The well-known coat-color mutant mouse dilute exhibits a defect in melanosome transport, and although various mutations in the myosin-Va gene, which encodes an actin-based motor protein, have been identified in dilute mice, why missense mutations in the globular tail of myosin-Va, a putative cargo-binding site, cause the dilute phenotype (i.e. lighter coat color) has never been elucidated. In this study we discovered that missense mutations (I1510N, M1513K and D1519G) in the globular tail (GT) of myosin-Va partially impair the binding of Slac2-a/melanophilin, a linker protein between myosin-Va and Rab27A on the melanosome. The myosin-Va-GT-binding site in Slac2-a was mapped to the region (amino acids 147-240) adjacent to the N-terminal Rab27A-binding site, but it is distinct from the myosin-Va-exon-F-binding site (amino acids 320-406). The myosin-Va-GT·Slac2-a interaction was much weaker than the myosin-Va-exon-F·Slac2-a interaction. The missense mutations in the GT found in dilute mice abrogated only the myosin-Va-GT·Slac2-a interaction and had no effect on the myosin-Va-exon-F·Slac2-a interaction. We further showed that expression of green fluorescence protein-tagged Slac2-a lacking the myosin-Va-GT-binding site (ΔGT), but not the wild-type Slac2-a, severely inhibits melanosome transport in melan-a cells, especially at the melanosome transfer step from microtubles to actin filaments (i.e. perinuclear aggregation of melanosomes). On the basis of our findings, we propose that myosin-Va interacts with Slac2-a·Rab27A complex on the melanosome via two distinct domains, both of which are essential for melanosome transport in melanocytes.
Genetic analysis of coat-color mutant mice (dilute, ashen and leaden) (Moore et al., 1988; Mercer et al., 1991; Wilson et al., 2000; Matesic et al., 2001; Hume et al., 2002; Provance et al., 2002) and patients with Griscelli syndrome (Pastural et al., 1997; Menasché et al., 2000; Bahadoran et al., 2001; Menasché et al., 2003; Bahadoran et al., 2003), as well as recent biochemical studies of their gene products (myosin-Va, Rab27A, and Slac2-a (synaptotagmin-like protein (Slp) homologue lacking C2 domains-a)/melanophilin, respectively) (Fukuda et al., 2002; Wu et al., 2002a; Strom et al., 2002; Menasché et al., 2003; Bahadoran et al., 2003), have provided evidence that a tripartite protein complex formed by myosin-Va, Slac2-a and Rab27A is essential for melanosome transport from the perinuclear region of melanocytes to their actin-rich cell periphery (for a review, see Marks and Seabra, 2001; Hammer and Wu, 2002; Fukuda, 2002a). Slac2-a simultaneously binds the GTP-bound form of Rab27A on the melanosome, via the N-terminal Slp homology domain (referred to as SHD or RBD27) (Fukuda and Mikoshiba, 2001; Fukuda et al., 2001a; Fukuda et al., 2002; Wu et al., 2002a; Strom et al., 2002; Kuroda et al., 2002a; Kuroda et al., 2002b; Fukuda, 2002a; Fukuda, 2002b; Fukuda, 2003), and myosin-Va, an actin-based motor protein, via the large C-terminal domain (Fukuda et al., 2002; Wu et al., 2002a; Strom et al., 2002). The interaction between Slac2-a and myosin-Va has been shown to be regulated by a melanocyte (MC)-specific alternative splicing in the tail domain of myosin-Va (Seperack et al., 1995; Huang et al., 1998; Wu et al., 2002a; Wu et al., 2002b): Slac2-a strongly interacts with the MC-type myosin-Va containing the MC-specific exon F (- exon B, + exon D and + exon F), and weakly with a brain-type myosin-Va lacking exon F (+ exon B, - exon D, and - exon F) (Fukuda et al., 2002; Wu et al., 2002a; Fukuda and Kuroda, 2002) (see also Fig. 2C).
We and others have recently mapped the binding site for MC-type myosin-Va to the middle region of Slac2-a (Strom et al., 2002; Fukuda and Kuroda, 2002; Nagashima et al., 2002) (see Fig. 1A, shaded box), but little is known about the binding site for brain-type myosin-Va in Slac2-a (i.e. the interaction between globular tail (GT) of myosin-Va and Slac2-a). Because the MC-specific exon F alone is insufficient for melanosome recognition by myosin-Va (Wu et al., 2002b; da Silva Bizario et al., 2002), an additional region (i.e. the globular tail) must also be required for melanosome transport. Moreover, several missense mutations have been found in the GT of myosin-Va in dilute mice (Huang et al., 1998), indicating a crucial role of the GT of myosin-Va in melanosome transport. Why such mutations cause a defect in melanosome transport, however, has never been elucidated at the molecular level, nor has the effect of dilute mutations on Slac2-a binding activity or on the GT-binding site in Slac2-a ever been determined.
In this study we determined the minimal essential binding site for the GT of myosin-Va in Slac2-a and discovered that missense mutations in the GT of myosin-Va observed in dilute mice (i.e. I1510N, M1513K and D1519G) impair binding to Slac2-a. We also found that expression of green fluorescence protein (GFP)-tagged Slac2-a lacking the myosin-Va-GT-binding site causes perinuclear aggregation of melanosomes in melan-a cells. The physiological importance of the GT of myosin-Va in melanosome transport is discussed on the basis of our findings.
Materials and Methods
Construction of deletion mutants of Slac2-a and myosin-Va
Deletion mutants of Slac2-a (pEF-T7-Slac2-a-Δ146/Δ241, pEF-T7-Slac2-a-ΔGT (Δ153-235) and pEGFP-C1-Slac2-a-ΔGT) and of myosin-Va (pEF-FLAG-myosin-Va-GT-ΔAf6 and -myosin-Va-GTAf6) were essentially constructed by conventional PCR as described previously (Fukuda et al., 1994; Fukuda and Mikoshiba, 2000; Fukuda et al., 2001b) using the following oligonucleotides with restriction enzyme sites (underlined) or stop codons (in bold): 5′-CGGATCCGGTGGAGGTGGATCTGAGCC-3′ (Slac2-a-Δ146 primer; sense), 5′-GCACTAGTCAGGGCTCCAGAGAGGTGGTGT-3′ (Slac2-a-Δ241 primer, antisense), 5′-CCAGAGAGGTGGTCTCAGATCCACCTCCAC-3′ (Slac2-a-ΔGT primer, antisense), 5′-CGGATCCCAGGTAGTCAAACAGATGTTC-3′ (myosin-Va-Af6 5′ primer, sense) and 5′-TCACTTGATTAGCTCCGGGT-3′ (myosin-Va-GT-ΔAf6 3′ primer, antisense). Other expression constructs (pEF-T7-Slac2-a deletion mutants, pEF-HA-Rab27A, pEF-FLAG-myosin-Va-F-GT (exon F + globular tail), pEF-FLAG-brain myosin-Va-tail, pEF-FLAG-MC myosin-Va-tail and pEF-FLAG-myosin-Va-GT) were prepared as described previously (Fukuda et al., 2002; Fukuda, 2002b; Fukuda and Kuroda, 2002).
Mutant myosin-Va plasmids carrying an Ile-to-Asn substitution at amino acid position 1510 (I1510N), M1513K, D1519G or S1650E were produced by two-step PCR techniques using the following mutagenic oligonucleotides with artificial SplI sites (or SpeI) (underlined) as described previously (Fukuda et al., 1995): 5′-TCGTACGCACATAAACAAGTTAT-3′ (I1510N primer), 5′-TCGTACGCACTTAAACAAGA-3′ (M1513K primer), 5′-TGCGTACGACATGCTGGCTA-3′ (D1519G primer), 5′-TGCGTACGACATGCTGACTA-3′ (SplI primer), 5′-CTCACTAGTTCGTTTTCTGAGCC-3′ (S1650E primer 1) and 5′-ACTAGTGAGATCGCAGATGAGGG-3′ (S1650E primer 2). The mutant myosin-Va-GT fragments were subcloned into the pEF-FLAG-tag expression vector as described previously (Fukuda et al., 1999; Fukuda and Mikoshiba, 2000).
Cell culture, transfections and immunocytochemistry
Cells of a murine immortalized melanocyte cell line, melan-a cells (generous gift of Dorothy C. Bennett, St George's Hospital Medical School, London, UK and Katsuhiko Tsukamoto, University of Yamanashi, Yamanashi, Japan), were cultured on glass-bottom dishes (35 mm dish; MatTek, Ashland, MA) (Bennett et al., 1987; Kuroda et al., 2003). Transfection of pEGFP-C1-Slac2-a-ΔGT, pEGFP-C1-Slac2-a or a control vector (pEGFP-C1 alone) was achieved with FuGENE 6 (Roche Molecular Biochemicals, Mannheim, Germany) according to the manufacturer's instructions. Two days after transfection, cells were fixed in 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, stained with anti-Rab27A mouse monoclonal antibody (BD Transduction Laboratories, Lexington, KY) and anti-myosin-Va rabbit polyclonal antibody (Fukuda et al., 2002), followed by anti-rabbit Aalexa Fluor 633 IgG and anti-mouse Aalexa Fluor 568 IgG (Molecular Probes, Eugene, OR), and then examined for fluorescence and by bright-field images with a confocal fluorescence microscope (Fluoview; Olympus, Tokyo, Japan) as described previously (Kuroda et al., 2003). The pigment distribution of the transfected cells (`dispersed' = normal peripheral distribution of melanosomes shown in Fig. 6E or `aggregated' = accumulation of melanosomes in the perinuclear regions shown in Fig. 6J) was evaluated as described previously (Strom et al., 2002; Kuroda et al., 2003). More than 50 cells transfected with GFP-Slac2-a mutants in one dish were counted for each construct, and the same experiments were independently repeated three times (n>150).
Transfection of plasmids into COS-7 cells (7.5·105 cells/10 cm dish, the day before transfection) was performed as described previously (Fukuda et al., 2001b). Proteins were solubilized at 4°C for 1 hour with a buffer containing 1% Triton X-100, 250 mM NaCl, 1 mM MgCl2, 50 mM HEPES-KOH, pH 7.2, 0.1 mM phenylmethylsulfonyl fluoride, 10 μM leupeptin and 10 μM pepstatin A. T7-Slac2-a mutants were immunoprecipitated with anti-T7 tag antibody-conjugated agarose (Novagen, Madison, WI) as described previously (Fukuda et al., 1999; Fukuda and Mikoshiba, 2000). SDS-PAGE and immunoblotting analyses with horseradish peroxidase (HRP)conjugated anti-FLAG tag (Sigma Chemical Co.; St Louis, MO), anti-HA tag (Roche Molecular Biochemicals) and anti-T7 tag antibodies (Novagen) were also performed as described previously (Fukuda et al., 1999; Fukuda and Mikoshiba, 2000). The intensity of the bands on x-ray film or gels was quantified with Basic Quantifier Software (version 1.0) (BioImage) or Lane Analyzer (version 3.0) (ATTO, Tokyo, Japan) as described previously (Fukuda and Mikoshiba, 2000). The protein concentration of Slac2-a (or myosin-Va-tail) in the SDS-polyacrylamide gel was then estimated using bovine serum albumin as a reference. The statistical analyses and curve fitting were performed with a GraphPad PRISM computer program (version 4.0). The blots and gels shown in this paper are representative of at least two or three independent experiments.
Globular tail and exon F of myosin-Va bind distinct domains of Slac2-a/melanophilin
A systematic deletion analysis was performed as described previously (Fukuda and Kuroda, 2002) to determine the minimal essential domain required for the binding of myosin-Va-GT to Slac2-a. As shown in Fig. 1A,B, myosin-Va-GT bound a previously uncharacterized region adjacent to SHD2 (amino acids 147-240; cross-hatched box in Fig. 1A). By contrast, myosin-Va-F-GT (i.e. exon F) has been shown to bind the middle region of Slac2-a (amino acids 320-406; shaded box in Fig. 1A) (Strom et al., 2002; Fukuda and Kuroda, 2002; Nagashima et al., 2002; Kuroda et al., 2003). Interestingly, twice the amount of myosin-Va-GT bound to Slac2-a·Rab27A-beads as bound to Slac2-a-beads alone (i.e. without Rab27A) (Fig. 1C, top panel, compare lanes 1 and 2), suggesting that myosin-Va-GT preferentially interacts with the Rab27A·Slac2-a complex rather than with Slac2-a alone.
The interaction between Slac2-a and myosin-Va-GT was less stable (Fig. 2) and much weaker (Fig. 3) than the interaction between Slac2-a and myosin-Va-F-GT. As shown in Fig. 2A, the Slac2-a·myosin-Va-GT interaction was highly sensitive to ionic strength (mild concentrations of NaCl), whereas the Slac2-a·myosin-Va-F-GT (+ exon F) interaction was resistant to high NaCl concentrations (up to 750 mM NaCl) (Fig. 2A,C). Similar results were obtained when the entire tail domain of brain-type myosin-Va and MC-type myosin-Va were used (Fig. 2B). Slac2-a was also found to bind the brain myosin-Va-tail with much lower affinity than the MC myosin-Va-tail (top panels in Fig. 3A). Because the dose-dependence curve of the MC myosin-Va-tail (open circles in Fig. 3B) was almost saturated, we used the curve-fitting program to calculate the EC50 (myosin-Va-tail concentration at half maximal binding = 0.15) and Bmax values (0.74 μg). Interestingly, under the maximal binding conditions, one molecule of Slac2-a (0.25 μg; approximately 3.6 pmol) was estimated to bind approximately two molecules of MC myosin-Va-tail (0.74 μg; approximately 6.7 pmol), which is consistent with the fact that myosin-Va functions as a dimer (Fig. 2C). The dose-dependence curve of brain myosin-Va-tail (closed circles in Fig. 3B), however, was not saturated under our experimental conditions, and only 0.15 μg of brain myosin-Va-tail bound Slac2-a at the highest concentrations of myosin-Va-tail (lane 7 in the upper right panel), suggesting that Slac2-a binds brain myosin-Va with more than ten times lower affinity than it binds MC myosin-Va (Fig. 3B, closed and open circles, respectively). Consistent with these findings, brain-type myosin-Va has been shown not to bind Slac2-a after extensive washing of the myosin-Va-beads (Wu et al., 2002a; Wu et al., 2002b).
Effect of dilute missense mutations in the GT of myosin-Va on Slac2-a binding
Because the GT of mouse class V myosins (myosin-Va, Vb and Vc) consists of two parts, a first part and a second part, which contains an Af6 homology domain (Zhao et al., 1996; Hock et al., 1998; Reck-Peterson et al., 2000; Rodriguez and Cheney, 2002) (see boxed in Fig. 4A), we attempted to identify the domain(s) responsible for binding to Slac2-a. As shown in Fig. 4B (middle panel), Slac2-a was found to bind the first part of GT (myosin-Va-GT-ΔAf6) but not the Af6 homology domain (myosin-Va-GT-Af6). It should be noted that several missense mutations were found in the first part of the myosin-Va-GT in dilute mice (I1510N, M1513K and D1519G; asterisks in Fig. 4A) (Huang et al., 1998), as well as the Ca2+/calmodulin-dependent protein kinase II phosphorylation site (Ser-1650; # in Fig. 4A) (Karcher et al., 2001), all of which are highly conserved among mouse class V myosins. Because the S1650E mutant of the mouse myosin-Va (mimics the phosphorylated form of myosin-Va) does not recruit to Xenopus egg melanosomes, phosphorylation of Ser-1650 is thought to be a crucial process in cargo recognition of, and detachment from, myosin-Va (Karcher et al., 2001). Surprisingly, all of the missense mutations almost completely abolished the Slac2-a binding activity of the myosin-Va-GT, whereas the S1650E mutant, which mimics the phosphorylated form, had no significant effect on binding to Slac2-a (Fig. 4C, lanes 2-6 in middle panel). The lack of effect of the S1650E mutation on binding to Slac2-a may be explained by the species difference or the presence of other myosin-Va-GT binding protein(s) in Xenopus eggs (El-Husseini and Vincent, 1999). By contrast, in the presence of exon F, the myosin-Va-FGT protein containing such missense mutations bound Slac2-a, the same as the wild-type protein (Fig. 4D, middle panel).
Expression of GFP-Slac2-a-ΔGT inhibits melanosome transport in melan-a cells
To investigate further the physiological significance of the myosin-Va-GT·Slac2-a interaction, we prepared a deletion mutant of Slac2-a lacking the myosin-Va-GT-binding site (named Slac2-a-ΔGT) (Fig. 5A). Because the mutant Slac2-a-ΔGT protein still contained both the Rab27A-binding site (amino acids residues 1-146) and the myosin-Va-exon-F-binding site (amino acid residues 320-406), the mutant protein formed a tripartite protein complex with HA-Rab27A and FLAG-MC myosin-Va-tail, the same as the wild-type protein (Fig. 5B, lane 5 at third and fourth panels), but it did not interact with FLAG-brain myosin-Va-tail (Fig. 5B, lane 2 at third panel). The mutant Slac2-a-ΔGT bound the MC myosin-Va-tail with lower affinity than the wild-type Slac2-a, probably because of the lack of myosin-Va-GT·Slac2-a interaction, but it completely lacked brain myosin-Va-tail binding activity (Fig. 3B, open squares and closed squares, respectively).
If the myosin-Va-GT-binding site was essential for melanosome transport in vivo, expression of the Slac2-a-ΔGT in melan-a cells should inhibit melanosome transport (Wu et al., 2002a; Strom et al., 2002), the same as the Slac2-a(EA) mutant lacking the MC myosin-Va-binding site (Kuroda et al., 2003). As expected, almost all of the cells expressing the Slac2-a-ΔGT protein exhibited `aggregated' melanosomes in their perinuclear region (Fig. 6J) (91.6±3.0% (mean±s.e.) cells exhibiting aggregated melanosomes, n=169). Slac2-a-ΔGT-expressing cells exhibited segregation of endogenous Rab27A from myosin-Va (Fig. 6I, Rab27A in green and myosin-Va in red) and decreased immunoreactivity of endogenous myosin-Va (compare panels C and H in Fig. 6). Similar attenuation of myosin-Va-immunoreactivity was observed in ashen- and leaden-mouse-derived melanocytes (Provance et al., 2002) as well as the Slac2-a(EA)-expressing melan-a cells (Kuroda et al., 2003). By contrast, expression of the wild-type Slac2-a protein with the GFP tag (Fig. 6A-E) or GFP alone had virtually no effect on melanosome transport (i.e. peripheral melanosome distribution), with both Rab27A and myosin-Va being present on the melanosomes (Fig. 6D, yellow) (12.1±3.3% or 4±2.6% cells exhibiting aggregated melanosomes, respectively).
We and others previously showed that Slac2-a functions as a receptor for myosin-Va in melanocytes and that formation of a tripartite protein complex composed of Rab27A, Slac2-a and myosin-Va is essential for normal melanosome distribution (Fukuda et al., 2002; Wu et al., 2002a; Strom et al., 2002; Kuroda et al., 2003). In the present study we discovered two forms of myosin-Va binding to Slac2-a: weak interaction between GT and Slac2-a (amino acids 147-240) and a strong interaction between exon F and the middle region of Slac2-a (Fig. 1A). The dilute missense mutations (I1510N, M1513K and D1519G) in the GT of myosin-Va impair only the former interaction, not the latter (Fig. 4), indicating that the Slac2-a·myosin-Va-GT interaction should be physiologically relevant. The physiological significance of the latter interaction is evident on the basis of the following observations: (1) deletion of exon F by mutations in the alternative splicing sites causes the dilute phenotype (i.e. perinuclear aggregation of melanosomes) (Seperack et al., 1995; Huang et al., 1998); (2) expression of brain myosin-Va lacking an exon F cannot rescue the phenotype of dilute-derived melanocytes (Wu et al., 2002b), and expression of MC myosin-Va-tail, but not of brain myosin-Va-tail, induces perinuclear aggregation of melanosomes (Wu et al., 2002b; da Silva Bizario et al., 2002; Westbroek et al., 2003); and (3) Slac2-a mutants lacking the exon-F-binding site cannot support normal melanosome distribution in melan-a cells (Kuroda et al., 2003). We therefore concluded that both interaction sites in Slac2-a identified in this study, the myosin-Va-GT-binding site and the myosin-Vaexon-F-binding site, are essential for melanosome transport in melanocytes.
Do the GT and exon F of myosin-Va function differently, synergistically or sequentially in melanosome transport? Although a previous analysis of yeast myosin V mutants showed that the tail domain has two distinct functions (Catlett et al., 2000), the GT and exon F of mammalian myosin-Va probably function synergistically in melanosome transport, given the following observations. First, neither the GT nor exon F alone (GST-myosin-Va-exon-F) recognized Slac2-a or melanosomes in melanocytes (data not shown) (Wu et al., 2002b), indicating that both domains are required for recognition of melanosomes in vivo. Second, both the mutant Slac2-a lacking the myosin-Va-GT-binding site and the mutant lacking the exon-F-binding site failed to mediate melanosome transport (Fig. 6J) (Kuroda et al., 2003). Third, the phenotype induced by expression of Slac2-a-ΔGT or Slac2-a(EA) lacking exon-F-binding activity in melanocytes is identical, indicating that both sites function at the same step of melanosome transport (Fig. 6J) (Kuroda et al., 2003). Both mutants induced melanosome aggregation in the perinuclear region of the melanocytes (i.e. inhibition of the melanosome transition from microtubules to actin filaments rather than actin-based transport itself) (Fig. 6J) and induced segregation of myosin-Va from Rab27A (Fig. 6I) and attenuation of myosin-Vaimmunoreactivity (Fig. 6H). Because deletion of the GT-binding site of Slac2-a slightly reduced the binding affinity to MC myosin-Va (Fig. 3B), we speculate that both the GT- and exon-F-binding sites of Slac2-a are essential for higher affinity or stable recognition of myosin-Va in vivo. Alternatively, the interaction between myosin-Va-GT and Slac2-a may be a prerequisite for starting unidirectional actin-based transport driven by the myosin-Va motor.
In summary, we have identified two distinct myosin-Va-binding sites in Slac2-a, the GT-binding site (amino acids 147-240) and the exon-F-binding site (amino acids 320-406), both of which are essential for melanosome transfer from microtubules to actin filaments. Because missense mutations in the GT of myosin-Va selectively impair the former interaction, the recognition mechanism of myosin-Va by Slac2-a is not so simple as previously thought. Three-dimensional structural analysis of the Slac2-a·MC myosin-Va complex will be necessary to fully understand the role of the tripartite protein complex in the melanosome transfer step at the molecular level.
We thank Dorothy C. Bennett (St George's Hospital Medical School, London, UK) and Katsuhiko Tsukamoto (University of Yamanashi, Yamanashi, Japan) for kindly donating melan-a cells, and Eiko Kanno and Yukie Ogata for expert technical assistance. This work was supported in part by Grants-in-Aid for Young Scientists (A) (15689006) from the Ministry of Education, Culture, Sports and Technology of Japan (to M.F.), and grants from Uehara Memorial Foundation (to M.F.).
- Accepted September 18, 2003.
- © The Company of Biologists Limited 2004