Materials and methods

Establishment of melanocyte cultures from neonatal mice

The dorsal skin of a neonatal mouse was removed aseptically and cut into several small pieces, rinsed in Ca²⁺,Mg²⁺-free phosphate-buffered saline (PBS), and incubated in 0.25% trypsin at 37°C and 5% CO₂ for 2 h. The dermis was then separated away from the epidermis and incubated in 0.05% trypsin containing 0.5 U per ml of collagenase for 30 min. The digested dermis was pipetted repeatedly in 0.05% trypsin and 0.02% ethylenediamine tetraacetic acid (EDTA) until the solution became turbid. The cells were pelleted and resuspended in Ham's F-10 medium. Melanocytes were cultured in Ham's F-10 medium containing 10% fetal bovine serum, 100 U per ml of penicillin, 100 g per ml of streptomycin, 2 mM L-glutamine, 0.1 mM dibutyryl cyclic adenosine-5'-monophosphate, 100 nM phorbol-12-myristate-13-acetate and 25 g per ml bovine pituitary extract-protein (Life Technologies, Rockville, MD). Melanocyte lines from C57BL/6+/+, p^J/p^J, and p^cp/p^6H were established from primary cultures. The last two cell lines are null for p gene function. The p^J allele is characterized by an intragenic deletion (^{Oakey et al. 1996}) and p^cp/p^6H is a compound heterozygote of two deletion alleles and lacks all protein encoding sequences of the p gene (^{Gardner et al. 1992;Nakatsu et al. 1993;Lehman et al. 1998}). After establishing pure cultures of these melanocytes, electron microscopy and immunofluorescent studies were performed, comparing cells of the same passages (passages 3–6).

Electron microscopy

Confluent flasks of melanocytes were treated with 0.05% trypsin and 0.02% EDTA in Ca²⁺,Mg²⁺-free PBS, pelleted, and washed twice in PBS. Cell pellets were fixed for at least 18 h at 4°C in 3% glutaraldehyde and 0.1 M phosphate buffer (pH 7.2). The fixed pellets were dissected to 1 mm³ pieces. Postfixation was in 1% osmium tetroxide in 0.1 M phosphate buffer for 1 h followed by en bloc staining for 30 min in 1% uranyl acetate in 50% ethanol. The tissues were then dehydrated using serial alcohol and acetone incubations and embedded in Spurr resin. A Sorvall MT-2B ultramicrotome was used to section the tissues to 80 nm (silver-gold). Sections were stained with uranyl acetate and lead citrate. Grids were viewed on a Philips 400 electron microscope at an accelerating voltage of 80 kV.

Tyrosinase activity

To visualize tyrosinase activity in melanocytes, L-DOPA conversion to melanin was assayed in vitro. Trypsinized cell pellets were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) for 1 h, and then washed three times in cacodylate buffer. These fixed pellets were dissected into 1 mm³ pieces and incubated in 0.1% L-DOPA in cacodylate buffer at 4°C overnight. The next day, cell pellets were incubated in fresh 0.1% L-DOPA cacodylate buffer at room temperature for 2 h, and then incubated a third time in 0.1% L-DOPA cacodylate buffer at 37°C for 2 h. Postfixation was done in 1% osmium tetroxide in 0.1 M cacodylate buffer, and all further steps of electron microscopy processing were performed as described above.

Immunofluorescent studies of mouse melanocytes using confocal laser scanning microscopy

Melanocytes (1  10⁵) from p^J/p^J or p^cp/p^6H mice were plated on glass coverslips in six-well plates. Cells were allowed to grow for 48 h at 37°C, and then were incubated with 30 M DAMP for 30 min at 37°C and washed with Ham's F10 medium. For monensin studies, monolayers were treated with 25 M of monensin (Sigma Chemical, St. Louis, MO) for 5 min after DAMP treatment. To localize DAMP, cells were fixed at room temperature for 15 min in 3% (wt/vol) paraformaldehyde in buffer A (10 mM sodium phosphate, 150 mM sodium chloride, 2 mM magnesium chloride; pH 7.4), and then were washed once with 50 mM ammonium chloride and twice with buffer A. Each monolayer was permeabilized with 0.1% (vol/vol) Triton X-100 in buffer A for 5 min at -10°C. Coverslips were blocked with 5% goat serum (Jackson ImmunoResearch Laboratories, West Grove, PA) in PBS for 30 min. They were then covered with 100 l of monoclonal mouse anti-dinitrophenol (DNP) IgG diluted 1:10 (Oxford Biomedical Research) and incubated at 37°C for 60 min. After three washes with buffer A, the cells were incubated with 100 l fluorescein-5-isothiocyanate (FITC) conjugated goat antimouse IgG diluted 1:200 (Organon Teknika Corporation) at 37°C for 60 min. After three additional washes with buffer A, melanocytes were blocked again with 5% goat serum for 30 min and incubated with rabbit polyclonal antiserum PEP1 diluted 1:500 (a kind gift of Vince Hearing, NIH) generated against the carboxyl terminus of murine Tyrp1 (^{Jimenez et al. 1991}) for 60 min. Coverslips were washed in buffer A, incubated with lissamine rhodamine-conjugated goat antirabbit IgG diluted 1:100 (Jackson ImmunoResearch Laboratories) for 60 min and washed with buffer A. These coverslips were then mounted in Vecta shield (Vector Laboratories, Burlingame, CA) and viewed under a Bio-Rad MRC-600 confocal laser scanning microscope (Bio-Rad, Richmond, CA) equipped with an argon krypton laser coupled to a Nikon Optiphot II fluorescence microscope and a 60 plan Apo oil objective. A standard k1/k2 filter set was used. Simultaneous two-channel recording was performed using excitation wavelengths of 488 and 568 nm. Images were processed and merged using a Voel view ultra 2.5 software (Vital Images, Airfield, IA). The numbers of vesicles that stained for Tyrp1, DAMP, or both were scored in four independent experiments. In each experiment, 25 independent fields (47 m²) were counted from cells established from C57BL/6+/+ and the p mutants. The data from 100 (4  25) representative fields of approximately equivalent vesicular densities of peripheral cytoplasm were used to calculate the percentage of vesicles stained with Tyrp1 and DAMP in all three cell lines.

Immuno-gold studies of wild-type melanocytes

C57BL/6 wild-type melanocytes were grown and treated with DAMP as detailed above, and pelleted in Ca²⁺,Mg²⁺-free PBS, pH 7.2. The cells were then fixed for 1 h in 1% glutaraldehyde in Ca²⁺,Mg²⁺-free PBS, pH 7.2, and washed for 5 min three times in Ca²⁺,Mg²⁺-free PBS. The remaining glutaraldehyde was quenched by incubation in 0.1 M NH₄Cl in Ca²⁺,Mg²⁺-free PBS, pH 7.2, for 30 min, followed by two 5 min washes in Ca²⁺,Mg²⁺-free PBS, pH 7.2. The treated cell pellets were then dehydrated by serial alcohol dilution and embedded in LRWhite at 50°C, under vacuum. The cell pellets were cut in 80 nm sections onto 200 mesh formvar/carbon coated nickel grids. The grids were floated on a drop of blocking buffer (0.5 M NaCl, 0.1% Na-azide, 1% ovalbumin, in 0.01 M Tris-HCl, pH 7.2) for 30 min at room temperature. The grids were then floated on a drop of blocking buffer plus mouse monoclonal anti-DNP (Oxford Biomedical Research) at 1:10 dilution for 16 h at 4°C, in a moisture chamber. The grids were then washed with rinse buffer (0.15 M NaCl in 0.01 M Tris-HCl, pH 7.2) and floated on rinse buffer plus 0.02% polyethylene glycol-20, 0.1% Na-azide, plus Protein A gold, 10 nm (Amersham), at 1:20 for 1 h at room temperature and washed again with rinse buffer. Fixation of the antigen-antibody complex was with 2% glutaraldehyde in PBS for 10 min followed by two washes with ddH₂O. The grids were then stained with 2% uranyl acetate (aqueous) for 10 min and lead citrate for 10 min and viewed on a Philips 400 electron microscope at 80 kV.

Results

Electron microscopy studies with melanocytes from C57BL/6^+/+ and p mutant mice

Cultures of melanocytes were established from C57BL/6^+/+, p^J/p^J, and p^cp/p^6H mice. In comparison with melanocytes established from wild-type C57BL/6^+/+ mice, both p null mutant cell lines exhibited minimal amounts of visible pigmentation in their cell pellets or culture media. Electron microscopy demonstrated that melanocytes from C57BL/6^+/+ (Figure 1a) contained numerous elliptical melanosomes that were highly melanized (stage III-IV). Melanocytes from p null mutants also contained numerous melanosomes, but these were poorly melanized (stage I-II) (Figure 1b, c); similar results were obtained by^{Rosemblat et al. (1998)}. In addition, we tested L-DOPA reactivity of the p^cp/p^6H melanocytes that we established in vitro. These melanocytes exhibited L-DOPA oxidase activity of tyrosinase that revealed the presence of this enzyme in the trans-Golgi network and in stage I-II melanosomes (Figure 1d).

Figure 1.

Ultrastructure of cultured dermal melanocytes established from wild type mice (C57BL/6^+/+) and p mutant mice (p^J/p^J and p^CP/p^6H). (a) C57BL/6^+/+ melanocytes with numerous stage III and IV melanosomes (m) around the nucleus (n). Melanocytes from p^J/p^J (b) and p^CP/p^6H (c) mice displaying primarily stage I-II melanosomes (m). Melanocytes from p^CP/p^6H (d) mice exhibit a DOPA oxidase reaction of tyrosinase on incubation with 0.1% DOPA in the trans-Golgi network and stage I and II melanosomes (m). All panels are shown at the same magnification. Scale bar: 1 m.

Full figure and legend (353K)

Immunofluorescent studies using confocal laser scanning microscopy

Normal melanosomes have been shown to be acidic organelles (^{Moellman et al. 1988;Bhatnagar et al. 1993}). To determine whether melanosomes from p mutant melanocytes are acidic, we used DAMP and anti-DNP staining to visualize acidic organelles.^{Anderson et al. (1984)} have shown that DAMP penetrates the membranes of living cells, accumulates in acidic vesicles within the cell, and can be retained in these vesicles after aldehyde fixation. Melanocytes from C57BL/6^+/+ and p mutant mice were treated with DAMP, fixed in paraformaldehyde, and processed for indirect immunofluorescence (Figure 2) using mouse monoclonal antibodies directed against dinitrophenol (anti-DNP) and visualized with goat antimouse antibodies coupled to FITC. The polyclonal antibody PEP1 directed against Tyrp1 was used to detect melanosomes and their precursors and visualized with goat antirabbit IgG conjugated with lissamine rhodamine.

Figure 2.

Confocal laser scanning microscopy of dermal melanocytes from C57BL/6^+/+ and from p mutant mice (p^J/p^J and p^CP/p^6H) immunostained with antibodies against DAMP and Tyrp1. Acidic compartments of melanocytes were detected by DAMP incorporation followed by treatment with anti-DNP antiserum; melanosomes were detected by anti-Tyrp1 antiserum. Parts (a)–(c) show C57BL/6^+/+ melanocytes (a) stained with anti-DNP antibodies visualized as green fluorescence; (b) stained with anti-Tyrp1 antibody visualized as red fluorescence; (c) stained with both anti-DNP and anti-Tyrp1 antiserum (merged image), demonstrating extensive colocalization of acidic compartments and melanosomes visualized as yellow fluorescence. Parts (d)–(f) and (g)–(i) are similarly stained and are from melanocytes from p^CP/p^6H and p^J/p^J mice, respectively. Melanocytes from p mutant mice show very little colocalization of DAMP and Tyrp1. Part (j) shows a merged image of C57BL/6^+/+ melanocytes treated with 25 m monensin for 5 min after DAMP incorporation, followed by immunostaining with DAMP and Tyrp1 antibodies. The latter control demonstrates the loss of acidic compartments after monensin treatment. Part (k) is a merged image of melanocytes from p^CP/p^6H mice similarly treated as in (j). All panels are reproduced at the same magnification. Scale bar: 13.8 m.

Full figure and legend (80K)

Tyrp1-staining vesicles were detected primarily in the perinuclear area but were also scattered throughout the cytoplasm in melanocytes from all three genotypes (Figure 2b, e, h). To ensure that our assay for DAMP accumulation and binding was specific for acidic vesicles, we included control experiments employing monensin treatment of melanocytes from both C57BL/6^+/+ and p mutants. The carboxylic ionophore monensin is known to dissipate proton gradients by exchanging protons for potassium ions across membranes (^{Pressman, 1976}). Melanocytes from all genotypes were allowed to take up DAMP, and were then washed and exposed to monensin for 5 min. After monensin treatment, DAMP was no longer visualized within the cells (Figure 2j, k). These results confirm that in our assays the accumulation of DAMP and its retention in cell organelles is dependent on a pH gradient across the membrane of these organelles.

To determine the number of vesicles that stained for Tyrp1, DAMP, or both, we counted a total of 100 fields of 47 m² (25 fields in each of four independent experiments) of peripheral cytoplasm. A total of 7846 vesicles were counted. The average number of stained vesicles per 47 m² field was 22.4 in C57BL/6^+/+ 27.9 in p^cp/p^6H; and 28.1 in p^J/p^J. The three cell lines contained roughly the same number of Tyrp1-staining vesicles per single 47 m² field (C57BL/6^+/+, 20.0; p^cp/p^6H, 18.4; p^J/p^J, 19.1). In C57BL/6^+/+ melanocytes 94.2% of melanosomes defined by Tyrp1 staining were also acidic by DAMP staining (i.e., colocalization of Tyrp1 and DAMP, Figure 2c; Table 1). In melanocytes from p^cp/p^6H and p^J/p^J mice, however, only 7.4% and 8.4%, respectively, of the Tyrp1 staining vesicles exhibited colocalization with DAMP (Figure 2f, i; Table 1).

Table 1 - Percentage of organelles in wild-type and p mutant melanocytes that are melanosomes (defined by Tyrp1 staining), acidic vesicles (defined by DAMP staining), or both.

Full table

To confirm that the acidic vesicles visualized by DAMP staining were indeed melanosomes, DAMP staining vesicles were revealed by electron microscopy using immuno-gold conjugated anti-DNP. As shown in Figure 3, structurally identifiable melanosomes in wild-type melanocytes were clearly stained in this assay. Early stage to mature melanosomes were stained, indicating that melanosomes are acidic even before melanin synthesis and deposition.

Figure 3.

Immuno-gold labeling of melanosomes from a wild-type C57BL/6 melanocyte line. Acidic vesicles were indirectly visualized following DAMP treatment using gold beads (see Materials and Methods). Gold beads were detected in structurally identifiable melanosomes (indicated by arrows) representing various stages of melanization. Scale bar: 1 m.

Full figure and legend (288K)

We also found a class of organelles that were acidic (DAMP staining) but were not melanosomes (not Tyrp1 staining) in wild-type and mutant cell lines. These acidic nonmelanosomal vesicles accounted for about 11% of the vesicles detected in wild type and for about 33% of the vesicles detected in the two mutant cell lines (Table 1). In addition, the mutant cell lines appear to have a 25% increase in the total number of staining vesicles per field (Table 1). These extra vesicles seen in mutant cell lines were primarily nonmelanosomal acidic vesicles and were not further characterized in this study.

Discussion

We have established melanocyte cell lines that recapitulate the in vivo phenotypes of melanocytes from wild-type and p null mutant mice. Electron microscopy revealed mature and highly pigmented stage IV melanosomes in the wild-type cell line, whereas only immature stage I and stage II melanosomes were detected in the p null mutant cell lines, although the mutant melanosomes did possess tyrosinase activity using L-DOPA as substrate (Figure 1). Similar observations have been made in vivo (^{Orlow & Brilliant, 1999}) and in vitro (^{Potterf et al. 1998;Rosemblat et al. 1998}). Like their wild-type derived counterparts, the p null derived melanocytes also possess Tyrp1 staining melanosomes and their precursors (Figure 2).

To evaluate the possible role of the p protein in the acidification of melanosomes, we characterized the intracellular compartments of melanocyte cell lines derived from wild-type mice and two p mutants, p^cp/p^6H and p^J/p^J. We stained acidic compartments using DAMP, a basic congener of DNP. DAMP has a primary and tertiary amino group that becomes protonated and positively charged at acidic pH. The primary amino group also allows the DAMP molecule to become covalently linked to proteins in the presence of aldehyde fixatives, such as formaldehyde, which allows it to be retained in acidic organelles after fixation (^{Anderson et al. 1984;Anderson & Pathak, 1985}). DAMP can be visualized after fixation with appropriately tagged anti-DNP antibodies via a dinitroarene group that reacts with monoclonal antibodies to DNP. Control experiments were conducted using the ionophore monensin that disrupts the proton gradient by exchanging protons for potassium across the membrane. No DAMP was detected in the melanocytes after monensin treatment indicating that DAMP retention is dependent on a pH gradient across vesicular membranes.

Our data demonstrate that in C57BL/6^+/+ melanocytes the major class of acidic organelles contains Tyrp1 that presumably includes melanosomes and their precursors. These results confirm the earlier results of^{Moellman et al. (1988)} who used DAMP to demonstrate that the pH of the melanosomes is acidic. Using other pH indicators,^{Bhatnagar et al. (1993)} determined the pH of the melanosomes to be in the range 3.0–4.6 and found that the pH of this organelle is inversely correlated with its degree of melanization. Our results indicated that the vast majority (94.2%) of melanosomes and their precursors from the C57BL/6^+/+ melanocyte cell line were acidic (i.e., showed colocalization of Tyrp1 and DAMP). In two p deficient mutant melanocyte lines derived from p^cp/p^6H and p^J/p^J mice, however, only 7.4% and 8.4%, respectively, of melanosomes were found to be acidic (Figure 2, Table 1). Thus, the major difference between the wild-type and p mutant melanocyte cell lines is the near absence of acidic melanosomes in p mutant melanocytes. This observation suggets that the p protein plays an important role in the acidification of melanosomes.

We found more acidic vesicles in the two mutant p melanocyte lines than in the single wild-type line examined (Table 1), possibly reflecting independent variables (genetic or in vitro) between these three cell lines. Alternatively, this may underscore an, as yet unknown, compensatory mechanism. One hypothesis is that these acidic nonmelanosomal vesicles in the mutant cells represent increased numbers of lysosomes that are required to degrade the resulting nonfunctional melanosome components.

Mutations in the P gene, the human homolog of the mouse p gene, have been shown to be responsible for oculocutaneous albinism type 2 or OCA2 (^{Gardner et al. 1992;Durham-Pierre et al. 1994;Rinchik et al. 1993}). Both the mouse and human genes encode a protein with 12 membrane-spanning domains (^{Gardner et al. 1992;Rinchik et al. 1993}) localized to the melanosome membrane (^{Rosemblat et al. 1994}). With its 12 membrane-spanning domains, the p protein shares significant homology to known protein transporters (^{Gardner et al. 1992;Rinchik et al. 1993}), in particular anion transporters (^{Lee et al. 1995}).

Tyrosinase catalyzes the first and rate limiting enzymatic reaction in the production of melanin from the precursor tyrosine (reviewed by^{Pawelek & Chakraborty, 1998}). Tyrosinase is associated with the melanosomal membrane (^{Potterf et al. 1998}) and its activity is increased under acidic conditions (^{Ramaiah, 1996}). In this context there are two potential roles for a transporter protein: tyrosine transport or ionic transport to regulate melanosomal pH.^{Sidman & Pearlstein (1965)} were the first to suggest that the p protein might play a role in the transport of tyrosine. In previous studies, however, we (^{Gahl et al. 1995}) found that the characteristics of tyrosine transport in normal melanocytes were indistinguishable from melanocytes from homozygous p mutants. This result has been recently confirmed in other studies (^{Potterf et al. 1998}). These studies show that the p protein is not a tyrosine transporter. The p protein could be an ion transporter that is involved in the process of acidification of melanosomes, however.

Melanosomes and lysosomes may be products of a common (endosomal) intracellular pathway, and both types of organelles are characterized by an acidic lumen (^{Orlow, 1995}). The identity of the molecular components responsible for maintenance of the acidic environment of endosomes is not known, but it is thought that an anion channel and an adenosine-5'-triphosphate (ATP)-driven proton pump are essential components in the acidification of endosomes. Anion (e.g., Cl^– or PO₄^–) conductance provides the compensating charge balance to electrogenic proton transport (^{Al-Awqati, 1995;Van Dyke, 1996}). It is possible that the p protein is an anion transporter based on its sequence similarity to known anion transporters (^{Chen et al. 1986;Lee et al. 1995;Bun-ya et al. 1996}) involved in arsenate resistance in bacteria and PO₄^– transport in yeast. Like lysosomes, melanosomes have been shown to contain proton-translocating ATPase that may be involved in their acidification (^{Bhatnagar et al. 1993;Orlow, 1995}). The absence of this (melanosomal) proton-translocating ATPase could result in melanosomes that are not acidic. It is unlikely, however, that the p protein is this melanosomal proton-translocating ATPase, as known ATPases do not share a similar protein structure (i.e., 12 membrane-spanning domains). Thus, although the specificity of the proposed transport function of the p protein remains to be determined, the protein homology data, combined with our results, strongly suggest that the p protein is a transporter involved in melanosomal acidification. As an almost exclusively melanocyte-specific gene product (^{Gardner et al. 1992}), it is conceivable that the p protein is part of a specialized system required to acidify melanosomes that is distinct from the system used to acidify lysosomes. For example, inclusion of the p protein in its membranes may commit an endosomal vesicle to becoming a melanosome through specific protein targeting or melanosome biogenesis, in addition to providing an acidic environment favorable for optimum tyrosinase function.

Mammalian tyrosinase activity in vivo is inversely related to pH and the kinetics of tyrosinase activity in vitro is also affected by pH with a significant decrease in lag time at more acidic pH (^{Bhatnagar et al. 1993;Ramaiah, 1996}). Thus, in the absence of p-protein-mediated acidification of the melanosome, it is very likely that there is a decrease in tyrosinase activity in vivo. This alone could result in a minimal amount of melanin produced in p mutants. Intriguingly, we have also observed that under low pH conditions p melanocytes grown in vitro make more melanin (unpublished observations). Tyrosinase exhibits Michaelis-Menten kinetics (^{Laskin & Piccinini, 1986}). As such, an increase in substrate can increase the initial velocity of the reaction (i.e., M of product per min per mg of protein) even under suboptimal (nonacidic) conditions. This seems to be true in the case of melanin production mediated by tyrosinase in p mutant melanosomes. Following exposure of p mutant melanocytes to high (above physiologic) concentrations of tyrosine in vitro, there is an increase in melanin and late stage melanosomes (stage III–IV) that are almost normal in size (^{Sidman & Pearlstein, 1965;Rosemblat et al. 1994,1998;Potterf et al. 1998}).

It is also possible that the p protein may play a role in melanosome biogenesis by contributing to luminal acidification in developing melanosomes. In normal pigmented melanocytes, a fraction of the p protein is present in an intracellular compartment distinct from those containing tyrosinase and Tyrp1 (^{Rosemblat et al. 1994}). It has also been noted that melanosomes lacking p protein are missing a high molecular weight complex of the p protein, tyrosinase, and Tyrp1, and possess characteristics of immature premelanosomes (^{Rosemblat et al. 1994}; 1998). Moreover, tyrosinase and Tyrp1 localization to melanosomes, mediated by peptide sorting signals (^{Vijayasaradhi et al. 1995}), may be impaired in p mutant melanosomes (^{Potterf et al. 1998}). Although the exact biochemical function of the p protein remains to be elucidated, our data indicate that it clearly plays a role in the acidification of melanosomes that may be direct or indirect, affecting the activity and/or routing of the rate limiting enzyme, tyrosinase.

References

References

Acknowledgments

The authors wish to thank Tracey Gales and Manfred Bayer for electron microscopy studies, Jonathan Boyd for confocal studies, Carole Meyer for manuscript preparation, and Vince Hearing for generously providing the PEP1 antiserum. We thank Fayez Ghishan, Setaluri Vijayasaradhi, J Newton, Nobuko Hagiwara, and Richard Gniewek for helpful comments on the manuscript. This work was supported by NIH/NIAMS grant RO1 AR45496 (to M.H.B.).