Dvl, an important component of the Wnt signalling pathway, is thought to be involved in synaptogenesis. In this study, we investigated whether Dvl regulates neurotransmitter release. Knockdown of Dvl in PC12 cells suppressed K+-induced dopamine release, and this phenotype was restored by expression of Dvl-1. We identified synaptotagmin (Syt) I, which is involved in neurotransmitter release, as a Dvl-binding protein. Dvl directly bound to the C2B domain of Syt I. Dvl colocalized with Syt I at the tip of neurites of differentiated PC12 cells and of neurons in the rat dorsal root ganglion. Dvl and Syt I was located in large dense-core vesicles, which contain dopamine. In addition, endocytosis of vesicles containing Syt I was suppressed in Dvl knockdown PC12 cells. Dvl inhibited the binding of Syt I to the complex consisting of syntaxin-1A and SNAP-25. Furthermore, µ2-adaptin of AP-2, which is known to play a role in endocytosis, formed a complex with Dvl and Syt I. Taken together, these results suggest that Dvl is involved in endo- and exocytotic processes through the binding to Syt I.
The Dvl protein family is a critical component of the Wnt signalling pathway (Wharton 2003). All Dvl family members, including those in mammals: Xenopus and Drosophila, contain three highly conserved domains: a DIX domain, a PDZ domain, and a DEP domain. Dvl interprets the Wnt-generated receptor activation and transmits its signal to at least two distinct pathways (the β-catenin and planar cell polarity (PCP) pathways), and thereby organizes pathway-specific subcellular signalling complexes, signal amplification, and dynamic control through feedback regulation (Wharton 2003). These functions of Dvl in the Wnt signal pathway could be mediated by Dvl-binding proteins, including casein kinase, Axin, Frat, Nkd, Frodo, Strabismus, and Daam1 (Wharton 2003). Dvl may act as a functional switch between distinct downstream pathways by different binding partners in the Wnt signalling pathway. However, some Dvl-binding proteins such as Notch (Axelrod et al. 1996) and Eps8 (Inobe et al. 1999) function in other signalling pathways. Notch is a transmembrane receptor for a lateral inhibitory signalling pathway of Delta. Eps8 is a substrate of EGF receptor kinase and mediates EGF receptor intracellular trafficking. Thus, it is possible that Dvl regulates not only Wnt signalling but also other signalling depending on the binding proteins.
In humans and mice, three Dvl genes, Dvl-1, Dvl-2, and Dvl-3 have been identified. Dvl-1 is ubiquitously expressed at the early stages of development (Wharton 2003). In the central nervous system Dvl-1 is highly expressed in areas of high neuronal density at embryonic and postnatal stages of development. A Dvl-1 knockout mouse study showed that Dvl-1 is not required for early development, but Dvl-1 null mice exhibit behavioral abnormalities and neurological deficits (Lijam et al. 1997). Dvl-2 has been shown to be highly expressed in the outer root sheath and hair precursor cells. Dvl-2 knockout mice show 50% lethality because of cardiovascular outflow tract defects and all Dvl-1/2 double knockout mutant embryos display craniorhachischisis, a completely open neural tube from the midbrain to the tail (Hamblet et al. 2002). Therefore, there is functional redundancy between Dvl-1 and Dvl-2 in neural tube closure and Dvls are suggested to be required for the formation and/or function of specific neuronal pathways.
It has been shown that Dvl-1 localizes to axonal microtubules and regulates microtubule stability through GSK-3 (Hall et al. 2000; Ciani & Salinas 2005). Dvl-1 activates Rho-kinase and inhibits nerve growth factor (NGF)-dependent neurite extension in PC12 cells (Kishida et al. 2004). Further, Dvl interacts with MuSK, a receptor tyrosine kinase, and this interaction is required for Agrin-induced acetylcholine receptor clustering at the neuromuscular junction (Luo et al. 2002). These results suggest that Dvl is involved in synaptogenesis including the regulation of neurite extension and synapse development. In addition, Dvl has been shown to localize to intracellular small vesicles by binding to lipids (Kishida et al. 1999; Capelluto et al. 2002) and to translocate to the plasma membranes in a ligand-dependent manner (Cliffe et al. 2003). Dvl also forms a complex with β-arrestin, which plays a role in receptor-mediated endocytosis (Chen et al. 2001). Thus, Dvl has been suggested to regulate the trafficking of intracellular vesicles, but the roles of Dvl in neurotransmitter release are not well understood.
Synaptotagmin (Syt) is a Ca2+-binding protein that contains an N-terminal transmembrane region and two C-terminal C2 domains and is thought to regulate membrane traffic (Chapman 2002; Südhof 2004). There are at least 15 Syt family members in mammalian cells. Syt I is evolutionally conserved and is the best characterized isoform of Syt. Syt I is enriched in synaptic vesicles and is known to be essential for synaptic vesicle exocytosis and endocytosis in neurons. Soluble N-ethylmaleimide-sensitive fusion factor (NSF) attachment protein receptor (SNARE) proteins are thought to form the minimal membrane fusion machinery that mediates exocytosis (Weber et al. 1998). In neurons, synaptobrevin is vesicle membrane SNARE (v-SNARE), and syntaxin-1A and SNAP-25 are target membrane SNARE (t-SNARE). Syntaxin-1A and SNAP-25 can form a complex on the plasma membrane to bind to synaptobrevin. Ca2+ promotes the interaction of Syt I with the syntaxin-1A and SNAP-25 complexes and with the ternary SNARE complex, suggesting that the interaction of Syt I with the SNARE complex plays a role in the reversible docking of secretary vesicles. It has also been reported that Syt I stimulates the membrane fusion of reconstituted v- and t-SNARE proteoliposomes in vitro by facilitating SNARE complex formation (Tucker et al. 2004). Furthermore, Syt I interacts with µ2- and α-adaptin of AP-2, which play an essential role in endocytosis, and this interaction is important in endocytosis, because perturbation of the interaction inhibits transferrin internalization in living cells (Haucke et al. 2000). Thus, Syt I is implicated in the regulation of endocytosis in addition to exocytosis.
In this study, we found that knockdown of Dvl in rat adrenal medulla-derived pheochromocytoma PC12 cells decreases K+-dependent dopamine release, and furthermore identified Syt as a Dvl-binding protein. Here, we show that Dvl is involved in endo- and exocytotic processes of PC12 cells and that this action of Dvl is mediated through the binding to Syt I.
Involvement of Dvl in dopamine release from PC12 cells
We have demonstrated that Dvl regulates the neurite extension of PC12 cells (Kishida et al. 2004), and PC12 cells is known to release dopamine by depolarization (Ritchie 1979). To examine whether Dvl is involved in the regulation of neurotransmitter release, the protein levels of Dvl were reduced by RNA interference (RNAi) in PC12 cells. The expression of Dvl-1, Dvl-2, and Dvl-3 was detected by their specific antibodies and all of them were decreased by RNAi (Fig. 1A). In this experiment, a small interfering RNA (siRNA) consisting of 21-mer duplexed RNA based on a sequence from rat Dvl-2 was designed. This sequence has one and two mismatches from Dvl-1 and Dvl-3, respectively. Since it has been reported that siRNA may target genes with sequence similarity (Jackson et al. 2003), it appeared that siRNA for Dvl-2 reduces the protein level of all of Dvls. The expression levels of Syt I, SNAP-25, syntaxin-1A, synaptobrevin, and β-actin were not changed (Fig. 1A).
A low concentration (4.7 mm) of KCl did not induce dopamine release, while a high concentration (60 mm) of KCl stimulated it (Fig. 1B). K+-induced dopamine release was greatly reduced in Dvl knockdown cells as compared with control PC12 cells (Fig. 1B). Expression of Flag-HA-Dvl-1 (wild type) (WT) encoded by a retrovirus vector rescued the inhibition of the dopamine release in Dvl knockdown PC12 cells (Fig. 1C,D). Knockdown or expression of Dvl-1 did not affect the neurite extension of PC12 cells maintained with growth medium used for dopamine release assay (10% fetal calf serum and 5% horse serum) (data not shown). Thus, it is likely that Dvl regulates dopamine release from PC12 cells without affecting morphology. These results prompted us to clarify how Dvl is involved in the release of neurotransmitter and to search for proteins that interact with Dvl-1 using the yeast two-hybrid method.
Identification of Syt as a Dvl-1-binding protein
A 1.8-kb cDNA insert was found to encode a sequence containing an open reading frame of Syt XI from 1 × 106 initial transformants of a mouse brain cDNA library. There are at least 15 Syt family members in mammalian cells. To determine which Syt family members interact with Dvl-1 in intact cells, haemaggulutinin (HA)-tagged Dvl-1 (HA-Dvl-1) was expressed with Flag-tagged Syt I (Flag-Syt I), III, IV, VII, IX, and XI in COS cells (Fig. 2B). When the lysates were immunoprecipitated with anti-Flag antibody, HA-Dvl-1 was found in the Flag-Syt I and XI immune complexes, but not in the Flag-Syt III, IV, VII, and IX immune complexes (Fig. 2B). Neither HA-Dvl-1 nor Flag-Syt I was immunoprecipitated from the lysates with non-immune immunoglobulin (data not shown). HA-Dvl-2 and Myc-Dvl-3 were also observed in the Flag-Syt I and XI immune complexes (Fig. 2C, D). Thus, each Dvl member specifically forms a complex with Syt I and XI. In addition, HA-Dvl-2 formed a complex with Flag-Syt IV faintly, but HA-Dvl-1 and Myc-Dvl-3 did not. At present, we do not know the reason for the differences between the bindings of Dvl members and Syt IV.
Interaction between Dvl and Syt I
Among the Syt family members, Syt I has been extensively studied and is known to play a role in exocytosis and endocytosis of synaptic vesicles (Chapman 2002; Südhof 2004). Therefore, we analyzed the biological function of the binding between Dvl-1 and Syt I. Syt I consists of 421 amino acids and has one transmembrane domain and two Ca2+ binding domains, C2A and C2B (Fig. 2A). To examine which regions of Syt I and Dvl-1 are responsible for their complex formation, various deletion mutants of T7-Syt I and HA-Dvl-1 were generated (Fig. 2A). T7-Syt I-(ΔC2A) was found in the Myc-Dvl-1 immune complex as efficiently as T7-Syt I (WT), while neither T7-Syt I-(ΔC2AB) nor T7-Syt I-(ΔC2B) was detected (Fig. 2E). HA-Dvl-1-(140–670), HA-Dvl-1-(Δ228–250), and HA-Dvl-1-(Δ251–336) formed a complex with T7-Syt I to the same extent as HA-Dvl-1 (WT). However, HA-Dvl-1-(281–670) was not observed in the T7-Syt I immune complex and HA-Dvl-1-(Δ141–227) was observed to a lesser extent (Fig. 2F). Thus, the regions containing the C2B domain of Syt I and amino acids 141–227 of Dvl-1 are important for the interaction of Dvl-1 with Syt I.
To examine whether the interactions of Dvl and Syt I are direct, in vitro binding assays using the purified proteins (Fig. 3A) were performed. Maltose-binding protein (MBP)-Syt I-(C2B), but not MBP-Syt I-(C2A), interacted with glutathione-S-transferase (GST)-Dvl-1 (Fig. 3B). C2B is known to bind to Ca2+ (Chapman 2002; Südhof 2004). However, the interaction of MBP-Syt I-(C2B) with GST-Dvl-1 was little influenced in the presence of free Ca2+ (Fig. 3B). Analyses by surface plasmon resonance (SPR) demonstrated that the static dissociation constant, Kd, is 140 nM for the binding of MBP-Syt I-(C2B) to GST-Dvl-1 (Fig. 3C). Approximately 0.6 mol of GST-Dvl-1 bound to 1 mol of Syt I-(C2AB) (Fig. 3D). The binding affinity of GST-Dvl-1-(Δ141–227) to Syt I-(C2AB) seemed to be similar to that of GST-Dvl-1, but the maximal binding activity of GST-Dvl-1-(Δ141–227) to Syt I-(C2AB) was reduced, which is consistent with the results in intact cells.
Colocalization of Dvl-1 and Syt I in PC12 cells and DRG neuron
Syt I is highly expressed in the brain and endocrine glands, and is localized to synaptic vesicles and secretory granules. Next, we examined the interaction between Dvl and Syt I in PC12 cells to know the functional significance of their binding. When Myc-Dvl-1 was expressed in PC12 cells using Sendai virus vector, it was co-immunoprecipitated with endogenous Syt I (Fig. 4A). It has been shown that the immunoprecipitates by anti-Syt I antibody contain synaptic vesicles or large dense core vesicles (LDCVs) (Lowe et al. 1988). Taken together with the previously reported observations that Dvl associates with cytoplasmic vesicles (Capelluto et al. 2002), Dvl may colocalize with Syt I on the vesicles. Endogenous Syt I was observed in the peripheral region of PC12 cells not treated with NGF, and endogenous Dvl was also detected in the peripheral region, where the two proteins were colocalized (Fig. 4B). Anti-Dvl antibody also recognized proteins throughout the cytoplasm (Fig. 4B). In PC12 cells induced to differentiate by NGF, Syt I was dominantly observed at the termini of extended neurites. Dvl was also detected in the neurite termini and colocalized with Syt I (Fig. 4B).
Rat dorsal root ganglion neuron is a well established cell model for sympathetic peripheral neuron as well as rat hippocampal neuron for CNS neuron (Banker & Goslin 1998). Syt I was enriched at the growth cone region of rat dorsal root ganglion neuron (Fig. 4C). Furthermore, Syt I was present in the cytoplasm with a punctate pattern (Fig. 4C). Consistent with the observations of PC12 cells treated with NGF, Dvl and Syt I were colocalized at the tips of the neurite termini. Taken together, these results indicate that Dvl and Syt I form a complex in neuroendocrine and neuronal cells.
Neuroendocrine and neuronal cells contain at least two types of vesicles, LDCVs containing catecholamine and neuropeptide and small synaptic vesicles (synaptic-like microvesicles (SMVs) in PC12 cells) containing acetylcholine (Winkler 1997). To characterize the subcellular distribution of Dvl and Syt in PC12 cells, the post-nuclear supernatants were fractionated by sedimentation through sucrose density gradients (Fig. 4D). Dopamine, which is present in LDCVs, was detected in fractions 12–16, and treatment of PC12 cells with NGF for 48 h increased the dopamine content (Fig. 4D). Syt I was also detected in fractions 12–16. NGF increased the total amount of Syt I and the amount of Syt I which overlapped with the fractions containing dopamine and LDCVs (Fig. 4D). Synaptophysin (Syp), which is a marker of SMVs, was detected in fractions 4–10. Taken together, these results demonstrate that Syt I is mainly localized to LDCVs (Fig. 4D).
Most Dvl was detected in the soluble fractions (fractions 1 and 2) and a small fraction of Dvl was present in fractions 4–10 and 12–16 (Fig. 4D). Although NGF did not change the total amount of Dvl (Fig. 4D, upper right panel), the amount of Dvl that was detected in fractions 10–16 was increased. This may reflect the increment in Dvl complexed with Syt I according to the synthesis of Syt I by NGF. Thus, Dvl could be localized to both LDCVs and SMVs in addition to the cytosolic fractions. We collected the fractions containing LDCVs, and LDCVs were immunoisolated by anti-Syt I antibody. However, endogenous Dvl was not detected in the immunoisolated LDCVs (data not shown). Since it was hard to show the physical interaction between Dvl and Syt I on LDCVs at the endogenous level, we next tried to prove the functional binding of Dvl and Syt I on LDCVs.
To examine whether the interaction of Dvl with Syt I is important for dopamine release, we expressed Myc-Dvl-1 (WT) or HA-Dvl-1-(Δ141–227) in wild-type PC12 cells using a Sendai virus vector (Fig. 4E). Over-expression of Dvl-1 (WT) did not affect K+-induced dopamine release, but that of HA-Dvl-1-(Δ141–227) impaired dopamine release (Fig. 4F). Taken together, these results indicate that Dvl is necessary for dopamine release by binding to Syt I and that Dvl-1-(Δ141–227) acts as a dominant-negative factor for it.
Involvement of Dvl in endocytosis
It is believed that endocytosis is coupled with exocytosis and this coupling is involved in the secretion of neuroendocrine cells (Südhof 2004). Antibody against lumenal domain of Syt I on the synaptic vesicles was used as a marker to study recycling and endocytosis of synaptic vesicles (Matteoli et al. 1992). The internalization of the antibody was increased in time-dependent and K+-dependent manners (Fig. 5A,B). The internalized antibody was observed not only into neurite tip but also into other portions of neurite and cell body (Fig. 5A). These results indicate that the internalization of Syt I-containing vesicles is coupled with exocytosis. When Dvl was reduced by RNAi in PC12 cells, the internalization of Syt I-containing vesicles was suppressed as compared with control PC12 cells (Fig. 5A,B). When Dvl-1 was over-expressed in PC12 cells using a Sendai virus vector, the internalization of anti-lumenal Syt I antibody was not affected, whereas over-expression of Dvl-1-(Δ141–227) in PCl2 cells decreased it (Fig. 5C). Thus, the decrease in the internalization of Syt I in Dvl knockdown PC12 cells might be a consequence of the blockade of exocytosis.
Formation of a complex between Dvl and exocytotic or endocytoic machineries
Finally, we found how Dvl is involved in transmitter release at the molecular level. We examined the effects of Dvl on the formation of a complex between Syt I and t-SNARE (syntaxin-1A and SNAP-25), which is important for neurotransmitter release. It has been reported that the C2AB domain of Syt I binds to t-SNARE, and that the affinity of the binding of Syt I to SNAP25 is about 200 nM in the presence of Ca2+ (Gerona et al. 2000). In vitro biochemical studies using recombinant proteins showed that syntaxin-1A binds to immobilized Syt I-(C2AB) (Fig. 6A). Stoichiometric binding was observed upon addition of Ca2+, with about 0.8 mol of syntaxin-1A bound to 1 mol of Syt I-(C2AB) (data not shown). GST-Dvl-1 (WT) inhibited the binding of Syt I-(C2AB) and syntaxin-1A in a dose-dependent manner, but GST-Dvl-1-(Δ141–227) inhibited it to a lesser extent (Fig. 6A). A binary SNARE complex including syntaxin-1A and SNAP-25 bound to MBP-Syt I-(C2AB) (Fig. 6B). GST-Dvl-1 inhibited the binding of lower concentrations (0.13 and 0.25 µM) of SNARE to MBP-Syt I-(C2AB), but it had little affect on the binding of higher concentrations (1 µM) of SNARE to MBP-Syt I-(C2AB). These results indicate that Dvl competes with t-SNARE for the binding to Syt I in vitro.
Since Syt I binds to µ2-adaptin of AP-2, which plays a critical role in endocytosis (Haucke et al. 2000), we also examined a complex between Dvl, Syt I, and µ2-adaptin. When HA-µ2 was expressed with Myc-Dvl-1 or T7-Syt I, HA-µ2 formed a complex with Myc-Dvl-1 in addition to T7-Syt I (Fig. 6C). When these proteins are expressed simultaneously, the amount of Dvl immunoprecipitated with µ2 was increased (Fig. 6C). Therefore, Dvl may bind to a complex of Syt I and µ2 more efficiently than to either of them.
In this study, we have for the first time demonstrated that Dvl is involved in neurotransmitter release and that this action is mediated by the binding to Syt I. Syts constitute a family of membrane traffic proteins that are characterized by an N-terminal transmembrane domain, a variable linker, and two C-terminal C2-domains (Chapman 2002; Südhof 2004). Based on their sequences and properties, at least 15 Syts can be grouped into six classes: class 1 (Syts I and II), class 2 (Syt VII), class 3 (Syts III, V, VI, and X), class 4 (Syts IV and XI), class 5 (Syts IX), and class 6 (Syts VIII, XII, XIII, XIV, and XV). We found that each of Dvl-1, -2, and -3 bound to Syts I and XI, but not to Syts III, IV, VII, and IX. Amino acids 141–227 of Dvl-1 were required for the binding to the C2B domain of Syt I. Since the corresponding regions of Dvl-2 and -3 have 83 and 90% similarities, respectively, with amino acids 141–227 of Dvl-1, this region is important for the binding of the Dvl protein family with Syts. Dvl colocalized with Syt I at the neurite termini in PC12 cells induced to differentiate by NGF and in dorsal root ganglion neurons. Furthermore, sucrose density gradient-subcellular fractionation indicated that Dvl is present in the fractions containing the large dense-core vesicles, which also contains Syt I, in addition to the cytosol. Thus, Dvl binds to Syt I, which is located to secretory vesicles.
In addition to the ability to bind to Syt I, the following evidence indicated that Dvl might contribute to dopamine release from PC12 cells. When Dvl was reduced by RNAi, depolarization-dependent dopamine release in PC12 cells was decreased. Moderate expression of Dvl-1 mediated by retrovirus vector in Dvl knockdown cells rescued the dopamine release to the WT level. Overexpression of Dvl-1 mediated by Sendai virus vector did not affect K+-induced dopamine release, suggesting that endogenous Dvl is not a limiting factor for dopamine release. In contrast, over-expression of Dvl-1-(Δ141–227), which has a reduced ability to bind to Syt I, inhibited dopamine release. The DIX domain of Dvl mediates the oligomerisation of Dvl (Kishida et al. 1999). If the oligomerisation of wild-type Dvl is necessary for dopamine release, Dvl-1-(Δ141–227) may inhibit the function of endogenous Dvl by forming a complex with it in a dominant-negative manner. Furthermore, consistent with the concept that endocytosis is coupled with exocytosis (Südhof 2004), K+-dependent internalization of Syt I was also suppressed in Dvl knockdown or Dvl-1-(Δ141–227)-expressing PC12 cells. Taken together, these results suggest that Dvl is involved in the depolarization-dependent dopamine release from PC12 cells by binding to Syt I.
Wnt-3a induces the accumulation of β-catenin and activates Rho in PC12 cells (Kishida et al. 2004). However, purified Wnt-3a did not affect K+-induced dopamine release (data not shown). Therefore, the activation of the β-catenin and Rho pathways do not affect dopamine release from PC12 cells. Dvl inhibits NGF-dependent neurite extension of PC12 cells (Kishida et al. 2004). However, it is unlikely that regulation of dopamine release by Dvl is due to effects on neurite outgrowth, because Dvl inhibited dopamine release from PC12 cells without NGF treatment. Dvl may have a role in regulating the neurotransmitter release independently of the Wnt-3a signalling pathway. Wnt-7a is expressed by granule cells during synaptogenesis with mossy fibers and increases synapsin I clustering in mossy fiber axons (Ciani & Salinas 2005). Therefore, some Wnt ligands may regulate neurotransmitter release through Dvl.
How does Dvl regulate neurotransmitter release through binding to Syt I? Syt I is involved in endo–exocytosis by binding to Ca2+, t-SNARE, and adaptor protein AP-2 (Shao et al. 1997; Haucke et al. 2000). Our in vitro data showed that Dvl inhibits the interaction of Syt I with t-SNARE. Since syntaxin-1A binds to the C2A domain of Syt I (Shao et al. 1997) and Dvl binds to the C2B domain of Syt I, the association of Syt I with Dvl may cause conformational change and reduce the affinity of Syt I for syntaxin-1A. Although we cannot draw conclusions about the actions of Dvl on the SNARE complex because of the difficulty of examining the interaction of Dvl, Syt I, and t-SNARE at the endogenous level, these observations suggest that Dvl may inhibit the ability of Syt I to stimulate the docking or fusion between vesicles and membranes. Alternatively, Dvl may function in releasing Syt I from t-SNARE.
The C2B domain of Syt I has been shown to bind to µ2- and α-adaptin, both of which are the subunits of the endocytotic adaptor complex AP-2, and this interaction is required for clathrin-dependent endocytosis (Haucke et al. 2000). Thus, Syt I is also known to regulate recycling or endocytosis of synaptic vesicles (Poskanzer et al. 2003) and to serve as a link between the exocytotic and endocytotic events of the synaptic vesicles (Zhang et al. 1994). Dvl has been shown to localize to intracellular vesicles (Capelluto et al. 2002) and to be involved in the internalization of Frizzled 4, a receptor of Wnt-5a (Chen et al. 2003). Our results show that Dvl binds to µ2-adaptin and that this complex formation is enhanced by the presence of Syt I also support that Dvl is involved in the regulation of endocytosis.
Although the exact roles of Dvl-1 in the endo- and exocytosis of Syt I-containing vesicles are not clear at present, it is likely that Dvl is involved in multiple steps of synaptic or synaptic like vesicle cycle (Südhof 1995). First, Dvl may play a role in intracellular translocating process of vesicles filled with transmitters to the release sites (the active zone), and Dvl may be released from Syt I to stimulate the binding of Syt I to t-SNARE before docking or fusion. Another possible model is that after exocytosis, Dvl may inhibit the association of Syt I to the SNARE complex and binds to the complex between Syt I and AP-2, thereby inducing endocytosis. Since syntaxin-1A is excluded from recycling synaptic vesicles at nerve terminals (Mitchell & Ryan 2004), Dvl may be involved in the release of syntaxin-1A from Syt I during endocytosis. In both cases, Dvl acts as a positive regulator for dopamine release from PC12 cells. Dvl-1 null mice exhibit behavioral abnormalities and neurological deficits (Lijam et al. 1997). Clarifying the roles of Dvl in endo- and exocytosis would contribute to understand the molecular mechanism of these phenotypes.
Materials and chemicals
pcDNA-HA-µ2 and pOZ-N were provided by Drs V. Haucke (George-August-University, Göttingen, Germany) and T. Ikura (Tohoku University, Sendai, Japan), respectively. Anti-Myc antibody was prepared from 9E10 cells. PC12 cells were grown in Dulbecco's modified essential medium (DMEM) supplemented with 10% fetal bovine serum, 5% horse serum, and 500 ng of tetracycline hydrochloride/mL. Rat dorsal root ganglion neurons were collected as described (Inoue et al. 1999), and cultured with DMEM containing 10% horse serum and 50 ng of NGF/mL. GST-Syt I-(C2A), -(C2B), -(C2AB), MBP-Syt I-(C2A), and -(C2B) were purified according to manufacturers’ instructions besides recombinant protein immobilized to the glutathione sepharose or amylose resin were treated with 10 µg/mL of DNase and RNase (Roche applied sciences) and washed with 400 mm KCl and 20 mm Tris–HCl (pH 7.5) to remove nucleotide contaminants (Bhalla et al. 2006). GST-Dvl-1 and GST-Dvl-1-(Δ141–227) were purified from Spodoptera frugiperda (Sf) 9 cells. Anti-Dvl antibody was prepared as described (Kishida et al. 1999). Anti-Dvl-1, -2, and -3 antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-T7 antibody was purchased from Novagen (Madison, WI). Anti-Flag M2, anti-synaptophysin (clone SVP38), and anti-syntaxin-1A (clone HPC-1) antibodies were from Sigma (Saint Louis, MO). Anti-Syt I and anti-synaptobrevin antibodies were from Stressgen (Victoria BC, Canada). Oyster®550-labelled-antibody against lumenal domain of Syt I and anti-SNAP-25 (clone CI71.2) antibodies, and NGF 7.5S were from Synaptic Systems (Göttingen, Germany) and Invitrogen Inc. (Carlsbad, CA), respectively. Other materials were purchased from commercial sources.
pCGN/hDvl-1, pCGN/hDvl-1-(140–670), pCGN/hDvl-1-(Δ141–227), pCGN/hDvl-1-(Δ228–250), pCGN/hDvl-1-(Δ251–336), pCGN/hDvl-1-(281–670), pVIKS/Dvl-1, and pEFBOS-Myc/hDvl-3 were constructed as described (Kishida et al. 1999; Hino et al. 2003). pEF-Flag/Syt I, pEF-Flag/Syt III, pEF-Flag/Syt IV, pEF-Flag/Syt VII, pEF-Flag/Syt IX, pEF-Flag/Syt XI, pEF-T7/Syt I, pEF-T7/Syt I-(ΔC2AB), pEF-T7/Syt I-(ΔC2A), pEF-T7/Syt I-(ΔC2B), pGEX-2T/Syt I-(C2A), and pGEX-2T/Syt I-(C2B) were constructed as described (Fukuda et al. 2001; Fukuda 2002). pMAL-c2/Syt I-(C2A), pMAL-c2/Syt I-(C2B), pMAL-c2/Syt I-(C2AB), pGEX-2T/syntaxin-1A, pGEX-4T/SNAP-25, pGEX-4T-3/synaptobrevin, pOZ/Dvl-1 (wildtype), pCGN/Dvl-2, and pVIKS/Dvl-1-(Δ141–227) were constructed using standard techniques.
The nucleotide sequences of rat Dvl-2 (accession number; XM_239254) and Dvl-3 (accession number; XM_221304) were predicted by automated computational analysis from an annotated genomic sequence (accession number; NW_047334). The siRNA for rat Dvl was derived from the sequence for mRNA nucleotides 970–990 of rat Dvl-2. The siRNA for rat Dvl-2 has one and two nucleotides differences from the siRNA for rat Dvl-1 and Dvl-3, respectively. To generate retroviruses, pSUPER-retro.neo+gfp/Dvl-2 was transfected into ampho-pack 293 cells (BD Biosciences Clontech, Palo Alto, CA). These cells were cultured for 2 days, and the conditioned medium containing retroviruses were collected and filtrated. Retrovirus-containing media were incubated with PC12 cells in the presence of 4 µg/mL polybrene for 48 h. Then, the cells were selected in the presence of 400 µg/mL G418 to obtain PC12 cells stably expressing siRNA for Dvl. Stably expressing cells were pooled from each transfection, rather than individual clones, to avoid the effects of clonal variations.
Dopamine release assay
PC12 cells were plated onto 35-mm-diameter dishes at a density of 1.5×106 cells per dish and cultured for 48 h. The cells were washed once with a low K+ solution (Buffer A (20 mm Hepes–NaOH, pH 7.4, 140 mm NaCl, 2.5 mm CaCl2, 1.2 mm MgSO4, 1.2 mm KH2PO4, and 1.1 mm glucose) containing 4.7 mm KCl), and incubated in the low K+ solution for 10 min. After the solutions were subsequently replaced with 1 mL of the low K+ solution or a high K+ solution (Buffer A containing 60 mm KCl) for the indicated periods of times, the solutions were collected and the cells were extracted with 0.5 mL of 0.4 % (w/v) perchloric acid. Contents of dopamine were determined by high pressure liquid chromatography using a reverse-phase column and an electrochemical detector system (Eicom, Kyoto, Japan) (Nagano et al. 2003). The rate of dopamine release was expressed as a percentage of secreted dopamine relative to the sum of dopamine secreted from cells and dopamine remaining in the cells.
Imaging of vesicle endocytosis using antibody against lumenal domain of Syt I
PC12 cells plated onto 18-mm glass cover slips coated with poly-d-lysine (Sigma, St Louis, MO) were cultured with 100 ng/mL NGF for 48 h. Syt I exposed to the cell surface and internalized as a result of exocytotic stimulation was labelled as described with modifications (Matteoli et al. 1992). Briefly, the cells were washed with the low K+ solution and incubated with the low K+ solution for 10 min. After stimulation with high K+ solution containing antibody against lumenal domain of Syt I conjugated with Oyster®550 (Synaptic Systems GmbH, Germany) for the indicated periods of time at 37°C, the cells were rinsed with ice cold 0.5% acetic acid (w/v) and 0.5 M NaCl for 10 min to remove the antibody which was not internalized and then were fixed with 4% paraformaldehyde in PBS for 10 min. The cells were viewed with the LSM510 confocal microscope system (Carl Zeiss, Germany).
To quantitate the internalization of Syt I, 50 cells from three independent experiments were selected randomly and all images were acquired under identical conditions. The integrated fluorescence of each cell was determined by taking the mean fluorescence of the whole cell (fluorescence of neurites and cell body/area) using Metamorph software (Molecular devices Corporation, Sunnyvale CA). As the total expression level of Syt I was not changed during the treatment with the high K+ or low K+ solution up to 20 min (by Western blot and immunostaining, data not shown), the mean fluorescence per cell (nurites and cell body) represented the endocytosis of the Syt I containing vesicles.
Binding of Dvl, Syt I, and µ2
To examine the direct interaction of Dvl and Syt I in vitro, GST-Dvl immobilized on glutathione-Sepharose 4B was incubated with MBP, MBP-Syt I-(C2A) or MBP-Syt I-(C2B) in 50 uL of reaction mixture (20 mm Hepes–NaOH, pH 7.4, 150 mm NaCl, 0.5% (w/v) Triton X-100, and 1 mm EGTA) for 1 h at 4°C. After the GST-fusion proteins were precipitated by centrifugation, the precipitates were resolved by electrophoresis and stained with Coomassie Brilliant Blue (CBB) or probed with anti-MBP antibody. In some experiments, Syt I-(C2AB), which was obtained by digestion of GST-fusion protein, bound to anti-Syt I antibody and protein A complex was incubated with GST-Dvl-1 (WT) or GST-Dvl-1-(Δ141–227) under the same conditions as above. The amounts of each protein were quantified with densitometry using NIH image software.
To examine the formation of a complex between Dvl and Syts in intact cells, COS cells (60-mm-diameter dishes) expressing epitope-tagged Dvls and Syts were lysed in 300 µL of NP-40 buffer (20 mm Tris–HCl (pH 7.5), 135 mm NaCl, 1% (w/v) Nonidet-P40, 20 µg/mL aprotinin, 20 µg/mL leupeptin, and 1 mm phenylmethylsulfonyl fluoride). After the proteins were immunoprecipitated with anti-FLAG, anti-Myc, or anti-T7 antibody, the immunoprecipitates were probed with anti-Dvl, anti-FLAG, anti-Myc, anti-T7, or anti-HA antibody. To examine the formation of a complex between Dvl, Syt, and µ2-adaptin, COS cells (60-mm-diameter dishes) expressing various combinations of Myc-Dvl-1, T7-Syt I, and HA-µ2 were lysed in 300 µL of NP-40 buffer. After the proteins were immunoprecipitated with anti-HA antibody, the immunoprecipitates were probed with anti-HA, anti-T7, or anti-Dvl antibody.
PC12 cells (four 100-mm-diameter dishes) cultured in the presence or absence of 100 ng/mL NGF for 48 h were harvested and washed with PBS once. The cells were homogenized in 1 mL of homogenization buffer (10 mm Hepes–NaOH, pH 7.4, 0.32 M sucrose, 1 mm phenylmethylsulfonyl fluoride, 20 µg/mL aprotinin, and 20 µg/mL leupeptin) with ten strokes in a Teflon/glass Potter homogenizer and the homogenates were centrifuged at 10 000 g for 10 min at 4°C. The supernatant was loaded on a 9-mL linear 0.6–1.6 M sucrose gradient and centrifuged at 150 000 g for 6 h in a P40ST rotor (Hitachi Koki, Co., Ltd., Hitachi, Japan). Twenty fractions of 0.5 mL each were collected from top of the gradient and the aliquots were analyzed by immunoblotting.
Binding of Syt I to syntaxin-1A and SNARE complex
The binding of Syt I to syntaxin-1A and SNARE complex was measured as described (Gerona et al. 2000). Syntaxin-1A was obtained by digestion of GST-fusion protein with thrombin. Syt I-(C2AB) (4 µg) was immunoprecipitated with the anti-Syt I antibody and protein A Sepharose for 1 h at 4°C and washed with PBS containing 0.5% Triton X-100 3 times. Syntaxin-1A (3 µm) was incubated with Syt I-(C2AB) immobilized on protein A Sepharose in the presence GST-Dvl-1 or GST-Dvl-1-(Δ141–227) in 100 µL of reaction mixture (20 mm Hepes–NaOH, pH 7.4, 150 mm NaCl, 0.5% (w/v) Triton X-100, 1 mm EGTA, and 1.1 mm CaCl2) for 1.5 h at 4°C. After Syt I-(C2AB) was precipitated by centrifugation, the precipitates were probed with anti-syntaxin antibody. In some experiments, binary SNARE complex including syntaxin-1A and GST-SNAP-25 was incubated with MBP-Syt I-(C2AB) immobilized to amylose resin in the presence or absence of GST-Dvl-1 (WT) in the same reaction mixture as above.
Yeast two-hybrid screening was carried out using Dvl-1 as a bait (Kishida et al. 1999). Immunoblotting, immunoprecipitation, and immunohistochemical analyses were performed as described (Hino et al. 2003; Kishida et al. 2004). Retrovirus carrying Flag-HA-Dvl-1 was produced as described (Ikura et al. 2000). Sendai virus carrying the Myc-Dvl-1 and HA-Dvl-1-(Δ141–227) gene were produced as described (Li et al. 2000). SPR spectrometry for the binding of MBP-Syt I-(C2B) to GST-Dvl-1 was performed by the methods as described (Oshiro et al. 2002).
We are grateful to Dr A. Mizoguchi (Mie University, Tsu, Japan) for helpful discussion, to Drs K. Yano-Aoki, A. Oshita, and A. Inoue (Hiroshima University, Hiroshima, Japan) for their technical assistances, and to T. Yamashita (DNAVEC Corporation, Tsukuba, Japan) for Sendai virus construction. We also thank Dr V. Haucke (George-August-University, Göttingen, Germany) for a plasmid. This work was supported by Grants-in-aid for Science Research from Ministry of Education, Culture, Sports, Science and Technology of Japan (2004, 2005, 2006).