Syndapin 3 modulates fusion pore expansion in mouse neuroendocrine chromaffin cells

Abstract

Adrenal neuroendocrine chromaffin cells receive excitatory synaptic input from the sympathetic nervous system and secrete hormones into the peripheral circulation. Under basal sympathetic tone, modest amounts of freely soluble catecholamine are selectively released through a restricted fusion pore formed between the secretory granule and the plasma membrane. Upon activation of the sympathoadrenal stress reflex, elevated stimulation drives fusion pore expansion, resulting in increased catecholamine secretion and facilitating release of copackaged peptide hormones. Thus regulated expansion of the secretory fusion pore is a control point for differential hormone release of the sympathoadrenal stress response. Previous work has shown that syndapin 1 deletion alters transmitter release and that the dynamin 1-syndapin 1 interaction is necessary for coupled endocytosis in neurons. Dynamin has also been shown to be involved in regulation of fusion pore expansion in neuroendocrine chromaffin cells through an activity-dependent association with syndapin. However, it is not known which syndapin isoform(s) contributes to pore dynamics in neuroendocrine cells. Nor is it known at what stage of the secretion process dynamin and syndapin associate to modulate pore expansion. Here we investigate the expression and localization of syndapin isoforms and determine which are involved in mediating fusion pore expansion. We show that all syndapin isoforms are expressed in the adrenal medulla. Mutation of the SH3 dynamin-binding domain of all syndapin isoforms shows that fusion pore expansion and catecholamine release are limited specifically by mutation of syndapin 3. The mutation also disrupts targeting of syndapin 3 to the cell periphery. Syndapin 3 exists in a persistent colocalized state with dynamin 1.

neuroendocrine adrenal chromaffin cells release catecholamine and peptide hormones in response to varied sympathetic excitation. Under basal sympathetic tone, modest amounts of catecholamine are released into the peripheral circulation to maintain homeostatic cardiovascular, as well as metabolic, status of the “rest-and-digest” sympathetic state. Elevated sympathetic activation (e.g., under acute stress) maximizes catecholamine release and facilitates peptide hormone secretion to meet the increased demand for peripheral blood circulation and elevated glucose of the “fight-or-flight” stress response. The fundamental mechanism of this differential hormone release is a shift in the chromaffin granule exocytic mode (27). Briefly, sympathetic input under basal tone activates modest Ca²⁺ influx to evoke fusion of hormone-containing secretory granules with the plasma membrane. Under these conditions, the fusion pore between the granule lumen and the extracellular space fails to expand and remains in a low-conductance state, resulting in a fractional release of highly soluble granule catecholamines (29). This fusion mode is followed, several hundred milliseconds later, by endocytosis (12, 40). Under increased sympathetic activation, driven by elevated cytosolic Ca²⁺ concentration, the fusion pore is actively driven to expand (47). Under these conditions, the granule remains fused with the cell surface for up to tens of seconds (14, 23, 59). This slower kinetic and expanded fusion pore diameter increases catecholamine quantal size and facilitates release of copackaged peptide hormones from the granule core (27, 50). Despite identification of several of the proteins involved, the molecular mechanism of fusion pore expansion and its regulation remains unclear.

Dynamin is a well-studied GTPase responsible for scission of nascent endocytic vesicles in a variety of secretory systems (25). Recently, several studies have indicated a role for dynamin in the exocytic process as well. Deletion of dynamins 1 and 3 results in decreased transmitter release in cultured mouse cortical neurons (43). In neuroendocrine cells, dynamin has been shown to play a direct role in formation and regulation of the fusion pore (3, 6, 26, 32, 33, 62). The syndapins (synaptic dynamin-associated protein, also known as PACSINs) represent a family of dynamin-binding partners (37). Syndapin is an Fes-CIP4 homology Bin-amphiphysin-Rvs (F-BAR) and Src-homology 3 (SH3) domain-containing protein. F-BAR domains express membrane deformation (bending) function and SH3 domains act as a binding partner with proline-rich domain (PRD)-containing proteins, including dynamins. Three isoforms of syndapin are expressed in mammalian cells: syndapin 1 is specifically expressed in neuronal cells (51, 53), and syndapin 2 is ubiquitously expressed (55) and syndapin 3 is mainly expressed in skeletal muscles, heart, and lung (45, 60). The syndapin family participates in several cellular processes involving membrane and cytoskeletal remodeling (19, 20, 37, 52, 54). Syndapin 1 has been extensively studied and is crucial for synaptic vesicle endocytosis (4, 16, 38, 41, 52). Syndapin 2 is involved in vesicle transport from the trans-Golgi network, caveolae biogenesis, and trafficking (34, 35, 58). Syndapin 3, the least characterized isoform (37), has been demonstrated to play a role in glucose transport (56) and transient receptor potential vanilloid 4 channel trafficking (17, 18). We recently showed that interaction between dynamin 1 and an at-the-time unidentified member of the syndapin family contributes to regulated expansion of the fusion pore and facilitation of catecholamine release from neuroendocrine chromaffin cells (57). However, there is no available report regarding the expression profile of syndapin isoforms and their specificity of function in neuroendocrine chromaffin cells. Nor is it known whether dynamin and syndapin must be brought together to modulate pore dynamics or whether they exist in a preformed complex.

Here we show message and expression of all three syndapin isoforms in the mouse adrenal medulla, with syndapin 3 message the greatest and syndapin 1 the least.¹ Our functional analysis shows that expression of a mutant syndapin 3 that lacks affinity for PRD-containing proteins decreases activity-dependent fusion pore expansion, while equivalent mutation of syndapins 1 and 2 is without effect. Of the isoforms associated with granule trafficking, syndapins 1 and 3 exhibit differential subcellular distribution in isolated chromaffin cells: syndapin 1 is distributed mainly in the cytoplasm, whereas syndapin 3 localizes to the cell periphery. We found that syndapin 3, but not syndapin 1, is consistently colocalized with dynamin 1 in mouse chromaffin cells.

MATERIALS AND METHODS

All chemicals and reagents were obtained from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA) unless otherwise noted.

Isolated chromaffin cell preparation.

C57BL/6 mice, >6 wk old, were obtained from the Jackson Laboratory (Bar Harbor, ME) and used for culture of primary neuroendocrine adrenal chromaffin cells. Mice were deeply anesthetized by isoflurane (USP, Halocarbon Products, River Ridge, NJ) inhalation and then killed by decapitation. Anesthesia and euthanasia protocols were approved by the Institutional Animal Care and Use Committee of Case Western Reserve University, an accredited oversight body (Federal Animal Welfare Assurance No. A3145-01). Adrenal glands were quickly removed and stored in a cooled dissociation solution (80 mM Na-glutamate, 55 mM NaCl, 6 mM KCl, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose, pH 7.0, osmolarity 300 mosM). Excess fat and adrenal cortex were trimmed from the medulla. Chromaffin cells were further isolated by proteinase and collagenase incubation, as previously described (27). Isolated cells were plated on 25-mm-diameter cover glasses and cultured in DMEM supplemented with insulin-transferrin-selenium (ITS), a defined serum substitute (Mediatech, Manassas, VA), and penicillin-streptomycin. The cells were incubated at 35°C in 10% CO₂ for 24–48 h before initiation of experiments.

Real-time quantitative PCR.

Adrenal medullae were dissected from adult (>6 wk old) mice, as described above. Then the cortex was carefully dissected, and the medullary RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA). Reverse transcription and real-time quantitative PCR detection were performed using the TaqMan RNA-to-C_T one-step kit on the StepOnePlus real-time PCR system (Applied Biosystems, Carlsbad, CA). Real-time primers, including syndapin 1 (Mm00478702_m1), syndapin 2 (Mm00449303_m1), syndapin 3 (Mm00453174_m1), and GAPDH (Mm99999915_g1), were purchased from Applied Biosystems. GAPDH was used as the reference message.

Tissue lysate preparation.

Brain, liver, and skeletal muscle were dissected from mice, and adrenal medullae were dissected as described above. Tissues were homogenized using a glass-Teflon homogenizer in lysis buffer [25 mM Tris-Cl, pH 7.6, 150 mM NaCl, 1 mM EDTA, and 2% Triton X-100 supplemented with a cocktail of protease inhibitors (Roche Applied Biosciences, Indianapolis, IN)]. The homogenates were centrifuged at 900 g for 10 min. Supernatants were maintained under constant agitation on an orbital shaker for 2 h and centrifuged at 14,000 g for 20 min and again at 16,000 g for 30 min. All procedures were performed on ice or at 4°C. Protein concentration of tissue lysates was determined by a bicinchoninic acid assay kit (Thermo Scientific, Pittsburgh, PA).

Western blot analysis.

Tissue lysates or purified GST fusion proteins (syndapins 1, 2, and 3) were resolved by SDS-PAGE on 10% Mini-PROTEAN TGX polyacrylamide gels (Bio-Rad, Hercules, CA; 50 μg per lane for tissue lysates and 0.3 μg per lane for purified proteins), transferred onto nitrocellulose membranes, and immunoblotted using the following primary antibodies (Santa Cruz Biotechnology, Dallas, TX): rabbit anti-PACSIN1 for syndapin 1 (M-46, 1:200 dilution), mouse anti-PACSIN2 for syndapin 2 (F-12, 1:100 dilution), and rabbit or goat anti-PACSIN3 for syndapin 3 (H-100 or K-16, 1:200 dilution). Secondary antibodies were horseradish peroxidase-conjugated anti-goat (1:2,500 dilution; Thermo Scientific), anti-mouse (1:5,000 dilution; Thermo Scientific), and anti-rabbit (1:2,500 dilution; Cell Signaling Technology, Danvers, MA). Western blots were developed using SuperSignal West Pico chemiluminescent substrate (Thermo Scientific).

Immunofluorescence labeling.

For immunohistological labeling, mice were deeply anesthetized by isoflurane (USP, Halocarbon Products) inhalation and fixed with 3.7% paraformaldehyde in PBS by transcardiac perfusion. Adrenal glands were removed and postfixed in the same fixative containing 30% sucrose overnight at 4°C, embedded in optimum cutting temperature compound, cut into 16-μm sections on a cryostat, and mounted on slides. For immunofluorescence labeling, sections containing medulla were washed with PBS and permeabilized with PBS containing 0.15% Triton X-100 for 30 min. Nonspecific background staining was blocked with 5% donkey, rabbit, or goat serum, to match the secondary antibody host species, for 30 min. Sections were immunolabeled with primary antibodies: mouse anti-dynamin Hudy 1 monoclonal IgG (1:200 dilution; Millipore) and rabbit anti-PACSIN1 (M-46), goat anti-PACSIN2 (M-19), or goat anti-PACSIN3 (K-16) (1:50 dilution; Santa Cruz Biotechnology). For visualization of dynamin 1 and syndapins, sections were incubated in species-matched secondary antibodies tagged with Alexa Fluor 488 and Alexa Fluor 594, respectively (Molecular Probes). Cells were washed multiple times with PBS between each antibody staining to completely remove excess unbound antibodies. Sections were then mounted with an aqueous mounting medium (Dako, Carpinteria, CA). For isolation of chromaffin cells, cultured chromaffin cells were washed with a Ringer solution (150 mM NaCl, 10 mM HEPES-H, 10 mM glucose, 2.8 mM CaCl₂, 2.8 mM KCl, and 2 mM MgCl_2, pH 7.2, osmolarity 320 mosM), fixed in PBS containing 4% paraformaldehyde for 30 min, and subjected to the labeling protocol described above. Cells that were stimulated in “high-K⁺” solutions were bathed in a Ringer solution of the following composition: 123 mM NaCl, 10 mM HEPES-H, 10 mM glucose, 2.8 mM CaCl₂, 30 mM KCl, and 2 mM MgCl₂ (pH 7.2, osmolarity 320 mosM). For adrenal cryosections, primary and secondary antibody incubation times were 2 h and 1 h, respectively, at room temperature. For isolated chromaffin cells, incubation time was 1 h at room temperature for primary and secondary antibodies.

Fluorescence imaging.

Fluorescence images were captured by a cooled charge-coupled device camera (Retiga EXi, QImaging, Surrey, BC, Canada) mounted on an Olympus IX-81 inverted microscope. A ×40 water-immersion objective (1.15 numerical aperture) was used to visualize and identify fluorescent-transfected isolated cells and to acquire images of adrenal gland sections. A ×100 oil-immersion objective (1.3 numerical aperture) was used to acquire high-resolution images for quantitative colocalization analysis. In the latter application, exposure time was optimized to yield the best signal for each image stack. Camera gain was set consistent throughout a data set. Excitation light was provided by a Polychrometer IV (T.I.L.L. Photonics, Pleasanton, CA) under control of SlideBook image acquisition software (version 4.1, Intelligent Imaging Innovation, Denver, CO). Image z-stack pairs (red filter for syndapins 1 and 3 and green filter for dynamin 1) were captured at an interplane distance of 300 nm over a total 20-μm stack.

Image preparation.

Collected images of fluorescence-labeled isolated cells were flat-field-corrected offline and deblurred by a constrained iterative deconvolution protocol (SlideBook) using a point-spread function measured for each fluorophore and objective pairing. Background subtraction was performed using mean background staining measured from an area outside the cell region. Regions of interest (ROIs) were chosen to cover the entire cell area excluding the nucleus. Fluorescence intensity was quantified in matched ROIs for each pair of images. The masks representing whole cell and peripheral cell area (0- to 1-μm depth from the cell edge) were created from the same ROIs for each image pair. Fluorescence intensity and ROI masks were used to further analyze colocalization (see below). Background subtraction, ROI selection, and fluorescence intensity measurement were performed using ImageJ software (developed by Wayne Rasband and available from the National Institutes of Health).

Quantitative colocalization analysis.

Colocalization was quantified using the intensity correlation analysis (ICA)/intensity correlation quotient (ICQ) algorithm, as previously described (42). The ICA/ICQ algorithm is based on covariance of signal intensity of two fluorophores in space. For analysis, paired pixel intensity and ROI masks of each data set were obtained as described in Image preparation and used to calculate ICA and ICQ. Briefly, ICA was calculated as the differences from the mean of both dyes' signal intensity: (A_i − a)(B_i − b), where A_i and B_i represent staining intensity at each pixel for dyes A and B, respectively, and a and b are the means of A_i and B_i. A positive value of the ICA product [(A_i − a)(B_i − b)] indicates that both dye signals vary in synchrony (meaning that the differences of pixel intensity from the mean of both channels are either below average or above average at the same location). Staining intensity of each dye (A_i or B_i) was plotted in separate scatter plots against their (A_i − a)(B_i − b) value and used to evaluate the staining patterns. A data distribution that skews to the right reflects dependent staining between image pairs (signals vary in synchrony). A symmetrical distribution (∼0 on the abscissa) indicates random staining, while a left-skewed distribution reflects segregated staining patterns between the image pairs. To complement the ICA, the ICQ was developed as a single quotient for statistical comparison purposes (42). The ICQ is the calculation of the ratio of the number of pixels that express a positive ICA value to the total number of pixels in the ROI. In the raw form, the ratio is distributed between 0 and 1. For consistency with the ICA values, a constant 0.5 is subtracted from the ICQ value to provide a range between −0.5 and +0.5, with −0.5 representing segregated staining, 0 representing random staining, and +0.5 representing colocalized staining between channels. A mean ICQ value obtained from more than six ROIs can be used for statistical evaluation: −0.05 to +0.05 indicates random staining, +0.05 to +0.1 indicates a moderate covariance, and >0.1 indicates a significant covariance (8, 11, 39). ICA and ICQ were calculated for each cell. Average ICQ was calculated from pooled ICQ of each condition. ICA/ICQ was calculated and intensity scatter plots were constructed using MATLAB R2012b (MathWorks).

Image presentation.

Image overlays were created and brightness and contrast were adjusted with ImageJ software. No nonlinear or partial-image adjustments were made. Representative images were cropped and resized with Paint.NET software for display, but no such processing was performed on any images used in quantification.

Plasmids.

Human syndapin 1 and rat syndapin 2 cDNAs were gifts from Holger Sondermann (Cornell University) and Sandra L. Schmid (University of Texas Southwestern Medical Center), respectively, to Rajesh Ramachandran. Human syndapin 3 cDNA was purchased from DNASU Plasmid Repository (Arizona State University). Human and mouse syndapin 1 and 3 proteins show 95% identity for the whole amino acid sequence alignment and 100% identity for SH3 domain amino acid sequence alignment, respectively. Additionally, rat and mouse syndapin 2 proteins show 98% and 100% identity for the whole amino acid sequence alignment and for SH3 domain amino acid sequence alignment, respectively. The cDNAs were amplified with appropriate primers (Eurofins MWG Operon, Huntsville, AL) and subcloned into pGEX-6P-1 vector (GE Healthcare Lifesciences, Pittsburgh, PA) to generate GST fusion proteins. Point mutations to generate syndapin 1 P437L, syndapin 2 P478L, and syndapin 3 P415L were introduced in GST-syndapin 1, GST-syndapin 2, and GST-syndapin 3, respectively, using the PCR-based QuikChange Site-Directed mutagenesis protocol and confirmed by DNA sequencing. Syndapin 1 wild-type (WT), syndapin 1 P437L mutant, syndapin 2 WT, syndapin 2 P478L mutant, syndapin 3 WT, and syndapin 3 P415L mutant cDNAs were amplified with appropriate primers (Eurofins MWG Operon) and subcloned into pEGFP-N1 (Clontech, Mountain View, CA) or pIRES2-EGFP (Clontech) vectors. The plasmids were amplified in NEB 5α competent Escherichia coli (New England Biolabs, Ipswich, MA) and purified by EndoFree Plasmid Maxi Kit (Qiagen, Valencia, CA; or 5PRIME, Gaithersburg, MD).

Transfection.

Immediately after cell isolation and plating, primary chromaffin cells were transfected with empty vector, syndapin 1 WT, syndapin 1 P437L mutant, syndapin 2 WT, syndapin 2 P478L, syndapin 3 WT, or syndapin 3 P415L mutant plasmids (pEGFP-N1 or pIRES2-EGFP vector) using Lipofectamine 2000 or Lipofectamine LTX and PLUS reagents (Invitrogen). Plasmids were mixed with 100 μl of Opti-MEM (GIBCO) and transfection reagents and then incubated for 30 min at room temperature. Regular culture medium (DMEM containing ITS and antibiotics) was replaced by 900 μl of Opti-MEM. The Opti-MEM-Lipofectamine-plasmid mixture was gradually added to each culture dish containing previously added Opti-MEM and gently shaken to thoroughly mix reagents with the medium. After 20 min of incubation at 35°C with 10% CO₂, the Opti-MEM-Lipofectamine-plasmid mixture was replaced by regular culture medium. Transfected culture dishes were incubated at 35°C in 10% CO₂ for 24–48 h before initiation of experiments.

Electrophysiology.

The perforated-patch configuration was used to preserve exo- and endocytic activity (10, 23, 59). Patch-clamping was performed as previously described (14). Patch pipettes (∼4- to 5-MΩ resistance) were pulled from borosilicate glass, partially coated with molten dental wax (Electron Microscopy Sciences, Hatfield, PA), and fire-polished. The pipettes were backfilled with internal solution containing 135 mM Cs-glutamate, 10 mM HEPES-H, 9.5 mM NaCl, 0.5 mM tetraethylammonium-Cl, and 0.53 mM amphotericin B (pH 7.2, osmolarity 320 mosM). Amphotericin B was freshly prepared as a 100× stock solution in DMSO and diluted into the internal solution. During the recordings, cells were constantly superfused with a HEPES-buffered Ringer solution (see Immunofluorescence labeling). Cells were voltage-clamped and held at −80-mV command potential and depolarized with trains of action potential-equivalent (APe) voltage templates. The APe mimics native action potentials and evokes ionic currents identical to native action potentials to simulate cell firing activity under sympathetic excitation (14). APe were delivered at 15 Hz in trains of 2,250 pulses to stimulate secretion through an expanded fusion pore (29). Only cells that showed patch perforation to <40-MΩ series resistance and leak currents less than −40 pA were included in the analysis. Voltage-clamp records were acquired by an amplifier (model EPC-10, HEKA Elektronik, Lambrecht, Germany) controlled by Pulse software (version 8.80, HEKA Elektronik).

Amperometry.

Electrochemical amperometric recordings were conducted as described previously (27). Catecholamine secretion from chromaffin cells was detected by 5-μm-diameter carbon fiber electrodes (ALA Scientific, Westbury, NY). Fibers were held at +650-mV potential, lowered into the bath, and moved to touch the cell membrane gently without membrane deformation. Recordings were performed after initial fiber oxidation, and the background current of the fiber electrode was stable. The fiber was recut or replaced if the resting current was unstable or >5 pA. Amperometric currents were recorded by a dedicated amplifier (model VA-10 with 1 GΩ head stage, ALA Scientific Instruments). The head stages of the VA-10 and EPC-10 amplifiers shared a common AgCl bath ground. To minimize the cross talk between the two amplifiers, a 10-Ω resistor was added into the ground wire of the VA-10 amplifier. The signal was passed through an analog 1-kHz Bessel filter and sampled at 20 kHz through a Digidata 1322A data acquisition system (Axon Instruments, Foster City, CA) into WinEDR software (version 2.8.9, John Dempster, University of Strathclyde).

Statistical analysis.

Amperometric spike kinetics were analyzed in IGOR Pro (WaveMetrics, Lake Oswego, OR) on a single-spike basis with a peak detection routine modified from the SPIKE 4.0 package initially published by Gomez et al. (31). Median values of each parameter were calculated from each cell. The average of the individual cell median for each condition was evaluated for significance with a Student's t-test for comparison. Statistical significance for mean analysis was determined at a 95% (P < 0.05) confidence level.

RESULTS

Syndapin expression in mouse adrenal medulla.

To assess the expression profiles of syndapin isoforms in mouse adrenal medulla, we used real-time quantitative PCR, Western blot analysis, and immunolabeling to measure message levels and protein expression. Real-time quantitative PCR was performed with specific primers for each syndapin isoform. The results showed expression of mRNAs encoding for syndapins 1, 2, and 3 in mouse adrenal medulla (Table 1). Among these, message for syndapins 2 and 3 is high compared with that for syndapin 1. Western blot analysis (Fig. 1A; 50, 65, and 48 kDa for syndapins 1, 2, and 3, respectively) and immunolabeling of native tissues (Fig. 1B) showed protein expression for all syndapin isoforms in mouse adrenal medulla. In the Western blot, syndapin 1 is expressed at a low level in adrenal medulla compared with brain tissue (Fig. 1A, top lane); therefore, we provide a longer-exposure for the syndapin 1 medulla lane.

Table 1.

Real-time quantitative PCR detection of syndapin isoforms in mouse adrenal medulla

	ΔCt	Relative Expression
Syndapin 1	7.76 ± 0.24	0.2
Syndapin 2	5.97 ± 0.30	0.9
Syndapin 3	5.82 ± 0.49	1.0

Adrenal medullae were obtained from adult mouse adrenal glands. Specific primers were used for real-time PCR detection of syndapin 1, 2, and 3 messages. ΔC_t values (means ± SE) represent the difference in cycle threshold (C_t) between expression of the housekeeping gene GAPDH and each syndapin isoform. Relative expression was calculated from mean C_t of each syndapin isoform, with syndapin 3 used as a reference. Data were pooled from 3 experiments, with each experiment consisting of 5 animals.

Fig. 1.

Syndapin expression in mouse adrenal medulla. A: Western blot analysis of syndapin isoform expression in adrenal medulla. All syndapin isoform expression was detected in mouse adrenal medulla: 50, 65, and 48 kDa for syndapins 1, 2, and 3, respectively. Syndapin 1, however, was detected at low levels in mouse adrenal medulla, which can be seen more clearly after a long exposure (long exp). Mouse brain cortex, skeletal muscle, and liver are controls for syndapin isoform expression. B: bright-field images of adrenal gland sections containing adrenal cortex and medulla. Adrenal gland sections were double-immunostained for dynamin 1 and syndapins 1, 2, and 3. All syndapin isoforms are present in mouse adrenal gland sections. Scale bar, 45 μm.

Syndapin 3 modulates fusion pore expansion and hormone release from chromaffin cells.

We previously showed that syndapin-dynamin interaction contributes to fusion pore expansion (57). Since syndapins 1, 2, and 3 are expressed in mouse adrenal medulla, we next tested whether each syndapin isoform specifically plays a role in regulating expansion of the fusion pore by electrochemical detection and analysis of catecholamine release. Syndapin-dynamin interaction occurs through an SH3-PRD binding interaction that depends on the central proline residue in the ligand-binding site of syndapin SH3 domain. Therefore, we generated syndapin mutants that contain a point mutation of the central proline residue: P437L for syndapin 1, P478L for syndapin 2, and P415L for syndapin 3. The mutants have been shown to disrupt SH3 domain-binding activity to PRD-containing proteins including dynamin (45, 53). We expressed syndapin 1 WT, syndapin 1 P437L, syndapin 2 WT, syndapin 2 P478L, syndapin 3 WT, or syndapin 3 P415L in primary cultured chromaffin cells. Transfection was confirmed by expression of enhanced green fluorescent protein (EGFP) as a separate protein, with a bicistronic vector (pIRES2) used to eliminate potential disruption of syndapin function resulting from a COOH-terminal EGFP fusion protein (61). We also transfected cells with the pIRES2 vector containing only the EGFP as an additional control. Transfected cells were identified by EGFP fluorescence, patch-clamped, and stimulated with a train of APe delivered at 15 Hz. This stimulation paradigm has been shown to drive active expansion of the exocytic fusion pore to a high-conductance state (29) that is mediated by a dynamin 1-syndapin interaction (57). Evoked catecholamine release was simultaneously measured by electrochemical amperometry at the single-granule level (15, 66). Representative raw amperometric traces recorded from the cells transfected with empty vector, syndapin 1 WT, syndapin 1 P437L mutant, syndapin 2 WT, syndapin 2 P478L mutant, syndapin 3 WT, and syndapin 3 P415L mutant are shown in Fig. 2A. In the physiological setting, the trigger for fusion pore expansion is an elevation of cytosolic Ca²⁺ in response to increased stimulation (29). We found no statistical difference in Ca²⁺ currents evoked between cells expressing empty vector and cells expressing syndapin constructs (data not shown). Thus any differences are not due to an altered Ca²⁺ influx.

Fig. 2.

Syndapin 3 P415L mutation decreases catecholamine quantal size. Isolated mouse chromaffin cells were transfected with empty vector (pIRES2), syndapin 1 wild-type (WT), syndapin 1 P437L, syndapin 2 WT, syndapin 2 P478L, syndapin 3 WT, or syndapin 3 P415L. Positive-transfected cells [determined by enhanced green fluorescent protein (EGFP) fluorescence signal] were held in perforated-patch voltage-clamp mode and stimulated with trains of action potential equivalents (APe) delivered at 15 Hz. Catecholamine secretion events were recorded by electrochemical amperometry. A: representative raw traces for each construct-transfection condition. B and C: amperometric parameters (spike charge and foot charge) for construct-transfected cells presented as category plots of average individual cell median. Sample sizes are as follows: 14 cells for empty vector, 11 cells for syndapin 1 WT, 9 cells for syndapin 1 P437L, 11 cells for syndapin 2 WT, 15 cells for syndapin 2 P478L, 12 cells for syndapin 3 WT, and 13 cells for syndapin 3 P415L. Statistical significance was determined by Student's t-test comparison of each condition with empty-vector condition for each parameter: *P < 0.05.

To determine the effect of syndapins 1, 2, and 3 on fusion pore expansion, we quantified the charge of single spike events. Spike charge correlates to catecholamine quantal size, a parameter that increases with facilitated granule emptying subsequent to fusion pore expansion (7, 27, 29, 33). Spike charge was quantified by integration of the area under each amperometric event (Fig. 2B, inset, gray area). Consistent with previous studies, we found that spike charge followed a skewed distribution (28). Therefore, we quantified the median value from individual cells and then averaged all median values for each condition (mean of median) (46). We found that expression of the syndapin 3 P415L mutant significantly decreased quantal size of catecholamine release compared with empty vector (Fig. 2B). In addition to spike charge, we also analyzed prespike foot charge (Fig. 2C, inset, gray area). This parameter correlates to a flux of catecholamines through the initial fusion pore and is a measure of initial pore opening prior to expansion (1, 2, 15, 64). A large body of accumulating evidence has established that the initial formation of the fusion pore is due to soluble N-ethylmaleimide-sensitive fusion attachment protein receptor (SNARE) complex-mediated vesicle-cell membrane interaction (24, 48, 65) and that its initial opening is triggered by synaptotagmin-dependent processes (63, 67). Thus syndapin mutation should not affect initial pore opening and should not alter foot charge. Data presented in Fig. 2C confirm this prediction and isolate the site of action for syndapin 3 to regulated expansion of the pore. Transfection with WT or SH3 mutation of syndapins 1 and 2, on the other hand, did not affect spike charge or foot charge. In addition, none of the constructs affected spike frequency (data not shown).

Differential localization of syndapin isoforms in adrenal chromaffin cells.

Quantification of message and protein expression confirms that mouse adrenal medulla expresses all syndapin isoforms. However, mutational disruption of binding between syndapin and PRD-containing proteins isolates functional regulation of fusion pore expansion to syndapin 3. Thus we further studied the subcellular localization of syndapin 3 and its binding partner dynamin 1 in isolated mouse adrenal chromaffin cells. Moreover, we compared syndapin 3 with syndapin 1, which is specifically expressed in neuronal tissue and involved in regulation of neurotransmitter release and subsequent synaptic vesicle endocytosis (4, 16, 38, 41, 52) and, thus, would serve as a closely related negative control signal for comparison with syndapin 3 in our neuroendocrine system.

We initiated this component of the study by confirming specificity of the primary antibodies for their intended epitope using purified GST-syndapin 1, 2, and 3 fusion proteins for Western blot analysis. As demonstrated in Fig. 3, the syndapin 1 antibody exhibits a very high degree of specificity for syndapin 1 over syndapins 2 and 3 (Fig. 3Ai). Similarly, the syndapin 3 antibody is highly specific for syndapin 3 over syndapins 1 and 2 (Fig. 3Bi). Next, using specific antibodies against either isoform, we performed immunocytochemistry in isolated chromaffin cells. Endogenous syndapin 1 immunoreactivity (IR) is distributed mainly in the cytoplasmic region and exhibits a punctate staining pattern (Fig. 3Aii). Syndapin 3 IR staining, on the other hand, localizes mainly at the periphery of the cell (Fig. 3Bii), the site of granule fusion. Pooled imaging data show a higher cytoplasmic-to-peripheral fluorescence intensity ratio for syndapin 1 than syndapin 3 and confirms a dissimilarity of syndapin 1 to syndapin 3 localization (Fig. 3C, Endogenous).

Fig. 3.

Comparison of endogenous syndapins and syndapin construct localization. Ai and Bi: syndapin 1 and 3 antibody specificity characterized using GST-syndapin 1 (Sdp1), GST-syndapin 2 (Sdp2), and GST-syndapin 3 (Sdp3) purified protein (26, 50, 65, and 48 kDa for GST, syndapin 1, syndapin 2, and syndapin 3, respectively). A and B: positive-transfected cells determined by EGFP fluorescence signal for each construct imaged at 24–48 h posttransfection. For endogenous syndapin protein, isolated chromaffin cells were immunostained with syndapin 1 or 3 antibody. For syndapin construct overexpression, isolated chromaffin cells were transfected with syndapin 1 WT-EGFP, syndapin 1 P437L-EGFP, syndapin 3 WT-EGFP, or syndapin 3 P415L-EGFP. Subcellular distribution profiles of endogenous syndapin 1 (Aii), syndapin 1 WT-EGFP (Aiii), syndapin 1 P437L-EGFP (Aiv), endogenous syndapin 3 (Bii), syndapin 3 WT-EGFP (Biii), and syndapin 3 P415L-EGFP (Biv) are shown in representative images (top) and line profile analysis plots (bottom). For line analysis, single-line profiles were drawn across the cells, and normalized fluorescence intensity along the lines was plotted against normalized line distance. C: fluorescence intensity of immunoreactivity in cytoplasm (Cyto) and cell peripheral (Peri) region (∼1 μm deep from cell edge) from 20 cells for endogenous syndapin 1, 7 cells for syndapin 1 WT-EGFP, 12 cells for syndapin 1 P437L-EGFP, 20 cells for endogenous syndapin 3, 12 cells for syndapin 3 WT-EGFP, and 10 cells for syndapin 3 P415L-EGFP. Mean values of cytoplasmic-to-peripheral intensity ratio were calculated from pooled ratio data of each cell. *P < 0.05 (by Student's t-test).

We also tested subcellular distribution of endogenous syndapins 1 and 3 in response to cell stimulation. We challenged cells with a Ringer solution with 30 mM KCl isosmotically substituted for NaCl (see materials and methods) for 5 min followed by fixation. After stimulation, the cytoplasmic-to-peripheral intensity ratio (as calculated in Fig. 3C) for syndapin 3 was increased (0.26 ± 0.03 in control vs. 0.39 ± 0.03 after stimulation, n = 20 for each data set, P = 0.01). No such stimulus-dependent movement was determined for syndapin 1 (cytoplasmic-to-peripheral intensity ratio = 1.38 ± 0.1 for control vs. 1.47 ± 0.09 after stimulation, n = 20, P = 0.54). These data indicate that only syndapin 3 exhibits an activity-dependent distribution, as expected for a protein associated with exocytosis-endocytosis membrane trafficking.

Mutation of the SH3 domain disrupts targeting of syndapin 3 to the cell periphery.

Next, we examined whether disruption of the SH3-PRD binding interaction alters syndapin localization in chromaffin cells. We compared the localization of syndapins 1 and 3 WT with point mutants syndapin 1 P437L and syndapin 3 P415L used in the functional studies presented in Fig. 2. In these experiments, we expressed syndapin WT and mutant as EGFP fusion proteins to track syndapin localization. Endogenous syndapin 1 (Fig. 3Aii), syndapin 1 WT-EGFP (Fig. 3Aiii), and syndapin 1 P437L-EGFP (Fig. 3Aiv) displayed a cytoplasmic distribution. In contrast, syndapin 3 WT-EGFP was highly localized to the cell periphery in a pattern similar to endogenous syndapin 3 (Fig. 3, Bii and Biii). Syndapin 3 P415L-EGFP showed a lower localization profile at the cell periphery, instead exhibiting a more cytosolic distribution (Fig. 3Biv). Quantitation of pooled imaging data comparing the cytoplasmic-to-peripheral intensity ratio in Fig. 3C confirms a higher cytoplasmic-to-peripheral fluorescence intensity ratio for syndapin 1 WT-EGFP, syndapin 1 P437-EGFP, and syndapin 3 P415L-EGFP than syndapin 3 WT-EGFP. These results indicate that syndapin 3 is targeted to the cell periphery through a functional SH3-PRD interaction.

Syndapin 3 is colocalized with dynamin 1 in adrenal chromaffin cells.

Since the syndapin-dynamin 1 interaction is critical for chromaffin secretory function (57), we further investigated endogenous syndapin isoform codistribution with dynamin 1. We double-immunostained isolated chromaffin cells for dynamin 1 and syndapin 1 or 3, respectively, and performed a quantitative colocalization analysis. We implemented the ICA approach initially developed by Li et al. (42). This analysis has been validated and used to determine protein colocalization in intact cells (8, 9, 42, 44). The rationale of this method is that if two fluorescent dye-labeled proteins are in the same complex, the paired pixel intensities should vary in synchrony over space (i.e., better correlation in labeling intensity at a given pixel). In contrast, asynchronous staining should be observed if they are not in the same complex (i.e., less correlation in labeling intensity). ICA was calculated as (A_i − a)(B_i − b), the difference of each paired pixel intensity for dye A and dye B (A_i and B_i) from its mean (a and b), respectively (see materials and methods). For statistical evaluation, the ICQ is calculated from the product of ICA as a number between −0.5 (segregated staining), 0 (random staining), and +0.5 (dependent staining). We tested the resolution of the ICA/ICQ method for quantitatively evaluating colocalization between two labeled proteins, i.e., to determine the statistical resolution of the technique in our experimental context. A protocol was developed to incorporate properties of our optics and the cell staining characteristics to calibrate the ICA/ICQ approach. We performed ICA and ICQ calculations on controlled image pairs: the original green channel (A) cell image and a duplicate of the image converted to the red channel (B). As expected, this positive control reports perfect colocalization in a typical “dye-overlay” method (Fig. 4A, top) as well as a perfect intensity correlation plot (Fig. 4A, bottom left). The normalized pixel intensity of image 1 or image 2 against ICA product [(A_i − a)(B_i − b) values] plots also shows dependent staining patterns, i.e., skewed toward positive values (Fig. 4A, bottom middle and bottom right). To calibrate the resolution of ICA/ICQ in our optical, analytic configuration and cell morphology, we adopted a simple displacement approach. We displaced the second (duplicate) image in steps and calculated ICA/ICQ. We tested linear and rotational displacement. For linear displacement, the second duplicate image was displaced in 25-nm steps for a total of 625 nm horizontally or vertically with respect to the first image position. ICQ was calculated from the overlapping (intersection) area of the two images. As expected, prior to displacement, ICQ was +0.5, indicating perfect colocalization. The ICQ value decreased immediately after only the first 25-nm step and showed additional reduction with further displacement. The resulting mean ICQ values from 20 such image pair experiments are provided in Fig. 4B. For rotational displacement, the center of the cell in each image pair was identified, then the second image was rotated about the center from 0 to 360 degrees in 1-degree increments. The ICQ value was calculated for each degree of rotation. As expected, ICQ showed the highest possible value (+0.5) at 0- and 360-degree rotation, where the two images were perfectly aligned. ICQ values were statistically decreased when the second image was rotated by only 1 degree, disrupting the colocalization of the duplicate image pairs. The more the second image was rotated, the lower the ICQ value. ICQ values increased again when the second image was rotated >180 degrees, approaching its original position. Again, the rotational calibration was repeated on 20 image pairs, with the mean ICQ values presented in Fig. 4C. The ICQ of disoriented images (linear shift or rotational shift, Fig. 4, B and C) never reported a value of 0, as expected for random orientation/covariance, probably because the calibrations are performed on actual images, where the cell itself defines image morphology and, thus, a degree of covariance. Rather, the minimum ICQ value obtained from duplicate pairs after displacement/rotation defines the random ICQ value in our experimental system as 0.24. Spatial calibration is also provided. In each calibration condition, the P value for statistical determination fell to <0.01 after just a 25-nm linear shift or 1-degree rotation. In our optical/imaging system, the 1-degree rotation provides an 87-nm arc length for a typical cell geometry of 5-μm radius. Thus the linear shift mode defines our limit of analytic resolution to be 25 nm.

Fig. 4.

Calibration of intensity correlation analysis (ICA)/intensity correlation quotient (ICQ). A: example of images used in calibrating resolution of ICA/ICQ in our experimental context. Image 1 is an original image, and image 2 is a duplicate of image 1 with immunofluorescence staining converted to red; therefore, overlay of images 1 and 2 (merge) shows a perfect colocalization determined by the “dye-overlay” method (green + red = yellow; top). Scatter plots of normalized dye intensity of image 1 vs. image 2 show high correlation (bottom left). Each individual dye intensity was plotted against (A − a)(B − b) value and also reflects a right-shifted dependent staining pattern (bottom middle and bottom right). B: horizontal and vertical shifting of images was used to calibrate resolution of ICA/ICQ in our experimental condition. The second duplicate images were shifted horizontally or vertically with respect to the first images in increments of 1/5th of a pixel dimension (25 nm) for a total of 10 pixels. ICQ was calculated at each step and averaged from 20 different sets of images. Mean ICQ was plotted against horizontal/vertical shifting distance (nm). C: radial ICA/ICQ sensitivity was also tested by image rotation. Center of cells was identified; then duplicate cell image was rotated according to cell center from 0 to 360 degrees in 1-degree increments. ICQ was calculated for each degree rotation. Mean ICQ values from 20 sets of images were plotted against rotation.

Next, we performed immunocytochemistry double-staining for dynamin 1 and syndapin 3 and the functional negative control (syndapin 1) and tested their colocalization using the ICA/ICQ algorithm. Representative images in Fig. 5, Ai and Bi, show immunolabeled dynamin 1-syndapin 3 and dynamin 1-syndapin 1. Syndapin 1 IR is punctuate and does not show obvious association with the plasma membrane, whereas syndapin 3 IR shows a strong signal at the cell periphery with little staining within the cell (see also Fig. 3). Dynamin 1 IR is highly concentrated at the cell periphery, similar to syndapin 3 IR. The dynamin 1-syndapin paired pixel intensity plots show that the intensity of the syndapin 1 signal is not well correlated to the dynamin 1 signal, while the syndapin 3 signal shows a positive correlation to the dynamin 1 signal (Fig. 5, Aii and Bii, left). The ICA products of dynamin 1-syndapin 1 and dynamin 1-syndapin 3 pairs also show that the staining pattern of the dynamin 1-syndapin 1 pair did not exhibit signal intensity covariance; the (A − a)(B − b) values clustered about the 0 line of the abscissa (Fig. 5Aii, middle and right). However, the dynamin 1-syndapin 3 pair shows signal intensity covariance; the (A − a)(B − b) values skewed to the right of 0 (Fig. 5Bii, middle and right). We calculated the ICQ in three areas of the cells: whole cell (excluding nucleus), interior cytoplasm (excluding 1 μm from the edge of the cell), and peripheral region (limited to 1 μm inside the edge). Mean ICQ was calculated from 20 cells for each condition and is shown in Fig. 5C. Mean ICQ was significantly higher for dynamin 1-syndapin 3 than dynamin 1-syndapin 1 double-staining in every subarea of the cell (Fig. 5C), indicating a stronger colocalization between dynamin 1-syndapin 3 than dynamin 1-syndapin 1. Moreover, in our calibration images, the ICQ value for a displaced signal pair was ∼0.24 (see shifted data in Fig. 4B and rotated data in Fig. 4C). The dynamin 1-syndapin 1 ICQ values are ∼0.1 (Fig. 5C), indicating that they are less colocalized than disoriented image pairs. This result supports the idea that dynamin 1 and syndapin 1 are not colocalized in neuroendocrine chromaffin cells and are likely segregated in the cytosol. Similarly, the ICQ values of dynamin 1-syndapin 3 staining are well above those calculated for disorientated image pairs and support the idea that these molecules are colocalized or preassembled.

Fig. 5.

Colocalization of dynamin 1 with syndapins 1 and 3. A and B: isolated chromaffin cells were double-stained with dynamin 1 and syndapin 1 or 3 antibodies. Ai and Bi: representative immunostained images of dynamin 1, syndapins 1 and 3, and overlay (merge). Scale bar, 7 μm. Aii and Bii: intensity correlation plots of dynamin 1-syndapin 1 and dynamin 1-syndapin 3 (left). Dynamin 1-syndapin 3 plot showed a positive correlation, while dynamin 1-syndapin 1 plot did not; blue dashed line represents the highest correlation coefficient (r = 1). Pixel intensity of dynamin 1-syndapin 1 and dynamin 1-syndapin 3 is plotted against (A − a)(B − b) values (middle and right); red dashed line indicates 0. ICQ = +0.13 for dynamin 1-syndapin 1 and ICQ = +0.37 for dynamin 1-syndapin 3. C: mean ICQ was analyzed from 20 cells for each group. ICQ was calculated from whole cell, cytoplasm, and peripheral region (∼1 μm deep from cell edge). Values are means ± SE. Statistical significance was determined by Student's t-test: *P < 0.05. D: rotational analysis as introduced for duplicate image pairs in Fig. 4C performed on image pairs labeled for dynamin (Dyn) 1-syndapin (Sdp) 3 and dynamin 1-syndapin 1. Syndapin image of a stained cell pairing was rotated from 0 to 360 degrees at 1-degree increments, and ICQ was calculated for each position. Values are means ± SE for each pairing and rotation. Gray dashed line shows minimum ICQ value for disoriented image pairs from Fig. 4C for reference.

Lastly, the rotational ICA/ICQ analysis utilized to calibrate duplicate image pairs in Fig. 4 was utilized to further test the colocalization of syndapins 1 and 3 to dynamin 1. If two fluorescence signals are truly colocalized, rotation of the second image of the pair would disrupt the colocalization and result in a decreased ICQ score. In a complementary sense, if two images do not represent colocalized staining patterns, initial ICQ values should be lower in absolute magnitude and rotation of one with respect to the other should not alter the ICQ to any significant degree. We analyzed all immunofluorescence-stained image pairs in this manner. We provide data from this analysis in Fig. 5D and show, as predicted, that rotation of syndapin 3 staining with respect to dynamin 1 does, indeed, result in an initially high ICQ score that decreases with rotation, only to increase again when the rotation nears a complete 360 degrees (blue trace, Fig. 5D). Also consistent with previous data, dynamin 1-syndapin 1 staining starts with a low ICQ value that does not appreciably change with rotation of the syndapin 1 signal (red trace, Fig. 5D). For reference, the gray dashed line indicates the 0.24 minimum ICQ value obtained for disorientated pairs generated in Fig. 4. These data quantitatively further support the idea that syndapin 3 colocalizes to dynamin 1 within our 25-nm resolution. The same data support the idea that syndapin 1 and dynamin 1 are not associated with one another within this resolution.

Taken together, our data demonstrate that all syndapin isoforms (1, 2, and 3) are expressed in mouse adrenal medulla. Among these, only mutation of the syndapin 3 SH3 domain disrupts fusion pore expansion and, thus, alters hormone release. The same mutation disrupts syndapin 3 targeting to the cell periphery of adrenal chromaffin cells. We further show that syndapin 3 and dynamin 1 exist in a persistent colocalized state.

DISCUSSION

Neuroendocrine chromaffin cells release hormone through the fusion of large dense core granules with the cell surface. The initial step in this process is the regulated, Ca²⁺-triggered opening of a secretory fusion pore. Subsequent to this step, there is a choice to be made. If Ca²⁺ levels are modest, as under sympathetic tone, the pore remains in a restricted state, filtering the specific release of soluble and mobile catecholamine prior to a rapid endocytic retrieval of the granule (13, 14). However, if the cell is under heightened stimulation, as under the acute sympathoadrenal stress reflex, elevated Ca²⁺ initiates calmodulin-activated calcineurin to dephosphorylate dynamin 1 at the PRD. This, in turn, reveals a binding site for the complementary SH3 domain of syndapin (57). This dephosphorylation step initiates the activation of a signaling cascade that ultimately leads to the regulated myosin-dependent expansion of the fusion pore (21, 47). The expansion of the fusion pore is considered a necessary step for the release of peptide transmitters sequestered in the dense proteinaceous granule core (5) and may play a role in further sorting of peptide transmitters (50). Endocytosis following this form of release is much slower, taking seconds or longer to complete (59). Thus activity-regulated expansion of the secretory fusion pore has been proposed as a key mechanistic component of the sympathoadrenal stress reflex, increasing catecholamine quantal size and facilitating peptide hormone secretion from the dense granule core (22, 27, 50). We previously showed that fusion pore expansion is regulated by a calcineurin-activated dynamin 1-syndapin-dependent process (57). Syndapin is an accessory protein that regulates several endocytic processes, including receptor-mediated internalization and recycling of neuronal membrane during transmitter release (4, 16, 38, 41, 52). Our current study provides the first report of syndapin isoform expression in neuroendocrine adrenal medullary tissue. We demonstrate expression of message and protein for all three syndapin isoforms in mouse adrenal medulla. These results were not surprising, since the adrenal medulla originates from neuronal tissue (49), which also expresses syndapin 1, as well as both other syndapin isoforms, at high levels (41). Syndapin 2 is ubiquitously expressed in all tissue types and is involved in general cellular processes, including Golgi transport (55). We also demonstrate expression of syndapin 3, which has been shown to be highly expressed in skeletal muscle, heart, and lung and present in neurons (41, 45). Our data show that, unlike neurons, where dynamin-syndapin 1 regulates synaptic vesicle function (41), neuroendocrine cells utilize a dynamin-syndapin 3 system for modulation of fusion pore expansion.

We previously demonstrated that dynamin 1-syndapin interaction and neural Wiskott-Aldrich syndrome protein (N-WASP) activation play a critical role in mediating fusion pore behavior in neuroendocrine cell secretion (57). The syndapin 1 P437L, syndapin 2 P478L, and syndapin 3 P415L mutants used in the current study have been shown to lose binding affinity to PRD-containing proteins including dynamin 1, synaptojanin 1, and N-WASP (36, 37, 45). Our functional amperometric assay showed that the syndapin 3 P415L mutant, but not the syndapin 1 P437L or syndapin 2 P478L mutant, decreased catecholamine quantal size, indicating reduced pore expansion. Moreover, this effect was specific; the mutant did not alter initial pore opening, Ca²⁺ current, or spike frequency. Therefore, syndapin 3 has specific function in mediating expansion of the neuroendocrine fusion pore. The normal dynamin-syndapin binding relieves an autoinhibitory conformation in syndapin (54) to initiate downstream signaling events. Thus the mutant is, indeed, predicted to disrupt the pore expansion process, as reported in Fig. 2. The mutants used in this study also lose binding affinity to synaptojanin 1 and N-WASP (45) and represent signaling partners that may contribute to activity-mediated fusion pore expansion.

The intracellular distribution profiles of endogenous syndapins 1 and 3 in isolated chromaffin cells shown here are similar to expression patterns in neurons and myotubes. Syndapin 1 shows punctate staining throughout the cytoplasm and is not targeted to the cell periphery, whereas syndapin 3 shows staining concentrated close to the edge of the cells (45, 53). Expression of mutant syndapin constructs shown in Fig. 3 reveals disruption of the syndapin 3 SH3 domain targeting to the cell periphery, possibly through loss of affinity for its PRD-containing partners (dynamin 1, synaptojanin 1, and N-WASP). Further experiments will reveal the participation of one or more of these PRD-containing proteins in subcellular syndapin 3 localization. Since syndapin 1 is involved in the dynamin 1-regulated endocytosis in neuronal systems (4, 16, 37), one would expect that syndapin 1 would be highly colocalized with dynamin 1 in neuroendocrine chromaffin cells as well. However, we demonstrate a persistent colocalization between syndapin 3 and dynamin 1 but segregated localization between syndapin 1 and dynamin 1. The dynamin 1-syndapin 3 colocalization we demonstrate in neuroendocrine chromaffin cells is dissimilar to that in other cell systems. Qualmann et al. (53) showed that syndapin 1 is colocalized to dynamin 1 in cultured neurons. Modregger et al. (45) showed that syndapin 1 colocalized with dynamin in primary hippocampal neuron and syndapin 3 with dynamin in C2F3 myotubes. Thus tissue-specific interaction of syndapin isoforms may fine-tune the regulation of membrane trafficking to tissue-specific tasks: syndapin 1 regulates neuronal synaptic vesicle function, syndapin 2 is involved in membrane trafficking in nonsecretory cells, and syndapin 3 is involved in modulation of fusion pore expansion in neuroendocrine cells. This raises the intriguing possibility that syndapin 3 represents a control point for enhanced catecholamine release and peptide transmitter release profiles under the physiological fight-or-flight sympathetic stress response.

Lastly, we show specific affinity of dynamin for syndapin 3 in our system. Yet previous studies have shown that dynamin 1 exhibits a binding affinity for syndapin 1 in vivo (45, 53) and in vitro (30). However, neither the WT nor the P437L mutant of syndapin 1 had an effect on fusion pore expansion (Fig. 2), nor did overexpressed syndapin 1 P437L mutant show an altered cytosolic distribution compared with overexpressed WT (Fig. 3). Thus there must be a mechanism for selective interaction between dynamin 1 and syndapin 3 in our neuroendocrine system. Future studies are needed to elucidate this mechanism.

GRANTS

This work was supported by the National Institute of General Medical Sciences Grant GM-102191.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.S. and C.S. are responsible for conception and design of the research; P.S., S.-A.C., and R.R. performed the experiments; P.S., K.L., and S.-A.C. analyzed the data; P.S., K.L., S.-A.C., and C.S. interpreted the results of the experiments; P.S. prepared the figures; P.S. drafted the manuscript; P.S., K.L., S.-A.C., R.R., and C.S. edited and revised the manuscript; P.S., K.L., S.-A.C., R.R., and C.S. approved the final version of the manuscript.