HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 24, 22025-22034, June 14, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/24/22025    most recent M201614200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Grishanin, R. N. Articles by Martin, T. F. J. Search for Related Content PubMed PubMed Citation Articles by Grishanin, R. N. Articles by Martin, T. F. J. Social Bookmarking                      What's this?

Membrane Association Domains in Ca²⁺-dependent Activator Protein for Secretion Mediate Plasma Membrane and Dense-core Vesicle Binding Required for Ca²⁺-dependent Exocytosis^*

Ruslan N. Grishanin, Vadim A. Klenchin, Kelly M. Loyet, Judith A. Kowalchyk, Kyoungsook Ann, and Thomas F. J. Martin

From the Department of Biochemistry, University of Wisconsin, Madison, Wisconsin 53706

Received for publication, February 17, 2002, and in revised form, April 1, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Ca²⁺-dependent activator protein for secretion (CAPS) is a cytosolic protein essential for the Ca²⁺-dependent fusion of dense-core vesicles (DCVs) with the plasma membrane and the regulated secretion of a subset of neurotransmitters. The mechanism by which CAPS functions in exocytosis and the means by which it associates with target membranes are unknown. We identified two domains in CAPS with distinct membrane-binding properties that were each essential for CAPS activity in regulated exocytosis. The first of these, a centrally located pleckstrin homology domain, exhibited three properties: charge-based binding to acidic phospholipids, binding to plasma membrane but not DCV membrane, and stereoselective binding to phosphatidylinositol 4,5-bisphosphate. Mutagenesis studies revealed that the former two properties but not the latter were essential for CAPS function. The central pleckstrin homology domain may mediate transient CAPS interactions with the plasma membrane during Ca²⁺-triggered exocytosis. The second membrane association domain comprising distal C-terminal sequences mediated CAPS targeting to and association with neuroendocrine DCVs. The CAPS C-terminal domain was also essential for optimal activity in regulated exocytosis. The presence of two membrane association domains with distinct binding specificities may enable CAPS to bind both target membranes to facilitate DCV-plasma membrane fusion.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Regulated secretion in neuroendocrine cells is mediated by the Ca²⁺-dependent merger of dense-core vesicles (DCVs)¹ with the plasma membrane. When reconstituted in broken cell (1) or purified plasma membrane preparations containing docked vesicles (2), DCV exocytosis requires MgATP and Ca²⁺, which act at sequential priming and triggering stages. Both stages require distinct cytosolic proteins, characterization of which has revealed essential molecular mechanisms underlying late steps in exocytosis (3). The cytosolic proteins required for MgATP-dependent priming consist of phosphatidylinositol transfer protein and PtdIns(4)P 5-kinase (4, 5), which mediate PtdIns(4,5)P₂ synthesis that is essential for Ca²⁺-triggered vesicle fusion (3). A single cytosolic factor, Ca²⁺-dependent activator protein for secretion (CAPS), is necessary and sufficient for Ca²⁺-triggered vesicle fusion after MgATP-dependent priming has been completed (1, 6, 7). CAPS binds membranes after MgATP-dependent priming in the absence of Ca²⁺ and is required during Ca²⁺-triggered fusion,² indicating that CAPS functions in parallel with the membrane fusion machinery consisting of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins (8).

Genetic studies in Caenorhabditis elegans (7, 9) and in Drosophila (10) confirm an essential role for CAPS in the regulated secretion of a subset of neurotransmitters. CAPS antibody inhibition studies in endocrine cells (11, 12) indicate that CAPS is required for the exocytosis of DCVs that undergo rapid release and that are tightly coupled to Ca²⁺ influx. Moreover, studies in permeable synaptosomes indicated that CAPS was selectively required for Ca²⁺-dependent DCV but not synaptic vesicle exocytosis (13). Consistent with this, membrane-bound CAPS in brain homogenates was found to reside on DCVs but not synaptic vesicles (14). The basis for the association of CAPS with membranes and its mechanism in DCV exocytosis remain to be clarified.

CAPS and its invertebrate orthologs form a unique family of novel proteins that exhibit limited sequence homology to members of the Munc13 protein family that function in synaptic vesicle fusion (15). The only functional domain annotated in CAPS with high confidence by sequence analysis tools such as SMART (24) is a pleckstrin homology (PH) domain located in a central region of the protein. In the current work, we studied the properties of the PH domain and its role in CAPS function. PH domains found in numerous proteins consist of ~100-amino acid, structurally related modules that bind phospholipids and serve as membrane association or regulatory domains (16-19). Our studies demonstrate through mutagenesis that the PH domain is required for CAPS function in regulated exocytosis. The properties of the CAPS PH domain that were essential for function were its charge-based interactions with acidic phospholipids and its ability to associate with specific cellular membranes (plasma membrane but not DCV membrane). A distinct membrane association domain was found to mediate the targeting of CAPS to DCVs. This domain comprising the C-terminal 177 amino acid residues was also essential for CAPS function in regulated exocytosis. The presence of two membrane association domains with distinct membrane binding specificities may enable CAPS to bind both target membranes to facilitate DCV-plasma membrane fusion.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell Culture and Transfection-- PC12 cells were cultured as described in Ref. 20. COS-1 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in a 5% CO₂ atmosphere. Cells were transfected by electroporation using conditions described by Van den Hoff et al. (21) with 50 µg of plasmid DNA/10⁷ cells. Electroporation was performed using an Invitrogen Electroporator II at 1000 microfarads/330 V for PC12 cells or 300 V for COS-1 cells. Electroporated cells were plated in 10-cm dishes in Dulbecco's modified Eagle's medium supplemented with 5% bovine serum, 5% horse serum, and 10% fetal bovine serum, and the cells were grown for 48 h before use in studies.

Plasmids and Site-directed Mutagenesis-- The cDNA encoding CAPS was fused with a triple HA tag at the 3' end of coding sequences and subcloned into a pcDNA3 mammalian expression vector (Invitrogen) to create pCMVCAPS(HA)₃. To fuse CAPS with triple HA at the stop codon in CAPS cDNA, an EcoRI site was introduced in place of a stop codon. A 4146-bp fragment of cDNA containing stop codon was amplified by PCR using GCTGCGCGTAACCACCACA and GGACCCAGAGAATGAATTCGT primers. Full-length CAPS cDNA was reconstituted by the ligation of a PCR product digested with SacI and EcoRI with the upper KpnI-SacI fragment of CAPS cDNA and cloned into KpnI-EcoRI sites of pBluescriptII SK+ (Stratagene). This construct was called pBCAPSEcoRI. DNA encoding the 12CA5 triple HA epitope tag cloned in pBluescriptII SK was obtained from M. D. Rose (22). The triple HA tag was amplified with CGGCGAATTCTTTTACCCATACG and GGACTGTCAAGCGTAATCTGGAA primers. The forward primer was designed to introduce an EcoRI site in frame with the EcoRI site at the end of CAPS coding sequence, and the reverse primer was designed to introduce a stop codon at the end of triple HA tag sequence. PCR amplified HA tag was cloned into pGEM-T vector (Promega) and excised by EcoRI and NotI. HA tag was cloned into EcoRI-NotI sites of pBCAPSEcoRI. Finally the CAPS(HA)₃ coding cassette was excised using KpnI and NotI and cloned into the KpnI-NotI sites of the pcDNA3 vector. To tag CAPS with Myc tag and His₆ tag at the C terminus, the cDNA of CAPS was excised from pCMVCAPS(HA)₃ with XhoI and EcoRI and cloned into XhoI-EcoRI sites of pcDNA3.1/myc-His()B vector (Invitrogen). This construct was called pCMVCAPSmyc6xHis. Recombinant C-terminally tagged CAPS proteins expressed in COS-1 cells and purified to homogeneity were as active in the secretion assay (see below) as native CAPS from rat brain cytosol (data not shown).

Plasmid pGEX4T-PHD-myc6xHis was constructed to provide a convenient way to purify the PH domain from Escherichia coli lysates with GST at the N terminus and His₆ at the C terminus. To construct the plasmid, a CAPS cDNA fragment encompassing nucleotides 1548-1880 of coding sequence was amplified by PCR using the primers CCTCAAAGGATCCCTCGCTGTCCGAATGGATAAG (containing a BamHI site) and GCTTGGGTACCTGCACCTGGGTAGGGGGCACAG (containing a KpnI site). The PCR product was digested with BamHI and KpnI and cloned into the BamHI-KpnI sites of pcDNA3.1/myc-His()B, which provided the PH domain with Myc-His₆ tag at the C terminus. This cassette containing the His-tagged PH domain was excised using BamHI and PmeI and cloned into BamHI-SmaI sites of the pGEX4T1 vector (Amersham Biosciences).

The plasmid pCAPSPH-GFP was constructed to express the CAPS PH domain fused at its C terminus with EGFP in mammalian cells. A CAPS cDNA fragment consisting of nucleotides 1548-1880 of the reading frame was produced by PCR using the primers CCTCGAGAACATGCTCGCTGTCCGAATGGATAAG (containing a XhoI site and a Kozak sequence) and GCTTGGGTACCTGCACCTGGGTAGGGGGCACAG. The PCR product was cloned into the EcoRV site of the pZero2.1 vector. The PH domain-encoding fragment was excised with XhoI and KpnI and cloned into the XhoI-KpnI sites of the pEGFP-N1 vector (CLONTECH).

The plasmid pCAPS-CT177-GFP was constructed to express the 177-residue CAPS C-terminal fragment fused at its C terminus with EGFP in mammalian cells. A CAPS cDNA fragment was produced by PCR using the primers AAAAAAGCTTAAAATGATCACACTCTTGGTGGCAAAGT (containing a HindIII site and a Kozak sequence) and agggcggactgggtgctca with the plasmid pCMVCAPS-GFP as a template. The fragment encoding last 177 residues of CAPS was digested with HindIII and EcoRI and cloned into the HindIII-EcoRI sites of the pCMVCAPS-GFP plasmid in place of excised full-length CAPS.

The plasmid pCMVCAPSC135-TEV6xHis was constructed to express CAPS with a deletion of C-terminal 135 residues fused with His₆ tag with an intervening tobacco mosaic TEV protease cleavage site. Oligonucleotides AATTCGAAAACCTGTATTTTCAGGGCCATCACCATCACCATCACTGAGTTTAAACC and TTAAGGTTTAAACTCAGTGATGGTGATGGTGATGGCCCTGAAAATACAGGTTTTCG were annealed into duplexes that were flanked with EcoRI- and AflII-compatible overhangs and this was ligated into EcoRI-AflII sites of the pcDNA3.1mycHis()B vector to obtain pcDNA3.1TEV6xHis. CAPS cDNA was excised from pCMVCAPSmyc6xHis using XhoI-EcoRI and cloned into XhoI-EcoRI sites of pcDNA3.1TEV6xHis in frame with C-terminal TEV-His₆ tag to obtain pcDNA3.1CAPS-TEV6xHis. A fragment of CAPS cDNA was amplified using the primers TCACAAACACCTGCAGGATCTGTTCGC and CAATAAGAATTCATACTTAGATGCCGCCTTCAC, and the PCR product was digested with SdaI and EcoRI for ligation into SdaI-EcoRI sites of pcDNA3.1CAPS-TEV6xHis in place of an excised fragment to yield the construct pCMVCAPSC1154-TEV6xHis.

To introduce mutations into the full-length CAPS cDNA, a fragment consisting of nucleotides 1170-2458 of the coding sequence was amplified by PCR and the product was cloned into the SmaI site of the pBluescript II KS vector. The construct was used as a template in site-directed mutagenesis performed using the QuikChange protocol (Stratagene). The CAPS fragments were checked by sequencing and then digested with AflII and Eco47III for insertion into the AflII and Eco47III sites of pCMVCAPS(HA)₃ in place of the corresponding wild type fragment. To mutagenize CAPS PH domain, the pGEX4T-PHD-myc6xHis and pCAPSPH-GFP plasmids were used as templates.

Protein Purification-- Purification of the recombinant CAPS PH domain containing two purification tags expressed in E. coli was conducted by a two-step affinity protocol. Expression of proteins in the E. coli strain DH5 was carried out, and cells were collected by centrifugation, resuspended in 50 mM Tris-HCl, pH 7.6, 1 mM phenylmethylsulfonyl fluoride, 2 mM ATP, and 10 mM MgCl₂, and homogenized by sonication followed by addition of 1% Triton X-100. The homogenate was incubated 10 min at room temperature to shift heat shock proteins from the recombinant protein, 10 mM DTT was added, and the clarified homogenate was passed through a column of glutathione-Sepharose 4B (Amersham Biosciences). Following a wash with 100 column volumes of 100 mM Tris-HCl, pH 7.6, 500 mM NaCl, 0.2% -mercaptoethanol, the column was equilibrated with 100 mM Tris-HCl, pH 8.0, and eluted with 20 mM reduced glutathione in 100 mM Tris-HCl, pH 8.0. Protein eluted from glutathione-Sepharose column was applied to a column of Ni-NTA-Sepharose (Qiagen), and the column was washed with 10 column volumes of 20 mM imidazole, pH 7.6, 250 mM NaCl, 10% v/v ethylene glycol and eluted with 150 mM imidazole, pH 7.6, 20% ethylene glycol.

To purify the recombinant CAPS expressed in COS-1 fibroblasts, COS-1 cells were transfected with pCMVCAPSmyc6xHis encoding wild type or mutant CAPS proteins. 48 h after transfection, cells were harvested using phosphate-buffered saline with 5 mM EDTA, washed into 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, and lysed in buffer containing 20 mM Tris-HCl, pH 7.5, 250 mM NaCl, 5 mM imidazole, 1%(v/v) Triton X-100, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride, pH 7.5. Lysates centrifuged at 80,000 × g were applied to a column of Ni-NTA-Sepharose (Qiagen), and the column was washed with 50 column volumes of lysis buffer, with 50 column volumes with buffer of the same composition as lysis buffer excluding detergent, and with 20 column volumes of 10 mM imidazole, pH 7.6, 250 mM NaCl, 10% v/v ethylene glycol, and eluted with 150 mM imidazole, pH 7.6, 250 mM NaCl.

Lipid Binding Assays-- Liposome binding assays with 0.178 mM immobilized recombinant glutathione S-transferase-CAPS PH domain fusion protein were performed essentially as described (23). Liposomes consisted of either PtdChol/PtdSer/phosphoinositide (0.39 mM, 0.183 mM, 0.137 mM), PtdChol/PtdSer/PtdIns(4,5)P₂ (0.275 mM, 0.137 mM, 0.137 mM) or PtdChol/PtdSer (0.366 mM, 0.183 mM). All liposomes also contained [³H]PtdChol (0.5 mCi/reaction). Incubations of liposomes with glutathione-Sepharose beads were terminated by sedimentation for 5 min at 500 × g, and the beads were washed with a 10-fold volume of buffer and extracted with 10% SDS (10-fold volume). The first supernatant (A), wash supernatant (B), and bead SDS extract (C) were analyzed by liquid scintillation counting, and liposome binding was calculated as (dpm in C)/(dpm A + B + C) × 100%. A nonlinear least squares fit to the data (Microcal Origin) was determined using the equation [bound] = B_max × [free]/(K_D + [free]), where [bound] is the concentration of PtdIns(4,5)P₂ in liposomes bound, [free] is the concentration of PtdIns(4,5)P₂ in free liposomes, B_max is the saturation binding, and K_D is the dissociation constant. Lipid binding was also assessed using a lipid-protein overlay assay. Nitrocellulose membranes with 100 pmol of 12 different lipids spotted and dried on them (PIP-Strips) were purchased from Echelon Inc. (Salt Lake City, UT). The assay was conducted using the manufacturer's protocol with minor modifications.

Secretion Assay-- COS-1 cells were transfected with pCMVCAPS(HA)₃ encoding wild type or PH mutant CAPS proteins. 48 h after transfection, cells were harvested using phosphate-buffered saline with 5 mM EDTA and washed three times with KGlu buffer (50 mM HEPES, pH 7.2, 120 mM potassium glutamate, 20 mM potassium acetate, 2 mM EGTA). Cell pellets were resuspended in KGlu buffer containing 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride, and cells were homogenized by multiple passes through a 10-µm clearance ball homogenizer. Cytosols were prepared by centrifugation at 80,000 × g and concentrated to 2-5 mg/ml by ultrafiltration using Centricon 50 (Amicon) devices. The relative content of wild type and mutant CAPS was estimated by Western blot and adjusted with cytosol from nontransfected COS-1 cells to equalize the amount of CAPS per milligram of cytosol protein. Analysis of the activity of COS-1 cytosols and purified CAPS in the secretion assay was performed as described (20). PC12 cells were labeled with 0.5 µCi/ml [³H]norepinephrine (Amersham Biosciences) plus 0.5 mM sodium ascorbate for 16 h at 37 °C, and the Ca²⁺-triggered release of [³H]norepinephrine was determined in 3-min incubations at 30 °C. Results are plotted as percentage of [³H]norepinephrine release by normalization to the total content of [³H]norepinephrine per incubation.

Immunocytochemistry-- To determine the membrane localization of CAPS and CAPS fragments, PC12 cells were transfected with pCMVCAPS(HA)₃, pCAPS-PH-GFP, pCMVCAPS-GFP, or pCMVCAPS-CT177-GFP plasmids, plated on tissue culture dishes, and incubated at 37 °C for 2 h. Cells were resuspended and plated onto poly-L-lysine-treated coverslips for 24 h prior to processing for microscopy. To visualize membrane structures to which CAPS or CAPS PH domains bound, cells were permeabilized with 0.05% saponin or 0.025% digitonin, fixed in 4% formaldehyde, and either further fixed in cold (20 °C) methanol/acetone (2:1 v/v) or extracted with 0.3% Triton X-100. Standard immunocytochemical methods were used with the following antibodies: CAPS polyclonal antibody generated to full-length protein and affinity-purified, GFP polyclonal or monoclonal antibody (Panvera) at 1:1000, TGN-38 monoclonal antibody (clone 2F7, generously provided by Dr. K. Howell) at 1:50, SNAP-25 monoclonal antibody (Sternberger) at 1:1000, mannosidase II polyclonal antibody (generously provided by K. Moremen) at 1:400, and synaptotagmin I luminal domain peptide antibody (Sigma) that we affinity-purified or synaptotagmin I monoclonal antibody (Synaptic Systems GmbH) at1:200. Secondary reagents used were fluorescein isothiocyanate goat anti-rabbit and Texas Red goat anti-mouse antibodies. Fluorescence was imaged with a Bio-Rad MRC 600 confocal imaging system.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The Central Region of CAPS Contains C2 and PH Domains-- Analysis of the full-length CAPS sequence using the SMART algorithm (24) revealed two putative functional domains, a C2 domain and a PH domain (Fig. 1A), in a central region of the protein. The presence of the PH domain was confirmed by Pfam and Blast programs. The CAPS PH domain is highly conserved among rat, Drosophila melanogaster (69.1% similarity), and C. elegans (59.1% similarity) proteins (Fig. 1C), suggesting a possible functional role. PH domains in other proteins mediate interactions with phosphoinositides, with other acidic phospholipids, or with proteins (16-19, 25, 26). Because CAPS functions in membrane fusion and is associated with DCVs and plasma membranes as a peripherally bound protein (14), we determined the properties of the PH domain and whether it is essential for CAPS function in exocytosis.

View larger version (36K): [in this window] [in a new window]   Fig. 1.   CAPS contains central C2 and PH domains. A, schematic depiction of CAPS domains predicted by the SMART comparative sequence analysis program (smart.embl-heidelberg.de/) (24). The C2 domain and PH domain represent residues 397-492 (E-value = 1.08e02) and 520-624 (E-value = 1.78e10), respectively, of the 1289-residue CAPS protein. MHD indicates a Munc13 homology domain. B, model of the three-dimensional structure of the CAPS PH domain. Coordinates for the model and the X-PLOR and MODELLER protocols used in the modeling were obtained on the web site www.nmr.embl-heidelberg.de/~blomberg/PHdomains/. The structure is presented using the Swiss PDB viewer 3.6b3. Residues targeted for mutagenesis are shown in white. C, conservation of mutated residues is shown on alignment of PH domain residues in CAPS orthologues from R. norvegicus, D. melanogaster, and C. elegans. The generated point mutations are marked on top of the R. norvegicus CAPS PH domain sequence. Blue arrows underneath represent predicted -strands, and red bar represents predicted -helix. D and E, a search of structural neighbors and structural alignment using DALI (40) revealed that the PH domains of pleckstrin (N-terminal), insulin receptor substrate, dynamin, spectrin, BTK, PLC1, and GRK-2 are most closely similar to the CAPS PH domain (Z score > 8.5, percentage of identity > 14%). Alignment of the CAPS PH domain with PLC1 (C) and BTK (D) PH domains are shown. Residues important for binding Ins(1,4,5)P3 in the PLC1 PH domain (30) and Ins(1,3,4,5)P4 in the BTK PH domain (29) are indicated in bold.

The CAPS PH Domain Exhibits Nonspecific Acidic Phospholipid Binding and Stereoselective PtdIns(4,5)P₂ Binding-- To assess the phospholipid- and membrane-binding properties of the CAPS PH domain, a purified recombinant GST-CAPS PH protein was immobilized for binding studies with liposomes of defined phospholipid composition. Whereas the CAPS-PH fusion protein did not bind PtdChol liposomes (data not shown), it did bind PtdChol/PtdSer liposomes (Fig. 2A) and PtdChol/PtdIns liposomes (data not shown). Structural modeling studies indicated that the CAPS PH domain, similar to other PH domains (36), exhibits a strongly polarized charge distribution with numerous positively charged residues at the face of the PH domain (Fig. 1B). To determine whether the electrostatically polarized CAPS PH domain interacts with the negatively charged surface of liposomes via nonspecific electrostatic interactions, binding studies with PtdChol/PtdSer liposomes were conducted at various ionic strengths (Fig. 2B). The electrostatic nature of the binding was confirmed by the finding that liposome interactions were significantly reduced at 400 mM KCl.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   The CAPS PH domain exhibits stereoselective interactions with PtdIns(4,5)P2. A, specificity of CAPS PH domain for phospholipid binding. Liposomes formed of PtdChol/PtdSer or PtdChol/PtdSer plus indicated phosphoinositides were incubated with GST-CAPS-PH beads and liposome retention was quantitated. Binding of liposomes to GST beads was used to determine nonspecific binding, which was subtracted from values shown. Error bars represent range of duplicate determinations. B, binding of GST-CAPS PH domain fusion protein to PtdChol/PtdSer vesicles (0.39 mM PtdChol, 0.195 mM PtdSer) in the presence of indicated concentrations of KCl () compared with binding to GST (). Error bars represent range of duplicate determinations. C, binding of GST-CAPS PH domain fusion protein to PtdChol/PtdSer vesicles (0.39 mM PtdChol, 0.137 mM PtdSer) containing indicated concentrations of PtdIns(4,5)P2. The data are fitted to a hyperbolic curve indicating a KD of ~14 µM. Error bars represent the range of duplicate determinations. D, binding of GST-CAPS PH domain fusion protein to phospholipids in an overlay assay. Nitrocellulose strips with indicated phospholipids were incubated with 10 nM GST-CAPS-PH fusion protein, and the bound protein was detected with a Myc antibody. The integrated density of the chemiluminescent signals (in parentheses) was determined using Scion Image software.

Full-length CAPS has been shown to interact with PtdIns(4,5)P₂ (23). To determine whether the CAPS PH domain exhibited phosphoinositide binding, liposomes consisting of PtdChol/PtdSer (2:1 molar ratio) or of PtdChol/PtdSer with phosphoinositides PtdIns, PtdIns(4)P, PtdIns(3,4)P₂, PtdIns(4,5)P₂, or PtdIns(3,4,5)P₃ (3:1:0.6 molar ratio) were incubated with the immobilized GST-CAPS-PH protein. Control liposomes containing relatively higher concentrations of PtdSer were used to approximate the electrostatic charge of phosphoinositide-containing liposomes. CAPS-PH domain binding to liposomes was significantly enhanced by inclusion of PtdIns(4,5)P₂ in the PtdSer-containing liposomes (Fig. 2A). In contrast, binding was not enhanced by the inclusion of PtdIns, PtdIns(4)P, PtdIns(3,4,5)P₃, or PtdIns(3,4)P₂ (Fig. 2A). To further characterize CAPS PH domain binding to PtdIns(4,5)P₂, liposomes comprising PtdChol/PtdSer (3:1 molar ratio) with graded concentrations of PtdIns(4,5)P₂ were used. The CAPS PH domain exhibited an apparent K_D of 14 µM for PtdIns(4,5)P₂ binding (Fig. 2C).

The specificity of phosphoinositide binding to the CAPS PH domain was further examined in a protein overlay assay in which binding to phospholipids immobilized on a nitrocellulose membrane was assessed (Fig. 2D). The synthetic phospholipids used share a common dioctonoyl glycerol composition, which eliminates the contribution of fatty acyl chains to the specificity of binding. The results confirmed that the CAPS PH domain preferentially binds PtdIns(4,5)P₂ (Fig. 2D). Lesser binding to PtdIns(5)P, PtdIns(3,5)P₂, and PtdIns(3)P was also detected in the overlay assay, but binding to PtdIns(3,4,5)P₃ was negligible. Binding of the CAPS PH domain to acidic phospholipids such as PtdSer was not detected, but this can be attributed to the non-bilayer configuration of phospholipids in the overlay assay. Compared with most other PH domains tested in the overlay assay (27), the CAPS PH domain exhibits a strong preference for binding PtdIns(4,5)P₂ over PtdIns(3,4,5)P₃. Overall, these results indicate that, in addition to nonspecific electrostatic interactions with negatively charged phospholipids, the CAPS-PH domain exhibits stereoselective interactions with phosphoinositides with a strong selectivity for PtdIns(4,5)P₂ over PtdIns(3,4,5)P₃.

The CAPS PH Domain Binds to Golgi and Plasma Membranes-- Studies were conducted to determine the cellular membrane-binding properties of the CAPS PH domain. We first determined the cellular membrane-binding properties of the full-length CAPS protein overexpressed in PC12 cells. Native CAPS is a predominantly soluble protein in neural and neuroendocrine cells but is also present as a peripheral membrane protein on DCVs (14). We confirmed that overexpressed CAPS was localized diffusely throughout the cytoplasm as a soluble protein (data not shown). However, when cells were detergent-permeabilized prior to fixation to extract soluble protein, CAPS retained on membranes was localized exclusively to DCVs (Fig. 3, A-C). Similar results were obtained expressing full-length CAPS in C-terminal fusion with GFP (data not shown). In cells expressing CAPS-GFP, localization to DCVs was evident above the diffuse cytoplasmic fluorescence, which indicates that detergent permeabilization does not artificially alter the membrane binding and cellular localization of CAPS.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   CAPS localizes to dense-core vesicles in transfected PC12 cells. CAPS proteins were expressed in PC12 cells by transfection and localized by immunocytochemistry following detergent permeabilization. PC12 cells expressing wild type protein (A-C) or the R558D/K560E/K561E mutant (D-F) are shown. A and D, fluorescein channel showing CAPS localization. B and E, rhodamine channel showing synaptotagmin I localization. C and F, merge of both channels. Both the wild type and mutant CAPS proteins were localized to DCVs indicated by co-localization with synaptotagmin I. Scale bar represents 10 µm.

To assess interactions of the CAPS PH domain with cellular membranes, a C-terminal GFP fusion protein was expressed in PC12 cells. The CAPS PH-GFP fusion protein also distributed as a soluble protein throughout the cytoplasm (data not shown) and exhibited a stable association with membranes following detergent permeabilization. However, unlike full-length CAPS, the CAPS PH-GFP fusion protein associated with the plasma membrane, indicated by co-localization with SNAP-25 (Fig. 4, D-F), and with Golgi membranes, indicated by co-localization with TGN-38 and mannosidase II (Fig. 4, G-I and J-L, respectively). Localization of the PH domain to Golgi was confirmed in studies with 3 µM brefeldin A, which resulted in the disappearance of Golgi-localized CAPS PH-GFP (data not shown). In contrast to the CAPS PH-GFP distribution, expression of the PtdIns(4,5)P₂-specific PH domain of PLC1 as a GFP fusion protein resulted in an exclusive localization to the plasma membrane but not Golgi as reported for other cell types (28). The results indicate that the membrane localization properties of the CAPS-PH domain protein are distinct.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   CAPS PH GFP fusion protein localizes to Golgi and plasma membranes in transfected PC12 cells. Confocal images of PC12 cells expressing CAPS PH domain GFP fusion protein. Cells were detergent-permeabilized prior to fixation. A, D, G, and J, localization of CAPS PH-GFP. B, localization of synaptotagmin I in same cells shown in A. E, localization of SNAP25 in the cells shown in D. H, localization of TGN38 in the cells shown in G. K, localization of mannosidase II in the cells shown in J. C, F, I, and L, merge of corresponding images. Arrows indicate secretory granules (SG), plasma membrane (PM), trans-Golgi network (TGN). Scale bar represents 5 µm.

Unlike the full-length CAPS-GFP (Fig. 3, A-C), the CAPS PH-GFP protein did not localize to DCVs (Fig. 4, A-C). A similar distribution at the plasma membrane and Golgi was observed for expression of the CAPS PH-GFP in COS-1 cells, which suggests that a neuroendocrine cell-specific molecule is not involved in recruiting the CAPS PH domain to Golgi and plasma membranes. The results indicate the CAPS PH domain does not mediate the localization of CAPS to DCVs; however, it may contribute to CAPS interactions with other membranes such as the plasma membrane.

Membrane Binding of the CAPS PH Domain Is Independent of PtdIns(4,5)P₂ Binding-- For studies to determine the role of the PH domain in CAPS function, we generated mutations in the PH domain and characterized the properties of mutant PH domain proteins. Site-directed mutagenesis was guided by a structural model of the CAPS PH domain (Fig. 1B) and structure-based alignments with the related PLC1 (Fig. 1D) and BTK (Fig. 1E) PH domains. We replaced basic residues and tryptophans in the 1/2 and 3/4 loops that are clustered and form a positively charged pocket (Fig. 1B). In the PLC1 and BTK PH domains, cognate residues form the binding site for phosphoinositide headgroups (Fig 1, D and E). The residues selected for mutagenesis are conserved in Rattus norvegicus, D. melanogaster, and C. elegans CAPS orthologs (Fig. 1C).

Within the 1/2 loop, conserved CAPS Lys-531 was replaced by Glu, Trp-534 by Ser, and Lys-535 by Glu. A triple replacement mutant W537S/K538Q/R540E was generated in the 2 strand in a motif conserved among PH domains including PLC1and BTK, where cognate residues participate directly in phosphoinositide binding (29). A K560E mutation was introduced into the 3/4 loop of the CAPS PH domain, which corresponds to Lys-57 of the PLC1 PH domain that is important for inositol trisphosphate binding (29, 30). Finally, a cluster of three basic residues in the 3/4 loop that contributes to a putative phosphoinositide binding pocket (Fig. 1B) was replaced with acidic residues (R558D/K560E/K561E) to generate a mutant in which the polarity of the membrane-binding face of the PH domain was reversed.

The K531E mutation did not significantly affect the binding of a PH domain fusion protein to PtdSer- or PtdIns(4,5)P₂-containing liposomes (Fig. 5). In contrast, the other mutant proteins exhibited significantly altered liposome-binding properties. The K560E mutation selectively affected interactions with PtdIns(4,5)P₂-containing liposomes consistent with similar effects of mutation in the PLC1 PH domain (30). The triple mutant W537S/K538Q/R540E was deficient in binding to both PtdSer- and PtdIns(4,5)P₂-containing liposomes. The R558D/K560E/K561E triple mutant was similar with greater impairment of binding to PtdIns(4,5)P₂-containing liposomes. The decreased binding of mutant PH domains to PtdIns(4,5)P₂ was also evident in the protein overlay assay (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effect of mutation on the lipid-binding properties of the CAPS PH domain. A, Coomassie-stained gel of purified fusion proteins expressed in E. coli. 4 µg of protein was loaded on each lane. B, effect of mutations on the lipid-binding properties of the CAPS PH domain. Liposomes consisting of PtdChol/PtdSer (2:1) or PtdChol/PtdSer/phosphoinositide (3:1:0.6) and [3H]PtdChol were incubated with GST-CAPS-PH proteins immobilized on glutathione-agarose beads, and the bound fraction of lipids was measured. Binding of liposomes to GST immobilized on glutathione-agarose beads was used to determine nonspecific binding (indicated by line). Error bars represent the range of duplicate determinations.

To assess the effect of PH domain mutations on interactions with cellular membranes, mutant CAPS-PH-GFP fusion proteins were expressed in PC12 cells and imaged after cell permeabilization (Fig. 6). Although the expression level of most of the mutants was similar to that of the wild type PH domain, the W537S/K538Q/R540E mutant was poorly expressed, possibly the result of reduced stability caused by three replacements in the 2-strand that affects structure. The K531E mutant, which exhibited unaltered liposome-binding properties, localized to Golgi and plasma membranes similarly to the wild type protein (Fig. 6). The K560E mutant, which exhibited a selective loss in binding to PtdIns(4,5)P₂-containing liposomes, also exhibited localization properties indistinguishable from the wild type protein. In contrast, both triple mutants, W537S/K538Q/R540E and R558D/K560E/K561E, which exhibited strong reductions in binding to PtdSer-containing liposomes, failed to associate with cellular membranes and were quantitatively extracted by detergent permeabilization (Fig. 6). Although the reduced expression of the W537S/K538Q/R540E mutant precluded definitive interpretation, the R558D/K560E/K561E mutant exhibited a clear loss of cellular membrane binding. The results indicate that the cellular membrane localization of the PH domain is not mediated by PtdIns(4,5)P₂ binding because the localization of the K560E mutant was similar to wild type. Cellular membrane binding is likely mediated through general electrostatic interactions with the PtdSer-rich cytoplasmic leaflet of cellular membranes. This is indicated by the strict correlation observed in the mutant PH domains for cellular membrane localization and competence for in vitro binding to PtdSer-containing liposomes. Interactions of the PH domain with other membrane determinants such as proteins cannot be eliminated as a possible basis for cell membrane localization.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effect of mutation on the cellular membrane association of the CAPS PH domain. Transfection studies similar to those described in the legend to Fig. 4 were conducted to express wild type and mutant CAPS PH-GFP in PC12 cells. Left upper panel shows Western blot analysis of expressed proteins detected with a peptide antibody generated against CAPS (574-588). Triton X-100 lysates of cells transfected with each of constructs (10 µg of protein/lane) were loaded on SDS-PAGE gel and analyzed. The upper band is the PH domain GFP fusion protein and the lower band is a product of its degradation. The level of expression of all mutants except W537S/K538Q/R540E was similar to that of wild type. Cells expressing the indicated proteins were detergent-permeabilized and fixed for confocal microscopy to image GFP fusion proteins (green) or SNAP25 (red). Scale bar represents 10 µm.

A Functional CAPS PH Domain Is Essential for Regulated Exocytosis-- With the above characterization of mutant PH domains as essential background, we determined whether PH domain-mediated interactions were essential for the activity of CAPS in regulated exocytosis by producing full-length CAPS proteins containing characterized mutations in the PH domain. Cytosols from nontransfected COS-1 cells fail to reconstitute the Ca²⁺-dependent triggering step of regulated DCV exocytosis in permeable PC12 cells (Fig. 7A) because COS-1 cells do not express neural/endocrine-specific CAPS (7, 8). The expression of the wild type CAPS protein in COS-1 cells confers activity on the cytosol in the reconstitution assay (Fig. 7A). Thus, we were able to test CAPS proteins containing mutant PH domains by expressing them in COS-1 cells and testing cytosols for the reconstitution of Ca²⁺-triggered secretion in permeable PC12 cells.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Reconstitution of exocytosis in permeable PC12 with cytosols from COS cells expressing wild type and mutant forms of CAPS. A, CAPS expression in COS cells confers reconstituting activity on cytosol. PC12 cells loaded with [3H]norepinephrine were permeabilized and incubated under MgATP-dependent priming conditions as described under "Materials and Methods." 5-min incubations for Ca2+-triggered norepinephrine release were conducted with indicated amounts of cytosol derived from rat brain () or COS-1 cells expressing GFP () or COS-1 cells expressing CAPS(HA)3 (). Inset shows Western blot analysis of CAPS in rat brain cytosol (RBC), cytosol from GFP-expressing COS-1 cells, and cytosol from CAPS-expressing COS-1 cells. Arrows indicate CAPS bands. B, mutations in the CAPS PH domain differentially affect the activity of CAPS in regulated exocytosis. COS-1 cells were transfected with constructs encoding wild type and mutant forms of CAPS(HA)3. Cytosols prepared from cells expressing CAPS were adjusted with cytosol from nontransfected COS-1 to equalize the amount of CAPS per milligram of cytosol protein. The Ca2+-triggered release of norepinephrine from permeable PC12 cells was tested with 0.4 mg/ml cytosol for each of the indicated CAPS proteins. Maximal cytosol-dependent activity from wild-type CAPS-expressing COS-1 cells was set at 100%. C, comparison of the reconstituting activity of wild type CAPS () and R558D/K560E/K561E mutant CAPS (). Proteins were expressed as C-terminal fusions with His6 tag and purified on Ni-NTA-agarose.

Several CAPS mutants exhibited wild type activity in reconstituting Ca²⁺-dependent secretion (Fig. 7B). These included the K531E mutation, which did not affect PH domain binding to liposomes containing PtdSer or cellular membrane binding. In addition, the K560E mutation, which specifically inhibited PH domain binding to PtdIns(4,5)P₂-containing but not PtdSer-containing liposomes and did not alter cellular localization, was fully active in the secretion assay. Two other mutants (W534S and K535E) that exhibited reduced PtdIns(4,5)P₂ binding in the overlay assay (data not shown) were also fully active in the secretion assay.

In contrast, several CAPS mutants exhibited strongly impaired activity in the secretion assay (Fig. 7B). These included the triple mutant in a conserved motif for PH domains (W537S/K538Q/R540E), which exhibited decreased PtdSer binding and failed to localize to cellular membranes. Although this protein was expressed well in COS-1 cells, the possibility that its loss-of-function in exocytosis results from PH domain misfolding cannot be entirely excluded. However, the triple mutant in the 3/4 loop, R558D/K560E/K561E, which is unlikely to be improperly folded (see below) and which exhibited decreased liposome and cellular membrane binding, also exhibited strongly impaired activity in the secretion assay. The results with these mutants establish that the CAPS PH domain is essential for the activity of the protein in regulated exocytosis. There was a striking correlation between CAPS activity in regulated exocytosis and the ability of the PH domain to associate with PtdSer-rich liposomes and cellular membranes. This suggests that the essential role of the PH domain in CAPS is to mediate membrane associations that are required at sites of Ca²⁺-triggered exocytosis (see "Discussion").

Because the basic mechanism by which CAPS enhances Ca²⁺-dependent exocytosis is unknown, we utilized the R558D/K560E/K561E triple mutant protein in studies to determine whether a functional membrane-binding PH domain was critical for the intrinsic activity of CAPS in secretion or whether it served to facilitate membrane interactions. Mutant and wild type CAPS proteins were purified and tested in the assay for Ca²⁺-dependent DCV exocytosis. The results (Fig. 7C) indicated that the inactivating mutation in the CAPS PH domain did not affect the intrinsic activity of CAPS because the R558D/K560E/K561E mutant stimulated Ca²⁺-dependent exocytosis to a similar extent as the wild type CAPS. However, the activity of the mutant protein was only observed at 50 times greater protein concentration than required for the wild type protein. These results are consistent with a role for the PH domain in facilitating membrane interactions.

We also determined whether full-length CAPS harboring inactivating mutations in the PH domain was properly localized to DCVs. We found that overexpression of the R558D/K560E/K561E mutant in PC12 cells resulted in a localization of the protein to DCVs that was indistinguishable from that of the wild type protein (Fig. 3, D-F). These results were consistent with the observation that the CAPS PH domain did not localize to DCVs (Fig. 4, A-C) and indicate that CAPS association with DCVs is not mediated by its PH domain.

CAPS Possesses PtdIns(4,5)P₂-binding Domains Other than the PH Domain-- In the preceding studies, CAPS proteins containing mutations that inactivate PtdIns(4,5)P₂ binding by the PH domain were found to be fully functional in regulated exocytosis. These results might indicate that PtdIns(4,5)P₂ binding is not essential for CAPS activity in regulated exocytosis. However, previous studies indicated that CAPS may have multiple PtdIns(4,5)P₂-binding sites (23). Thus, we determined whether inactivating mutations in the PH domain altered PtdIns(4,5)P₂ binding by full-length CAPS. We compared the lipid-binding activities of purified full-length wild type CAPS and CAPS containing the PH domain-inactivating mutations R558D/K560E/K561E. These mutations inhibit binding of the PH domain to PtdSer- as well as PtdIns(4,5)P₂-containing liposomes (see Fig. 5). We found that full-length CAPS containing the R558D/K560E/K561E mutations did not exhibit reduced PtdIns(4,5)P₂ binding, nor was the binding specificity for PtdIns(4,5)P₂ altered (Fig. 8). The results indicate that PtdIns(4,5)P₂-binding by full-length CAPS is principally mediated by domains other than the PH domain.

View larger version (44K): [in this window] [in a new window]   Fig. 8.   The CAPS R558D/K560E/K561E mutant binds PtdIns(4,5)P2. Wild type full-length CAPS and the R558D/K560E/K561E mutant fused at the C terminus with Myc-His6 tag were expressed as recombinant proteins and purified. The phospholipid-binding properties of the recombinant proteins were determined in the overlay assay. Nitrocellulose strips containing the indicated phospholipids were incubated with 10 nM protein, and protein binding was detected with Myc antibody.

A CAPS C-terminal Domain Mediates Dense-core Vesicle Associations-- The previous studies indicated that the PH domain did not mediate CAPS associations with DCVs. Consistent with the conclusion that a distinct domain is responsible for DCV targeting, we identified a C-terminal domain that is sufficient for DCV association. C-terminal fragments of CAPS fused to GFP localize exclusively to DCVs when expressed in PC12 cells. The shortest domain identified to date with this property comprises 177 C-terminal amino acid residues (Fig. 9, A-C). The C-terminal-GFP fusion protein did not associate with either plasma membrane or Golgi, indicating that the cellular membrane-binding properties of the C-terminal domain are distinct from those of the PH domain.

View larger version (24K): [in this window] [in a new window]   Fig. 9.   CAPS C-terminal fragment is sufficient to localize to DCVs and is essential for CAPS activity. Upper panel, confocal images of PC12 cells expressing a CAPS-(1113-1289) C-terminal GFP fusion protein. Cells were fixed in 4% formaldehyde, extracted with 0.3% Triton X-100, and immunostained with synaptotagmin I antibodies. A, localization of C-terminal GFP fusion protein (green). B, localization of synaptotagmin I (red) in same cells shown in A. C, merge of A and B. The CAPS C-terminal GFP fusion protein co-localizes with synaptotagmin I on DCVs. Scale bar represents 5 µm. Lower panel (left), comparison of the reconstituting activity of cytosols of COS-1 cells expressing wild type CAPS or a mutant truncated at residue 1154. Cytosols were tested at 0.8 mg/ml and maximal activity from wild type CAPS was set at 100%. Error bars represent the range of triplicate determinations. Lower panel (right), Western blots for CAPS in cytosols (10 µg/lane) from COS-1 cells expressing wild type CAPS or CAPSC135.

Because a portion of cytosolic CAPS is peripherally bound to DCVs (14) and the protein functions in DCV exocytosis (13), we determined whether the C-terminal domain is required for CAPS activity in regulated exocytosis. A truncation mutant lacking 135 C-terminal amino acids was found to lack activity in Ca²⁺-dependent exocytosis when tested over a concentration range where wild type CAPS is active (Fig. 9D). These results indicate that the C-terminal domain, which is sufficient for DCV targeting, is also necessary for CAPS function in exocytosis. However, as was the case for PH domain mutants (i.e. Fig. 7C), C-terminal truncation mutants were able to activate Ca²⁺-dependent exocytosis when tested at concentrations 50-100-fold greater than maximally effective wild type CAPS concentrations (data not shown). This indicates that deletion of the C-terminal domain does not eliminate the core activity of CAPS in exocytosis. Compensation by higher concentrations of the mutant protein is consistent with the specific role of the C-terminal domain in mediating DCV membrane association. Determining the mechanism by which the C-terminal domain mediates DCV associations should provide insight into the selective activity of CAPS in DCV exocytosis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CAPS is a cytosolic protein that acts at late step in Ca²⁺-dependent DCV exocytosis close to or at the point of membrane fusion. It is a novel protein, possibly distantly related to Munc13 proteins, and its molecular mechanism of action in regulated exocytosis is unclear. Active in membrane fusion, it is anticipated that CAPS interacts with target membranes and previous work has characterized the membrane-binding activity of the protein (14, 23). The results of the current study indicate that CAPS contains two distinct membrane association domains, the PH and C-terminal domains, respectively, that mediate CAPS binding to plasma membranes and DCV membranes. Because both membrane association domains were essential for CAPS function in Ca²⁺-dependent exocytosis, these findings suggest that CAPS may bridge donor and acceptor membranes at a critical point during DCV exocytosis.

PH domains have been detected in over 100 protein sequences (31). Despite low overall sequence conservation, the structures of characterized PH domains are extremely well conserved (32). Almost all of the PH domains that have been studied to date bind phosphoinositides or inositol phosphates, although the role, if any, of phosphoinositide binding in protein function is unknown for the majority of these proteins (19, 27, 33). The results of the present study indicate that the CAPS PH domain has unique properties that differ from previously characterized PH domains in exhibiting relatively high specificity in binding PtdIns(4,5)P₂ but with only moderate affinity (K_D ~ 14 µM). This contrasts with the higher affinity (K_D ~ 1 µM) of the PLC1 PH domain, which exhibits strongly stereoselective interactions with 4- and 5-phosphorylated inositides (34, 35). Other PH domains such as those found in pleckstrin, spectrin, and -adrenergic receptor kinase exhibit affinities for PtdIns(4,5)P₂ similar to or lower than that of the CAPS PH domain, but these show little stereoselectivity in phosphoinositide binding (31, 36). The CAPS PH domain binds PtdIns(4,5)P₂ with specificity and discriminates strongly against PtdIns(3,4,5)P₃. Although these features are conserved among CAPS proteins throughout evolution, they do not appear to be the key features of the PH domain that are critical for the function of CAPS in regulated DCV exocytosis.

We found that a series of mutations in the CAPS PH domain that reduce PtdIns(4,5)P₂ binding (K531E, W534S, K535E, K560E; see Fig. 5B) failed to adversely affect either the cellular localization of the PH domain or CAPS activity in regulated exocytosis. These results indicate that properties unrelated to PtdIns(4,5)P₂-binding mediate the activity of the CAPS PH domain that is required for membrane association or regulated exocytosis. Similar results were reported for the Sos1 PH domain, where a mutation similar to the CAPS K560E mutation that disrupts PtdIns(4,5)P₂ binding did not affect the localization of Sos1 in cells or its function (37). For many other PH domains that exhibit lower affinity or less highly selective interactions with PtdIns(4,5)P₂ than does the CAPS PH domain, the functional role of phosphoinositide binding is unclear (19, 33). For CAPS, the stereoselective phosphoinositide binding by the PH domain does not appear to contribute to the function of the protein in membrane fusion.

PtdIns(4,5)P₂ synthesis is required for Ca²⁺-triggered DCV exocytosis (5). Because CAPS functions at a stage following PtdIns(4,5)P₂ synthesis (1) and binds PtdIns(4,5)P₂ selectively (23), it was proposed that CAPS may function as an effector for PtdIns(4,5)P₂ in regulated exocytosis. The current studies indicate that PtdIns(4,5)P₂ binding by the CAPS PH domain is not essential for CAPS activity in regulated exocytosis, but they do not eliminate a role for PtdIns(4,5)P₂-binding in CAPS function because PtdIns(4,5)P₂ binding by full-length CAPS was not altered by mutations that inactivate the PH domain (Fig. 7). This is compatible with recent studies indicating that there are other PtdIns(4,5)P₂-binding domains in CAPS including the C2 domain adjacent to the PH domain (Fig. 1A).³ Additional studies will be required to evaluate the functional significance of PtdIns(4,5)P₂ binding by CAPS mediated by regions of the protein other than the PH domain.

Of seemingly greater significance for CAPS function, the CAPS PH domain also exhibited non-stereoselective, charge-based binding to acidic phospholipids. The polarized charge distribution within PH domains (36) may mediate long range electrostatic interactions that orient the positively charged face of the PH domain relative to the negatively charged membrane surface. Electrostatic effects have been shown to be a major determinant in membrane binding for the pleckstrin N-terminal and dynamin PH domains (38, 39). Interactions of PH domains with PtdSer have also been reported for spectrin, RasGAP, and diacylglycerol kinase (19).

Our results with the CAPS PH domain mutations are consistent with a prevalent role for long range electrostatic interactions in mediating membrane associations of the PH domain. Thus, the R558D/K560E/K561E mutant, where significant reduction in PH domain polarization was achieved by charge inversion mutations, exhibited strongly reduced interactions with PtdSer-containing liposomes and an inability to stably associate with cellular membranes. Based on the full-length CAPS mutants analyzed, mutations in this class uniquely resulted in a loss-of-function for CAPS in regulated exocytosis. CAPS containing these mutations retained intrinsic activity in regulated exocytosis but exhibited a 50-fold reduction in potency (Fig. 7C), which is consistent with the properties of a protein deficient in membrane association. The localization of the CAPS-PH domain GFP fusion protein to the plasma membrane (as well as Golgi) suggests that the PH domain could mediate CAPS interactions with the plasma membrane during exocytosis.

The CAPS PH domain is not involved in targeting CAPS to the membranes of DCVs. This was evident from the indistinguishable localization of wild type and R558D/K560E/K561E mutant CAPS to DCVs and from the finding that a CAPS-PH domain GFP fusion protein did not localize to DCVs. This indicates that a distinct membrane association domain is involved in targeting CAPS to DCVs, and the current work identified a C-terminal domain in CAPS that was sufficient for DCV targeting. Although the C-terminal domain was important for CAPS activity in exocytosis, deficiencies in the specific activity of the C-terminal truncation mutant were compensated by increasing protein concentrations, which is consistent with a role for the C terminus in membrane association. The distal C-terminal region of CAPS contains an acidic cluster consisting of MKDSDEEDEEDD, which resembles recently identified sorting motifs in several proteins. Acidic cluster sorting motifs are present in the cytosolic domains of several proteins with different membrane trafficking itineraries (41), including proprotein convertases furin and PC6B, mannose 6-phosphate receptors, herpes virus envelope glycoproteins, and human immunodeficiency virus type 1 Nef (42-47). Such clusters often contain consensus sites for phosphorylation by casein kinase II, as does that for CAPS, and phosphorylation by casein kinase II regulates sorting. Acidic cluster motifs have also been shown to be important for sorting transmembrane proteins such as the vesicular monoamine transporter (48) and peptidylglycine-amidating monooxygenase (49) to DCVs. Although CAPS is a cytosolic protein that associates with DCVs, it may exploit the same mechanism employed by transmembrane proteins for recruitment to DCVs, a possibility that will be addressed in future studies.

In summary, the current studies identify two membrane association domains in CAPS with distinct membrane-binding specificities that are each essential for optimal CAPS activity in Ca²⁺-dependent DCV exocytosis. The essential role of the PH domain is mediated by interactions of its electrostatically polarized face with the negatively charged cytoplasmic leaflet of the plasma membrane. A C-terminal domain mediates stable associations of CAPS with DCVs, although the binding determinants on DCVs remain to be identified. Tethered to the DCV by the C-terminal domain, CAPS could associate with the plasma membrane via its PH domain at the point of contact where DCVs dock with the plasma membrane. Interactions with both DCV and plasma membrane may play an important role in the mechanism by which CAPS facilitates the fusion of DCVs with the plasma membrane.

	FOOTNOTES

^* This work was supported by a grant from the United States Public Health Service (to T. F. J. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Biochemistry, University of Wisconsin, Madison, WI 53706. Tel.: 608-263-2427; Fax: 608-262-3453; E-mail: tfmartin@facstaff.wisc.edu.

Published, JBC Papers in Press, April 1, 2002, DOI 10.1074/jbc.M201614200

² V. A. Klenchin, unpublished data.

³ K. M. Loyet and R. N. Grishanin, unpublished data.

	ABBREVIATIONS

The abbreviations used are: DCV, dense-core vesicle; CAPS, Ca²⁺-dependent activator protein for secretion; PtdIns, phosphatidylinositol; PtdChol, phosphatidylcholine; PtdSer, phosphatidylserine; PH, pleckstrin homology; TEV, tobacco etch virus; BTK, Bruton's tyrosine kinase; Ni-NTA, nickel-nitrilotriacetic acid; PLC, phospholipase C; HA, hemagglutinin; GST, glutathione S-transferase; GFP, green fluorescent protein; EGFP, enhanced green fluorescent protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES