The Pmel 17/SiZver Locus Protein CHARACTERIZATION AND INVESTIGATION OF ITS MELANOGENIC FUNCTION*

The silver mutation in mice causes progressive gray- ing of hair due to the loss of functional follicular melanocytes. Recently the silver locus gene (called Pml17) has been cloned; its encoded product shares homology with a chick melanosomal matrix protein and a bovine retinal pigment epithelial protein. Although the se- quence of the silver gene and the correlation of its expression with pigment production have been reported, its function in melanogenesis is still unknown. In an effort to characterize that function, we have synthesized the predicted carboxyl-terminal peptide of the mouse Pmell7 protein and generated a rabbit polyclonal anti- body (aPEP13) to it; that antibody recognized the silver protein specifically. The inununoaffinity-purified silver protein lacked all of the known melanogenic catalytic activities which other tyrosinase-related proteins (TRP) have, nor did it appear to modulate any of those TRP activities. Metabolic labeling experiments demon- strated that the silver protein disappears in vivo within a few hours, indicating that it is rapidly degraded, or quickly processed to

mapped to the silver locus (3,4). The expression of Pmel 17 mRNA was shown to be melanocyte-specific and both p-melanotropin and isobutylmethylxanthine could stimulate its expression concurrent with increases in melanin production (5).
The predicted amino acid sequence of the product of the silver locus (termed the silver protein in this report) has potential glycosylation sites and a putative transmembrane region close to its carboxyl terminus; the predicted molecular mass of the protein backbone is about 70 kDa, which is very similar to the tyrosinase-related proteins (TRPs).' Further, the silver protein has a unique high content of serine and threonine, and three repeats of a 26-amino acid motif in the middle of its sequence (3). The recently identified melanosomal matrix protein MMP115 of chicken retinal pigment cells (6) and the bovine retinal pigment epithelial protein RPE-1 (7) share significant sequence homology with the silver protein (8).
However, the function of the silver protein in mammalian melanogenesis remains unclear. The silver mutation has been reported to result in the graying of coat hairs by causing the premature loss of functional melanocytes in hair follicles (9, 10). The slower growth of silver mutant melanocytes (melan-si) in culture compared to wild type melanocytes has also been observed (11). Therefore, the silver mutation seems likely to induce toxic effects to melanocytes analogous to those caused by the phenotypically similar B" mutation at the brown locus (10, 12). Melanocytes are continually exposed to the stress of potentially toxic intermediates of melanogenesis, particularly 5,6-dihydroxyindole (DHI) (13). The silver protein might protect melanocytes against such intermediates through a catalytic function (e.g. by their rapid removal or selective modulation of their production), or, alternatively, it might function as a structural melanosomal protein restricting melanogenesis to that subcellular compartment.
To characterize the silver protein, and to determine its function, we have prepared a rabbit polyclonal antibody (termed aPEP13) against the synthetic peptide corresponding to the mouse Pmel 17 carboxyl terminus (14): a technique we have previously used to characterize the TRPs encoded by the albino, brown, and slaty loci (15)(16)(17)(18). In a recent collaborative study (191, we found that an antibody ( a m ) generated against melanosomal matrix components recognized a protein encoded at the silver locus, although additional melanosomal matrix proteins are also recognized by the aMX antiserum. In this The abbreviations used are: TRP, tyrosinase-related protein; DHI, 5,6-dihydroxyindole; DHICA, 5,6-dihydroxyindole-2-carboxylic acid; PBS-, phosphate-buffered saline without CaC1, and MgC1,; a M X , antimelanosomal matrix protein; MES, 4-morpholinethanesulfonic acid.
B. S. Kwon, S. Ponnazhagan, K. Kim, C. Chintamaneni, D. C. Bennett, E. Rinchik, R. Pickard, and R. Halaban, submitted for publication. study, we show that aPEPl3 specifically recognizes the silver protein, and we have used it to immunopurify the silver protein for analysis of possible catalytic functions and to examine its synthesis, processing and subcellular distribution. EXPERIMENTAL PROCEDURES Cells a n d Culture Conditions-B16 F10 murine melanoma cells were cultured as described previously (16,17); the B16 F10 subline used in this study was virtually amelanotic but could be induced by a-melanotropin to increase production of tyrosinase and melanin (20). Melanocyte cell lines cultured from genetically defined mice (melan-a, melan-b, melan-c, and melan-si cells) were gifts from Dr. Dorothy Bennett, Department of Anatomy, St. George's Hospital Medical School, London, U K they were grown as reported previously (11,21,22). NIH 3T3 murine fibroblasts and Meth A murine sarcoma cells (23) were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. B16 F10 melanoma tumors were also passaged in vivo following subcutaneous injection in the rear flanks of female 6-week-old C57BU6J mice.
Peptide Synthesis and Antibody Production-The peptide named PEP13 (CGLGENSPLLSGQQV-C0,H) corresponds to the predicted carboxyl terminus of the protein encoded by the murine Pmel 17 gene ( 14),; it was synthesized by Dr. Mark Thompson (LCBNIH) and coupled to bovine serum albumin Supercarrier (Pierce). Immunization and serum collection from the rabbit, and the titration and specificity of antibody production as determined by enzyme-linked immunosorbent assay, was as described previously (16,17).
Metabolic Labeling a n d Immunoprecipitation-These techniques were performed as reported previously (15)(16)(17)20). Briefly, semiconfluent cells were preincubated in methionine-free medium containing dialyzed fetal bovine serum, pulsed for 30 min to 4 h with [35Slmethionine (0.4 -1.0 mCi/ml) (Du Pont-NEN) in methionine-free medium, and chased in complete medium for 0-48 h, as detailed in the figure legends. The cells were harvested with trypsin/EDTA, washed with phosphatebuffered saline without CaCI, and MgCI, (PBS-) and solubilized at 4 "C for 60 min in Nonidet P-4OISDS buffer (1% Nonidet-P40,0.01% SDS, 0.1 M Tris-HC1, pH 7.2, 100 p~ phenylmethylsulfonyl fluoride, 1 pg/ml aprotinin). The resulting 35S-labeled extracts were then precleared with 10 pl of normal rabbit serum and 100 pl of GammaBind G Sepharose (Pharmacia Biotech Inc.). 2 x lo6 trichloroacetic acid-precipitable countsImin of precleared extracts were incubated with 5 pl of antibodies for 1 h at 4 "C, and then complexed with 30 1. 11 of GammaBind G Sepharose for 30 min at 4 "C. The immune complexes were washed six times with Nonidet P-4OISDS buffer, then eluted in SDS sample buffer containing 100 mM dithiothreitol at 95 "C for 5 min and analyzed by SDSgel electrophoresis (241, followed by fluorography. Western Immunoblotting Analysis-Proteins from Nonidet P-40ISDS solubilized cells or tumors, or immunopurified proteins as detailed below, were separated on 7.5% SDS gels, then transferred to nitrocellulose membrane sheets or polyvinylidene difluoride membranes (Immobilon-P, Milipore Corp., Bedford, MA) and incubated with primary antibodies (1/1000 dilution) as noted in the figure legends. Subsequent visualization of antibody binding was carried out with Enhanced ChemiLuminescence (Amersham Corp.) according to the manufacturer's instructions.
Immuno-affinity Purification-Immune-affinity columns were prepared by covalently linking 1 mg of purified anti-peptide IgG to 2 ml of protein A-Sepharose columns using IgG Orientation Kits (Pierce), according to the manufacturer's instructions and as detailed elsewhere (16)(17)(18). B16 F10 melanoma cells growing subcutaneously in mice were excised, homogenized gently, and lysed in 155 mM NH,CI, 10 mM KHCO,, 0.1 mM EDTA. The cells were washed in ice-cold PBS' and then solubilized in Nonidet P-40ISDS buffer overnight a t 4 "C. The samples were centrifuged at 10,000 x g for 30 min at 4 "C, and identical aliquots (-50 mg of protein) of the soluble supernatant fraction were bound to each immune-affinity column; nonabsorbed proteins were washed through with 30-40 ml of Nonidet P-4OISDS buffer. Specifically absorbed proteins were then eluted from each column with 7 ml of the Elution buffer supplied with the columns, immediately neutralized with phosphate buffer and dialyzed against 0.1% Nonidet-P40, 25 mM phosphate buffer, pH 6.8, 0.1 mM EDTA, 100 PM phenylmethylsulfonyl fluoride, and 1 pg/ml aprotinin overnight a t 4 "C. They were subsequently concentrated (-20x1 using Centripep 30 (Amicon Corp., Beverly, MA).
Preparative Electrophoresis-For further purification of the silver protein, we used preparative SDS-gel electrophoresis. Immunopurified silver protein, obtained a s detailed above, was reduced in SDS sample buffer a t 100 "C and separated using a Prepcell model 491 (Bio-Rad) according to the manufacturer's instructions. The eluted fractions containing the silver protein were identified by Western immunoblotting, collected, dialyzed and concentrated by freeze-drying.
Isolation of Melanosomes and Coated Vesicles-Isolation of melanosomes was carried out as described previously (25). All steps were performed at 4 "C. B16 F10 melanoma cells were washed twice with PBS-, collected by centrifugation, and resuspended in homogenization buffer (0.25 M sucrose, 50 mM phosphate buffer, pH 6.8, 1 mM EDTA, 100 p~ phenylmethylsulfonyl fluoride, and 1 pg/ml aprotinin), homogenized with twelve strokes of a Potter-Elvehjem grinder. Following centrifugation for 10 min a t 1000 x g, the supernatant was centrifuged for 30 min at 10,000 x g . The pellet was resuspended in homogenization buffer and applied to a discontinuous gradient of 1.0-2.0 M sucrose and centrifuged for 1.5 h at 100,000 x g . Melanosome-rich fractions were collected at the 1.8-2.0 M and at the 1.6-1.8 M interfaces.
Isolation of coated vesicles was performed as reported previously (26). B16 F10 melanoma cells were isolated, lysed, washed with ice-cold phosphate-buffered saline, and mixed with a n equal volume of MES buffer (0.1 M MES-NaOH, pH 6.5, 1 mM EGTA, 0.5 mM MgCl,, 0.02% NaN,, 100 p~ phenylmethylsulfonyl fluoride, and 1 pg/ml aprotinin). The mixture was homogenized with a Potter-Elvehjem grinder and centrifuged for 40 min a t 19,000 x g. The supernatant was centrifuged for 70 min at 43,000 x g, and the pellet was resuspended in the MES buffer, mixed with a n equal volume of 12.5% sucrose and 12.5% Ficoll in MES buffer, and centrifuged again for 70 min at 43,000 x g. The supernatants were diluted in MES buffer, centrifuged again at 43,000 x g, and coated vesicles were collected in the pellet.
Melanogenic Assays-Assays for melanogenic catalytic activities as described below were carried out at pH 6.8, 37 "C for 60 min unless otherwise noted in the text. 1) Tyrosine hydroxylase activity was measured using the [,Hltyrosine assay (27,28). This method specifically measures the tritiated water produced during the hydroxylation of tyrosine to DOPA. 2) DOPA oxidase activity was measured using incorporation of [3-l4C1DOPA into acid-insoluble melanin as detailed previously (20). 3: DOPAchrome tautomerase activity was measured by high performance liquid chromatography as the disappearance of DOPAchrome substrate and the production of 5,6-dihydroxyindole-2-carboxylic acid (DHICA) rather than DHI; data are converted to picomoles of products by comparison with known standards; this assay was detailed previously (17,29). 4) DHI oxidase activity was measured by high performance liquid chromatography as the disappearance of DHI substrate from reaction mixtures compared to controls for spontaneous auto-oxidation; data are converted to pmol by comparison with known standards. 5) DHICAoxidase activity was measured by MBTH assay, as reported previously (18). 6) Melanin production was measured using incorporation of [14Cltyrosine into acid-insoluble melanin as detailed previously (27,28); picomoles of melanin produced are calculated from the radioactive product. Tyrosine and DOPA used as standards and reaction substrates in these assays were obtained from Sigma; [~-3,5-,HItyrosine, [3-14ClDOPA, and [U-'4Cltyrosine were obtained from Du Pont-NEN; DOPAchrome was prepared using the silver oxide method originally described by Korner and Pawelek (30). DHI and DHICA were kindly provided by Dr. Giuseppe Prota (University of Naples, Naples) IL). and purchased commercially from Regis Chemical Co. (Morton Grove, Miscellaneous Methods-Protein concentrations were determined with the BCA assay kit (Pierce) using bovine serum albumin as a standard.

RESULTS
Preparation and Specificity of Antibody to the Silver Protein-We synthesized the 15 amino acid peptide (named PEPlS), which corresponds to the carboxyl terminus of the predicted protein encoded by the murine Pmel 17 gene (14),2 which has been mapped to the silver locus (3,4). The rabbit polyclonal antibody termed aPEP13, generated against PEP13, recognized that peptide specifically. The specific reactivity of aPEP13 IgG to PEP13, without cross-reactivity to any other peptides, is shown by enzyme-linked immunosorbent assay ( Fig. 1). Each of the antibodies was immunizing peptide-specific, and aPEP13 IgG had almost the same titer against the immunizing peptide as the other antibodies had. Note that aPEP1, aPEP7, and aPEP8 did not react with PEP13, nor did the aPEP13 recognize peptide PEP1, PEP7 or PEP8; aPEP1, aPEP7, and aPEP8 recognize distinct melanogenic proteins ?A). As with the TRP family, the protein precipitated by aPEP13 was observed only in cells of melanocytic origin (i.e. B16 F10 and melan-a). By Western immunoblotting analysis, we observed a similar pattern of cell-specific expression of aPEP13 antigen and the other melanogenic proteins (Fig. 2B ).
Although aPEP13 reacted with several minor low molecular weight bands on the immunoblot filter, it recognized the major band only in the extracts of normal and transformed melanocytes, at the same molecular size as in the metabolic labeling study (Fig. 2 A ) . The observed molecular mass of the protein identified by aPEP13,85 kDa, was somewhat larger than that expected from the amino acid sequence predicted for human Pmel 17 (i.e. 70 kDa). This was a single band, very sharp and thin as compared with the broadly migrating ones of the TRPs. Fig. 2, A and B, also reveal that the highly pigmented melan-a melanocytes expressed more silver protein than did the unpigmented B16 F10 melanoma cells, consistent with the previous report correlating Pmel 17-1 expression with pigmentation (6). When the unpigmented B16 F10 cells were stimulated with a-melanotropin (201, levels of tyrosinase and the silver protein, as well as melanin production, were significantly increased (data not shown). When various organs of newborn C57BLJ6J mice were analyzed for expression of the silver protein by Western immunoblotting, the 85-kDa protein identified by aPEP13 was restricted to the melanogenic organ (skin) where tyrosinase activity (tyrosine hydroxylase and melanin formation activity) could also be detected (data not shown).
It has been reported that the silver mutation is the result of a single base insertion which alters the predicted carboxyl terminus sequence of the protein ( 14).2 The mutated silver protein would therefore not be expected to be recognized by aPEP13, whose immunizing peptide was at the carboxyl terminus. When we examined cell extracts of melan-si cells (derived from mice homozygous for the silver mutation) by Western immunoblotting, the 85-kDa band was not detected although the other TRPs (tyrosinase, TRP1, and TRPS) were detected in both melan-a cells and melan-si cells (Fig. 31, consistent with our previously reported results (19).
These results, combined with the cell-and organ-specific expression, and lack of reactivity to the mutated protein, clearly demonstrate that aPEP13 identified the protein encoded by the silver locus. Immuno-afinity Purification of the Silver Protein and Enzymatic Analysis-Since there was a good correlation of silver gene expression with visible pigmentation and since the silver mutation may result in toxicity to melanocytes, the silver protein has been anticipated to have some specific catalytic function(s) comparable to the other TRPs, or alternatively, an inhibitory effect on the production of harmful intermediates produced during melanogenesis. In order to determine if the protein encoded by the silver locus has any known melanogenic catalytic function, we purified the silver protein from B16 F10 tumors using immuno-affinity chromatography and then examined its potential melanogenic activity. TRPs and the silver protein binding to immuno-affinity columns were eluted, dialyzed and concentrated as detailed in the experimental procedure section and their purities were analyzed by Western immunoblotting. The fraction immuno-purified by the aPEP13 IgG-linked column was recognized only by aPEP13 without any cross-reactivity to other anti-peptide antibodies (Fig. 4). The immunopurified protein had the identical size described in the radiolabeling experiments above. The other low molecular weight band (arrow) was observed in all blots and all fractions, and is IgG eluted from the columns. Western immunoblotting aPEP1 aPEP7 aPEP8 aPEP13 also showed that the other immunopurified proteins had little or no cross-contamination, except the TRP2 fraction which contained a minor contamination with tyrosinase.
Melanogenic assays of these purified proteins demonstrated that the silver protein had no known catalytic function attributed to the members of the TRP family (Table I). As previously reported (17,18), the TRPs each have specific enzymatic functionb); tyrosine hydroxylase, DOPA oxidase, DHI oxidase and melanin formation activities of tyrosinase; DHICA oxidase activity of TRP1; DOPAchrome tautomerase activity of TRPB (Table I). TRPB had detectable levels of tyrosinase catalytic activities, which were consistent with the slight contamination of this immunopurified fraction with tyrosinase. The immunopurified silver protein, however, had just the baseline, negative control level of each of those melanogenic activities. Although the purified silver protein appears to have a slight DHI oxidase activity, this is not statistically significant above background.
In separate experiments, we found that the specific melanogenic activities of immunopurified tyrosinase, TRPl and TRPB were not modulated by coincubation with a n equal amount of the purified silver protein (data not shown).
Characterization and Subcellular Fractionation of the Silver Protein-The immunopurification experiments reported above suggested that the silver protein does not have a demonstrable known melanogenic enzymatic function. In further studies, we examined how it is synthesized, processed and degraded, and where it is delivered and exists within the melanocyte.
It is well known that tyrosinase, TRPl and TRP2 are highly glycosylated proteins (31)(32)(33). When melan-a melanocytes or B16 F10 melanoma cells were pulse-labeled by ["'Slmethionine and then chased, the presence of inhibitors of glycoprotein processing (swainsonine or deoxymannojirimycin) decreased the molecular weight of TRPs, from those of the glycosylated forms seen in the normal chase for 1 h (Fig. 51, as has been reported previously (31). However, identical treatments had no effect on the size of band precipitated by aPEP13 (Fig. 5). The immunoprecipitated band of the silver protein also was not affected when the labeled cell extract was treated with neuraminidase ! Silver Locus Protein 29201 under conditions in which TRPl was shifted to the de novo type completely (data not shown). These data demonstrate that the silver locus product is not subject to late Golgi processing of Asn-linked carbohydrates although it has several potential sites for N-glycosylation (3).
When the 30-min pulse labeling of cells was followed by longer chases in the unlabeled medium and immunoprecipitation analysis, the TRPs were observed to be fully glycosylated from the de novo type and then degraded (Fig. 6, left), as previously reported by our group (17). The silver protein, however, was not similarly processed to a higher molecular weight form and disappeared completely within 4 h. Similar results were obtained with B16 F10 melanoma cells. A Western immunoblotting experiment (Fig. 6, right) shows that majority of the silver protein in cell lysates could be recognized by aPEP13 even after incubation for 48 h a t 37 "C. These results indicate that the silver protein is inherently stable (as are the TRPs) but that it is actively processed to lose its carboxyl terminus (the aPEP13 epitope) within a few hours in vivo within melanocytes. When the 35S-labeled and 4-h chased cell extracts were treated by neuraminidase (conditions, under which the molecular weight of immunoprecipitated TRPs are decreased) or were denatured in 1% SDS, 20 mM DTT, aPEP13 still could not precipitate a band from those extracts (data not shown).
Since these data suggest that the silver protein must have lost its carboxyl terminus during post-translational processing, we tested other antibodies which might recognize the silver protein at an epitope different from aPEP13. We found that aMX, which has been recently described (34), also recognized the silver protein along with other or yet unidentified matrix constituents (19). That study showed that, when "'S-labeled cell extracts were precleared with either aMX or aPEP13, reactivity with the other antibody was lost. We now report that the specificity of aMX was identical to that of aPEP13; the silver protein recognized by aMX is specific to transformed and normal melanocytes (Fig. 7, A and B ) . However, aMX reacted with the silver protein of melan-si cells by Western immuno blotting (Fig. 7B), suggesting that aMX recognizes epitope(s1 distinct from aPEP13. To map the epitope recognized by aMX on the silver protein, we used partial digestion with V8 protease, as described previously (16,35). We purified the silver protein by preparative electrophoresis, digested it with V8 protease, separated the fragments by SDS-polyacrylamide gel electrophoresis, and analyzed the reactivity of aPEP13 or aMX by Western blotting (data not shown). aPEP13 reacted with a 25-kDa fragment generated by digestion of the silver protein with V8 protease whereas aMX did not recognize that fragment, indicating that aMX recognizes an epitope of the silver protein a t least 25 kDa apart from its carboxyl terminus.
Orlow and collaborators (34) have found, using pulse-chase metabolic labeling, that as the melanosomal matrix protein recognized by a M X disappeared, a band of 53 kDa appeared after long chases. That result was reproduced in this study, but since aMX was generated against a mixture of melanosomal matrix proteins extracted in SDS buffer, it was possible that the 53-kDa band was a separate, unrelated protein, and not a processed form of the silver protein. To examine this possibility, we tested whether the recognition of the 53-kDa protein by aMX could be removed by incubation with the silver protein.
Although the purified silver protein had the specific ability to compete with the binding of aPEP13 to its ligand and of a M X to the 85-kDa band, it had no similar ability to compete with binding of aMX to the 53-kDa band (Fig. 8). Therefore we conclude that the silver protein was not processed to the 53-kDa form.
Based on the cross-reactivity experiments of aPEP13 and aPEP1 aPEP7 aPEP8 aPEP13 a M X , the silver protein seemed not to be lost completely, but to be delivered and localized in melanosomes as an internal, matrix protein (19). We therefore isolated subcellular fractions of melanocytic cells and examined the localization of the silver protein therein. In fact, the silver protein was recognized by aPEP13 in the fractions of isolated melanosomes as were the other TRPs (Fig. 9). Considering the rapid loss of the aPEP13 epitope in vivo (shown above), the newly synthesized silver protein must be delivered to melanosomes very quickly, probably within a few hours. Tyrosinase, TRP1, and TRPP have been proposed to be transferred to melanosomes via coated vesicles and were highly represented in the purified coated vesicle fraction, but little silver protein was detected by aPEP13 in that fraction (Fig. 9). Studies in progress using immunoelectron microscopy support the concept that the silver protein is localized to spherical vacuoles and is also present within the internal structure of stage I1 or I11 melanosomes, but is not detectable in fully melanized stage IV melanosomes (data not shown). DISCUSSION We have generated an antibody (aPEP13) against the predicted carboxyl terminus of the protein encoded by the silver locus which recognizes that protein specifically. Our data on the expression of the silver protein using aPEP13 demonstrate that the putative protein is in fact expressed and that its expression is restricted to cells of melanocytic origin. Our studies show further that the silver protein has no known melanogenic activities and that while it is an internal component of melanosomal matrix, it is delivered to melanosomes in a manner distinct from the tyrosinase-related proteins. Due to recent advancements in molecular biology and biochemistry, many Solubilized proteins (2.5 pgflane) of B16 F10 cells were analyzed by   FIG. 8. 53-kDa protein recognized by aMX and silver protein. Western immunoblotting as detailed for Fig. 2B. Each antibody was incubated with the purified 85-kDa silver protein as noted 1 h before and during the immunostaining. genes which affect murine coat color have now been cloned and characterized. Some of those genes specifically regulate melanogenesis in melanocytes; albino, brown, slaty, dilute, pinkeyed dilution, silver, and so on. The proteins encoded by the albino, brown, and slaty loci (tyrosinase, TRP1, and TRP2, respectively) have distinct enzymatic functions. Tyrosinase, the activity most critical to melanogenesis, has tyrosine hydroxylase, DOPA oxidase and DHI oxidase activities (27,36). The DOPAchrome tautomerase activity of TRP2 (17) and the DHICA oxidase activity of TRPl (18) are known to modulate melanogenesis as post-tyrosinase factors. In this study, however, we could not attribute any of the established melanogenic catalytic functions to the silver protein. Two recent studies have suggested that the silver protein may have a novel catalytic function at a more distal point in the melanogenic pathway than TRP1; such an activity may be required for the polymerization of DHICA metabolites to melanin (37,38). On the basis of these observations, and the distinctive pattern of its subcellular localization, the silver protein seems likely to be an integral component of melanosomes, with a function distinct from that of the TRP family. These studies further show that the silver protein is synthesized, and transported quickly to the melanosome fraction. It is subsequently quickly processed to lose its carboxyl terminus (and thus its recognition by aPEP13) although it is not completely degraded and remains within the melanosome. Since the reactivity with aPEP13 is lost relatively quickly due to rapid processing of the silver protein wherein the carboxyl terminus is cleaved, we are unable at this time to determine whether the processed silver protein remains an integral part of the melanizing melanosome or is lost to the cytoplasm and/or to the extracellular milieu. In light of the rapidly emerging literature on the importance of the silver protein (termed gpl00 in humans), as well as the tyrosinase related proteins, as specific targets for humoral and cellular immune responses to melanoma, such information takes on an added significance (39,40). Further studies of the silver protein using the various antibodies now available should provide valuable new information about the formation of premelanosomal structural proteins, the origin of stage I melanosomes and the intracellular trafficking of TRPs to melanosomes.
It is known that the silver mutation has a genetic interaction with the brown locus (1,41). The brown protein (i.e. TRP1) is multifunctional and may play a structural role in addition to its catalytic role. Since some mutations at the brown and silver loci have similar phenotypes, both proteins might act in series in aPEP1 aPEP7 aPEP8 aPEP13 aMX the melanogenic pathway so that a mutation in either gene affects the same type of melanin production. In melanosomes, TRPl might interact with the silver protein. Although the brown mutation has no effect on the expression or degradation of the silver protein (data not shown), their products might interact at the protein level, a possibility that will be studied in the future.
We hypothesize that the silver protein functions as a component of the melanosomal matrix, which limits melanogenesis to that intracellular compartment and protects melanocytes from harmful melanogenic intermediates. This in no way obviates a potential novel melanogenic function for the silver protein; on the contrary, as noted above, preliminary studies in other laboratories have suggested that the silver protein may function as a polymerase for DHICA or one of its metabolites (37, 38). Mutations at the silver locus might result in the loss of that function and elicit effects toxic to melanocytes. However, other mechanisms could be proposed to explain such toxic effects. First, the silver protein may have a novel enzymatic function which excludes or decreases the production of harmful melanogenic intermediates (e.g. an anti-oxidant function), although, in this study, we could observe no such melanogenic enzymatic function. Mutations at the silver locus might then lead to the loss of those activities and result in cell death. Because the silver mutation is recessive, however it is unlikely that the mutated silver protein actively produces a toxic substance. Secondly, i t is possible that the mutated silver protein could be an antigen recognized by cytotoxic T-lymphocytes, which might then attack and kill the melanocytes. It has long been thought that disorders such as vitiligo result from the unscheduled destruction of melanocytes by immune mechanisms, and the presence of immune effector mechanisms which target peptides of the silver protein give added credence to this possibility (42,43). Aberrant trafficking or processing of the mutated silver protein could result in its presentation to the immune system as an antigen, thus prompting a progressive autodestruction of melanocytes.