Melanosomes are transported to the cell periphery of melanocytes by coordination between bidirectional microtubule-dependent movements and unidirectional actin-dependent movement. Although both the mechanism of the actin-dependent melanosome transport and the mechanism of the microtubule-dependent retrograde melanosome transport in mammalian skin melanocytes have already been determined, almost nothing is known about the mechanism of the microtubule-dependent anterograde melanosome transport. Small GTPase Rab proteins are common regulators of membrane traffic in all eukaryotes, and in this study we performed genome-wide screening for Rab proteins that are involved in anterograde melanosome transport by expressing 60 different constitutive active (and negative) mutants, and succeeded in identifying Rab1A, originally described as a Golgi-resident Rab, as a prime candidate. Endogenous Rab1A protein was found to be localized to mature melanosomes in melanocytes, and its functional ablation either by siRNA-mediated knockdown or by overexpression of a cytosolic form of Rab1A-GTPase-activating protein/TBC1D20 induced perinuclear melanosome aggregation. The results of time-lapse imaging further revealed that long-range anterograde melanosome movements were specifically suppressed in Rab1A-deficient melanocytes, whereas retrograde melanosome transport occurred normally. Taken together, these findings indicate that Rab1A is the first crucial component of the anterograde melanosome transport machinery to be identified in mammalian skin melanocytes.
Melanocytes are a special type of cell in the skin that synthesize and store melanin pigments in specialized organelles called melanosomes. Melanosomes are formed and mature around the nucleus of the melanocyte and are transported to the cell periphery along two components of the cytoskeleton, microtubules and actin filaments. Mature melanosomes are first transported to the peripheral area of the cell by long-range, bidirectional, microtubule-dependent movements and then transferred to actin filaments and carried by unidirectional, local movement at the cell periphery (reviewed by Marks and Seabra, 2001; Aspengren et al., 2009; Ohbayashi and Fukuda, 2012). These melanosome movements are thought to be essential for skin pigmentation in mammals, and defects in the melanosome transfer step between microtubules and actin filaments have actually been shown to cause human Griscelli syndrome, a rare autosomal recessive disorder characterized by pigmentary dilution in hair and skin (reviewed by Van Gele et al., 2009). Genetic and biochemical analyses of Griscelli syndrome patients and of a murine model of Griscelli syndrome in the past decade have revealed that the molecular mechanism of actin-dependent melanosome transport is mediated by a tripartite protein complex composed of Rab27A, its specific effector Slac2-a/melanophilin, and an actin-based motor myosin Va (Fukuda et al., 2002; Provance et al., 2002; Strom et al., 2002; Wu et al., 2002) (reviewed by Fukuda, 2005).
Most studies of the molecular mechanism of microtubule-dependent melanosome transport have been conducted on amphibian melanophores, and involvement of kinesin motors (i.e. kinesin II) and cytoplasmic dynein in microtubule-dependent anterograde and retrograde melanosome transport, respectively, has been reported (Tuma et al., 1998; Deacon et al., 2003; Aspengren et al., 2009). More recently, involvement of a dynein–dynactin motor complex in retrograde melanosome transport in cultured mouse melanocytes has also been reported. Melanoregulin, a dilute suppressor (dsu) gene product (Sweet, 1983; O'Sullivan et al., 2004), regulates retrograde melanosome transport through interaction with RILP–p150Glued (a subunit of the dynein–dynactin motor complex) (Ohbayashi et al., 2012). However, very little is known about the involvement of specific kinesin motors, e.g. kinesin II/Kif3, in microtubule-dependent anterograde melanosome transport (referred to simply as anterograde melanosome transport below) in mammalian skin melanocytes. To investigate the molecular mechanism of anterograde melanosome transport, we focused on the small GTPase Rabs, conserved membrane trafficking proteins in all eukaryotic cells (Fukuda, 2008; Stenmark, 2009), because several members of the mammalian Rab family function as a cargo receptor for kinesin motors (reviewed by Akhmanova and Hammer, 2010; Horgan and McCaffrey, 2011). For example, Kif20A/Rabkinesin-6 interacts with Rab6A/B to regulate the movement of Rab6-bearing vesicles (Echard et al., 1998) and KIF16B interacts with Rab14 to regulate trafficking of the fibroblast growth factor receptor 2 (FGFR2) (Ueno et al., 2011). Moreover, several Rab isoforms, e.g. Rab3A, Rab9, Rab11, and Rab27, indirectly interact with particular kinesins via their effector molecules (Jackson et al., 2008; Niwa et al., 2008; Arimura et al., 2009; Schlager et al., 2010). These observations led us to hypothesize that a certain as yet unidentified Rab isoform is also involved in anterograde melanosome transport in mammalian skin melanocytes.
In this study we performed a genome-wide analysis of members of the mouse Rab family and succeeded in identifying Rab1A, originally reported as a Golgi-resident Rab involved in ER-to-Golgi transport (Allan et al., 2000; Satoh et al., 2003), as a candidate for the melanosome-resident Rab that regulates microtubule-dependent transport. We showed that knockdown of Rab1A by specific siRNAs or overexpression of its GTPase-deficient mutant in cultured mouse melanocytes induced perinuclear aggregation of melanosomes. We also showed that anterograde melanosome transport was dramatically suppressed in Rab1A-deficient melanocytes. Our findings indicate that Rab1A is a novel component of the anterograde melanosome transport machinery in mammalian skin melanocytes.
Screening for Rabs that affect melanosome distribution
We used the two methods described below to search for novel Rab proteins involved in anterograde melanosome transport. We first screened for constitutive active/negative (CA/CN) mutants of Rabs, because such mutants are generally thought to modulate particular membrane trafficking events (Fukuda, 2010). We transiently expressed CA/CN mutants of 60 different Rab proteins (i.e. Rab1–43) with an enhanced green fluorescent protein (EGFP) tag in melan-a cells and selected Rab proteins that were capable of inducing perinuclear melanosome aggregation. Of the 60 Rab proteins tested, we selected the 15 Rab proteins whose CA or CN mutants induced perinuclear melanosome aggregation as initial candidates (Fig. 1; Table 1). They included Rab8A and Rab27A whose CA mutants have previously been shown to cause perinuclear melanosome aggregation in melanocytes because of the defect in the melanosome transfer step between microtubules and actin filaments (Chabrillat et al., 2005; Bahadoran et al., 2003). We also observed an interesting phenotype called ‘enlarged melanosomes’ that was induced by the expression of EGFP–Rab5(CA) or EGFP–Rab22(CA) (Table 1 and supplementary material Fig. S1). Enlarged melanosomes are reminiscent of the enlarged early endosomes that are induced by the CA mutant of Rab5 in HeLa cells as a result of an increased rate of homotypic fusion between early endosomes (Stenmark et al., 1994). At this stage, however, we do not know whether enlarged melanosomes are large single melanosomes or aggregates of small melanosomes that have adhered to each other. Then, because expression of a Rab(CA) mutant can affect membrane trafficking even though the corresponding Rab protein is not endogenously expressed in the cells [e.g. even though Rab27B is not endogenously expressed in melanocytes (Barral et al., 2002), expression of its CA mutant in melanocytes induced perinuclear melanosome aggregation (Fig. 1)], we used the second method, RNA interference technology, to screen these 15 candidate Rabs by knocking down expression of each of them in melan-a cells. To do so, shRNA-expressing plasmids for each candidate Rab were constructed, and the knockdown efficiency of the shRNA was evaluated by co-transfecting shRNA-expressing plasmids and EGFP–Rab-expressing plasmids into COS-7 cells (supplementary material Fig. S2A). The shRNA-expressing plasmids that were found to be effective were then introduced into melan-a cells to knock down endogenous Rab proteins. As shown in Fig. 2, knockdown of endogenous Rab1A or Rab27A induced perinuclear melanosome aggregation, whereas knockdown of the other 13 Rab proteins had no effect on melanosome distribution in melan-a cells. Although Rab27A is a well-known melanosome-resident Rab whose deficiency results in perinuclear melanosome aggregation in melanocytes (Wilson et al., 2000; Kuroda and Fukuda, 2004), Rab1A was originally reported as a Golgi-resident Rab involved in ER-to-Golgi transport (Allan et al., 2000; Satoh et al., 2003). However, Rab1A has recently been shown to regulate other intracellular trafficking events, including trafficking of early endocytic vesicles (Wang et al., 2010; Mukhopadhyay et al., 2011; Kicka et al., 2011), and its presence on early melanosomes has been demonstrated by a proteomic analysis of melanosomes (Chi et al., 2006). We therefore selected Rab1A as the prime candidate for the Rab protein that is involved in melanosome transport in melanocytes, and we attempted to characterize the function and localization of Rab1A in melanocytes in greater detail.
|Rab name||Perinuclear aggregation||Enlarged melanosomes||Rab name||Perinuclear aggregation||Enlarged melanosomes|
|Rab name||Perinuclear aggregation||Enlarged melanosomes||Rab name||Perinuclear aggregation||Enlarged melanosomes|
Each EGFP–Rab(CA/CN) protein was transiently expressed in melan-a cells, and its effects on melanosome distribution and size were assessed by bright-field microscopy.
Melanosome distribution: +, ++ and +++ indicate that 20-40%, 40-60%, >60%, respectively, of the EGFP–Rab-expressing cells exhibited perinuclear melanosome aggregation; n>50 (Kuroda et al., 2003); melanosome size: +, most of the melanocytes contained enlarged melanosomes; ±, only a few melanocytes contained enlarged melanosomes; −, no phenotype was observed.
The only CN mutant that affected melanosome distribution.
Localization of Rab1A protein on mature melanosomes and its requirement for normal peripheral melanosome distribution
To rule out the possibility that N-terminal EGFP-tagging of Rab1A(CA) artificially induces the perinuclear melanosome aggregation phenotype (Fig. 1), we expressed monomeric strawberry (mStr)-tagged Rab1A wild-type (WT), Q70L (CA mutant = GTPase-deficient mutant), and S25N (CN mutant) in melan-a cells, and investigated their effect on melanosome distribution. As expected, more than 60% of the mStr–Rab1A(Q70L)-expressing cells exhibited the perinuclear melanosome aggregation phenotype, the same as the EGFP–Rab1A(Q70L)-expressing cells did (Fig. 3A, second row of panels, and Fig. 3B), whereas neither mStr–Rab1A(WT) nor mStr–Rab1A(S25N) affected melanosome distribution. It should be noted that both Rab1A(WT) and Rab1A(Q70L) were clearly present on mature melanosomes (Fig. 3A, insets) and that Rab1A(S25N) was not associated with mature melanosomes at all, suggesting that Rab1A is recruited to mature melanosomes in a GTP-dependent manner. To confirm the melanosomal localization of endogenous Rab1A protein, we initially attempted to perform an immunofluorescence analysis with anti-Rab1A-specific antibody (supplementary material Fig. S2D), but, unfortunately, our antibody did not work in melan-a cells. To overcome this problem, we biochemically isolated mature melanosomes with anti-Rab27A IgG (Kuroda and Fukuda, 2004) and analyzed the melanosomal fraction for the presence of Rab1A by immunoblotting. Consistent with the results of the immunofluorescence analysis of mStr–Rab1A (Fig. 3A), Rab1A was clearly co-purified with four melanosome markers, Rab27A, tyrosinase, Tyrp1, and melanoregulin (Ohbayashi et al., 2012), and was not co-purified with other organelle markers, including EEA1 (an early endosome marker), calreticulin (an ER marker), and GM130 (a Golgi marker; Fig. 4). We therefore concluded that Rab1A is a novel mature melanosome-resident Rab, the same as Rab27A.
We adopted two different strategies to confirm the effect of Rab1A shRNA on melanosome distribution (Fig. 2). In the first strategy, we prepared two additional independent siRNAs (no. 1 and no. 2) against mouse Rab1A and introduced them into melan-a cells. Both siRNAs depleted endogenous Rab1A protein in the melan-a cells (Fig. 5B, lanes 2 and 3), and ∼30% of the Rab1A siRNA-treated cells exhibited the perinuclear melanosome aggregation phenotype (Fig. 5A,C). That effect was very unlikely to have been attributable to the disruption of the Golgi complex, where Rab1A is also localized, or to an abnormality of the cytoskeletal components, because disruption of the Golgi structure by brefeldin A (BFA) had no effect on melanosome distribution in melan-a cells (supplementary material Fig. S3) and the distribution of actin filaments and microtubules seemed to be intact even in Rab1A-deficient cells (supplementary material Fig. S4). Furthermore, knockdown of Rab1A in melan-a cells had no clear effect on Golgi morphology under our experimental conditions (supplementary material Fig. S5A), thereby ruling out the possibility that disruption of the Golgi complex causes the perinuclear melanosome aggregation phenotype. In the second strategy, we prepared siRNA-resistant Rab1A (supplementary material Fig. S2B,C) to rescue the phenotype induced by Rab1A shRNA (Fig. 5D,E). Approximately 30% of the Rab1A shRNA-treated cells exhibited the perinuclear melanosome aggregation phenotype, the same as the Rab1A siRNA-treated cells had, and the phenotype was completely rescued by co-expression of Rab1ASR. Taken together, these results clearly indicated that the observed effects shown in Fig. 5 were not caused by an off-target effect of shRNA/siRNAs. The relatively low rate of induction of the perinuclear aggregation phenotype by Rab1A siRNA/shRNA may have been attributable to insufficient knockdown of endogenous Rab1A protein (see next section).
Inactivation of Rab1A by a specific Rab1A-GTPase-activating protein (TBC1D20) caused perinuclear aggregation of melanosomes
If Rab1A is actually required for melanosome transport, forced inactivation of Rab1A on mature melanosomes should cause perinuclear melanosome aggregation. To demonstrate that it did, we focused on TBC (Tre-2/Bub2/Cdc16) domain-containing Rab-GAPs (GTPase-activating proteins), because their overexpression often inhibits particular membrane trafficking events (e.g. expression of Rab27A-GAPα/EPI64 inhibits actin-dependent melanosome transport by inactivating endogenous Rab27A protein on mature melanosomes) (Itoh and Fukuda, 2006; Fukuda, 2011). Since a Rab1A-specific GAP, TBC1D20, had already been identified by other groups (Haas et al., 2007; Sklan et al., 2007), we initially overexpressed Myc-tagged wild-type TBC1D20 (Myc–TBC1D20) in melan-a cells to inactivate endogenous Rab1A protein. However, the Myc–TBC1D20 was mainly localized at the ER-like structures of melan-a cells, consistent with the previous results obtained in BSC-1 cells and HeLa cells (Sklan et al., 2007), and it did not efficiently induce perinuclear melanosome aggregation (Fig. 6A, right panels, B). Since TBC1D20 contains a transmembrane domain, which is required for its ER localization (Sklan et al., 2007), we hypothesized that the ER localization of TBC1D20 restricts its access to Rab1A on mature melanosomes and that the restricted access in turn decreases the rate of inactivation of Rab1A in melanocytes. To test our hypothesis, we prepared a TBC1D20 mutant lacking a transmembrane domain (named TBC1D20ΔTM) and overexpressed the mutant in melan-a cells. Consistent with our hypothesis, the Myc–TBC1D20ΔTM was dispersed throughout the cytosol, and ∼60% of the TBC1D20ΔTM-expressing cells exhibited perinuclear melanosome aggregation (Fig. 6A, left panels, B). This phenotype would have been directly caused by the inactivation of Rab1A, because a GAP-activity-deficient mutant (R105A) of TBC1D20ΔTM had no effect on melanosome distribution (Fig. 6A, right panels, B). Since the rate of perinuclear melanosome aggregation induced by TBC1D20ΔTM or by Rab1A(Q70L) (∼60%) was approximately twice as high as induced by Rab1A knockdown (∼30%), the Rab1A shRNA/siRNAs used in this study may have been insufficient to decrease the endogenous protein level and induce the phenotype in a high percentage of the cells. These results collectively indicated that the Rab1A on mature melanosomes is actually required for normal peripheral melanosome distribution in melanocytes.
Involvement of Rab1A in anterograde melanosome transport in melanocytes
Although functional ablation of Rab1A either by knockdown with specific siRNAs/shRNA (Fig. 5) or by overexpression of TBC1D20ΔTM (Fig. 6) induced perinuclear melanosome aggregation in melanocytes, it is unclear how Rab1A regulates peripheral melanosome distribution. We think that there are three possibilities in regard to the function of Rab1A in melanosome transport. The first possibility is that Rab1A is involved in the melanosome transfer step between microtubules and actin filaments, the same as Rab27A, whose deficiency induces perinuclear melanosome aggregation because of increased activity of retrograde transport and, as a result, causes the diluted coat color of ashen mice and silvery hair of type 2 Griscelli syndrome patients (Fukuda, 2005; Van Gele et al., 2009). One of the other two possibilities is that Rab1A is involved in anterograde melanosome transport along microtubules, and the third possibility is that Rab1A is involved in retrograde melanosome transport along microtubules. A deficiency of Rab1A may inhibit anterograde melanosome transport or promote retrograde melanosome transport. To determine which of these possibilities actually occurs in melanocytes, we performed siRNA-mediated double-knockdown experiments on Rab27A-deficient melan-ash cells, which normally exhibit melanosome clustering around the nucleus (Wilson et al., 2000; Ali et al., 2004). We previously reported that knockdown of endogenous p150Glued (i.e. disruption of the dynein–dynactin motor complex for retrograde melanosome transport) in melan-ash cells caused the melanosomes to re-disperse to the cell periphery (Ohbayashi et al., 2012). Thus, if Rab1A were involved in anterograde transport, knockdown of both p150Glued and Rab1A in melan-ash cells should suppress the re-dispersion of melanosomes to the cell periphery. On the other hand, if Rab1A were also involved in microtubule-dependent retrograde transport (or melanosome transfer from microtubules to actin filaments), knockdown of both p150Glued and Rab1A in melan-ash cells should have no effect on the re-dispersion of melanosomes to the cell periphery. As shown in Fig. 7A, knockdown of p150Glued alone caused the melanosomes in melan-ash cells to disperse from around the nucleus to the cell periphery as reported previously (left bottom panel) (Ohbayashi et al., 2012), whereas simultaneous knockdown of p150Glued and Rab1A significantly inhibited the dispersion of melanosomes (right bottom panel). Thus, depletion of Rab1A is likely to affect anterograde melanosome transport and not to affect retrograde transport or actin-dependent transport.
Finally, we directly analyzed melanosome movements in Rab1A-deficient melan-a cells and Rab27A-deficient melan-a cells, whose microtubule movements are intact, as a control to determine whether knockdown of Rab1A actually decreases anterograde melanosome transport. First, we defined ‘anterograde movements’ and ‘retrograde movements’ as movements toward the cell periphery and movements toward the nucleus, respectively (see Fig. 8A) (Wu et al., 1998), and tracked all melanosome movements as described in the Materials and Methods. As shown in Fig. 8B and supplementary material Fig. S6a–d, the Rab1A-deficient melanocytes exhibited shorter-range anterograde melanosome movements than the Rab27A-deficient melanocytes did. Actually, the average distance traveled during the anterograde movements in the Rab1A-deficient cells was significantly shorter than the average distance traveled in the Rab27A-deficient cells, and long-range anterograde movements were rarely observed in the Rab1A-deficient cells in comparison with the Rab27A-deficient cells (Fig. 8C,E and supplementary material Fig. S6a–d). By contrast, however, no clear differences in the properties of the retrograde melanosome movements, including their average distance, were observed between the Rab1A-deficient cells and Rab27A-deficient cells (Fig. 8D,F and supplementary material Fig. S6e–h). The maximum speed of anterograde transport was also slightly, but significantly, slower in the Rab1A-deficient cells (1.42±0.58 µm/sec; P<0.01, Student's unpaired t-test) than in the Rab27A-deficient cells (1.68±0.58 µm/sec). On the other hand, no significant difference in the average maximum speed of retrograde transport was observed between the Rab1A-deficient cells (1.73±0.69 µm/sec) and Rab27A-deficient cells (1.87±0.65 µm/sec). These results collectively indicate that Rab1A is functionally involved in long-range microtubule-dependent anterograde melanosome transport rather than in retrograde melanosome transport or actin-dependent melanosome transport.
In the present study we performed genome-wide screening for Rab proteins involved in microtubule-dependent anterograde melanosome transport in melanocytes for the first time and succeeded in identifying Rab1A as a prime candidate (Figs 1, 2). Both exogenously expressed Rab1A protein and endogenous Rab1A protein were found to be localized on mature melanosomes (Figs 3, 4), and functional ablation of endogenous Rab1A protein by specific siRNAs or by overexpression of TBC1D20ΔTM, a cytosolic mutant of Rab1A-GAP, strongly induced the perinuclear melanosome aggregation phenotype (Figs 5, 6). In addition, a GTPase-deficient, constitutive active mutant of Rab1A(Q70L) also induced the same aggregation phenotype, suggesting that a proper GTP–GDP exchange cycle is required for the function of Rab1A in melanosome transport (Fig. 3). A similar dominant interfering effect has previously been reported for the CA mutant of Rab27A, which is crucial for actin-dependent melanosome transport, in melan-a cells (Fig. 1) (Bahadoran et al., 2003). We then demonstrated two lines of evidence that Rab1A is involved in the microtubule-dependent anterograde transport of melanosomes rather than in their retrograde transport or actin-dependent transport. The first line of evidence is that knockdown of Rab1A in p150Glued-deficient melan-ash cells suppressed melanosome dispersion, which is induced by inhibition of the retrograde transport machinery (i.e. knockdown of p150Glued; Fig. 7), and the second line of evidence is that the rate of long-range anterograde melanosome movements in Rab1A-deficient melan-a cells was dramatically reduced in comparison with Rab27A-deficient melan-a cells (Fig. 8C,E). By contrast, retrograde movements in Rab1A-deficient melan-a cells seemed to be unaffected, and no significant difference was observed in the distance traveled during retrograde melanosome transport (Fig. 8D,F).
Our finding in regard to the localization of Rab1A on mature melanosomes and its involvement in their anterograde transport were surprising, because Rab1A was originally reported as a Golgi-resident Rab that regulates ER-to-Golgi transport (Allan et al., 2000; Satoh et al., 2003). However, since recent studies have shown that Rab1A also regulates intracellular trafficking events other than ER-to-Golgi transport, e.g. the motility of early endocytic vesicles (Wang et al., 2010; Mukhopadhyay et al., 2011; Kicka et al., 2011), Rab1A was suspected of being involved in the microtubule-dependent movement of organelles, including melanosomes. Consistent with our findings, it has recently been reported that defects in TRAPPC6A, a subunit of a Rab1-specific guanine nucleotide exchange factor (mTRAPPII complex) (Yamasaki et al., 2009), resulted in mosaic loss of coat pigment in mice (Gwynn et al., 2006). Further work will be necessary to determine whether a defect in melanosome transport actually occurs in TRAPPC6A-deficient mice.
How Rab1A regulates anterograde melanosome transport is an open question, but we speculate that Rab1A associates with a microtubule-based motor kinesin directly or indirectly via an effector molecule, as has been proposed in regard to motor-associated Rab proteins (Akhmanova and Hammer, 2010; Horgan and McCaffrey, 2011). Possible interactions between Rab1A and kinesin motors are now under investigation in our laboratory, but the results of our preliminary experiments have indicated that Rab1A does not directly interact with several kinesins, including kinesin II/Kif3, which is critical for anterograde melanosome transport in amphibian melanophores (Tuma et al., 1998) (M. I. and M. F., unpublished observations). Very recently, we identified SKIP/PLEKHM2 (SifA and kinesin-interacting protein) (Boucrot et al., 2005) as a specific GTP-Rab1A-binding protein (Fukuda et al., 2011). As its name indicates, SKIP directly interacts with the kinesin-1 motor (Boucrot et al., 2005; Dumont et al., 2010), and the SKIP–kinesin-1 complex regulates transport of lysosomes to the cell periphery in HeLa cells via an interaction between the RUN domain of SKIP and the GTP-bound form of Arl8 (Arf-like small GTPase) (Rosa-Ferreira and Munro, 2011). Since Rab1A also interacts with the RUN domain of SKIP (Fukuda et al., 2011) and melanosomes are lysosome-related organelles, it is highly possible that the Rab1A–SKIP–kinesin-1 complex transports melanosomes to the cell periphery in melanocytes (schematic model is shown in supplementary material Fig. S7). Further work will be necessary to determine whether this complex is actually involved in microtubule-dependent anterograde melanosome transport.
In addition to Rab1A(CA), we found that expression of 13 other CA Rab mutants in melan-a cells also induced perinuclear melanosome aggregation (Fig. 1). Although knockdown of each of these Rabs individually had no effect on melanosome distribution (Fig. 2), we cannot completely rule out the possibility that several of them contribute to microtubule-dependent melanosome transport in a redundant fashion. Since inhibition of the function of endogenous Rab1A by TBC1D20ΔTM or overexpression of Rab1A(Q70L) caused perinuclear melanosome aggregation in a maximum of 60% of the transfected cells, it is possible that a Rab1A-independent anterograde melanosome transport system exists and that certain other Rabs may be involved in it. It is also possible that several of these Rabs regulate microtubule-dependent retrograde melanosome transport, because activation of retrograde transport (e.g. overexpression of melanoregulin) also induced perinuclear melanosome aggregation (Ohbayashi et al., 2012). Actually, our most recent data indicate that Rab36 is involved in retrograde melanosome transport (Matsui et al., 2012). We also found that Rab5(CA) and Rab22(CA) induced enlarged melanosomes (supplementary material Fig. S1). Since these Rab isoforms regulate endosomal trafficking (Stenmark et al., 1994; Kauppi et al., 2002) and endosomal transport systems are thought to be required for melanosome formation (or biogenesis) (Raposo and Marks, 2007; Ohbayashi and Fukuda, 2012), Rab5 and/or Rab22 may regulate the size of melanosomes during melanosome formation. Extensive research will be necessary in the future to determine the exact function of each of these Rabs in melanosome biogenesis and transport.
In summary, we discovered that small GTPase Rab1A is present on mature melanosomes in cultured melanocytes and that it is required for long-range microtubule-dependent anterograde melanosome transport. So far as we know, Rab1A is the first key factor in anterograde melanosome transport in mammalian skin melanocytes to be identified. We anticipate that future elucidation of the Rab1A effector protein that associates with a kinesin motor(s) will reveal the full machinery of the anterograde melanosome transport complex.
Materials and Methods
Anti-tyrosinase rabbit antibody, anti-Rab27A rabbit antibody, and anti-melanoregulin (Mreg) rabbit antibody were prepared as described previously (Saegusa et al., 2006; Beaumont et al., 2011; Ohbayashi et al., 2012). The following antibodies used in this study were obtained commercially: anti-actin antibody (ABM, Richmond, BC, Canada); anit-calreticulin rabbit polyclonal antibody (Thermo-Fisher Scientific, Waltham, MA, USA); anti-EEA1 rabbit antibody (Cell Signaling Technology, Beverly, MA, USA); horseradish peroxidase (HRP)-conjugated anti-GFP (green fluorescent protein) antibody (MBL, Nagoya, Japan); anti-GM130 mouse monoclonal antibody and anti-p150Glued mouse monoclonal antibody (BD Biosciences, San Jose, CA, USA); anti-Rab1A rabbit antibody (Abnova, Taipei, Taiwan); anti-β-tubulin mouse antibody (TUB2.1) and ant-ERGIC53 rabbit antibody (Sigma-Aldrich Corp., St. Louis, MO, USA); and anti-Tyrp1 goat antibody and anti-Myc mouse antibody (9E10) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Texas-Red-conjugated phalloidin was obtained from Invitrogen Corp. (Carlsbad, CA, USA).
Constitutive active and negative mutants of 60 different human or mouse Rabs were prepared by site-directed mutagenesis (Tsuboi and Fukuda, 2006; Itoh et al., 2006; Tamura et al., 2009) and subcloned into the pEGFP-C1 vector (BD Biosciences Clontech, Mountain View, CA, USA). cDNAs of Rab1A, Rab1A(Q70L), and Rab1A(S25N) were also subcloned into the modified pmStr-C1 vector (Ohbayashi et al., 2012). pEGFP-C1-Rab1ASR and pmStr-C1-Rab1ASR (siRNA-resistant mutants) were produced by the two-step PCR technique as described previously (Tamura et al., 2011) using the following mutagenic oligonucleotides (substituted nucleotides are shown in italics): 5′-CACAAAAAAGGTGGTGGATTATACAACAG-3′ (Rab1A SR-5′ primer, sense) and 5′-CTGTTGTATAATCCACCACCTTTTTTGTG-3′ (Rab1A SR-3′ primer, antisense). The siRNA-resistance of Rab1ASR was checked by co-expression of Rab1A shRNA with EGFP–Rab1ASR in COS-7 cells (supplementary material Fig. S2C). cDNA encoding the human TBC1D20 was cloned as described previously (Itoh et al., 2006), and its catalytically inactive mutant (R105A, Arg-to-Ala substitution at amino acid position 105) was produced by the conventional PCR technique using the following mutagenic oligonucleotide (substituted nucleotides are shown in italic): 5′-TGGCATGCCAGGAGGGAACCGCCGCAATGACGCCCG-3′. TBC1D20 mutants lacking a transmembrane domain, i.e. TBC1D20ΔTM and TBC1D20ΔTM(R105A), were similarly produced by PCR using the following oligonucleotide (a stop codon is shown in bold type): 5′-CTACTTGGTCAGGACATCTTTTG-3′. The cDNA fragments of TBC1D20, TBC1D20(R105A), TBC1D20ΔTM and TBC1D20ΔTM(R105A) were subcloned into the pEF-Myc vector (Mori et al., 2012). pSilencer 2.1-U6 neo vectors (Ambion, Austin, TX, USA) encoding shRNA of the following Rab isoforms were constructed essentially as described previously (Kuroda and Fukuda, 2004): pSilencer-Rab1A (19-base target site: 5′-AGAAAGTAGTAGACTACAC-3′), pSilencer-Rab1B (19-base target site: 5′-TTTGCAGACTCTCTGGGTG-3′), pSilencer-Rab2A (19-base target site: 5′-GTACATCATCATCGGCGAC-3′), pSilencer-Rab2B (19-base target site: 5′-GTCATGTCTCCTCCTTCAG-3′), pSilencer-Rab3D (19-base target site: 5′-GCAGGTGTTCGAGCGCCTG-3′), pSilencer-Rab8A (19-base target site: 5′-TCACGACAGCCTACTACAG-3′), pSilencer-Rab8B (19-base target site: 5′-ATCCTTTGACAATATTAAA-3′), pSilencer-Rab10 (19-base target site: 5′-GTGGCTTAGAAACATAGAT-3′), pSilencer-Rab11A (19-base target site: 5′-TCTGGAAAGCAAGAGTACC-3′), pSilencer-Rab11B (19-base target site: 5′-GCATTCAAGAACATCCTCA-3′), pSilencer-Rab12 (19-base target site: 5′-TAGCATCCTTTCTCTACAA-3′), pSilencer-Rab19 (19-base target site: 5′-CTTGGTCATCATGCTGATC-3′), pSilencer-Rab27A (19-base target site: 5′-AAGAGAGTGGTGTACAGAG-3′), pSilencer-Rab27B (19-base target site: 5′-ACGTGTGGTTTATGACACA-3′), and pSilencer-Rab36 (19-base target site: 5′-AGACTAGCCTCATTCACAG-3′). siRNAs against mouse Rab1A no. 1 (19-base target site: 5′-CAAGTTGTTGGTAGGGAAC-3′), Rab1A no. 2 (19-base target site: 5′-CCACAAAGAAAGTAGTAGA-3′), and Rab27A (19-base target site: 5′-AAGAGAGTGGTGTACAGAG-3′) were chemically synthesized by Nippon EGT Corp., Ltd (Toyama, Japan). Unless otherwise stated, Rab1A siRNA means Rab1A siRNA no. 1 throughout this paper. Stealth RNA oligonucleotide against mouse p150Glued was obtained from Invitrogen Corp. (catalog number MSS203510).
Immunofluorescence analysis and melanosome distribution assays
The immortal mouse melanocyte cell lines melan-a, derived from a black mouse, and melan-ash, derived from an ashen mouse (generous gift of Dorothy C. Bennett, St George's Hospital Medical School, London, UK), were cultured on glass-bottom 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 melan-a/ash cells and B16-F1 cells by using FuGENE 6 (Roche Applied Science, Mannheim, Germany) and Lipofectamine-2000 reagents (Invitrogen Corp.), respectively, 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 anti-Myc mouse monoclonal antibody (1:100 dilution), anti-GM130 antibody (1:200 dilution) and/or anti-ERGIC53 antibody (1:100 dilution). The antibody was visualized with anti-mouse Alexa-Fluor-488/633-conjugated IgG (Invitrogen Corp.), and the cells were examined for fluorescence with a confocal fluorescence microscope (Fluoview; Olympus, Tokyo, Japan). siRNAs were transfected into melanocytes by using Lipofectamine RNAiMAX (Invitrogen Corp.). Three days after transfection, cells were fixed in 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, and co-stained with anti-β-tubulin antibody (1:100 dilution) and Texas Red-conjugated phalloidin (1:200 dilution) or with anti-GM130 antibody (1:200 dilution) and anti-ERGIC53 antibody (1:100 dilution). The antibody was visualized with anti-mouse Alexa-Fluor-488/594-conjugated IgG, and the cells were examined for fluorescence with the confocal fluorescence microscope. Melan-a cells were treated with brefeldin A (BFA) (Sigma-Aldrich Corp.; 10 µg/ml) or DMSO alone for 1 hour and co-stained with anti-GM130 antibody (1:200 dilution) and anti-ERGIC53 antibody (1:100 dilution). The antibodies were visualized with anti-mouse/rabbit Alexa-Fluor-488/594-conjugated IgG, and the cells were examined for fluorescence with the confocal fluorescence microscope. The images were processed with Adobe Photoshop software (CS4).
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 the data are expressed as means and standard deviation (s.d.). Student's unpaired t-test was used to perform the statistical analyses.
Immunoaffinity purification of mature melanosomes
Immunoaffinity purification of Rab27A-bound melanosomes with anti-Rab27A IgG-conjugated magnetic beads was performed as described previously (Kuroda and Fukuda, 2004). In brief, Dynabeads M-280 (20 µl volume, wet volume) coated covalently with sheep anti-rabbit IgG (Invitrogen Corp.) were incubated overnight at 4°C with anti-Rab27A rabbit antibody or control rabbit IgG (4 µg) in phosphate-buffered saline (PBS) containing 0.1% BSA. B16-F1 cells (two confluent 6-cm dishes) were homogenized in a homogenization buffer (5 mM HEPES-KOH at pH 7.2, 5 mM EDTA, 0.03 M sucrose and appropriate protease inhibitors), and after centrifugation at 800 g for 10 minutes, the supernatant was incubated with the primary-antibody-coated beads for 2 hours at 4°C in the homogenization buffer containing 10% fetal bovine serum and 2 mM EDTA. After washing the beads twice with PBS containing 2 mM EDTA, the bound fractions were analyzed by 10% SDS-PAGE followed by immunoblotting with anti-Rab27A rabbit antibody (2 µg/ml), anti-Rab1A rabbit antibody (1:500 dilution), anti-Tyrp1 goat antibody (1:1000 dilution), anti-tyrosinase rabbit antibody (0.2 µg/ml), anti-EEA1 rabbit antibody (1:500 dilution), anti-GM130 mouse antibody (1:300 dilution), and anti-Mreg rabbit antibody (0.4 µg/ml). Immunoreactive bands were visualized by enhanced chemiluminescence (ECL).
Analysis of melanosome movements
Melan-a cells that had been treated with specific siRNAs were maintained at 37°C under 10% CO2 in an incubator (Olympus), and images of living cells were acquired at 0.5-second (or 1.5-second) intervals for 5 minutes with a time-lapse microscope (FV500, Olympus). Melanosome movements were analyzed with the Manual tracking plugin for ImageJ software (version 1.44o; National Institutes of Health, Bethesda, MD, USA). Melanosome movements over a distance of more than 1 µm were selected, and the moved distance was measured until the melanosome stopped or moved more than 1 µm in the opposite direction (see also Fig. 8A,B). Almost all melanosome movements in the region of interest (42.68×42.68 µm area; n>110 from three independent cells) were measured and three independent experiments were performed. Data are expressed as the means and s.d. of the data from the three independent experiments. The maximum velocity of anterograde (or retrograde) melanosome transport is also expressed as the means and s.d. of the velocity of more than 110 melanosome movements in three independent cells.
We thank Megumi Aizawa for technical assistance and members of the Fukuda Laboratory for valuable discussions.
This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, and Technology of Japan [grant number(s) 20113006 to M. F., 24570206 to N. O.].