Evidence for Distinct Membrane Traf®c Pathways toMelanosomes and Lysosomes in Melanocytes

Hideaki Fujita,1 Emi Sasano,1 Kumiko Yasunaga, Koh Furuta,* Sadaki Yokota,² Ikuo Wada,³ andMasaru HimenoGraduate School of Pharmaceutical Sciences, Kyushu University, Fukuoka, Japan; *University of Occupational and Environmental Medicine,

Kita-Kyushu, Japan; ²Yamanashi Medical University, Yamanashi, Japan; ³Sapporo Medical University, Hokkaido, Japan

We report here morphologic and biochemicalevidence that melanosomes are distinct from lyso-somes. Immuno¯uorescence analysis revealed thatTRP-1, a melanosomal membrane protein, did notcolocalize with lysosomal membrane proteinsLAMP1 and LGP85 in melan-a cells. Wortmannintreatment of melanocytes enhanced the distinct com-partmentalization of these melanosomal/lysosomalmembrane proteins by the swelling of the endo-somal-lysosomal systems. The heavily melanizedmelanosomes did not have an altered shape, whichsuggests a lesser degree of membrane dynamics of

stage IV melanosomes. Terminal lysosomes loadedwith TR-dextran are also distinct from melano-somes, thus indicating that melanosomes are isolatedfrom the endocytic pathway that is a representativeroute to lysosomes. Because AP-3 mutation leads tomistargeting of both melanosome and lysosomemembrane proteins, we propose that there is a latesorting step for melanosomes and lysosomes in mela-nocytes after AP-3 sorting. Key words: protein sorting/targeting signal/AP-3. Journal of InvestigativeDermatology Symposium Proceedings 6:19±24, 2001

Lysosomes are general organelles that are expressed in mostcells, while melanosomes are speci®cally expressed inmelanocytes. There are many common features betweenlysosomes and melanosomes, such as acidic compartmentsand with acid hydrolases (Boissy et al, 1987; Diment et al,

1995), and it has been suggested that melanosomes are specializedlysosomes (Orlow, 1995, 1998).

Morphologic analyzes suggest the melanosomes biogenesismodel, in which the membrane of melanosomes originate fromsmooth ER, and this compartment receives melanin-synthesizingenzymes from the TGN-derived vesicles (Jimbow et al, 1997;Orlow, 1998). Immature melanosomes gradually mature intoheavily melanized melanosomes and are ®nally transferred toadjacent cells, keratinocytes, by cytophagocytosis (Okazaki et al,1976); however, there is little biochemical evidence to support thishypothesis and the origin of the membrane and how themelanosomes construct the ®lamentous matrix inside of themembrane during maturation are not clear.

Other researchers reported that membrane proteins are targeted tothe lysosomes and melanosomes using the same targeting motif, so-called tyrosine or di-leucine based motifs (Williams and Fukuda,1990; Vega et al, 1991a, b; Vijayasaradhi et al, 1995; Calvo et al,1999). Both lysosomal and melanosomal membrane proteins contain

targeting signals in their cytoplamic tails. Tyrosinase and TRP-1,melanosomal membrane proteins, contain more than two potentialtargeting motifs (Simmen et al, 1999; Vijayasaradhi et al, 1995),whereas lysosomal membrane proteins have one of each. Currentlyit is not clear whether melanosomal membrane proteins require twotargeting motifs or if one of each is suf®cient for targeting.

Genetic analysis revealed that gene products of coat colormutation in mice and eye color mutation in the fruit ¯y,Drosophilla, encoded various membrane transport factors (Lloydet al, 1998; Odorizzi et al, 1998), mutation in one of the AP-3subunits was identi®ed as mocha in mice and garnet in the ¯y(Simpson et al, 1997; Kantheti et al, 1998). Moreover, VPS41(light in ¯y) (Warner et al, 1998), pallidin (syntaxin 13 bindingprotein; pallid in mice) (Huang et al, 1999), and myosin-V(dilute in mice) (Wu et al, 1998) are possibly implicated invesicle transport and sorting to lysosomes and melanosomes.AP-3 recognizes the Di-leucine-based motif of membraneproteins (Honing et al, 1998) and myosin-V is required formelanosomes interactions with actin ®laments at the peripheralsite of melanocytes (Wu et al, 1998). This genetic andbiochemical evidence suggests that both melanosomes andlysosomes utilize the same mechanism for membrane transport;however, there are speci®c features in melanosomes, melanin-synthesizing enzymes, and matrix ®laments or the laminarmatrix inside the membrane. Furthermore, melanosomes aretransferred to adjacent cells, keratinocytes, by cytophagocytosis(Okazaki et al, 1976). Therefore, it is likely that melanosomesdiffer from lysosomes and have a speci®c protein-targetingmechanism.

Our immunolocalization analysis revealed that melanosomes aredistinct from lysosomes and we propose that melanosomes are notjust another lysosome but have a speci®c membrane transportroute.

1087-0024/01/$15.00 ´ Copyright # 2001 by The Society for Investigative Dermatology, Inc.

19

Manuscript received June 14, 2001; accepted for publication June 14,2001.

Reprint requests to: Dr. Masaru Himeno, Graduate School ofPharmaceutical Sciences, Kyushu University, Fukuoka, Japan. Email:himeno@bioc.phar.kyushu-u.ac.jp

Abbreviations: AP, adaptor protein; GFP, green ¯uorescent protein;LAMP1, lysosome-associated membrane protein-1; LGP, lysosomalglycoprotein; PFA, paraformaldehyde; PLC, prelysosomal compartment;TGN, trans-Golgi network; TRP-1, tyrosinase-related protein-1.

1These authors contributed equally to this work.

MATERIALS AND METHODS

Cell culture Murine melanoma, melan-a cells were maintained inDMEM supplemented with 10% fetal bovine serum and phorbol 12-myristate 13-acetate (Life Technology, Gaithersburg, MD). Forimmuno¯uorescence experiments, the cells were grown on the polylysine (10 mg per ml) coated coverslips.

Immuno¯uorescence analysis Cells were ®xed with 4% PFA inphosphate-buffered saline (PBS) and then permeabilized with methanolat 4°C. Fixed cells were blocked with PBS containing 1% (wt/vol)bovine serum albumin and processed for indirect immuno¯uorescenceusing the following primary antibodies: anti-TRP-1 tail antibody (af®nitypuri®ed rabbit polyclonal, 5 mg per ml), anti-LGP85 tail antibody(af®nity puri®ed rabbit polyclonal, 5 mg per ml), antirat cathepsin Lantibody (af®nity puri®ed rabbit polyclonal, 25 mg per ml), antimouseLAMP1 antibody (rat monoclonal ascites, 1:100 dilution). All thesecondary antibodies (Cy3-goat antirat, Alexa 488-goat antirabbit) wereused at 1:200±300 dilution.

Antibodies Synthetic peptide corresponding to the middle ofthe deduced cytoplasmic portion of mouse TRP-1 (509±528:EANQPLLTDHYQRYAEDY EEC) and LGP85 tail (457±477:CRGQGSTDEGTADERAPLIRT, the tail sequences are identicalamong mouse, rat, and human) were synthesized and coupled to bovineserum albumin as immunogen (Sawady Technology, Tokyo, Japan).Rabbit polyclonal antibodies against each peptide were generated byKBT Oriental (Tosu, Japan) and af®nity puri®cation was performed withthe peptide column, respectively. Rabbit polyclonal anti-rat cathepsin Lwas generated with GST-mature cathepsin L fusion protein as theimmunogen and the speci®c antibody was prepared by the sequentialaf®nity puri®cation with GST and a GST-fusion one coupled toSepharose 4B (Pharmacia Biotech, Uppsala, Sweden). Rat monoclonalantimouse LAMP1 antibody (1D4B) was a generous gift from Dr. T.August (Johns Hopkins University, Baltimore, MD). Secondaryantibodies coupled with Cy3 or Alexa488 were purchased from Jackson(West Grove, PA) and Molecular Probes (Eugene, OR), respectively.

Plasmid construction and transfection Plasmid encoding thechimeric protein, human TRP-1, and GFP was generated by thepolymerase chain reaction (PCR) with oligonucleotides corresponding tothe amino- and carboxyl-terminal of human TRP-1. Primers (HokkaidoSystem Science, Sapporo, Japan) were designed to obtain an Eco RI siteat the 5¢ end (5¢-CTCTGAATTCAAGCAGAATGAGTGCTCC-3¢)and Sal I site at the 3¢ end (5¢-CTGGAGTCGACCATATTCTTTCT-TCAGC-3¢). The PCR fragment was directionally subcloned intopHSG399 (TAKARA, Kyoto, Japan), sequenced, and re-cloned intopEGFP-N1 vector (Clontech, Palo Alto, CA). Plasmid transfection wasperformed with Fugen6 Transfection Reagent (Boehringer Mannheim,Indianapolis, IN) according to the manufacturer's protocol.

Imaging Immuno¯uorescence images were observed on a LeicaDMRB microscope (Wetzlar, Germany), acquired through cooled CCDcamera, MicroMAX (Princeton Instruments, Trenton, NJ) and digitally

processed using IPlab Software (Scanalytics, Fairfax, VA). All imagesfrom the real time movie, immuno¯uorescence microscopy, and westernblotting were subsequently assembled and labeled using AdobePhotoshop (Adobe Systems, Mountain View, CA).

Immunoisolation of melanosomes and lysosomes Preparation ofheavy granular fractions was done essentially according to Gahl's methodfrom cultured melan-a cells (Gahl et al, 1995). Granular fractions wereincubated with af®nity puri®ed tail speci®c antibodies (anti-TRP-1 tailor anti-LGP85 tail) or control IgG (anti-GST) at 4°C for 16 h.Subsequently, bovine serum albumin blocked protein A agarose(Boehringer Mannheim, Indianapolis, IN) was added and followed by afurther incubation for 3 h. The sedimented beads were washed fourtimes with 0.25 M sucrose/Hepes/EDTA buffer and bound organellewas solubilized with Laemmli's sample buffer for the western blotting.

TR-dextran labeling of melan-a Melan-a cells were labeled with2 mg per ml of TR-dextran (Molecular Probes) for 15 min at 37°C,chased for the indicated time, and then ®xed for immuno¯uorescence.

RESULTS

Wortmannin-induced vacuole formation in melan-acells Wortmannin, a phosphatidyl inositide-3-kinase inhibitor,causes vacuolation of endosomes and lysosomes because oftransport inhibition from prelysosomal compartments (PLC) tolysosomes (Reaves et al, 1996). We examined the effect ofwortmannin on melanosomes in mouse melanocyte, melan-acells. Wortmannin (10 mM)-treated melan-a cells were monitoredevery 3 min in living cells (Fig 1). Small black granules (thinarrows at 0 min) within the cytoplasm are melanosomes. Sphericalsmall vacuoles began to appear near the nucleus 6 min afterwortmannin treatment (thick arrows) and these small vacuolesmoved around, increased in size, and sometimes fused; however,the small black granules (melanosomes) never got into the vacuoles.This ®nding suggested that although vacuolation of endosomes andlysosomes had been observed in the melanocytes, the heavilymelanized melanosomes (stage IV) did not change in shape andnever reached wortmannin-induced vacuoles in melan-a.

TRP-1 did not colocalize with either lamp1 or LGP85 inmelanocytes The staining of melanosomes with anti-TRP-1antibody had the cytoplasmic punctuate pattern with strongconcentration at the perinuclear in melan-a cells (Fig 2B).Monoclonal antibody against Lamp1, a lysosomal membraneprotein, also showed strong perinuclear staining representing bothlate endosomes and lysosomes, in addition to the peripheralcytoplasmic signals indicating early endosomes (Fig 2A). Althoughthey shared a partial overlap at the perinucleous region (probably atthe late endosomes), cytoplasmic and peripheral staining wasdistinct (Fig 2C). In the presence of wortmannin, both

Figure 1. Wortmannin-induced vacuole formation in melan-a cells. Wortmannin (10 mM)-treated melan-a cells were monitored every 3 min inliving cells. Small black granules (thin arrows at 0 min) within the cytoplasm are melanosomes. Spherical small vacuoles begin to appear near the nucleusat 6 min after wortmannin treatment (thick arrows) and these small vacuoles moved around, increased in size, and sometimes fused.

20 FUJITA ET AL JID SYMPOSIUM PROCEEDINGS

melanosomes and lysosomes were swollen, probably because of animbalance of membrane retrieval or due to inhibition of the fusion(Fig 2D, E). There were few vacuoles stained with both TRP-1and LAMP1 but many vacuoles were positive for one but not the

other (Fig 2F). The wortmannin-induced vacuolation had anexaggerated distinctive compartmentalization in the cytoplasm.LGP85, another lysosomal membrane protein (Fujita et al, 1991),revealed perfect colocalization with LAMP1 (data not shown),

Figure 2. Localization of LAMP1 and TRP-1 in melan-a cells. Cells were treated without (A±C) or with (D±F) 10 mM wortmannin for 1 h at37°C. TRP-1 did not colocalize with LAMP1. Large arrows indicate solely LAMP1-positive structures and small arrows represent TRP-1-speci®ccompartments, respectively. Wortmannin-induced vacuolation exaggerated the distinct compartmentalization of each membrane protein.

Figure 3. Localization of hTRP-1-GFP and LGP85 in melan-a cells. hTRP-1-GFP plasmid was transfected and the cells were treated without(A±C) or with (D±F) 10 mM wortmannin for 3 h at 37°C. Small arrows indicate solely TRP-1-GFP-positive structures and large arrows representLGP85-speci®c compartments, respectively. Both proteins were distributed into the distinct vacuole in the presence of wortmannin.

VOL. 6, NO. 1 NOVEMBER 2001 MEMBRANE TRAFFIC TO MELANOSOMES AND LYSOSOMES 21

thereby suggesting that the distinct staining pattern between TRP-1 and LAMP1 was speci®c and signi®cant. Interestingly, GFP-tagged TRP-1 behaved in the same way as endogenous TRP-1 inmelanocytes (data not shown), but did not colocalize with LGP85in the absence or presence of wortmannin (Fig 3A±F).

Immunoseparation of lysosomes from melanosomessuggesting the distinct but partial overlap of melanosomeswith lysosomes To observe the distinct compartmentalizationof melanosomes and lysosomes in melanocytes, we wanted toimmunoseparate melanosomes from lysosomes, using tail-speci®cantibodies, rabbit polyclonal antibodies against the cytoplasmic tailsof both LGP85 and TRP1, respectively, following protein-A

agarose beads sedimentation. The isolated fractions were analyzedby western blotting (Fig 4). The subcellular fraction isolated withthe antibody to LGP85 tail was devoid of TRP-1, which suggestedthat the isolated fraction was highly enriched in lysosomes but notin melanosomes, i.e., melanosomes were distinct from lysosomes.The reciprocal experiment, the separation of melanosomes fromlysosomes with the antibody against TRP-1, was not successful.TRP-1 positive fractions contained signi®cant amounts of LGP85,which may mean that lysosomes were more abundant thanmelanosomes in melan-a. Therefore, the anti-LGP85 antibodywas primarily saturated with lysosomes, whereas the TRP-1antibody fully interacted with both melanosomes and late-endosomes, the sorting place of TRP-1 and LGP85.

Melanosomes are distinct from terminal lysosomes Wethen asked if melanosomes are the terminal of the endocyticpathway as lysosomes. In melan-a cells, TR-dextran wasinternalized and chased for the indicated time, then the ®xedcells were immunostained with anti-LAMP1, cathepsin L, andTRP-1 antibodies, respectively. Cathepsin L is one of the majoracid hydrolases in lysosomes and is well characterized as the markerenzyme of terminal lysosomes with dense and acidic compartments.

At an early time point (30 min or 1 h), TR-dextran waslocalized in the early endocytic compartments and did notcolocalize with LAMP1 or TRP-1 (Fig 5A, E). TR-dextrancolocalized with cathepsin L at 24 h, which suggested that TR-dextran was delivered to the terminal lysosomes in melanocytes atthis time point (Fig 5D). LAMP1 colocalized with TR-dextran atlater time points, 8 and 24 h after the chase period; however, theextent of colocalization was slightly less than that seen withcathepsin L at 24 h (Fig 5B, C). This may mean that LAMP1localizes on not only lysosomes but also the late/early endosomes.In contrast, there was no colocalization of TR-dextran with TRP-1 positive compartments at any time point (Fig 5E±G). These®ndings suggested that melanosomes are not the terminal of theendocytic pathway as lysosomes. The endocytic molecules neverencountered the biosynthetic route of melanosomes.

Figure 4. Immunoisolation of melanosomes and lysosomes usingtail-speci®c antibodies. LGP85-positive compartments and TRP-1-positive compartments were immunoisolated with cytoplasmic tail-speci®c antibodies, respectively. The isolated fractions were analyzed bywestern blotting. LGP85-positive compartments are devoid of TRP-1,whereas TRP-1-positive fractions contained signi®cant amounts ofLGP85.

Figure 5. Relationship between the endocytic routes and melanosomes. Melan-a cells were labeled with TR-dextran, chased for the indicatedtimes at 37°C, and then processed for the immuno¯uorescence analysis with antibodies against LAMP1 (A±C), cathepsin L (D), and TRP-1 (E±G),respectively. At an early time point, TR-dextran was internalized via ¯uid phase endocytosis and delivered to early endosomes those are devoid ofLAMP1 (A). At a later time point, TR-dextran reached late endosomes and lysosomes and colocalized with LAMP1 (B, C) and cathepsin L (D);however, TRP-1 never colocalized with TR-dextran at any time point (E±G).

22 FUJITA ET AL JID SYMPOSIUM PROCEEDINGS

DISCUSSION

Melanosomes and lysosomes are highly related organelles and theprotein targeting to these compartments is based on the vesiculartransport. Accumulating evidence suggests targeting signals andmolecular mechanisms that regulate the vesicle formation, trans-port, and fusion are common; however, it is not clear to whatextent melanosomes are related to lysosomes in melanocytes.

To characterize the membrane property of melanosomes weapplied wortmannin to melanocytes. Wortmannin treatment(10 mM) caused the appearance of the spherical vacuoles as withother types of cells; however, the black granules (heavily melanizedmelanosomes) never got into these vacuoles. Ultrastructural analysisalso indicated that the heavily melanized melanosomes did notchange shape and size in the presence of this drug (data not shown).As they have an organized internal matrix (i.e., ®laments) within amembrane and are rigid, they have little membrane retrieval. Whenmelanosomes have matured, they are separated from any membranetransport pathway and lose the membrane dynamics.

We have found that melanosomes and lysosomes are distinct inmelanocytes and wortmannin-induced vacuolation exaggerated thedistinct compartmentalization of each membrane protein. LAMP1and LGP85 colocalized either in the presence or in the absence ofwortmannin, whereas TRP-1 indicated a little overlap withlysosome markers only at the perinuclear region where TGN orlate endosomes are located. The appearance of swollen TRP-1-speci®c vacuoles in the presence of wortmannin suggests thatTRP-1 is not restricted on the mature melanosomes and hassigni®cant distribution on the immature melanosomes, which arewortmannin. Interestingly, TRP-1 redistributed into the distinctvacuoles from that of lysosomes in the presence of wortmannin. Alikely explanation for the appearance of swollen TRP-1-speci®cvacuoles is that TRP-1 is not restricted on the mature melanosomesand has signi®cant distribution on the immature melanosomes,where they are wortmannin sensitive but distinct from both heavilymelanized melanosomes and lysosomes. This compartment re¯ectsthe existence of premelanosomal compartments, which are equiva-lent to PLC in nonmelanocytes. Orlow et al have shown thatTRP-1 is limited to a less dense melanosomal compartment incontrast to tyrosinase, which is eventually localized to heavilymelanized melanosomes. Their observation also supports theheterogeneity of melanosomes and the existence of the pre-melanosomal compartments (Orlow et al, 1993); however, they didnot distinguish melanosomes from lysosomes and indicated thecolocalization of LAMP1 and TRP-1 in melanocytes. Thediscrepancy between our results and theirs might be explained bythe difference of cell types or the degree of melanization. Because

TRP-1 and LAMP-1/LGP85 are already segregated at thepremelanosomal compartment, the ®nal step of the sorting couldbe taking place before this, possibly at late endosomes.

Lysosomes are regarded as the terminal compartment of theendocytic pathway, which contain acid hydrolases but have lessmembrane retrieval in the presence of wortmannin (Reaves et al,1996; Bright et al, 1997). The PLC is thought to be an intermediatecompartment between lysosomes and late endosomes. Transiently,this compartment is formed by direct fusion of lysosomes with lateendosomes. Contents are rapidly mixed and the membranes areretrieved and redistributed to their original positions. ThereforePLC, so-called late endosome-lysosome hybrid organelle, is verydynamic and unstable (Mullock et al, 1998; Luzio et al, 2000; Pryoret al, 2000). Wortmannin inhibits antetrograde movement fromlate endosomes to lysosomes, and consequently enhances formationof hybrid compartments. We assume that in melanocytes there areboth PLC and pre-melanosomal compartments, whereas non-melanocytes contain only PLC. In melanocytes, wortmannin mayinduce vacuolation of both compartments. Immunoelectronmicroscopic analysis strongly supported our assumption, as it isclear that heavily melanized melanosomes are devoid of bothLGP85 and LAMP1 (data not shown). We also found a co-localization of both TRP-1 and LAMP1 in immature melanosomesand late endosomes (data not shown), which might be the sortingplace for the melanosome proteins.

A selective isolation of the subset of lysosomes from melano-somes indicates direct evidence that lysosomes are distinct frommelanosomes; however, it is controversial as to whether the TRP-1-positive compartment contains signi®cant amounts of LGP85.One interpretation of this discrepancy is that the number oflysosomes is more abundant than that of melanosomes inmelanocytes. Therefore, the LGP85 tail antibody easily reacts andis saturated with lysosomes, whereas the TRP-1 tail antibody pullsdown the mixture of melanosomes and late endosomes, which arethe sorting sites of melanosomes and lysosomes. Further carefulanalysis needs to be carried out to support this.

Based on the endocytic analysis in melanocytes, melanosomes arealso distinct from the endocytic pathway. When terminal lysosomeswere labeled with TR-dextran for 24 h chase, however, TRP-1never colocalized with endocytic markers at any time point.Therefore, we propose that melanosomes are segregated from theendocytic route. Because heavily melanized melanosomes havelittle membrane retrieval, it is unlikely that they receive endocyticmolecules from the outside of the cells. Interestingly theendocytosed TR-dextran colocalized well with LAMP-1 at anearly time point, such as 8 h after the internalization, whereas theytook longer to reach the cathepsin L-positive terminal lysosomes.

Figure 6. Model of the membrane traf®c tomelanosomes and lysosomes in melanocytes.In biosynthetic pathways, the traf®c route ofmelanosome membrane protein (TRP-1) partiallyoverlaps but is distinct from that of the endo/lysosomal one (LAMP1, LGP85). The melano-some pathway is separated from the lysosome oneafter the AP-3-dependent sorting steps, probablyat late endosomes and/or PLC. Melanosomesgradually maturate into heavily melanizedmelanosomes (from stage I to stage II±IV) withthe accumulation of melanin inside themembrane. In the presence of wortmannin, thereis membrane retrieval from the lysosomes orimmature melanosomes, but not from the heavilymelanized melanosomes. The melanosomepathway is also segregated from the endocyticpathway (TR-dextran).

VOL. 6, NO. 1 NOVEMBER 2001 MEMBRANE TRAFFIC TO MELANOSOMES AND LYSOSOMES 23

This result also represents the existence of the subpopulation andthe heterogeneity of lysosomes. LAMP1 is known as a recyclingmolecule between lysosomes and plasmamembrane, therefore thiscolocalization is found at early and late endosomes rather than atlysosomes. There is evidence for retrograde traf®c betweenterminal lysosomes and PLC (Reaves et al, 1996); however, thereis no evidence for the presence of premelanosomal compartments.Our results are a demonstration of the presence of the intermediatecompartments and the membrane.

In summary, we propose the following model of membraneprotein transport to the lysosomes and melanosomes in melanocyte(Fig 6). Both melanosome and lysosome membrane proteinsfollow the AP-3 pathway until the late endosomes and/or PLC,because AP-3 mutation causes mistargeting of both membraneproteins to the cell surface (Le Borgne et al, 1998); however, afterleaving the late endosome and/or PLC, they must take differentroutes to their ®nal destinations. Whereas lysosome membraneproteins are delivered to dense lysosomes, melanosome membraneproteins reach to the immature melanosomes (stage I). At this stage,it is likely that melanosome membrane protein, TRP-1 can beretrieved from there to the former sorting stations (PLC and/or lateendosomes). Stage I melanosomes gradually maturate into heavilymelanized melanosomes (stages II, III, and IV), lose the membranedynamics, and then are transferred to the adjacent cells,keratinocytes by the cytophagocytosis.

It is reasonable that melanosomes and lysosomes have differentcontributions to the cell homeostasis; melanosomes are for thepigmentation and lysosomes are for the degradation. To performthese distinct roles, they need to have a speci®c protein-sorting andtransport system. Our ®ndings support the idea that the melanocytehas a speci®c sorting place and a mechanism to have melanosomesindependent from lysosomes. Furthermore, heavily melanizedmelanosomes have unique membrane structures and proteins/lipidscomposition to be speci®cally transferred to the adjacent cells. Thisis the only example of the intercellular organelle transfer phenom-ena. Identi®cation the molecular mechanism of this phenomenon isvery important to understand the pigmentation in melanocytes.How is the sorting of melanosome and lysosome performed orwhat kind of molecules are implicated in these steps? The geneticanalysis of coat color mutation in mouse (Odorizzi et al, 1998) andeye color mutation in the fruit ¯y Drosophilla (Lloyd et al, 1998)may give us the answer to these questions.

We thank Dr. J. T. August for kindly providing antimouse LAMP1 monoclonal

antibody, Dr. Y. Tanaka for helpful assistance on the subcellular fractionation, M.

Ohara for comments on the manuscripts, and Dr. K. Jimbow for helpful comments

and discussion. This work was supported in part by a CREST grant from the

Sciences and Technology Corporation of Japan, the Kao Foundation For Arts and

Sciences, and Grants-in-Aid for Scienti®c Research from the Ministry of Education,

Science, Sports, and Culture of Japan.

REFERENCES

Boissy RE, Moellmann GE, Halaban R: Tyrosinase and acid phosphatase activities inmelanocytes from avian albinos. J Invest Dermatol 88:292±300, 1987

Bright NA, Reaves BJ, Mullock BM, Luzio JP: Dense core lysosomes can fuse withlate endosomes and are re-formed from the resultant hybrid organelles. J CellSci 110:2027±2040, 1997

Calvo PA, Frank DW, Bieler BM, Berson JF, Marks MS: A cytoplasmic sequence inhuman tyrosinase de®nes a second class of di-leucine-based sorting signals forlate endosomal and lysosomal delivery. J Biol Chem 274:12780±12789, 1999

Diment S, Eidelman M, Rodriguez GM, Orlow SJ: Lysosomal hydrolases are presentin melanosomes and are elevated in melanizing cells. J Biol Chem 270:4213±4215, 1995

Fujita H, Ezaki J, Noguchi Y, Kono A, Himeno M, Kato K: Isolation and sequencingof a cDNA clone encoding 85kDa sialoglycoprotein in rat liver lysosomalmembranes. Biochem Biophys Res Commun 178:444±452, 1991

Gahl WA, Potterf B, Durham-Pierre D, Billiant MH, Hearing VJ: Melanosomaltyrosine transport in normal and pink-eyed dilution murine melanocytes.Pigment Cell Res 8:229±233, 1995

Honing S, Sandoval IV, von Figura K: A di-leucine-based motif in the cytoplasmictail of LIMP-II and tyrosinase mediates selective binding of AP-3. Embo J17:1304±1314, 1998

Huang L, Kuo YM, Gitschier J: The pallid gene encodes a novel, syntaxin 13-interacting protein involved in platelet storage pool de®ciency. Nat Genet23:329±332, 1999

Jimbow K, Gomez PF, Toyofuku K, Chang D, Miura S, Tsujiya H, Park JS:Biological role of tyrosinase related protein and its biosynthesis and transportfrom TGN to stage I melanosome, late endosome, through gene transfectionstudy. Pigment Cell Res 10:206±213, 1997

Kantheti P, Qiao X, Diaz ME, et al: Mutation in AP-3 delta in the mocha mouselinks endosomal transport to storage de®ciency in platelets, melanosomes, andsynaptic vesicles. Neuron 21:111±122, 1998

Le Borgne R, Alconada A, Bauer U, Ho¯ack B: The mammalian AP-3 adaptor-likecomplex mediates the intracellular transport of lysosomal membraneglycoproteins. J Biol Chem 273:29451±29461, 1998

Lloyd V, Ramaswami M, Kramer H: Not just pretty eyes. Drosophila eye-colourmutations and lysosomal delivery. Trends Cell Biol 8:257±259, 1998

Luzio JP, Rous BA, Bright NA, Pryor PR, Mullock BM, Piper RC: Lysosome-endosome fusion and lysosome biogenesis. J Cell Sci 113:1515±1524, 2000

Mullock BM, Bright NA, Fearon CW, Gray SR, Luzio J: Fusion of lysosomes withlate endosomes produces a hybrid organelle of intermediate density and is NSFdependent. J Cell Biol 140:591±601, 1998

Odorizzi G, Cowles CR, Emr SD: The AP-3 complex: a coat of many colours.Trends Cell Biol 8:282±288, 1998

Okazaki K, Uzuka M, Morikawa F, Toda K, Seiji M: Transfer mechanism ofmelanosomes in epidermal cell culture. J Invest Dermatol 67:541±547, 1976

Orlow SJ: Melanosomes are specialized members of the lysosomal lineage oforganelles. J Invest Dermatol 105:3±7, 1995

Orlow SJ: The biogenesis of melanosomes. In: The Pigmentary System (Nordlund JJBR, Hearing VJ, eds). New York: Oxford University Press, 1998, pp 97±106

Orlow SJ, Boissy RE, Moran DJ, Pifko-Hirst S: Subcellular distribution of tyrosinaseand tyrosinase-related protein-1: implications for melanosomal biogenesis. JInvest Dermatol 100:55±64, 1993

Pryor PR, Mullock BM, Bright NA, Gray SR, Luzio JP: The role of intraorganellarCa(2+) in late endosome-lysosome heterotypic fusion and in the reformationof lysosomes from hybrid organelles. J Cell Biol 149:1053±1062, 2000

Reaves BJ, Bright NA, Mullock BM, Luzio JP: The effect of wortmannin on thelocalisation of lysosomal type I integral membrane glycoproteins suggests a rolefor phosphoinositide 3- kinase activity in regulating membrane traf®c late inthe endocytic pathway. J Cell Sci 109:749±762, 1996

Simmen T, Schmidt A, Hunziker W, Beermann F: The tyrosinase tail mediatessorting to the lysosomal compartment in MDCK cells via a di-leucine and atyrosine-based signal. J Cell Sci 112:45±53, 1999

Simpson F, Peden AA, Christopoulou L, Robinson MS: Characterization of theadaptor-related protein complex, AP-3. J Cell Biol 137:835±845, 1997

Vega MA, Rodriguez F, Segui B, Cales C, Alcalde J, Sandoval IV: Targeting oflysosomal integral membrane protein LIMP II. The tyrosine-lacking carboxylcytoplasmic tail of LIMP II is suf®cient for direct targeting to lysosomes. J BiolChem 266:16269±16272, 1991a

Vega MA, Segui-Real B, Garcia JA, Cales C, Rodriguez F, Vanderkerckhove J,Sandoval IV: Cloning, sequencing, and expression of a cDNA encoding ratLIMP II, a novel 74-kDa lysosomal membrane protein related to the surfaceadhesion protein CD36. J Biol Chem 266:16818±16824, 1991b

Vijayasaradhi S, Xu Y, Bouchard B, Houghton AN: Intracellular sorting andtargeting of melanosomal membrane proteins: identi®cation of signals forsorting of the human brown locus protein, gp75. J Cell Biol 130:807±820, 1995

Warner TS, Sinclair DA, Fitzpatrick KA, Singh M, Devlin RH, Honda BM: Thelight gene of Drosophila melanogaster encodes a homologue of VPS41, a yeastgene involved in cellular-protein traf®cking. Genome 41:236±243, 1998

Williams MA, Fukuda M: Accumulation of membrane glycoproteins in lysosomesrequires a tyrosine residue at a particular position in the cytoplasmic tail. J CellBiol 111:955±966, 1990

Wu X, Bowers B, Rao K, Wei Q, Hammer JAR: Visualization of melanosomedynamics within wild-type and dilute melanocytes suggests a paradigm formyosin V function In vivo. J Cell Biol 143:1899±1918, 1998

24 FUJITA ET AL JID SYMPOSIUM PROCEEDINGS