Slac2-a/Melanophilin Contains Multiple PEST-like Sequences That Are Highly Sensitive to Proteolysis*

The synaptotagmin-like protein homologue lacking C2 domains-a (Slac2-a)/melanophilin was recently identified as the “missing link” between the small GTPase Rab27A and the actin-based motor protein myosin Va. Although formation of a tripartite protein complex by three molecules had been shown to be required for proper melanosome distribution in melanocytes (Kuroda, T. S., Ariga, H., and Fukuda, M. (2003) Mol. Cell. Biol. 23, 5245-5255), the regulatory mechanisms of the complex (i.e. assembly and disassembly of the complex) had never been elucidated. In this study, we discovered that Slac2-a and a closely related isoform, Slac2-c/MyRIP, contain multiple PEST-like sequences (potential signals for rapid protein degradation) in the myosin Va- and actin-binding domains at the C terminus. We found that the C-terminal domain of Slac2-a is highly sensitive to low concentrations of proteases, such as trypsin and calpain, in vitro, whereas the N-terminal Rab27A-binding domain is highly resistant to these proteases. We further found that endogenous calpains selectively cleave Slac2-a, but not Rab27A or myosin Va, in melanocytes. A mutant Slac2-a lacking one of the PEST-like sequences located at the interface between the myosin Va- and actin-binding domains (ΔPEST; amino acids 399-405) is more stable than the wild-type protein, both in vitro and in melanocytes. Expression of the mutant Slac2-a-ΔPEST with an N-terminal green fluorescence protein tag often induced perinuclear aggregation of melanosomes (∼40% of the transfected cells) compared with the wild-type Slac2-a. Our findings suggest that protein degradation of Slac2-a is an essential process for proper melanosome distribution in melanocytes.

Melanosomes are the specialized organelles that produce and store melanin pigments and are responsible for the pigmentation of mammalian hair and skin. Mature melanosomes are transported from the cell body of melanocytes to the tips of their dendrites by two distinct motors (i.e. a microtubule-dependent motor and an actin-dependent movement). Finally, melanosomes are translocated from the dendrites of the melanocytes into adjacent epidermal keratinocytes (reviewed in Refs. 1 and 2).
Melanosome transfer from the microtubles to actin filaments must be a crucial step for melanosome transport in melano-cytes, because defects in this step cause pigment dilution in the skin and hair in human diseases (e.g. Griscelli syndrome) and the corresponding coat-color mutant mice (e.g. dilute, ashen, and leaden) (1,(3)(4)(5)(6)(7)(8). The results of recent genetic analyses of these mutant animals and biochemical studies of their gene products have indicated that a tripartite protein complex formed by myosin Va (dilute gene product), Slac2-a/melanophilin (GS3/leaden gene product), 1 and Rab27A (ashen gene product) is essential for melanosome transfer from the perinuclear region of melanocytes to their actin-rich cell periphery (9 -18; reviewed in Refs. 19 and 20). Slac2-a functions as a linker protein between myosin Va and Rab27A and directly and simultaneously binds the GTP-bound form of Rab27A on the melanosome via the N-terminal synaptotagmin-like protein (Slp) 1 homology domain (referred to as SHD or RBD27 (Rab binding domain specific for Rab27 isoforms)) (9 -11, 14, 17, 21-26) and myosin Va, an actin-based motor protein, via the large C-terminal domain (9 -11, 14, 17, 18, 27). The interaction between Slac2-a and myosin Va is strengthened by the presence of a melanocyte-specific exon F in the tail domain of myosin Va (18,(27)(28)(29)(30)(31). Although the formation of a tripartite protein complex is widely believed to be essential for melanosome transfer from microtubules to actin filaments in mammalian melanocytes, almost nothing is known about the mechanism of disassembly of the complex after the transfer of melanosomes to actin filaments. Phosphorylation of myosin Va may be involved in detachment from melanosomes (or Slac2-a), as has been shown in Xenopus laevis eggs (32), or certain GTPase-activating protein for Rab27A may be involved in this process through conversion of GTP-Rab27A to GDP-Rab27A.
In this study, we discovered that Slac2-a contains multiple PEST-like sequences in the C-terminal domain that are known to be sensitive to proteases such as calpains (33,34) and we found that Slac2-a, but not Rab27A or myosin Va, is selectively degraded by endogenous calpains in melanocytes. We also found that expression of green fluorescence protein (GFP)tagged Slac2-a lacking one of the PEST-like sequences (⌬PEST) often caused perinuclear aggregation of melanosomes in melan-a cells. Based on our findings, together with the fact that yeast Vac17p, a linker protein of the yeast class V myosin Myo2p contains a functional PEST sequence (35,36), we discuss the convergent evolution of the cargo receptor for class V myosins in membrane trafficking.

MATERIALS AND METHODS
Materials--Calpain, calpastatin peptide, and calpain inhibitor III were obtained from Calbiochem-Novachem Corp. Trypsin, thrombin, and horseradish peroxidase (HRP)-conjugated anti-FLAG tag (M2) * This work was supported in part by Grant-in-aid for Young Scientists (A) 15689006 from the Ministry of Education, Culture, Sports, and Technology of Japan (to M. F.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Limited Proteolysis of Slac2 and Slp3-a by Trypsin and Thrombin-Recombinant T7-tagged Slac2 (or Slp3-a) protein was expressed in COS-7 cells and purified with the anti-T7 tag antibody-conjugated agarose (Novagen) essentially as described previously (42,43), but all procedures were performed in the absence of any protease inhibitors. Purified proteins were incubated at 25°C for 30 min with various concentrations of trypsin (see Fig. 1A) or 1 unit of thrombin in 50 mM HEPES-KOH, pH 7.2. Digested proteins were then analyzed by 10% SDS-PAGE, followed by immunoblotting with HRP-conjugated anti-T7 tag (or anti-FLAG tag) antibody as described previously (43).
Digestion of Slac2-a by Calpains-Recombinant T7-Slac2-a-FLAG protein was incubated at 30°C for 15 min with various concentrations of -calpain in the presence or absence of 750 M Ca 2ϩ (Fig. 3, A and B). Digested proteins were then analyzed by 10% SDS-PAGE followed by immunoblotting with HRP-conjugated anti-T7 tag (or anti-FLAG tag) antibody (43).
Melan-a cells (one 10-cm dish at confluence) were homogenized in 500 l of a buffer (50 mM HEPES-KOH, pH 7.2, 1 mM MgCl 2 , and 150 mM NaCl) without any protease inhibitors in a glass-Teflon Potter homogenizer with 10 strokes at 1000 rpm, and proteins were solubilized with 1% Triton X-100 at 4°C for 1 h. After removing the insoluble material by centrifugation at 15,000 rpm for 10 min, cell lysates were incubated at 30°C for 15 min, 30 min, or 1 h with either 2 mM EGTA or 750 M Ca 2ϩ in the presence or absence of calpain inhibitors (10 M calpastatin peptide or 10 M calpain inhibitor III). Reactions were terminated by adding SDS sample buffer and boiling for 3 min. Protein expression levels of Rab27A, Salc2-a, myosin Va, and actin were analyzed by immunoblotting with specific antibodies as described previously (9,27). The intensity of the bands on x-ray films was quantified with Lane Analyzer (version 3.0) (ATTO, Tokyo, Japan).
Miscellaneous Procedures-The immortalized melanocyte cell line melan-a (generous gift of Dorothy C. Bennett) was cultured as described previously (17,44). Transfection of pEGFP-C1-Slac2-a into melan-a cells with FuGENE 6 (Roche Molecular Biochemicals), immunocytochemistry with Texas-Red phalloidin, and the melanosome aggregation assay were performed as described previously (17,18). The cells were examined for fluorescence and by bright-field images with a confocal fluorescence microscope (Fluoview; Olympus, Tokyo, Japan). The melanosome distribution of the transfected cells ("dispersed" indicates a normal peripheral distribution of melanosomes (Fig. 6A, top right); "aggregated" means accumulation of melanosomes in the perinuclear regions (Fig. 6A, bottom right) was evaluated as described previously (17). Plasmids were transfected into COS-7 cells (7.5 ϫ 10 5 cells/10 cm-dish, the day before transfection) with LipofectAMINE Plus reagent (Invitrogen) according to the manufacturer's notes. T7-Slac2-a proteins were immunoprecipitated with anti-T7 tag antibody-conjugated agarose as described previously (42,43). SDS-PAGE and immunoblotting analyses with HRP-conjugated anti-FLAG tag, anti-HA tag, anti-actin, and anti-T7 tag antibodies were also performed as described previously (17,27,43). The blots shown in this article are representative of at least two or three independent experiments.

RESULTS
Slac2-a Is Highly Sensitive to Proteases-In a previous study, we found that both recombinant Slac2-a and Slac2-c proteins usually contained a variety of degradation products FIG. 1. Limited proteolysis of the recombinant Slac2-a and Slp3-a proteins by trypsin and thrombin. After incubation of T7-Slac2-a-FLAG with the amounts of trypsin indicated (lanes 1-6) or thrombin (lane 7) at 25°C for 30 min, the reactions were stopped by addition of SDS sample buffer and boiling for 3 min. Samples were then analyzed by 10% SDS-PAGE and immunoblotting with HRP-conjugated anti-T7 tag antibody (1/10,000 dilution) (A) or HRP-conjugated anti-FLAG tag antibody (1/10,000 dilution) (B). Note that three major tryptic fragments were detected with the anti-T7 tag antibody (open arrowheads; sites 1-3), whereas only a single band (i.e. full-length protein) was detected with the anti-FLAG tag antibody, indicating that the C-terminal domain of Slac2-a was rapidly degraded by trypsin. The positions of the Slac2-a deletion mutants described previously (Slac2a-⌬ABD and Slac2-a-SHD) are indicated by closed arrowheads (9,17). C, limited proteolysis of the recombinant Slp3-a by trypsin (lanes 1-6) and thrombin (lane 7) as revealed by anti-T7 tag antibody. Note that Slp3-a was highly resistant to both enzymes. The positions of the molecular mass markers (ϫ 10 Ϫ3 ) are shown on the left. when expressed in cultured cells (27) (Figs. 1A, lane 1, and 3C, lane 1). Because the recombinant Slp1-5 proteins were expressed and purified with little contamination by degradation products (Fig. 1C, lane 1, and Refs. 23 and 24), Slac2-a proteins should be less stable than Slp proteins in living cells and more sensitive to proteases, and we performed limited proteolysis experiments with trypsin to test this hypothesis (25,45). As shown in Fig. 1, A and B, recombinant T7-Slac2-a-FLAG protein was easily digested with 100 ng/ml trypsin at 25°C for 30 min. To our surprise, however, monitoring the degradation of T7-Slac2-a-FLAG protein with anti-T7 tag antibody (i.e. detection of the N-terminal domain of Slac2-a) yielded three major tryptic fragments (Fig. 1A, sites 1-3, open arrowheads), whereas monitoring the degradation of T7-Slac2-a-FLAG protein with anti-FLAG tag antibody (i.e. detection of the C-terminal domain of Slac2-a) yielded only a single band corresponding to the full-length protein (Fig. 1B, lanes 1-4). The recombinant Slp3-a protein, on the other hand, was highly

FIG. 2. Identification of the PEST-like sequences in Slac2-a and Slac2-c.
A, schematic representation of three functional domains of mouse Slac2-a and content of four amino acids (Pro, Glu, Ser, and Thr) in each domain. The SHD is composed of two potential ␣-helical regions (SHD1 and SHD2; black boxes) separated by two zinc finger motifs (indicated by Zn 2ϩ ) and functions as a specific Rab27A/B binding domain (9 -11, 14 -18, 25). The MBD and ABD are indicated by the shaded and hatched boxes, respectively (17,27). GT and exon-F indicate the myosin Va globular tail-binding domain and exon-F-binding domain, respectively (10,11,14,17,18,27). Note that the content of the four amino acids (PEST) is very high in the MBD and ABD, which are highly sensitive to proteolysis (see Fig. 1). Putative proteolysis-sensitive sites are indicated by open arrowheads. B, sequence alignment of mouse Slac2-a and Slac2-c. Pro, Glu, Ser, and Thr (PEST) residues in the sequences that are conserved are shown against a black background, and all other non-conserved PEST residues are shown against a shaded background. The SHD, MBD-exon-F, and ABD are boxed. The solid lines indicate two SHDs. Asterisks indicate the conserved Cys residues corresponding to two zinc finger motifs; # indicate the positions of a putative PEST-like sequence conserved between Slac2-a and Slac2-c, which was deleted in this study (⌬PEST; see resistant to trypsin (up to 250 ng/ml). Similar results were obtained when the anti-Slac2-a-SHD and the anti-Slac2-a-actin-binding domain (ABD) antibodies were used to detect the N-terminal and C-terminal domains, respectively, of Slac2-a (data not shown). These results indicated that the C-terminal ABD was highly sensitive to and rapidly degraded by trypsin, whereas the N-terminal SHD of Slac2-a was highly stable. It was interesting that the recombinant T7-Slac2-a-FLAG protein was also highly sensitive to thrombin digestion (Fig. 1A, lane  7), and the digestion pattern of T7-Slac2-a-FLAG with thrombin was very similar to the pattern obtained with trypsin. Again, we detected three major fragments with the anti-T7 tag antibody and no signals with the anti-FLAG tag antibody (Fig.  1, compare lane 7 in A and B), suggesting that both trypsin and thrombin cleave the same Arg-X residues in Slac2-a. The recombinant Slp3-a, on the other hand, was highly resistant to thrombin digestion (Fig. 1C, lane 7).
Slac2-a Contains PEST-like Sequences Cleaved by -calpain in Vitro-The results of the limited proteolysis experiments described above clearly indicated that Slac2-a (and possibly Slac2-c) protein should contain specific sequences that are highly sensitive to proteases. After a careful sequence comparison between Slac2-a and Slac2-c combined with data base searching, we found that the C-terminal domain of Slac2-a (i.e. myosin Va-binding domain (MBD) and ABD) contains multiple PEST-like sequences (potential signals for rapid protein degradation) (33,34) that consist of four amino acids: Pro, Glu, Ser, and Thr (Fig. 2, A and B). It should be noted that the Pro, Glu, Ser, and Thr residues account for more than 40% of the total number of amino acids in the MBD and that the putative proteolytic cleavage sites would be mapped within the MBD ( Fig. 2A, arrowheads). By contrast, the four residues amounted to only 21% of the residues in the SHD, a finding that is consistent with the fact that the SHD is highly resistant to the proteases used in this study. Because PEST sequences are often cleaved by calpains (reviewed in Refs. 33, 34, and 46), we next tested the effect of Ca 2ϩ -dependent -calpain on the stability of Slac2-a in vitro (Fig. 3, A and B). As expected, -calpain attacked the C-terminal domain of T7-Slac2-a-FLAG in a Ca 2ϩ -dependent manner; 50 nM -calpain was sufficient to degrade the C-terminal domain of Slac2-a, whereas a concentration of -calpain (500 nM) 10 times higher was required for degradation of the N-terminal domain (Fig. 3A, site 1). Likewise, recombinant Slac2-c was also easily cleaved by -calpain in the presence of Ca 2ϩ (Fig. 3C) Selective Degradation of the Native Slac2-a Molecule by Endogenous Calpains in Melanocytes-To further determine whether Slac2-a protein is a molecular target of endogenous calpains in melanocytes, we prepared total cell lysates of melan-a cells without using protease inhibitors. When the total cell lysates were incubated at 30°C for 15 or 30 min in the presence or absence of 750 M Ca 2ϩ , endogenous Slac2-a protein was rapidly degraded within 15 min only in the presence of Ca 2ϩ (Fig. 4A, closed arrowhead; compare lanes 1-3 and 4 -6). Consistent with the results of the limited proteolysis experiments described above (Figs. 1A and 3A), we detected immunoreactive bands corresponding to the N-terminal domain of native Slac2-a molecule (i.e. Slac2-a-SHD is highly resistant to proteases) (Fig. 4A, asterisk). By contrast, however, the intensity of the immunoreactive bands of Rab27A, myosin Va, and actin was unaltered by treatment with Ca 2ϩ for 1 h under our experimental conditions (Fig. 4, B, top, third, and fourth panels, respectively, and C), although the tail domain of myosin Va is known to contain a PEST sequence (47). In addition, we did not observe any significant differences in the protein composition of the total cell lysates visualized with Amido Black stain-ing (Fig. 4B, bottom). Because the selective loss of Slac2-a protein as a result of exposure to Ca 2ϩ was completely abolished by simultaneous exposure to two different calpain inhibitors (i.e. calpastatin peptide and calpain inhibitor III), we concluded that endogenous Ca 2ϩ -dependent calpains selec- tively and rapidly cleave the Slac2-a molecule in melan-a cells (Fig. 4, A, lanes 7-9, and B, second panel, lanes 3 and 4).
Expression of GFP-Slac2-a-⌬PEST Induces Perinuclear Aggregation of Melanosomes in Melan-a Cells-Among the three putative protease cleavage sites (sites 1-3), we particularly focused on site 3 for further analysis for the following reasons. First, site 3 was highly sensitive to both trypsin and -calpain and seemed to be cleaved earlier than other sites. Second, because sites 1 and 2 were located within the MBD, deletion of either the site 1 or site 2 sequence seemed to reduce (or eliminate) myosin Va-binding activity, as shown in the myosin Va-bindingdefective mutants described previously (i.e. Slac2-a(EA) or Slac2a-⌬GT) (17,18). By contrast, because site 3 was located exactly at the interface between the MBD and ABD, deletion of the putative site 3 sequence (⌬PEST; Figs. 5A, underlined, and 2B, #) was expected not to disturb myosin Va-or actin-binding activity. As anticipated, the mutant Slac2-a-⌬PEST normally interacted with myosin Va-tail, Rab27A, and actin, the same as the wildtype protein (Fig. 5C, compare lanes 4 and 5). The results of the limited proteolysis experiment showed that the full-length mutant Slac2-a-⌬PEST protein was clearly more resistant to trypsin than the full-length wild-type protein (Fig. 5B, compare lanes 3  and 6, boxed). The mutant Slac2-a-⌬PEST protein was still cleaved around site 3, probably by another PEST-like sequence adjacent to site 3 (Fig. 2B, sequence surrounding #; many Pro, Glu, Ser, and Thr residues are still present).
Finally, we investigated the effect of expression of the GFP-Slac2-a-⌬PEST on melanosome transport in melanocytes. Because of the deletion of the PEST-like sequence the GFP-Slac2a-⌬PEST expression level was approximately twice as great as that of the wild-type protein (Fig. 6C, closed and open arrowheads), and it was resistant to Ca 2ϩ /calpain treatment, compared with the wild-type protein (Fig. 6D, compare lanes 2 and  4). In addition, expression of GFP-Slac2-a-⌬PEST protein induced aggregation of melanosomes in a significantly higher proportion of cells (ϳ40% of the transfected cells) than the wild-type protein (ϳ20% of the transfected cells) (Fig. 6B). It is noteworthy that GFP-Slac2-a-⌬PEST or even the wild-type Slac2-a protein (with high expression levels) colocalized more with actin filaments than with melanosomes (Fig. 6A, center and bottom middle  panels, respectively, insets, yellow), in contrast to the wild-type GFP-Slac2-a (with low expression levels), which was present on melanosomes rather than with actin filaments (i.e. green and red signals were mostly separate; Fig. 6A, inset, top center). DISCUSSION We and others had previously shown that formation of a tripartite protein complex composed of Rab27A, Slac2-a, and   FIG. 4. Native Slac2-a molecule is rapidly and selectively cleaved by endogenous calpain in melan-a cells. A, rapid degradation of endogenous Slac2-a molecule by Ca 2ϩ -dependent calpains. Total cell lysates of melan-a cells were prepared in the absence of protease inhibitors as described under "Materials and Methods." After incubation of the cell lysates at 30°C for 15 or 30 min in 50 mM HEPES-KOH, pH 7.2, 150 mM NaCl, and 1 mM MgCl 2 in the absence (lanes 1-3) or presence of 750 M Ca 2ϩ (lanes 4 -9), the reactions were stopped by addition of SDS sample buffer and boiling for 3 min. Samples were then analyzed by 10% SDS-PAGE and immunoblotting with anti-Slac2-a-SHD antibody (2 g/ml). Note that endogenous Slac2-a molecule was rapidly degraded within 15 min only in the presence of Ca 2ϩ (compare lanes 1-3 and 4 -6) and that this activity was reversed by a calpain inhibitor (lanes 7-9). Under our experimental conditions, the N-terminal domain of endogenous Slac2-a molecule was also resistant to Ca 2ϩ /calpains (asterisk in lanes 5 and 6), same as the recombinant Slac2-a molecule (see site 1 in Fig. 3A). myosin Va is indispensable for normal melanosome transport in melanocytes in vivo (10,11,17,18). However, the molecular mechanism of the disassembly of the complex after melanosome transport had been poorly understood. Although phosphorylation of the myosin Va globular tail (Ser-1650) has been suggested to be involved in detachment from melanosomes in X. laevis eggs (32), the phosphorylation-mimic form of myosin Va carrying the S1650E mutation normally binds Slac2-a in vitro (18), suggesting that phosphorylation of the globular tail of myosin Va is not involved in the disassembly of the tripartite protein complex. A certain GTPase-activating protein for Rab27A may disassemble the complex by inactivating Rab27A (i.e. conversion of GTP-Rab27A to GDP-Rab27A), which leads to dissociation of Slac2-a from Rab27A; however, no specific Rab27A-GTPase-activating protein has ever been identified. In the present study, we discovered that the C-terminal domain of Slac2-a (or Slac2-c) contains multiple PEST-like sequences (Fig. 2), and we have presented several lines of evidence that these sequences are highly sensitive to proteases in vitro and in intact melanocytes. We demonstrated that Slac2-a, but not Slp3-a, is highly sensitive to several proteases, including -calpain, in vitro ( Figs. 1 and 3) and that the native Slac2-a molecule in melanocytes is actually cleaved by Ca 2ϩ -dependent calpains (Fig. 4). We also demonstrated that deletion of one of the PEST-like sequences (⌬PEST) enhances the stability of mutant Slac2-a protein against proteases and that expression of GFP-Slac2-a-⌬PEST in melanocytes induces melanosome aggregation in the perinucleus at a significantly higher rate than the wild-type protein (Fig. 6). Because the Slac2-a-⌬PEST mutant normally forms the complex with Rab27A, myosin Va, and actin (Fig. 5C), the effect of expression of Slac2-a-⌬PEST on melanosome transport should be different from that of the dominant-negative effects of the Slac2-a(E14A), Slac2-a(EA), Slac2-a(KA), or Slac2-a-⌬GT mutants described previously (i.e. formation of an incomplete complex and induction of melanosome aggregation in 90% of the transfected cells) (17,18). Our results strongly suggest that protein degradation of the Slac2-a molecule is an essential process for melanosome transport in melanocytes. If that is true, melanosome transport should be affected by expression of excess wild-type Slac2-a protein. In fact, ϳ20% of the GFP-Slac2-a-expressing cells (which often displayed strong GFP fluorescence; Fig. 6A) exhibited perinuclear aggregation of melanosomes, and this value was clearly higher than in the control GFP-expressing cells (Fig. 6B). Although the precise molecular mechanism of how excess amounts of Slac2-a molecule on actin filaments inhibits melanosome transport is currently unknown, we speculate that overexpressed Slac2-a molecule induces excessive formation of actin bundles in the cell periphery via the C-terminal ABD, which may inhibit melanosome transfer from microtubules to actin filaments, because expression of GFP-Slac2-a-ABD seemed to induce actin bundles and inhibited neurite outgrowth in PC12 cells (17,27).
Yeast Vac17p, a vacuole-specific Myo2p (class V myosin) receptor, has recently been shown to contain a functional PEST sequence, and it was suggested that regulated degradation of Vac17p might be involved in proper vacuole transport from mother cell to daughter cell (35). Although mammalian Slac2-a and yeast Vac17p show no significant homology, these two class V myosin receptors share several features. First, Vac17p simultaneously binds Vac8p (a vacuole membrane protein) and Myo2p, and formation of the Vac8p⅐Vac17p⅐Myo2p complex is required for vacuole transport in yeast (36), resembling the function of the Rab27A⅐Slac2-a⅐myosin Va complex in melanosome transport in mammals. Second, both Vac17p and Slac2-a contain PEST-like sequences that are highly sensitive to pro- The same blots were then reprobed with HRP-conjugated anti-T7 tag antibody to ensure that the same amounts of T7-Slac2-a proteins had been loaded (bottom panel; Blot: anti-T7). Input means 1/80 volume of the reaction mixture (lanes 1-3). Note that Slac2-a-⌬PEST mutant normally interacted with myosin Va-tail, Rab27A, and actin (compare lanes 4 and 5). teolysis (ref. 36 and this study). Third, deletion of the PESTlike sequence of Vac17p or Slac2-a causes their stabilization in intact cells (Ref. 35 and this study). Therefore, it is highly possible that regulated degradation of the Slac2-a molecule (i.e. disassembly of the transport complex) is involved in the proper distribution of melanosomes in melanocytes, although nothing is known about the degradation machinery in melanocytes, and not even in the yeast-bud-specific degradation machinery has anything been discovered thus far. Because endogenous calpains in melanocytes selectively and efficiently degrade the Slac2-a molecule (Fig. 4) and there are several reports claiming that calpain is activated around the plasma membrane in other cell types (46, 48 -50), the calpain system may be responsible for the regulated degradation of the Slac2-a molecule in vivo. When melan-a cells were stimulated with the Ca 2ϩ ionophore A23187 (1 M) to activate endogenous Ca 2ϩ -dependent calpains, the A23187 treatment induced perinuclear aggregation of melanosomes and significant reduction of the Slac2-a immunoreactivity (data not shown). However, such effects cannot be attributable to the specific degradation of endogenous Slac2-a molecule by Ca 2ϩ -dependent calpains, because we also observed down-regulation of Rab27A, another essential component of melanosome transport and calpain inhibitors did not rescue this phenotype (data not shown). Because Ca 2ϩ activates a variety of cellular events, it is virtually impossible to activate calpain specifically by the A23187 treatment. Further work is necessary to determine how calpain is involved in the degradation of Slac2-a molecule in vivo and melanosome transport.
Slac2-c was originally identified as the closest isoform of Slac2-a and suggested to be involved in retinal melanosome transport in retinal pigment epithelium cells (27,51). Very recently, however, Slac2-c has been shown to control regulated secretion from endocrine (52,53) and exocrine cells (38), independent of myosins. Because the C-terminal domain of Slac2-c also contains conserved PEST-like sequences (Fig. 2B), regulated degradation of Slac2-c protein may occur during Ca 2ϩ -dependent secretory vesicle exocytosis. Consistent with this notion, we recently found that native Slac2-c protein is selectively and efficiently degraded by endogenous proteases, including calpains. 2 We therefore speculate that the function of Slac2-c in retinal melanosome transport or secretory vesicle exocytosis may be regulated by proteolysis.
In summary, we have discovered that Slac2-a contains PEST-like sequences that are highly sensitive to proteolysis, and we further demonstrated that deletion of one of the PESTlike sequences (⌬PEST) increases the protein stability and induces perinuclear aggregation of melanosomes, suggesting that degradation of the Slac2-a molecule is essential for proper melanosome distribution in melanocytes. Our findings provide new insight into the mechanism of the disassembly of the transport complex (Rab27A⅐Slac2-a⅐myosin Va) in melanosome transport. Future study will show whether machinery for regulated degradation is present in mammalian melanocytes. 2 M. Fukuda, manuscript in preparation.
FIG. 6. Deletion of the PEST-like sequence around site 3 induces melanosome aggregation in wild-type melanocytes. A, melan-a cells were transfected with a vector encoding GFP-tagged wild-type Slac2-a (top and middle rows) or mutant GFP-Slac2-a-⌬PEST (bottom row). After cells were fixed and permeabilized with 0.3% Triton X-100, they were stained with Texas Red-conjugated phalloidin. The cells were examined by confocal microscopy for fluorescence of GFP-Slac2-a proteins (green in left column) and actin (red in middle column). Bright-field images (right column) show the melanosome distribution in the cells. Note that the mutant GFP-Slac2-a-⌬PEST or the wild-type protein (with high expression levels) colocalized with actin (yellow in inset of the bottom row, center, or middle row, center, respectively) and induced perinuclear aggregation of melanosomes, whereas the wildtype Slac2-a protein (with low expression levels) was present on melanosomes rather than on actin filaments (top row, middle, inset). Scale bar, 10 m. B, summary of the results of the melanosome distribution assay. Images of transfected cells were captured at random by using GFP fluorescence as a marker, and we judged whether melanosomes had aggregated in the perinucleus by examining the corresponding bright-field images as described previously (17,18). The results are expressed as percentages of cells exhibiting perinuclear melanosome aggregation, and the values are means Ϯ S.D. of data from three independent experiments (n Ͼ 150). *, p Ͻ 0.05 (Student's unpaired t test). C, the expression level of GFP-Slac2-a-⌬PEST protein (closed arrowhead) in melan-a cells was twice as great as that of GFP-Slac2-a protein (open arrowhead), even when the same amount of plasmids were used for transfection. Total cell lysates of the transfected cells were analyzed by 7.5% SDS-PAGE and immunoblotting with anti-Slac2-a-SHD (2 g/ml; top) or anti-actin antibody (1/300 dilution; bottom). The arrow indicates endogenous Slac2-a molecules, whose expression was unaltered by overexpression of GFP-Slac2-a mutants or GFP alone (data not shown). D, mutant GFP-Slac2-a-⌬PEST protein is more resistant to endogenous calpain than wild-type GFP-Slac2-a protein.
Total cell lysates of melan-a cells expressing GFP-Slac2-a mutants were prepared in the absence of protease inhibitors as described under "Materials and Methods." After incubation of the cell lysates in 50 mM HEPES-KOH, pH 7.2, 150 mM NaCl, and 1 mM MgCl 2 in the presence or absence of 750 M Ca 2ϩ , the reactions were stopped by addition of SDS sample buffer and boiling for 3 min. Samples were then analyzed by 7.5% SDS-PAGE and immunoblotting with anti-Slac2-a-SHD antibody (2 g/ml). The results shown are representative of two independent experiments. The positions of the molecular mass markers (ϫ10 Ϫ3 ) are shown on the left.