INTRODUCTION

We have isolated Doc2 as a novel protein having two C2 domains that interact with Ca²⁺ and PL¹ (1). Doc2 consists of two isoforms, Doc2 and Doc2 (1, 2). Doc2 is specifically expressed in neuronal cells, whereas Doc2 is ubiquitously expressed (1-3). Both isoforms have at least the N-terminal Doc2-specific region and C-terminal two C2 domains. We have moreover shown that overexpression of the N-terminal fragment of Doc2 or its C-terminal fragment including the C2 domains in PC12 cells inhibits Ca²⁺-dependent exocytosis (4). These results suggest that Doc2 is involved in Ca²⁺-dependent exocytosis and interacts with another component of Ca²⁺-dependent exocytotic machinery. To clarify the mode of action of Doc2 in Ca²⁺-dependent exocytosis, it is important to isolate its interacting protein(s). We have attempted here to isolate a Doc2-interacting protein from a rat brain cDNA library by use of the yeast two-hybrid system and isolated Munc13 as a Doc2-interacting protein.

Munc13 has been isolated as a mammalian homologue of Caenorhabditis elegans unc-13, which is implicated in Ca²⁺-dependent neurotransmitter release (5, 6). Munc13 has three isoforms, Munc13-1, -2, and -3. All the isoforms have two C2 domains and Munc13-1 has another atypical C2 domain. They have moreover one C1 domain that interacts with PE or DAG and PL (5-7). Munc13 is specifically expressed in neuronal cells, and Munc13-1 is localized at the presynaptic plasma membrane (6).

We describe here that Doc2 directly interacts with Munc13-1 in a DAG-dependent manner and that the Doc2-Munc13-1 interactions play an important role in Ca²⁺-dependent exocytotic machinery.

EXPERIMENTAL PROCEDURES

The N-terminal fragment (1-90 aa) of human Doc2 cDNA (1) was inserted into the pBTM116 (pLexA-Doc2N). The yeast reporter strain L40 was transformed with pLexA-Doc2N and a rat brain cDNA library constructed in pGAD10 (CLONTECH). Library plasmids from positive clones were analyzed by transformation tests and DNA sequencing. Overlapping clones containing the full-length coding region of Munc13-1 were isolated by screening the rat brain cDNA library. The cDNA fragments encoding several Munc13-1 deletion mutants were constructed from the overlapping clones and inserted into pGAD424. The cDNA fragments encoding several Doc2 deletion mutants were inserted into pBTM116. After co-transformation into yeast strain L40, -galactosidase activity was assayed by liquid and filter assays (8, 9).

The cDNA fragments encoding the N-terminal fragment (1-90 aa) of human Doc2 (1) and Munc13-1-Did (851-1461 aa) were inserted into pGEX vectors, expressed in Escherichia coli as GST fusion proteins, and purified on glutathione-Sepharose 4B columns (Pharmacia Biotech Inc.).

Mammalian expression plasmids pEFBOS-HA and pEFBOS-myc were generated to express fusion proteins with the N-terminal HA and myc epitopes, respectively (4, 10). In vitro and in vivo expression plasmids pGEM-HA and pBluescript-myc were generated to express fusion proteins with the N-terminal HA and myc epitopes, respectively. The cDNA fragments encoding human Doc2 (1) and its deletion mutants were inserted into pEFBOS-HA and pGEM-HA. The cDNA fragments encoding Munc13-1 and its deletion mutants were inserted into pEFBOS-myc and pBluescript-myc.

Assay for Doc2-Munc13-1 Interactions in a Cell-free System

The cDNA fragments, which were inserted into pGEM-HA or pBluescript-myc, were translated in vitro using TNT T7-coupled reticulocyte lysate system (Promega). 2 µg of GST-Doc2 (1-90 aa) or GST-Munc13-1-Did (851-1461 aa) were immobilized onto 20 µl of glutathione-Sepharose 4B beads. The immobilized beads were added to 500 µl of Buffer A (150 mM NaCl, 50 mM HEPES, pH 7.4, and 1 mM EGTA) containing in vitro translated products and gently mixed for 4 h at 4 °C in the presence or the absence of 100 nM TPA or PDBu. The beads were washed four times with Buffer A and the bound proteins were eluted by addition of 100 µl of Buffer A containing 20 mM glutathione. The eluates were subjected to SDS-PAGE followed by autoradiography.

Assay for Doc2-Munc13-1 Interactions in an Intact Cell System

PC12 cells were plated at a density of 5 × 10⁵ cells/60-mm dish and were incubated for 18 h. PC12 cells were infected for 30 min with T7 RNA polymerase recombinant vaccinia virus (LO-T7) and then co-transfected with 2 µg of pGEM-HA encoding Doc2 or its deletion mutants and 2 µg of pBluescript-myc encoding Munc13-1 or its deletion mutants, by use of LipofectAMINE reagent (Life Technologies, Inc.). Immunoprecipitation was performed 5 h after the transfection. In some cases, before performing the immunoprecipitation, PC12 cells were washed with PSS (140 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 20 mM HEPES, pH 7.4, and 11 mM glucose) and incubated for 10 min with a high K⁺ solution (PSS containing 60 mM KCl and 85 mM NaCl), a low K⁺ solution (PSS containing 4.7 mM KCl and 140 mM NaCl), or low K⁺ solution containing 100 nM TPA. The cells were washed with phosphate-buffered saline twice and lysed in a lysis buffer containing 20 mM Tris/HCl at pH 7.5, 150 mM NaCl, and 1% Nonidet P-40. The cell lysate was subjected to immunoprecipitation with 3 µg of the anti-HA monoclonal antibody bound to 20 µl of protein A-Sepharose. Comparable amounts of the pellets were subjected to SDS-PAGE, followed by immunoblot analysis with the biotinated mouse anti-myc antibody.

PC12 cells were co-transfected with 2 µg of pXGH5 encoding human GH (11) and 2 µg of pEF-BOS bearing the indicated cDNA by use of LipofectAMINE reagent (4). After 48 h, PC12 cells were then stimulated by each agonist for 10 min. The amounts of the GH released into the medium and retained in the cells were measured using a radioimmunoassay kit (Nichols Institute).

RESULTS

We first attempted to isolate a Doc2-interacting protein by use of the yeast two-hybrid system with the N-terminal region (1-90 aa) of Doc2 as a bait from a rat brain cDNA library. Screening of 1 × 10⁶ transformants yielded nine independent positive clones that interacted with Doc2. The seven positive clones had cDNA inserts ranging from 2.0 to 3.1 kilobase pairs, all of which encoded the parts of the sequences corresponding to 840-1743 aa residues of Munc13-1, except that the isolated clones containing the C-terminal portion of Munc13-1 lacked the sequences corresponding to 1560-1578 aa residues of Munc13-1 (6). This region might be subjected to alternative splicing. The minimum regions of the Doc2-Munc13-1 interactions were 851-1461 aa residues of Munc13-1 (Munc13-1-Did = Doc2-interacting domain of Munc13-1) (Fig. 1a) and 13-37 aa residues of Doc2 (Doc2-Mid = Munc13-1-interacting domain of Doc2) (Fig. 1b). Doc2 also interacted with Munc13-1 (Fig. 1c). Munc13-2 also interacted with Doc2 and Doc2. Doc2-Mid showed striking sequence homology to Doc2-Mid (92% identity). Munc13-1-Did also showed striking sequence homology to Munc13-2-Did (81% identity). These interactions were estimated by the yeast two-hybrid system.

Doc2-Munc13-1 interactions in the yeast two-hybrid system.

The Doc2-Munc13-1 interactions were confirmed by the binding of in vitro translated, [³⁵S]methionine-labeled Doc2 and Munc13-1 to GST-tagged recombinant Munc13-1-Did and the GST-tagged Doc2 Mid-containing region, respectively (Fig. 2, a and b). The Doc2-Mid and Munc13-1-Did deletion mutants did not interact with the respective partner proteins.

Doc2-Munc13-1 interactions in cell-free and intact cell systems.

The Doc2-Munc13-1 interactions were furthermore confirmed by co-immunoprecipitation from cultured PC12 cells of myc-tagged full-length Munc13-1 and HA-tagged full-length Doc2 and of myc-tagged Munc13-1-Did and HA-tagged full-length Doc2 (Fig. 2, c and d). The Doc2-Mid and Munc13-1-Did deletion mutants were not co-immunoprecipitated with the respective partner proteins. The co-immunoprecipitation of full-length Munc13-1 with Doc2 from PC12 cells was markedly enhanced when the cells were stimulated by TPA or high K⁺ in the presence of extracellular Ca²⁺, which induced Ca²⁺-dependent exocytosis (Fig. 3a). However, the co-immunoprecipitation was not enhanced when the cells were stimulated by TPA or high K⁺ in the absence of extracellular Ca²⁺ (data not shown). The TPA-induced or high K⁺-induced co-immunoprecipitation was not observed with the Did deletion mutant of Munc13-1. With the C1 domain deletion mutant, the co-immunoprecipitation of full-length Munc13-1 with Doc2 was observed even without the stimulation of PC12 cells by high K⁺ or TPA. Consistent with these cell level experiments, the interactions of in vitro translated, [³⁵S]methionine-labeled Munc13-1 and the recombinant GST-tagged Doc2-Mid-containing region were also stimulated by TPA or PDBu in a cell-free binding assay system (Fig. 3b). The Doc2-Munc13-1 interactions were not observed with the Doc2-Mid and Munc13-1-Did deletion mutants of the respective proteins irrespective of the presence or absence of TPA (data not shown).

High K⁺- or TPA-induced Doc2-Munc13-1 interactions.

It was finally examined whether the Doc2-Munc13-1 interactions are functionally relevant for Ca²⁺-dependent exocytosis. For this experiment, we took advantage of the GH coexpression assay system of PC12 cells (12). In this system, human GH and a sample to be tested were coexpressed. Expressed GH is known to be stored in dense core vesicles and to be released in response to high K⁺ and TPA in the presence of extracellular Ca²⁺ (13, 14). The Northern blot analysis indicated that both Doc2 and Munc13-1 were expressed in PC12 cells (data not shown). Overexpression of Doc2 enhanced not only the high K⁺-induced GH release (4) but also the TPA-induced GH release (Fig. 4). Overexpression of Munc13-1-Did reduced both the high K⁺- and TPA-induced GH release. Co-expression with Doc2 suppressed this reduction.

Involvement of the Doc2-Munc13-1 interactions in Ca²⁺-dependent exocytosis from PC12 cells.

DISCUSSION

We have shown here that Doc2 interacts with Munc13-1 in a cell-free system and that these interactions are stimulated by PE. These results, together with the earlier findings that PE directly interacts with the C1 domain of unc-13 (5, 7), indicate that the binding of PE to the C1 domain of Munc13-1 causes the Doc2-Munc13-1 interactions. We have moreover shown here that the Doc2-Munc13-1 interactions are observed in intact PC12 cells and enhanced during the high K⁺ or TPA-induced Ca²⁺-dependent exocytosis,and that these interactions are observed even without the stimulation of PC12 cells by high K⁺ or TPA when the C1 domain deletion mutant of Munc13-1 is used. These results, together with the earlier findings that high K⁺ induces DAG formation (15), suggest that the Doc2-Munc13-1 interactions are induced by DAG produced during Ca²⁺-dependent exocytosis through the C1 domain of Munc13-1. Finally, we have demonstrated by use of the GH co-expression assay system of PC12 cells that Doc2 and Munc13-1 functionally interact with each other during Ca²⁺-dependent exocytosis.

Many systems and components are implicated in Ca²⁺-dependent exocytosis, such as neurotransmitter release. These include N-ethylmaleimide-sensitive factor/SNAP/SNARE, Rab, protein kinase C, and Ca²⁺-binding protein systems (for reviews see Refs. 16 and 17). Of these systems, the SNARE system is implicated in docking of synaptic vesicles with the presynaptic plasma membrane through the v-SNARE (vesicle-associated membrane protein)-t-SNARE (syntaxin and SNAP-25) interactions (for a review see Ref. 18). In C. elegans, unc-13 belongs to a group of genes defined by mutations with a paralytic phenotype and accumulation of acetylcholine (19), suggesting that Munc13-1 is also involved in neurotransmitter release in mammals. Doc2 is involved in Ca²⁺-dependent exocytosis from PC12 cells (4). Munc13-1 is located on the presynaptic plasma membrane (6), and Doc2 is concentrated on synaptic vesicles (1). Our present results together with these earlier findings suggest that the Doc2-Munc13-1 system is another docking machinery controlled by DAG. It has recently been shown that Munc13-1 interacts directly with syntaxin (20) and that Munc18, a mammalian homologue of C. elegans unc-18 (21), directly interacts with Doc2 (22). Munc18 directly interacts with syntaxin, and Munc18 is dissociated from syntaxin when syntaxin forms a complex with vesicle-associated membrane protein and SNAP-25 (23). Therefore, the Doc2-Munc13 and Doc2-Munc18 systems may function in cooperation with syntaxin in docking process. It is likely that the mutual interactions among syntaxin, Munc18, Munc13, and Doc2, play a crucial role in docking process. It is important to examine the effects of the DAG-induced Doc2-Munc13 interactions on the Doc2-Mun18, Mun13-syntaxin, and Munc18-syntaxin interactions.

Another recent analysis indicates that PE increases the size of the readily releasable pool of secretory granules in bovine adrenal chromaffin cells (24). It has been suggested that this action of PE is mediated through protein kinase C, but the properties of Munc13-1 suggest that it is a better candidate for this action of PE. Moreover, because Doc2 and Munc13-1 interact with Ca²⁺, they may serve as Ca²⁺ sensors for Ca²⁺-dependent exocytosis in cooperation with other Ca²⁺-binding proteins. Many proteins having two C2 domains have been identified. These include Doc2 (1), Munc13 (6), synaptotagmin (25), and rabphilin-3A (26), all of which are implicated in Ca²⁺-dependent exocytosis. The C2 domain of PKC has been shown to interact with membrane PL, particularly phosphatidylserine, in the presence of Ca²⁺ (for a review see Ref. 27). The precise role of Ca²⁺ in Ca²⁺-dependent exocytosis still remains unclear, but it could be speculated that the proteins having two C2 domains constitute a big complex to form a scafold-like structure and play a critical role in the fusion process in cooperation with Ca²⁺ and membrane PL in addition to the docking process.