Slac2-c (synaptotagmin-like protein homologue lacking C2 domains-c), a novel linker protein that interacts with Rab27, myosin Va/VIIa, and actin.

Slac2-a (synaptotagmin-like protein (Slp) homologue lacking C2 domains-a)/melanophilin is a melanosome-associated protein that links Rab27A on melanosomes with myosin Va, an actin-based motor protein, and formation of the tripartite protein complex (Rab27A.Slac2-a.myosin Va) has been suggested to regulate melanosome transport (Fukuda, M., Kuroda, T. S., and Mikoshiba, K. (2002) J. Biol. Chem. 277, 12432-12436). Here we report the structure of a novel form of Slac2, named Slac2-c, that is homologous to Slac2-a. Slac2-a and Slac2-c exhibit the same overall structure, consisting of a highly conserved N-terminal Slp homology domain (about 50% identity) and a less conserved C-terminal myosin Va-binding domain (about 20% identity). As with other Slac2 members and the Slp family, the Slp homology domain of Slac2-c was found to interact specifically with the GTP-bound form of Rab27A/B both in vitro and in intact cells, and the C-terminal domain of Slac2-c interacted with myosin Va and myosin VIIa. In addition, we discovered that the most C-terminal conserved region of Slac2-a (amino acids 400-590) and Slac2-c (amino acids 670-856), which is not essential for myosin Va binding, directly binds actin and that expression of these regions in PC12 cells and melanoma cells colocalized with actin filaments at the cell periphery, suggesting a novel role of Slac2-a/c in capture of Rab27-containing organelles in the actin-enriched cell periphery.

Recent studies have suggested that several Rabs (i.e. Rab6, Rab11, and Rab25) are involved in the movement of transport vesicles from their site of formation to their site of fusion, because these Rabs have been found to interact directly with specific microtubule-or actin-based motor proteins (see Refs. 7-9, and reviewed in Ref. 10). For instance, Rab6 interacts with the C-terminal domain of Rabkinesin-6 (7), whereas Rab11 and Rab25 interact with the C-terminal domain of the myosin Vb tail (8). Very recently, another type of Rab-motor protein interaction has been discovered in melanosome transport (11)(12)(13)(14) and plasma membrane recycling systems (15); myosin Va indirectly recognizes Rab27A on melanosomes via Slac2-a (synaptotagmin-like protein (Slp) 1 homologue lacking C2 domains-a) (also called melanophilin), a linker protein that interacts specifically and directly with the GTP-bound form of Rab27A and myosin Va (11,12,16,17), whereas myosin Vb interacts with Rab11 (and Rab25), as well as Rab11-FIP2 (15). Although five members of the Rab11-FIP family have been described to date (15,18,19), no information is available for the existence of a Slac2-a homologue (or Slac2 family). Because Rab27A and myosin Va are known to be expressed in tissues other than melanocytes, we hypothesized that other linker protein(s) must exist in the body.
In this paper, we report on a novel Slac2-a homologue (named Slac2-c) that interacts specifically with Rab27A/B and myosin Va/VIIa by means of EST database searches and biochemical experiments. We also discovered that the conserved most C-terminal region of Slac2 functions as a novel actinbinding site. Based on our findings, we discuss the role of Slac2 in Rab27A-dependent membrane trafficking.

EXPERIMENTAL PROCEDURES
Molecular Cloning of Mouse Slac2-c cDNAs-cDNAs encoding the N-terminal region of Slac2-c were amplified from Marathon-Ready adult mouse brain cDNA (Clontech Laboratories, Inc.; Palo Alto, CA) by 5Ј-rapid amplification of cDNA ends (RACE) as described previously (20). First 5Ј-RACE reactions were carried out using the adapter primer 1 (5Ј-CCATCCTAATACGACTCACTATAGGGC-3Ј) and Slac2-c-stop primer (5Ј-TTAGTACATCACAGCTGACT-3Ј; termination codon is shown in boldface letters) designed on the basis of rat and mouse EST sequences (GenBank TM accession numbers BF287121 and BG869374). Second RACE reactions were carried out using the internal adapter primer 2 (5Ј-ACTCACTATAGGGCTCGAGCGGC-3Ј) and Slac2-c C1 primer (5Ј-GTGCTGGACCGGGAATTCTG-3Ј). The purified PCR products were inserted directly into the pGEM-T Easy vector (Promega; Madison, WI), and both strands were sequenced completely with a Hitachi SQ-5500 DNA sequencer as described previously (20). cDNAs encoding the open reading frame of the mouse Slac2-c were amplified by reverse transcriptase (RT)-PCR from Marathon-Ready adult mouse brain cDNA with Slac2-c Met (5Ј-GGATCCATGGGGAGGAAGCTG-GACCT-3Ј; BamHI site is underlined) and Slac2-c-stop primers. The purified PCR products were subcloned into the pGEM-T Easy vector (named pGEM-T-Slac2-c) and verified by DNA sequencing. Addition of the T7 tag to the N terminus of Slac2-c and construction of the mammalian cell expression vector (named pEF-T7-Slac2-c) were performed as described previously (20 -22). The human Slac2-c cDNA was determined by database searching (standard BLAST search) using the mouse Slac2-c as a query.
RT-PCR Analysis-Mouse first-strand cDNAs prepared from various tissues and developmental stages were obtained from Clontech Laboratories, Inc. (mouse MTC Panel I) (28,29). PCRs were carried out in the presence of Perfect Match PCR enhancer (Stratagene; La Jolla, CA) for 30 cycles (for G3PDH), 35 cycles (for Slac2-a), or 40 cycles (for Slac2-c), each consisting of denaturation at 94°C for 1 min, annealing at 55°C for 2 min, and extension at 72°C for 2 min. Slac2-a-⌬146 primer (16) and Slac2-a-⌬321 primer and Slac2-c-N1 primer (5Ј-AGAGACTGACATCAGCAACG-3Ј) and Slac2-c-stop primer were used for amplification. The PCR products were analyzed by 1% agarose gel electrophoresis followed by ethidium bromide staining. The authenticity of the products was verified by subcloning into a pGEM-T Easy vector and DNA sequencing as described above.
Melanoma cells (B16-F1; 10-cm dish) were homogenized in a buffer containing 1 ml of the F-buffer and protease inhibitors, and proteins were solubilized with 1% Triton X-100 at 4°C for 1 h. After removal of insoluble material by centrifugation, the supernatant was incubated with either anti-Slac2-a-SHD IgG (10 g/ml) or control rabbit IgG for 1 h at 4°C followed by incubation with protein A-Sepharose beads (Amersham Biosciences) for 1 h at 4°C. After washing the beads five times with the F-buffer containing 0.2% Triton X-100 and protease inhibitors, the immunoprecipitates were subjected to 10% SDS-PAGE followed by immunoblotting with anti-actin (1/200 dilution) and antimyosin Va goat antibodies (1/100 dilution; Santa Cruz Biotechnology, Inc.) as described previously (11,20).

Identification of Slac2-c, A Novel Homologue of Slac2-a-
Slac2 was identified originally as a protein that contains an N-terminal SHD without tandem C2 domains (29,(35)(36)(37). Two forms of Slac2 (Slac2-a and Slac2-b/KIAA0624) have been reported in humans, but they do not show any significant homology except for the SHD (16). Recently, Slac2-a was identified as melanophilin, and a mutation in the mlph gene was discovered to cause defects in melanosome transport in leaden mice exhibiting a lighter coat color (i.e. accumulation of melanosomes in the perinuclear region) (38). Slac2-a/melanophilin has been shown to function as a linker protein that bridges between Rab27A on melanosomes (39 -42) and myosin Va (43)(44)(45), an actin-based motor protein, and formation of the tripartite protein complex (Rab27A⅐Slac2-a⅐myosin Va) is essential for melanosome transport (11)(12)(13)(14). To determine whether Slac2-a is the only protein that links Rab27 with myosin Va, we searched for novel linker protein(s) in the mouse, rat, and human EST databases by standard blastn (www.ddbj.nig.ac.jp/E-mail/ homology-j.html) using Slac2-a cDNA sequence as a query. We found several EST clones in the mouse, rat, and human EST databases that encode a novel protein exhibiting a significant homology to the C terminus of Slac2-a (GenBank TM accession numbers BG869374 (mouse), BF287121 (rat), and BE550329 (human)). cDNA cloning of this putative linker protein by 5Ј-RACE revealed that the 2571-base sequence comprises a single open reading frame encoding 856 amino acids. Interestingly, the corresponding protein also contained the highly conserved SHD with two zinc-finger motifs at the N terminus (more than 40% identity to Slac2-a), but not tandem C2 domains (Fig. 1), indicating that it should be classified as a third member of the Slac2 family, and we designated it as Slac2-c. The C terminus of this protein does not contain any known protein motifs and did not exhibit homology to any known proteins reported thus far except the putative myosin Va-binding site of the Slac2-a C terminus (about 25% identity) (Fig. 1), suggesting that Slac2-c functions as a linker protein between Rab27 and myosin Va.
Database searching in the human genome draft sequences revealed that the human ortholog of mouse Slac2-c consists of 859 amino acids (79.7% identity), and the highest homology between the mouse and human Slac2-c was observed in the SHD (90.3% identity). The human slac2-c gene was mapped to 3p21.33-32, whereas the human slac2-a/mlph and slac2-b were mapped to 2q37.3 and 11q23.1, respectively. The locations of mouse slac2-a/mlph, slac2-b, and slac2-c were predicted to be on mouse chromosomes 1, 9, and 9, respectively, by Human-Mouse Homology Map analysis (www.ncbi.nlm.nih.gov/ Homology/). The exon-intron boundaries of the slac2-c and slac2-a genes were exactly the same in the SHD and similar in the most C-terminal region but not in the middle of the C-terminal domain, suggesting that the slac2-a and slac2-c genes may have evolved from the same ancestor. FIG. 1. Identification of Slac2-c as a novel homologue of Slac2-a. A, comparison of mouse Slac2-c with mouse Slac2-a/melanophilin. The SHD is composed of SHD1 (black boxes), two zincfinger motifs (indicated as Zn 2ϩ ), and SHD2 (black boxes). The amino acid identities between domains are indicated as percentages. Note that the SHD is well conserved (higher than 42% identity) but that the C-terminal domain is not very well conserved (about 20% identity; shaded and hatched boxes). B, sequence alignment of mouse Slac2-a and Slac2-c. Residues in the sequences that are conserved and similar are shown against a black background and a shaded background, respectively. The solid lines indicate two SHDs. The number signs (#) indicate the sequence corresponding to the Rab3-binding site (SGAWFF) of rabphilin (61). The conserved Cys residues corresponding to two zinc-finger motifs are indicated by asterisks (*). Amino acid numbers are indicated at the right of each line.
RT-PCR analysis revealed the highest expression of Slac2-c mRNA in brain, lung, and testis, weak expression in heart, skeletal muscle, and kidney, and absence of expression in spleen and liver (Fig. 2, top panel). By contrast, Slac2-a mRNA was expressed almost ubiquitously in all tissues tested (Fig. 2, middle panel) (38). It should be noted that the mRNAs of Slac2-a and the Slp1-3 were expressed highly at embryonic day 7 (E7) (29), whereas the Slac2-c mRNA was absent completely at this stage (Fig. 2, top and middle panels).
Slac2-c Functions as a Linker Protein between Rab27 and Myosin Va/VIIa-In the next set of experiments, we investigated whether Slac2-c functions as a linker protein between Rab27 and myosin Va both in vitro and in intact cells. We first investigated the specific interaction of full-length Slac2-c with 22 different Rab proteins (Rab1, Rab2, Rab3A, Rab4A, Rab5A, Rab6A, Rab7, Rab8, Rab9, Rab10, Rab11A, Rab17, Rab18, Rab20, Rab22, Rab23, Rab25, Rab27A, Rab27B, Rab28, Rab34, or Rab37) in intact cells as described previously (11,16). As expected, T7-tagged Slac2-c protein coimmunoprecipitated specifically with the FLAG-tagged Rab27A and Rab27B isoforms but not with any of the other Rabs tested (lanes 18 and 19 in Fig. 3B). Deletion mutant analysis further indicated that the SHD alone coimmunoprecipitated efficiently with Rab27A (lane 2 in Fig. 3C) but that the C-terminal domain lacking the SHD (⌬SHD) did not (lane 3 in Fig. 3C). In addition, the SHD of Slac2-c bound a Rab27A(Q78L) mutant (dominant active form that mimics GTP-bound form) the same as the wild-type protein (Fig. 3D, lanes 1 and 3) but not a Rab27A(T23N) mutant (dominant negative form that mimics GDP-bound form), even when more than five times the amount of T23N mutant was used for immunoprecipitation (Fig. 3D, lane 2). These results indicated that the SHD of Slac2-c should be classified as the GTP-Rab27A/B-specific binding domain, the same as the SHD of Slac2-a, Slac2-b, Slp1, Slp2-a, Slp3-a, and Slp5 (11,16,46).
We then investigated the interaction between the Slac2-c C terminus with full-length myosin Va in intact cells (11,12) to determine whether Slac2-c functions as a linker protein between Rab27 and myosin Va. The same as Slac2-a, the fulllength Slac2-c and Slac2-c-⌬SHD, but not Slac2-c-SHD, interacted with BR-type myosin Va (Fig. 4B, lanes 1-3). Although the binding of Slac2-c to myosin Va was enhanced greatly by the presence of exon F, a melanocyte-specific exon, in myosin Va (MC-myosin Va) (12), Slac2-c can interact with BR-myosin Va (ϩ exon B, Ϫ exons D and F) (24, 25) (Fig. 4E, compare lanes   2 and 3 in the middle panel). Direct interaction of GST-Slac2c-⌬SHD with FLAG-myosin Va-GT was also demonstrated by in vitro binding experiments using the purified proteins (Fig. 4C). Formation of the tripartite protein complex (Rab27A⅐Slac2-c⅐myosin Va-GT) from the purified components in FIG. 2. Tissue distribution of mouse Slac2-a and Slac2-c. RT-PCR analysis of Slac2-a (middle panel) and Slac2-c (top panel) expression in various mouse tissues (heart, brain, spleen, lung, liver, skeletal muscle, kidney, and testis) and on embryonic days 7 (E7), 11, 15, and 17. Note that the Slac2-c mRNA is expressed highly in brain, lung, and testis but not on E7, whereas Slac2-a mRNA is expressed in all tissues tested and on E7. RT-PCR analysis of G3PDH expression was also performed (bottom panel) to ensure that equivalent amounts of first strand cDNA were used for RT-PCR analysis. Ϫ, without templates as a negative control. The size of the molecular weight markers (/StyI) is shown at the left of the panel. vitro was demonstrated further in Fig. 4D. These results indicated strongly that Slac2-c functions as a linker protein that bridges between Rab27 and myosin Va, the same as Slac2-a.
While this manuscript was being reviewed, Slac2-c has also been identified as MyRIP (myosin-VIIa-and Rab-interacting protein), which may be involved in transport of retinal melanosomes to the actin cytoskeleton (47). Consistent with this finding, Slac2-c interacted with the tail domain of myosin VIIa and myosin Va (Fig. 4E, lanes 1-3). By contrast, however, Slac2-a interacted only with the tail domain of myosin Va but not myosin VIIa (Fig. 4E, lanes 4 and 5).
The C Terminus of Slac2 Interacts Directly with Actin-Although the minimal essential myosin Va-binding site in Slac2-a was mapped to the middle of the Slac2-a molecule (amino acid residues 241-400) (Fig. 5, A and C, bottom panel) (a minimal determinant of myosin Va-binding site has recently been mapped to amino acid residues 367-400 of Slac2-a; see Ref. 48), the C-terminal 200 amino acids of Slac2-a were also conserved in Slac2-c. Accordingly, in the final set of experiments we attempted to determine the function of the most C-terminal region of Slac2-a/c. First, to investigate its possible function as a protein-protein interaction site, recombinant T7tagged Slac2-a C terminus (T7-Slac2-a-⌬400) was coupled to anti-T7 tag antibody-conjugated agarose beads, and the COS-7 cell lysates were loaded. After washing the beads extensively, proteins that bound the beads were analyzed by 10% SDS-PAGE and Amido Black staining. We found a single protein with an apparent molecular mass of 45,000 daltons (named p45) that was enriched specifically and dramatically by the C terminus of Slac2-a column (Fig. 5C, lane 5, top panel), and use of a specific anti-actin antibody subsequently showed that the p45 protein was actin (Fig. 5C, second panel). Similar results were obtained for the C terminus of Slac2-c (Slac2-c-495/856) (Fig. 5D, top and middle panels, lanes 2 and 4). Direct interaction of the C terminus of Slac2-a (GST-Slac2-a-⌬400) with purified G-actin and F-actin in vitro was demonstrated in Fig.  5E, indicating that Slac2-a binds both filamentous and globular actin. By contrast, GST alone hardly interacted with both Gand F-actin beads under our experimental conditions (Fig. 5E,  lanes 3 and 4). Because the myosin Va-binding site (Fig. 5C,  bottom panel) and actin-binding site (Fig. 5C, top panel) were different, Slac2-a bound F-actin and MC-myosin Va C terminus (Fig. 5F) simultaneously. In vivo interaction between Slac2-a and actin in melanoma cells was confirmed further by immunoprecipitation with anti-Slac2-a-SHD antibody (Fig. 5G).
Expression of the Slac2-a/c C Terminus Inhibits Neurite Out-growth of PC12 Cells-To investigate further whether the C terminus of Slac2-a/c functions as an actin-binding site in intact cells, we expressed either GFP-tagged Slac2-a-⌬400 or T7-tagged Slac2-c-495/856 in nerve growth factor-differentiated PC12 cells. As shown in Fig. 6, most of the Slac2-a-⌬400 (A-C) and Slac2-c-495/856 proteins (D-F) were present at the cell periphery and colocalized closely with actin filaments. It should be noted that PC12 cells abundantly expressing the Slac2-a/c C terminus hardly extended any neurites compared with the control cells and PC12 cells expressing the N-terminal domain of Slac2-a (Slac2-a-SHD) (Fig. 6, G-I). The inhibitory effect of the most C-terminal region of Slac2-a/c on neurite outgrowth may be attributable to inhibition of actin filament remodeling during neurite outgrowth. By contrast, the Slac2a-SHD proteins were present in the cell body, as well as the tips of neurites, where Rab27A was enriched (i.e. dense-core vesicles (49,50); see arrowheads in Fig. 6I). Interestingly, when Slac2-a-SHD was expressed in PC12 cells, Rab27A signals were reduced dramatically compared with non-transfected cells, especially in the cell body (arrow in Fig. 6H). Expression of the full-length Slac2-a with GFP-tag in melanoma cells also colocalized with actin at the cell periphery (Fig. 6, J-L). These results strongly suggest a novel role of Slac2-a/c in capture of Rab27-containing organelles in the actin-enriched cell periphery. DISCUSSION We and others showed recently that Slac2-a/melanophilin is a linker protein that bridges between the GTP-bound form of Rab27A and myosin Va and that the tripartite protein complex (Rab27A⅐Slac2-a⅐myosin Va) regulates melanosome transport to the actin-enriched cell periphery (11,12). In the present study, we isolated a novel homologue of Slac2-a, named Slac2-c, that was distributed differently from Slac2-a in mouse tissues (Fig. 2). Slac2-c contains the N-terminal SHD that specifically binds the GTP-bound activated form of Rab27A/B isoforms (Fig. 3), the same as other Slac2 proteins and the Slp family (11,16,17,46), and the C-terminal domain directly binds myosin Va and myosin VIIa (Fig. 4) (47). These results strongly support the idea that Slac2-c also functions as a linker protein between Rab27A/B and myosin Va/VIIa and that Rab27⅐Slac2-c⅐myosin Va (or VIIa) regulates membrane transport in specific tissues. We also demonstrated a novel function of the Slac2-a/c C terminus, direct binding of actin both in vitro and in intact cells (see Figs. 5 and 6). The actin binding ability of Slac2 is useful for tethering melanosomes to the actin network at the cell periphery after detaching from myosin Va (or VIIa). The putative actin-binding site of Slac2-a/c did not show any significant homology to the known actin-binding motifs; it should be classified as a novel actin-binding motif. Further work is necessary to determine critical residue(s) responsible for actin binding.
Are the functions of Slac2-c and Slac2-a redundant? If not, what is the cargo of the Rab27⅐Slac2-c⅐myosin Va (or VIIa) a tripartite protein complex? Genetic analysis of coat color mutation in mouse has indicated that mutations in rab27A (ashen mice; see Ref. 41), slac2-a/mlph (leaden mice; see Ref. 38), and myoVa (dilute mice; see Ref. 45) resulted in the same lighter coat color (i.e. defects in melanosome transport to the actinenriched cell periphery) but that their phenotypes, other than in melanocytes, are different. For instance, dilute mice and type II Griscelli syndrome patients (mutation in myoVa gene) exhibit neurological defects (e.g. transport of endoplasmic reticulum to dendrites in the cerebellum) (51,52), whereas ashen mice and type I Griscelli syndrome patients (mutation in rab27A gene) do not. By contrast, Rab27A-deficient T cells exhibited reduced cytotoxicity and cytolytic granule exocytosis, whereas myosin Va-defective T cells did not (39,53,54). These observations suggest that, at least in some tissues, closely related isoforms (Rab27B, myosin Vb, Vc, VIIa, and Slac2-c) may compensate for the function of Rab27A, Slac2-a, or myosin Va (55)(56)(57)(58). Because the expression of Slac2-c mRNA was highest in brain, and myosin Va and Rab27B proteins, but not Rab27A protein, were expressed abundantly in brain (59), it is tempting to speculate that the Rab27B⅐Slac2-c⅐myosin Va complex functions in specific membrane transport in brain (e.g. in endoplasmic reticulum membrane that contains inositol 1,4,5trisphosphate receptor) (51,52,60). Further work is necessary to identify the specific cargo of Slac2-c-containing vesicles/organelles. An attempt to identify cargoes is now under way in our laboratory.
In summary, we have identified a novel linker protein (named Slac2-c) that bridges between a small GTP-binding protein, Rab27, and a motor protein, myosin Va (or VIIa). We also showed that the most C-terminal region of Slac2-a and Slac2-c directly binds actin and that it may be involved in the tethering of cargo (e.g. melanosomes) in actin-enriched cell periphery.