for the HBL line, and inhibited proinflammatory cytokine-stimulated activation of the NF-kappaB transcription factor. However, A375-SM and C8161 cells did not respond to alpha-MSH. Immunofluorescent microscopy and Western blotting identified melanocortin-1 receptor (MC-1R) expression for all three lines and MC-2R on HBL and A375-SM lines. Receptor binding identified a similar affinity for alpha-MSH for all three lines with the highest number of binding sites on HBL cells. Only the HBL melanoma line demonstrated a detectable cyclic adenosine monophosphate (cAMP) response to alpha-MSH, although all three lines responded to acute alpha-MSH addition (+(-)-N(6)-(2-phenylisopropyl)-adenosine (PIA)) with an elevation in intracellular calcium. The nonresponsive lines displayed MC-1R polymorphisms (C8161, Arg (wt) 151/Cys 151; A375-SM, homozygous Cys 151), whereas the HBL line was wild type. Stable transfection of the C8161 line with wild-type MC-1R produced cells whose invasion was significantly inhibited by alpha-MSH. From this data, we conclude that alpha-MSH can reduce melanoma cell invasion and protect cells against proinflammatory cytokine attack in cells with the wild-type receptor (HBL).

Thr40 and Met122 are new partial loss-of-function natural mutations of the human melanocortin 1 receptor.

Activation by melanocortins of the melanocortin 1 receptor (MC1R), expressed in epidermal melanocytes, stimulates melanogenesis. Human MC1R gene loss-of-function mutations are associated with fair skin, poor tanning and increased skin cancer risk. We identified two natural alleles: Ile40Thr, probably associated with skin types I-II, and Val122Met. Val122Met bound [(125)I][Nle(4), D-Phe(7)]-alpha-melanocyte stimulating hormone with lower affinity than the wild-type. Dose-response curves of cAMP accumulation were right-shifted for both forms. The Val122Met form failed to achieve maximal cAMP responses comparable to the wild-type or Ile40Thr receptors. Thus, the Ile40Thr and Val122Met variants are partial loss-of-function natural mutations of MC1R.

Inhibition of melanogenesis in response to oxidative stress: transient downregulation of melanocyte differentiation markers and possible involvement of microphthalmia transcription factor.

H(2)O(2) and other reactive oxygen species are key regulators of many intracellular pathways. Within mammalian skin, H(2)O(2) is formed as a byproduct of melanin synthesis, and following u.v. irradiation. We therefore analyzed its effects on melanin synthesis. The activity of the rate-limiting melanogenic enzyme, tyrosinase, decreased in H(2)O(2)-treated mouse and human melanoma cells. This inhibition was concentration- and time-dependent in the B16 melanoma model. Maximal inhibition (50-75%) occurred 8-16 hours after a 20 minute exposure to 0.5 mM H(2)O(2). B16 cells withstand this treatment adequately, as shown by a small effect on glutathione levels and a rapid recovery of basal lipid peroxidation levels. Enzyme activities also recovered, beginning to increase 16-20 hours after the treatment. Inhibition of enzyme activities reflected decreased protein levels. mRNAs for tyrosinase, tyrosinase-related protein 1, dopachrome tautomerase, silver protein and melanocortin 1 receptor also decreased after H(2)O(2) treatment, and recovered at different rates. Downregulation of melanocyte differentiation markers mRNAs was preceded by a decrease in microphthalmia transcription factor (Mitf) gene expression, which was quantitatively similar to the decrease achieved using 12-O-tetradecanoylphorbol-13-acetate. Recovery of basal Mitf mRNA levels was also observed clearly before that of tyrosinase. Therefore, oxidative stress may lead to hypopigmentation by mechanisms that include a microphthalmia-dependent downregulation of the melanogenic enzymes.

The 5,6-dihydroxyindole-2-carboxylic acid (DHICA) oxidase activity of human tyrosinase.

Melanin synthesis in mammals is catalysed by at least three enzymic proteins, tyrosinase (monophenol dihydroxyphenylalanine:oxygen oxidoreductase, EC 1.14.18.1) and tyrosinase-related proteins (tyrps) 1 and 2, whose genes map to the albino, brown and slaty loci in mice, respectively. Tyrosinase catalyses the rate-limiting generation of L-dopaquinone from L-tyrosine and is also able to oxidize L-dopa to L-dopaquinone. Conversely, mouse tyrp1, but not tyrosinase, catalyses the oxidation of the indolic intermediate 5,6-dihydroxyindole-2-carboxylic acid (DHICA) into the corresponding 5,6-indolequinone-2-carboxylic acid, thus promoting the incorporation of DHICA units into eumelanin. The catalytic activities of the human melanogenic enzymes are still debated. TYRP1 has been reported to lack DHICA oxidase activity, whereas tyrosinase appears to accelerate DHICA consumption, thus raising the question of DHICA metabolism in human melanocytes. Here we have used two different approaches, comparison of the catalytic activities of human melanocytic cell lines expressing the full set of melanogenic enzymes or deficient in TYRP1, and transient expression of TYR and tyr genes in COS7 cells, to demonstrate that human tyrosinase actually functions as a DHICA oxidase, as opposed to the mouse enzyme. Therefore, human tyrosinase displays a broader substrate specificity than its mouse counterpart, and might be at least partially responsible for the incorporation of DHICA units into human eumelanins.

New insights on the structure of the mouse silver locus and on the function of the silver protein.

The melanosomal proteins encoded by the silver locus play important roles in melanogenesis. The human locus yields two proteins, PMEL17 and GP100, by alternative mRNA splicing. The mouse si locus was reported to encode a Pmel17 protein, and later gp87, a GP100 homologue. When we re-examined the products of wild-type and silver-mutant mouse si loci, RT-PCR of wild-type RNA and genomic DNA sequence accounted for gp87 but excluded the occurrence of Pmel17. Analysis of cDNA from the silver (si/si) melanocyte line, melan-si, showed that the pathogenic mutation is a G to A substitution at nt 1808, which yields a premature stop codon and a predicted protein truncated in the C-terminus. This was confirmed by reaction of a specific anti-gp87 antiserum with si/si melanocyte extracts. To further explore gp87 function, we compared the DHICA oxidase activity of extracts from B16, melan-si (heterozygous for the brown mutation and homozygous for the silver mutation) and Cloudman S91 cells (homozygous for the brown mutation), since both TRP1 and gp87 are thought to be involved in DHICA oxidation/polymerization. Cloudman extracts do not oxidize significantly DHICA and its methyl ester, supporting the involvement of native mouse TRP1 in DHICA oxidation. Extracts from B16 and melan-si do not show significant differences for the oxidation of free acid and methylated dihydroxyindoles, indicating that the mechanism is not decarboxylative. Melan-si extracts are very efficient in catalyzing dihydroxyindole oxidation, in spite of being heterozygous for the TRP1 mutation, consistent with a stablin effect for the wild-type gp87 protein. On the other hand, aggregated and degraded forms of that mutant gp87 protein are found in the cytosolic fraction of melan-si, suggesting that misrouting and aberrant processing of the gp87 and tyrosinase may also be related to the high DHICA oxidase activity of these melanocytes.

Regulation of the murine silver locus product (gp87) by the hypopigmenting cytokines TGF-beta1 and TNF-alpha.

The melanosomal proteins encoded by the silver (si, SILV, or PMEL17) locus play important roles in melanogenesis and are actively investigated as targets for melanoma immunotherapy. The human silver locus yields two proteins, gp100 and PMEL17, by alternative splicing of a common mRNA precursor. Mouse melanocytes exclusively express the gp100 orthologue, here termed gp87, thus providing a simpler model with which to study the silver locus products. We have analyzed the effects of [Nle4, D-Phe7]-alpha melanocyte-stimulating hormone (alphaMSH) and two hypopigmenting cytokines, tumor necrosis factor (TNF)-alpha and transforming growth factor (TGF)-beta1, on the expression of gp87 in B16 mouse melanoma cells. TNF-alpha and TGF-beta1 (at saturating doses for 48 hr) decreased gp87 mRNA by 50%. The gp87 protein was almost undetectable by Western immunoblotting after TNF-alpha treatment, but was not affected by TGF-beta1. alphaMSH increased the mRNA and the gp87 protein approximately 2-fold. Moreover, the amount of gp87 was not reduced by TNF-alpha in the presence of the hormone, in spite of a 50%, decrease in its mRNA. Therefore, the levels of mRNA and gp87 protein did not correlate after treatment with the cytokines. Overall, our data suggest that the silver locus product is not regulated exclusively at the transcriptional level, and highlight the importance of still-uncharacterized regulatory translational and/or post-translational events.

Expression of the MC1 receptor gene in normal and malignant human melanocytes. A semiquantitative RT-PCR study.

Alpha melanocyte-stimulating hormone (MSH) and related proopiomelanocortin-derived (POMC) peptides bind to the melanocortin 1 receptor (MC1-R) of mammalian melanocytes and stimulate proliferation and melanogenesis. POMC transcripts and alpha-MSH-like immunoreactivity have been found in melanoma cells and a possible autocrine loop involving MC1-R and POMC-derived products has been proposed. Therefore, the alpha-MSH/MC1-R system plays a major role in the biology of melanocytes, and provides targets for melanoma diagnosis and therapy. However, the relative levels of MC1-R expression in normal melanocytes (NM) and melanoma cells are unknown, and it is still debated whether or not all human melanomas express the MC1-R. We describe a semiquantitative RT-PCR assay for MC1-R expression, using a competition vector generated by deleting 164 bp of the native gene. The competitor was employed to analyse a panel of human melanoma cells, tumour samples, giant congenital nevus cells (CNM) and normal melanocytes (NM). All samples were positive for MC1-R expression, but expression of the receptor gene did not correlate with that of tyrosinase. Expression levels were about 10 and 20 times higher for surgical specimens and cultured melanoma cells, respectively, than for NM, but comparable for CNM and NM. Thus, high MC1-R expression is a frequent event in malignant melanocytes, and might lead to a higher activity of the alpha-MSH/MC1-R system in melanoma cells as compared to normal melanocytes, for equal local concentrations of the hormone or related melanocortins.

The melanogenic system of Xenopus laevis.

Melanin pigments in lower vertebrates are often found in locations other than the skin, thus forming an extracutaneous pigmentary system of unknown function. The cellular and biochemical structure of this system is still poorly characterized. This paper deals with the ultrastructural and biochemical features of the melanogenic system of Xenopus laevis. Melanin containing cells were identified in the dorsal and ventral skin, and in the lung, spleen, liver and connective tissue surrounding blood vessels. The pigment cells in the skin and the lungs appeared to be typical melanocytes. The spleen contained isolated melanocyte-like cells, but most of the pigment cells present in this organ were associated with melanomacrophage centers. Conversely, the liver appeared devoid of melanocytes and only displayed melanomacrophage centers. Tyrosinase activity was found in all pigment-containing organs except the liver. All organs containing tyrosinase activity also displayed melanin formation potential from L-tyrosine. Therefore, tyrosine hydroxylase and melanin formation activities could be detected only in those organs containing typical melanocytes but not in locations such as the liver, where only melanomacrophages centers were found.

Mechanisms of melanogenesis inhibition by tumor necrosis factor-alpha in B16/F10 mouse melanoma cells.

The inflammatory cytokine tumor necrosis factor-alpha (TNF-alpha) is present in the dermal and epidermal layers of normal skin [Kilgus, O., Payer, E., Schreiber, S., Elbe, A., Strohal, R. & Stingl, G. (1993) J. Invest. Dermatol. 100, 674-680]. Its local concentrations are modified by several stimuli, including wound healing and ultraviolet irradiation. Moreover, TNF-alpha inhibits melanogenesis in normal melanocytes [Swope, V., Abdel-Malek, Z., Kassem, L. & Norlund, J. (1991) J. Invest. Dermatol. 96, 180-185], and is, therefore, a potential autocrine/paracrine regulator of pigmentation. We have analyzed the mechanisms of this effect using B16/F10 melanoma cells as a model. Nanomolar concentrations of TNF-alpha inhibit the tyrosine hydroxylase and dopa oxidase activities of B16/F10 melanocytes, to less than 30% control levels, without effects on tyrosinase-related protein 2/dopachrome tautomerase (TRP2/DCT). The 50% inhibition was obtained at 1 nM TNF-alpha and 48 h treatment. The effect of TNF-alpha was noticeable after 6 h treatment, and maximal after 24 h. This inhibition is explained by decreased intracellular levels of tyrosinase and tyrosinase-related protein 1 (TRP1), but not of TRP2/DCT as detected by Western blotting. Northern-blot experiments showed that the inhibitory effect is partially explained by a reduced accumulation of the corresponding mRNAs, that dropped to about 50% of control values (48 h treatment, 5 nM TNF-alpha). Moreover, the tyrosine hydroxylase and dopa oxidase activities decreased more rapidly in TNF-alpha-treated cells than in control cells, under conditions of inhibition of protein synthesis. This suggests a TNF-mediated reduction of tyrosinase half-life. However, the possibility of an inhibitory post-translational modification of the enzyme induced by TNF cannot be ruled out. Therefore, the inhibitory effect of TNF-alpha on tyrosinase and TRP-1 results from combined effect on mRNA levels and enzymatic activity or protein stability.

Activation by thyroid stimulating hormone of nerve growth factor-induced gene-B expression in thyrocytes in culture: relation with proliferation and specific gene expression.

Nerve growth factor-induced gene-B (NGFI-B) is an immediate early gene first found as a part of the PC12 cell response to NGF (Milbrandt, J., Science 238 (1987) 797-799). We have previously reported that NGFI-B mRNA is strongly upregulated by thyroid-stimulating hormone (TSH) in dog thyrocytes in culture (Pichon et al., Endocrinology 137 (1996) 4691-4698). In this study, we have analyzed the regulation of NGFI-B mRNA expression by a variety of agents acting on thyrocytes proliferation and/or differentiation. We show that: (1) the induction of NGFI-B mRNA is stronger after stimulation of the cAMP cascade, but it is not restricted to this signaling pathway; (2) the powerful mitogens for thyroid cells EGF and HGF have little or no effect on NGFI-B mRNA induction; (3) NGFI-B mRNA is induced by anisomycin at a subinhibitory concentration for protein synthesis, and is superinduced by the combination of TSH and anisomycin; this treatment decreases the TSH-induced proliferation levels, but does not inhibit the induction of some differentiation markers; and (4) both in dog and in pig thyrocytes, NGFI-B mRNA induction is observed after a variety of treatments stimulating differentiation, but without proliferative effects. Our results therefore suggest that NGFI-B mRNA induction might not be related to TSH-induced thyrocyte proliferation, but could participate in the differentiation program triggered by TSH.

Molecular interactions within the melanogenic complex: formation of heterodimers of tyrosinase and TRP1 from B16 mouse melanoma.

Melanin synthesis in mammals is catalyzed by three structurally related, membrane-bound proteins, tyrosinase, and the tyrosinase-related proteins 1 and 2 (TRP1 and TRP2). Current evidence suggests that in vivo these proteins may form a multienzyme complex. However, neither the precise composition of the complex, nor the specific interactions between its components have been characterized. This study used purified preparations of tyrosinase and TRP1 to analyze their interactions in non ionic detergent solution. Purified tyrosinase and TRP1 behaved as homodimers as judged by gel filtration chromatography and electrophoresis. Upon mixing of the purified proteins, the preferential formation of heterodimers was detected by: i) coelution in gel filtration chromatography with a shift to a common partition coefficient for both proteins, and ii) the occurrence of fluorescent energy transfer between fluorescein-labeled tyrosinase and rhodamine-labeled TRP1. However, the formation of heterodimers did not cause changes in the tyrosine hydroxylase activity of the enzymes, at least under standard assay conditions. Thus, tyrosinase and TRP1 interact strongly and specifically in detergent solution to form an heterodimer that might contribute to the formation of the melanogenic complex.

Comparison of TRPs from murine and human malignant melanocytes.

Most of our knowledge of the mammalian tyrosinase related protein (TRP) activities is derived from studies using murine melanoma models, such as B16 or Cloudman S-91 melanocytes. Owing to the high degree of homology between the murine and human enzymes, it has been assumed that their kinetic behaviour could be similar. However, the protein sequences at the metal binding sites of the murine and human enzymes show some differences of possible functional relevance. These differences are more significant in the metal-A site than in the metal-B site. By using three human melanoma cell lines (HBL, SCL, and BEU), we have studied the catalytic abilities of the human melanogenic enzymes in comparison to those obtained for the counterpart murine enzymes isolated from B16 melanoma. We have found that TRP2 extracted from all cell lines show dopachrome tautomerase activity, although the activity levels in human malignant melanocytes are much lower than in mouse cells. Reconstitution experiments of the human enzyme indicate that TRP2 has Zn at its metal binding-sites. Although mouse tyrosinase does not show DHICA oxidase activity, and this step of the melanogenesis pathway is specifically catalyzed by mouse TRP1, the human enzyme seems to recognize carboxylated indoles. Thus, human tyrosinase could display some residual DHICA oxidase activity, and the function of human TRP1 could differ from that of the murine protein. Attempts to clarify the nature of the metal cofactor in TRP1 were unsuccessful. The enzyme contains mostly Fe and Cu, but the reconstitution of the enzymatic activity from the apoprotein with these ions was not possible.

Melanin formation in the inner ear is catalyzed by a new tyrosine hydroxylase kinetically and structurally different from tyrosinase.

Detergent solubilized extracts of the cochleae of adult gerbils (Meriones unguiculatus) contain a tyrosine hydroxylase activity measurable by the radiometric method of Pomerantz. This activity is not related to Fenton-type reactions, since it is not inhibited by free radical scavengers and is heat and protease sensitive. It does not appear to be related to a peroxidase (EC 1.11.1.7) since it is neither dependent on H2O2, nor inhibited by catalase (EC 1.11.1.6). The involvement of a tyrosine hydroxylase (EC 1.14.16.2) related to catecholamine synthesis is also unlikely, since the activity is highly sensitive to 2-mercaptoethanol and is not increased by addition of tetrahydrobiopterin. The activity in crude inner ear extracts displayed an unusual maturation behaviour, with a slow activation upon aging at 4 degrees C. Fully active enzyme displayed Michaelis-Menten kinetics, with a Km for L-tyrosine of 47 microM. Cochlear tyrosine hydroxylase, but not melanoma tyrosinase (EC 1.14.18.1), was inhibited by o-phenanthroline, and was not dependent on L-DOPA as cofactor for full enzymatic activity. Crude extracts were also able to catalyze L-DOPA oxidation and melanin formation from either L-tyrosine or L-DOPA. The tyrosine hydroxylase, DOPA oxidase and melanin formation activities most probably resided in the same molecule, as suggested by inhibition studies. A tyrosine hydroxylase and melanin formation activity with identical properties was found in primary cultures of stria vascularis melanocytes. Immunochemical evidence confirmed the absence of either the tyrosinase encoded for by the albino locus, or the tyrosinase isoenzyme TRP1, encoded for by the brown locus. Conversely, an immunorreactive band of molecular weight 70 kDa was specifically recognized by a tyrosinase polyclonal antiserum in Western blot experiments. These results prove that melanogenesis in the cochlea, and likely in other extracutaneous locations such as the brain, is catalyzed by enzymatic systems different from, but related to tyrosinase.