Melanoregulin regulates retrograde melanosome transport through interaction with the RILP–p150^Glued complex in melanocytes

Melanosomes are specialized organelles that synthesize and store melanin pigments in pigment cells. Pigmentation of mammalian hair and skin requires proper formation and transport of melanosomes along two cytoskeletal elements, namely microtubules and actin filaments. Unlike other organelles, melanosomes are exceptionally visible under a light microscope, and thus are considered a good model system for analyzing organelle formation and transport (Marks and Seabra, 2001; Raposo and Marks, 2007). Three types of movements by melanosomes have been defined on the basis of their speed, direction and location in the cell: fast anterograde movement towards the plus-ends of microtubules at the periphery of the cell; fast retrograde movement towards the minus-ends of microtubules at the cell center; and slow movement on actin filaments at the cell periphery (Wu et al., 1998).

During the past decade, a tripartite protein complex composed of the small GTPase Rab27A, its specific effector Slac2a (also called melanophilin, MLPH) and the actin-based motor myosin Va, has been shown to mediate actin-based melanosome transport in mammalian epidermal melanocytes (Fukuda et al., 2002; Provance et al., 2002; Strom et al., 2002; Wu et al., 2002; Fukuda, 2005), and the Rab-myosin transport system now appears to be a fundamental mechanism of organelle transport in many cell types (Seabra and Coudrier, 2004; Akhmanova and Hammer, 2010). Disruption of the tripartite protein complex, e.g. upon deficiency of Rab27A, causes the pigmentary dilution in human Griscelli syndrome patients and the corresponding mouse model [(Van Gele et al., 2009) and references therein] as a result of melanosome clustering around the nucleus (i.e. a defect in the transfer of melanosomes from microtubules to actin filaments). Interestingly, the diluted coat color phenotype of dilute (myosin Va-deficient), ashen (Rab27A-deficient) and leaden (Slac2a-deficient) mice was found to be restored by a second mutation in the Dsu (dilute suppressor) locus in 1983 (Sweet, 1983; Moore et al., 1988). However, although the Dsu gene product, melanoregulin (Mreg; also called Dsu) (O'Sullivan et al., 2004; Boesze-Battaglia et al., 2007), has been implicated in the regulation of microtubule-dependent melanosome transport (Marks and Seabra, 2001), whose molecular machinery also remains to be elucidated in mammalian epidermal melanocytes, the function of Mreg in melanosome transport has remained unknown for more than 25 years.

In this study, we investigated the involvement of Mreg in the retrograde melanosome transport in melanocytes that is dependent on the dynein–dynactin motor complex. We showed that Mreg forms a complex with Rab-interacting lysosomal protein (RILP) and p150^Glued (also known as dynactin subunit 1, DCTN1), a component of the dynein–dynactin motor complex. A possible function of Mreg as a cargo receptor for the dynein–dynactin motor complex is discussed on the basis of our findings.

To investigate the possible role of Mreg in microtubule-dependent melanosome transport, we first overexpressed EGFP-tagged Mreg protein in black-mouse-derived melan-a cells (Bennett et al., 1987). As shown in Fig. 1A, EGFP–Mreg was clearly targeted to mature melanosomes (arrows in the inset of Fig. 1A; see also Fig. 2; endogenous Mreg is localized on mature melanosomes), and ~50% of the Mreg-expressing cells exhibited the perinuclear melanosome aggregation phenotype (Fig. 1E). Although Damek-Poprawa and co-workers have previously reported that Mreg–EGFP is localized to late endosomes and lysosomes, but not to mature melanosomes, in melanocytes (Damek-Poprawa et al., 2009), this discrepancy might be attributable to the use of different Mreg constructs, i.e. EGFP–Mreg (N-terminal EGFP tag in this study) and Mreg–EGFP (C-terminal EGFP tag in the previous study), suggesting that C-terminal EGFP-tagging prevents the melanosomal localization of Mreg. To confirm the melanosomal localization of endogenous Mreg protein, mature melanosomes were affinity-purified using anti-Rab27A IgG and the presence of Mreg in the melanosomal fractions was investigated by immunoblotting. As expected, Mreg clearly co-purified with three melanosome markers, Rab27A, tyrosinase and Tyrp1, but was not co-purified with other organelle markers (Fig. 2).

Consistent with a previous report on Mreg-null mutant mice (Moore et al., 1988), knockdown of endogenous Mreg in melan-a cells with specific shRNA had no effect on peripheral melanosome distribution [Fig. 1B,D(a),F], whereas the same shRNA is able to re-disperse melanosomes to the cell periphery in Rab27A-defective melan-ash cells (Ali et al., 2004), which normally exhibit melanosome clustering around the nucleus (Bahadoran et al., 2001; Hume et al., 2001) [Fig. 1C,D(b),G; supplementary material Fig. S1]. It was evident, however, that the melanosomes in the Mreg-deficient melan-ash cells were not attached to the plasma membrane (Fig. 1C, insets in the middle two panels) because of a defect in the Rab27A–Slp2a complex that is required for anchoring of melanosomes to the plasma membrane (Kuroda and Fukuda, 2004), in contrast to normal melan-a cells and Rab27A-re-expressing melan-ash cells, whose peripheral melanosomes were often observed just beneath the plasma membrane (Fig. 1C, inset in the far right panel).

The above results prompted us to hypothesize that Mreg is involved in the regulation of microtubule-dependent retrograde melanosome transport, and to test our hypothesis we investigated the involvement of the dynein–dynactin motor complex in the melanosome aggregation phenotype in melan-ash cells. As expected, disruption of the dynein–dynactin motor complex either by expression of p50^dynamitin (also known as dynactin subunit 2, DCTN2), a negative regulator of dynein function (Burkhardt et al., 1997), or knockdown of p150^Glued, a subunit of dynactin, in melan-ash cells, phenocopied Mreg-deficient melan-ash cells (i.e. there was re-dispersion of melanosomes to the cell periphery) (Fig. 3A,B,D,E,G). It should be noted that the Mreg-induced perinuclear melanosome aggregation phenotype in melan-a cells was almost completely eliminated by coexpression with p50^dynamitin (Fig. 3C,F), strongly suggesting that Mreg is involved in retrograde melanosome transport through regulation of the dynein–dynactin motor complex.

The results of a biochemical screening for the Mreg-binding partner in the dynein–dynactin motor complex or among its associated proteins in a coexpression assay in COS-7 cells indicated that RILP (Fig. 4A), one of the p150^Glued-binding proteins involved in lysosomal transport to the cell center (Cantalupo et al., 2001; Jordens et al., 2001), interacted with Mreg (Fig. 4B, lane 7 in the middle row). Interestingly, Mreg was unable to interact with RILP-L1, a close homologue of RILP (Wang et al., 2004) (Fig. 4A) (Fig. 4B, lane 8 in the middle row). Consistent with these findings, knockdown of endogenous RILP, but not of RILP-Ll, in melan-ash cells also phenocopied Mreg-deficient melan-ash cells (Fig. 4C–E; supplementary material Fig. S2), but the same RILP shRNA had no effect on the peripheral melanosome distribution in melan-a cells, in the same manner as the Mreg shRNA had no effect (supplementary material Fig. S2D).

Deletion analysis of RILP indicated that Mreg binds the C-terminal half of RILP (called RILP-C), which contains a second coiled-coil domain (CC2) (Fig. 4F, lane 3 in the top panel). We also attempted to determine the location of the RILP-binding site in Mreg and prepared two deletion constructs, ΔN and ΔC (Fig. 5A). Interestingly, we found that RILP binds the C-terminal portion of Mreg, ΔN (Fig. 5B, lane 3), although the C-terminal portion of Mreg is not involved in melanosomal localization (Fig. 5C, middle panels). By contrast, the N-terminal portion of Mreg (ΔC; Fig. 5C, bottom panels) was sufficient to mediate melanosomal localization, although this portion lacked RILP-binding activity (Fig. 5B, lane 4).

Because RILP was originally described as a Rab7 effector (Fig. 4B, lane 3 in the middle row) and the C-terminal half of RILP is responsible for Rab7-binding activity (Cantalupo et al., 2001; Wu et al., 2005), we proceeded to explore the relationship between Rab7 binding and Mreg binding of RILP in vitro. A competition assay indicated that Rab7 and Mreg compete with each other for RILP binding (i.e. Mreg binding to RILP was clearly suppressed in the presence of an increasing amount of Rab7; Fig. 4G, compare the top two panels) indicating that Mreg occupies the same or overlapping binding site as Rab7, which does not allow for simultaneous binding. We did not directly compare the affinity of the RILP–Rab7 interaction with that of the RILP–Mreg interaction, but, as judged from the intensity of the bands in (Fig. 4B), Rab7 seemed to bind RILP preferentially (Fig. 4G, compare lanes 3 and 7 in the middle panel).

Although Rab7 has been shown to function as a RILP–p150^Glued receptor in lysosomal transport (Jordens et al., 2001), on the basis of the following observations Rab7 itself is unlikely to function in the process of retrograde melanosome transport. First, Rab7 is not present on mature melanosomes (supplementary material Fig. S3A, insets) and has already been shown to be involved in the transport of ‘early stage’ melanosomes and not of mature melanosomes in primary human melanocytes (Jordens et al., 2006). Second, in contrast to Mreg overexpression, overexpression of a dominant-active Rab7 mutant [i.e. Rab7(Q67L)] in melan-a cells had no effect on melanosome distribution (supplementary material Fig. S3A). Third, RNAi-mediated knockdown of Rab7 in melan-ash cells failed to restore peripheral melanosome distribution (supplementary material Fig. S3B–D). To determine whether Mreg functions as a melanosomal receptor for RILP–p150^Glued in the retrograde transport of mature melanosomes, by analogy to the function of Rab7 as a receptor for RILP–p150^Glued in lysosomal transport, we performed in vitro binding experiments using purified components (see Materials and Methods for details). As shown in Fig. 5D, Mreg formed a complex with purified RILP and p150^Glued (lane 2 in the top panel) and Mreg alone did not directly interact with p150^Glued (lane 1 in the top panel). Furthermore, the results of the co-immunoprecipitation assays using anti-Mreg-specific antibody indicated that endogenous Mreg is actually associated with p150^Glued in cultured melanocytes (Fig. 5E, lane 4 in the second panel; supplementary material Fig. S4). The association between Mreg and p150^Glued is likely to be dependent on RILP, because the knockdown of endogenous RILP caused a reduction in the amount of p150^Glued that was associated with Mreg (Fig. 5E, compare lane 4 with 5 and 6 in the second and third panels).

If Mreg actually functions as a RILP–p150^Glued receptor, expression of the Mreg-binding site of RILP (i.e. RILP-C) in melan-ash cells should cause re-dispersion of the melanosomes by disrupting the endogenous Mreg–RILP interaction (i.e. a dominant-negative effect). As expected, monomeric strawberry (mStr)-tagged RILP-C was targeted to mature melanosomes in melan-ash cells (Fig. 6A, arrows, upper inset in the bottom right panel), and its expression resulted in re-dispersion of the melanosomes to the cell periphery, as occurs upon Mreg knockdown in melan-ash cells (Fig. 6A,B). By contrast, mStr–RILP-N was mostly present in the cytosol and did not rescue the perinuclear melanosome aggregation phenotype in melan-ash cells (Fig. 6A,B). Moreover, coexpression of mStr–RILP-C with EGFP–Mreg in melan-a cells also rescued the perinuclear melanosome aggregation phenotype induced by EGFP–Mreg (Fig. 6C,D), whereas expression of mStr–RILP-C alone in melan-a cells had no effect on peripheral melanosome distribution (supplementary material Fig. S5). By contrast, mStr–RILP-N again had no effect on melanosome distribution in control melan-a cells or EGFP–Mreg-expressing melan-a cells (Fig. 6C,D; supplementary material Fig. S5). In contrast to the truncated RILP mutants, coexpression of full-length RILP with EGFP–Mreg in melan-a cells further increased the proportion of melanocytes showing perinuclear melanosome aggregation to ~80%, a much greater proportion than induced by solo expression of mStr–RILP-full or EGFP–Mreg (~40%) (Fig. 6C,D).

In the present study, we have identified RILP as a binding partner of Mreg on mature melanosomes and demonstrated that Mreg is associated with the dynein–dynactin motor complex through interaction with RILP, a p150^Glued-binding partner (Fig. 4). In vitro binding experiments using purified components indicated that Mreg is able to directly interact with RILP (Fig. 5D). However, because the in vitro interaction between Mreg and RILP appeared to be weak, we cannot completely rule out the possibility that an additional unidentified factor might stabilize the Mreg–RILP–p150^Glued interaction in melanocytes. RNAi-mediated knockdown of Mreg, RILP or p150^Glued (or overexpression of p50^dynamitin) in melan-ash cells resulted in the same melanosome dispersion phenotype (Figs. 1, 3 and Fig. 4C). This phenomenon is not a unique event in melan-ash cells, because the knockdown of Mreg, RILP or p150^Glued in cytochalasin D-treated melan-a cells also caused melanosomes to disperse from around the nucleus to the cell periphery (supplementary material Fig. S6). By contrast, expression of both Mreg and RILP in melan-a cells strongly induced perinuclear aggregation of melanosomes (Fig. 6). These results allowed us to conclude that the Mreg on mature melanosomes [i.e. lysosome-related organelles (Marks and Seabra, 2001; Blott and Griffiths, 2002)] recruits the dynein–dynactin motor complex through RILP, which is analogous to the recruitment of the dynein–dynactin motor complex to lysosomes (or secretory lysosomes) through Rab7 (Jordens et al., 2001; Johansson et al., 2007; Daniele et al., 2011) (a schematic model is shown in supplementary material Fig. S7). In contrast to our finding that the knockdown of Mreg in melan-ash cells restores the peripheral melanosome distribution (Fig. 1C), however, O'Sullivan et al. previously reported that dsu, dilute double mutant did not reverse the dilute phenotype of perinuclear melanosome aggregation (O'Sullivan et al., 2004). This discrepancy might be reconciled by the presence of an alternative compensatory machinery. An additional dynein–dynactin cargo receptor other than Mreg might be present on mature melanosomes. Future extensive study will be necessary to address this issue.

Given that both Mreg and RILP are expressed in a variety of mammalian tissues in addition to melanocytes (O'Sullivan et al., 2004; Wang et al., 2004; Damek-Poprawa et al., 2009), they might also be involved in the retrograde transport of Mreg-containing organelles to the cell center in other cell types. Interestingly, it has recently been reported that lysosome-dependent phagosome degradation activity is diminished in the retinal pigment epithelial cells of Mreg-knockout mice (Damek-Poprawa et al., 2009). If Mreg is present on cathepsin D transport vesicles or engulfed materials in retinal pigment epithelial cells, this phenotype might be explained by inefficient retrograde transport of the vesicles to the cell center, where lysosomes are usually abundant. Further work is needed to determine whether Mreg actually functions as a cargo receptor for RILP–p150^Glued in retrograde transport in other cell types, including retinal pigment epithelial cells.

In conclusion, this study is the first to demonstrate that Mreg regulates microtubule-dependent retrograde melanosome transport through regulation of the dynein–dynactin motor complex, probably as a cargo receptor on mature melanosomes. Therefore, functional loss of Mreg in melan-ash cells caused re-distribution of melanosomes from the perinucleus to the cell periphery by suppression of retrograde melanosome transport.

The following antibodies used in this study were obtained commercially: anti-actin goat polyclonal antibody and horseradish peroxidase (HRP)-conjugated anti-GST antibody (Santa Cruz Biotechnology); anit-calreticulin rabbit polyclonal antibody (Thermo-Fisher Scientific); anti-EEA1 (early endosomal antigen-1) rabbit monoclonal antibody (Cell Signaling Technology); anti-EGFR (epidermal growth factor receptor) sheep polyclonal antibody (Fitzgerald Industries International, Concord, MA, USA); anti-GM130 mouse monoclonal antibody, anti-γ-adaptin mouse monoclonal antibody, anti-Lamp1 (lysosomal-associated membrane protein 1) rat monoclonal antibody (1D4b), and anti-p150^Glued mouse monoclonal antibody (BD Biosciences); anti-Lamp1 rabbit polyclonal antibody (Abcam); anti-Pmel17 mouse monoclonal antibody (HMB45; Dako, Carpinteria, CA, USA); anti-TfR (transferrin receptor) mouse monoclonal antibody (Invitrogen); anti-Tyrp1 (tyrosinase-related protein 1) mouse monoclonal antibody (Ta99; ID Labs, London, ON, Canada); anti-GFP (green fluorescent protein) rabbit polyclonal antibody (MBL, Nagoya, Japan); and HRP-conjugated anti-T7 tag mouse monoclonal antibody (Merck Biosciences Novagen). Anti-tyrosinase rabbit polyclonal antibody and anti-Rab27A rabbit polyclonal antibody were prepared as described previously (Saegusa et al., 2006; Beaumont et al., 2011). Anti-FLAG rabbit polyclonal antibody, HRP-conjugated anti-FLAG tag (M2) mouse monoclonal antibody and anti-FLAG tag antibody-conjugated agaroase were obtained from Sigma-Aldrich.

GST–Mreg, GST–RILP-N (amino acids 1–199), and GST–Rab7 were expressed in bacteria and purified by standard protocols as described previously (Kuroda and Fukuda, 2005). Japanese White rabbits were immunized with the purified GST–Mreg (GST–RILP-N or GST–Rab7), and specific antibodies were affinity-purified by exposure to antigen-bound Affi-Gel 10 beads (Bio-Rad Laboratories) according to the manufacturer's instructions (Fukuda and Mikoshiba, 1999).

cDNAs encoding the open reading frame of the mouse Mreg, RILP, p50^dynamitin and p150^Glued were amplified from Marathon-Ready adult mouse brain, liver and/or testis cDNA (BD Biosciences Clontech) by PCR with specific pairs of oligonucleotides as described previously (Fukuda et al., 1999). Truncated mutants of RILP (i.e. RILP-N, amino acids 1–199; and RILP-C, amino acids 200–369; see Fig. 4F; and of Mreg, i.e. Mreg-ΔN, amino acids 76–214; and Mreg-ΔC, amino acids 1–139; Fig. 5A) were also prepared by conventional PCR techniques. The oligonucleotide sequences used in this study are available from the authors on request. The cDNA fragments amplified were subcloned into the pEGFP-C1 vector (BD Biosciences Clontech), pmStr-C1 vector, which was obtained by replacing the EGFP insert of pEGFP-C1 with mStr, pEF-T7/T7-GST/FLAG/HA vectors (Fukuda et al., 1999), and/or pGEX-4T-3 vector (GE Healthcare). pSilencer 2.1-U6 neo vector was from Ambion. pSilencer-Mreg-st1 (#1) (19-base target site: 5′-CTGCACTGCCTTCCATTTC-3′), pSilencer-Mreg-st2 (#2) (19-base target site: 5′-TCCGTATTCCTCCTTTGGA-3′), pSilencer-RILP-st1 (#1) (19-base target site: 5′-CAGAGCTTGGAACCTGATG-3′), pSilencer-RILP-st2 (#2) (19-base target site: 5′-GTCCAAGGTGTTTCTGCTG-3′), pSilencer-RILP-L1-st1 (#1) (19-base target site: 5′-GAATGAGGACGTCGAGGCT-3′), pSilencer-RILP-L1-st2 (#2) (19-base target site: 5′-GGAGGTGGTGGACAAGCAG-3′), pSilencer-Rab7-st1 (#1) (19-base target site: 5′-TTCCCTGAACCCATCAAAC-3′) and pSilencer-Rab7-st2 (#2) (19-base target site: 5′-GAAAGTGTTGCTGAAGGTC-3′) were constructed essentially as described previously (Kuroda and Fukuda, 2004). Stealth RNA oligonucleotides against mouse p150^Glued was obtained from Invitrogen (catalogue numbers MSS203510 and MSS203511; designated as #1 and #2, respectively). Because similar results were obtained with two independent shRNAs and/or siRNAs against Mreg, RILP, RILP-L1, p150^Glued or Rab7, the effect of the siRNAs and shRNAs reported in this study are not likely to be an off-target effect. pEGFP-C1-Mreg^SR and pEGFP-C-1-RILP^SR (siRNA-resistant mutants) were prepared by two-step PCR techniques as described previously (Tamura et al., 2011). Other expression plasmids, including pEF-FLAG-Rab7, pEF-FLAG-Rab27A, pEF-FLAG-Rab34, pEGFP-C1-Rab7, pEGFP-C1-Rab7(Q67L), and pEF-FLAG-RILP-L1, were prepared as described previously (Fukuda, 2003; Tsuboi and Fukuda, 2006; Fukuda et al., 2008).

The immortal mouse melanocyte cell lines melan-a, derived from a black mouse, and melan-ash, derived from an ashen mouse (a generous gift of Dorothy C. Bennett, St George's Hospital Medical School, London, UK), were cultured on glass-bottomed dishes (35-mm dish; MatTek, Ashland, MA, USA) as described previously (Bennett et al., 1987; Kuroda et al., 2003; Ali et al., 2004). B16-F1 cells were also cultured as described previously (Kuroda et al., 2002). Plasmids (e.g. pmStr, pEGFP and pSilencer vectors) were transfected into melanocytes by using FuGENE 6 (Roche Applied Science) according to the manufacturer's instructions. Two days after transfection, cells were fixed in 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, and stained with the following antibodies: anti-p150^Glued mouse monoclonal antibody (1:100 dilution), anti-Mreg rabbit antibody (18.7 μg/ml), anti-FLAG rabbit polyclonal antibody (1:100 dilution) and anti-T7 tag mouse monoclonal antibody (1:300 dilution). The antibodies were visualized with anti-mouse or rabbit Alexa-Fluor-488 or -594-conjugated IgG (Invitrogen), and examined for fluorescence with a confocal fluorescence microscope (Fluoview; Olympus, Tokyo, Japan). The images were processed with Adobe Photoshop software (CS3). Melanosome distribution assays (i.e. perinuclear aggregation versus peripheral dispersion) were performed as described previously (n>50 from three independent dishes) (Kuroda et al., 2003), and data are expressed as means+s.d. Statistical analyses were performed with Student's unpaired t test.

The total RNA of mouse melan-a cells (or melan-ash cells) that had been transfected with pSilencer vectors twice was prepared using TRI-reagent (Sigma-Aldrich Corp.) and was subjected to reverse transcription (RT) using ReverTra Ace (Toyobo, Osaka, Japan) according to the manufacturer's instructions. The cDNA encoding mouse Mreg, RILP or RILP-L1 was amplified by PCR with the following pairs of oligonucleotides: 5′-GGATCCAATTTATGGAGCATGCCT-3′ (Mreg-5′ primer, sense) and 5′-GTCGACCATTTGGCTAGCAATCTT-3′ (Mreg-3′ primer, antisense) for Mreg; 5′-TCCAGAAGCTGTCCAGTGTG-3′ (RILP-5′ primer, sense) and 5′-TTGCTGTGGGGTCTCTTCTC-3′ (RILP-3′ primer, antisense) for RILP; and 5′-GGATCCATGGAGGAGCCGCTAGGGTC-3′ (RILP-L1-5′ primer, sense) and 5′-GTCGACCTCCTCCTCCCCATT-3′ (RILP-L1-3′ primer, antisense) for RILP-L1. The mouse G3PDH (glyceraldehyde 3-phosphate dehydrogenase; GAPDH) was amplified using mouse G3PDH-5′ and G3PDH-3′ primers (BD Biosciences Clontech) as a reference.

Immunoaffinity purification of Rab27A-bound melanosomes with anti-Rab27A-IgG-conjugated magnetic beads was also performed as described previously (Kuroda and Fukuda, 2004). In brief, B16-F1 cells (one confluent 10-cm dish) were homogenized in a homogenization buffer [5 mM HEPES-KOH pH 7.2, 5 mM EGTA, 0.03 M sucrose and protease inhibitors (Complete; Roche Applied Science)]. After centrifugation at 800 g for 10 minutes, the supernatant obtained was incubated for 2 hours at 4°C with anti-Rab27A rabbit polyclonal antibody or control rabbit IgG in the presence of 1% BSA (bovine serum albumin), and then with Dynabeads M-280 (Invitrogen) for 30 minutes at 4°C. After washing the beads twice with PBS (phosphate-buffered saline), the bound fractions were subjected to SDS-PAGE (10% gels) followed by immunoblotting with the antibodies indicated in Fig. 2.

Co-immunoprecipitation assays in COS-7 cells (Fig. 4B; Fig. 5B) and GST pull-down assays (Fig. 4F) were performed as described previously (Fukuda et al., 1999; Fukuda and Kanno, 2005). In brief, T7–GST–RILP-N or T7–GST–RILP-C was transiently expressed in COS-7 cells and affinity-purified by glutathione–Sepharose beads (GE Healthcare). Beads coupled to either T7–GST–RILP-N or T7–RILP-C were incubated for 2 hours at 4°C with COS-7 cell lysates containing HA–Mreg. After washing the beads three times with 1 ml of washing buffer (50 mM HEPES-KOH pH 7.2, 150 mM NaCl, and 1 mM MgCl₂), the proteins bound to the beads were analyzed by SDS-PAGE (10% gels) followed by immunoblotting with HRP-conjugated anti-T7 tag antibody and anti-Mreg antibody. Similarly, FLAG–Mreg, FLAG–Mreg-ΔN or FLAG–Mreg-ΔC was transiently expressed in COS-7 cells and affinity-purified by anti-FLAG tag antibody-conjugated agarose. Beads coupled to FLAG–Mreg truncated proteins were incubated for 2 hours at 4°C with COS-7 cell lysates containing T7–GST–RILP, and the proteins bound to the beads were analyzed by SDS-PAGE (10% gels) followed by immunoblotting with HRP-conjugated anti-T7 tag antibody and HRP-conjugated anti-FLAG tag antibody.

Competition experiments between Mreg and Rab7 (Fig. 4G) was performed essentially as described previously (Fukuda et al., 2004). In brief, T7–GST–RILP-C was transiently expressed in COS-7 cells and affinity-purified by glutathione–Sepharose beads. Beads coupled to T7–GST–RILP-C or beads alone were incubated for 2 hours at 4°C with COS-7 cell lysates containing T7–Mreg and/or increasing amount of FLAG–Rab7 in 50 mM HEPES-KOH pH 7.2, 150 mM NaCl, 1 mM MgCl₂ and 0.5 mM GTPγS. After washing the beads three times with 1 ml of the washing buffer, the proteins bound to the beads were analyzed by SDS-PAGE (10% gels) followed by immunoblotting with anti-Mreg antibody, HRP-conjugated anti-FLAG tag antibody and HRP-conjugated anti-T7 tag antibody.

For in vitro direct-binding experiments, T7–GST–RILP and FLAG–p150^Glued were transiently expressed in COS-7 cells and affinity-purified by glutathione–Sepharose beads and anti-FLAG tag-conjugated agarose beads, respectively, as described previously (Fukuda and Kanno, 2005). GST–Mreg was prepared from bacteria as described above. Beads coupled to FLAG–p150^Glued were incubated for 2 hours with purified GST–Mreg alone or GST–Mreg and T7–GST–RILP in 50 mM HEPES-KOH pH 7.2, 150 mM NaCl, 1 mM MgCl₂ and protease inhibitors (Complete; Roche Applied Science). After washing the beads three times with 1 ml of the washing buffer, the proteins bound to the beads were analyzed by SDS-PAGE (10% gels) followed by immunoblotting with HRP-conjugated anti-GST antibody and HRP-conjugated anti-FLAG tag antibody.

Immunoprecipitation of the endogenous molecules from melanocyte lysates (Fig. 5E; supplementary material Fig. S4) was performed essentially as described previously (Fukuda and Kanno, 2005). In brief, the lysates from B16-F1 cells transfected with pSilencer-RILP-st1 or -st2 (#1 and #2) or pEF-FLAG-RILP were incubated for 2 hours at 4°C with either anti-Mreg rabbit polyclonal antibody described above or control rabbit IgG, and then with protein-A–Sepharose beads for 1 hour at 4°C. After washing the beads three times with 1 ml of washing buffer, the proteins bound to the beads were analyzed by SDS-PAGE (10% gels) followed by immunoblotting with the indicated antibodies in Fig. 5E and supplementary material Fig. S4.

We thank Dorothy C. Bennett for kindly donating melan-a and melan-ash cells, Megumi Aizawa for technical assistance and members of the Fukuda Laboratory for valuable discussions.