Golgi-associated long coiled-coil proteins, often referred to as golgins, are involved in the maintenance of the structural organization of the Golgi apparatus and the regulation of membrane traffic events occurring in this organelle. Little information is available on the contribution of golgins to Golgi function in cells specialized in secretion such as endocrine cells or neurons. In the present study, we characterize the intracellular distribution as well as the biochemical and functional properties of a novel long coiled-coil protein present in neuroendocrine tissues, NECC1 (neuroendocrine long coiled-coil protein 1). The present study shows that NECC1 is a peripheral membrane protein displaying high stability to detergent extraction, which distributes across the Golgi apparatus in neuroendocrine cells. In addition, NECC1 partially localizes to post-Golgi carriers containing secretory cargo in PC12 cells. Overexpression of NECC1 resulted in the formation of juxtanuclear aggregates together with a slight fragmentation of the Golgi and a decrease in K+-stimulated hormone release. In contrast, NECC1 silencing did not alter Golgi architecture, but enhanced K+-stimulated hormone secretion in PC12 cells. In all, the results of the present study identify NECC1 as a novel component of the Golgi matrix and support a role for this protein as a negative modulator of the regulated trafficking of secretory cargo in neuroendocrine cells.
- Golgi matrix
- Golgi trafficking
- neuroendocrine cell
- neuroendocrine long coiled-coil protein 1 (NECC1)
- regulated secretory pathway
Long α-helical coiled-coil proteins comprise a family of structurally related proteins with a predicted elongated rod-like conformation that display diverse biological functions . Motor proteins, membrane tethering and vesicle transport proteins are the dominant eukaryote-specific long coiled-coil proteins . Among those involved in membrane tethering, there are the peripherally associated or tail-anchored Golgi-localized coiled-coil proteins, commonly referred to as golgins, which mediate membrane–membrane and membrane–cytoskeleton interactions at the Golgi thus regulating membrane transport, cisternae formation and cisternal stacking .
One of the best-characterized golgins is GM130 (130 kDa Golgi matrix protein), which is targeted to cis-Golgi membranes through its binding to a member of the GRASP (Golgi reassembly stacking protein) family of Golgi stacking proteins, GRASP65 . GM130 also binds to p115 [4,5], a peripheral membrane golgin recruited to ER (endoplasmic reticulum)-to-Golgi carriers and the cis-Golgi  through its interaction with the small GTPase Rab1 . In turn, p115 binds to giantin, a tail-anchored transmembrane golgin located on COPI (coatamer protein I) vesicles . These golgins can form different tethering complexes that have been proposed to enable efficient capture of ER-to-Golgi carriers at the entry face of the Golgi [9,10]. Accordingly, they have been shown to participate in anterograde trafficking of cargo . In addition, a role for GM130, p115 and giantin in the stacking of cisternae during Golgi assembly has also been proposed , and the two former golgins have been shown to mediate the biogenesis of the Golgi ribbon [13,14]. The cis-Golgi golgin family also includes the peripheral membrane protein GMAP210 (Golgi-associated microtubule-binding protein 210) , which is involved in the maintenance of Golgi ribbon integrity and positioning  as well as in the regulation of polarized secretion . Peripheral membrane golgins, together with their regulatory GTPases [Rab, ARF (ADP-ribosylation factor) and ARL (ARF-like) proteins] and GRASPs, have been proposed to form a scaffold that helps organize Golgi structure and trafficking, the so-called Golgi matrix [9,10].
Other golgins associated with membrane transport in mammalian cells include the peripheral membrane coiled-coil proteins GCC (Golgi coiled-coil) 88, GCC185, golgin-97 and p230/golgin-245, which contain a conserved GRIP (golgin-97/RanBP2/Imh1p/p230) domain that serves as a targeting sequence for specific recruitment of these proteins to TGN (trans-Golgi network) membranes [18,19]. In addition to serve as vesicle tethers, these golgins have also been postulated to maintain the structural organization of the TGN [19–22].
Previously, we identified by differential screening a cDNA coding for a novel long coiled-coil protein whose expression was up-regulated in endocrine cells (i.e. pituitary melanotropes) displaying low secretory activity . This protein, which we named NECC1 (neuroendocrine long coiled-coil protein 1) [also known as Jakmip 2 (Janus kinase and microtubule-interacting protein 2)]  on the basis of its preferential expression in endocrine and neural tissues, was found to distribute at a juxtanuclear position when expressed in HEK (human embryonic kidney)-293 AD cells . However, the intracellular localization and functional role of NECC1 in cells specialized in secretion remain unknown. We have now determined the molecular characteristics and subcellular distribution of NECC1 in the neuroendocrine cell line PC12 and provided the first experimental evidence supporting the involvement of a Golgi-associated long coiled-coil protein in the regulated secretion of hormones.
A polyclonal anti-NECC1 antiserum was raised by rabbit immunization with a synthetic peptide corresponding to amino acid residues 783–801 (RIRDLEDKTDIQKRQIKDL) of human NECC1 conjugated to keyhole limpet haemocyanin. Purified anti-NECC1 antibody was obtained by means of immunoaffinity chromatography using the immobilized peptide. Monoclonal anti-cMyc and anti-β-actin antibodies were purchased from Serotec and Sigma respectively. Monoclonal antibodies against GM130, TGN38 and EEA1 (early endosome antigen 1) were obtained from BD Transduction Laboratories. Goat polyclonal anti-Sec23 antibody was purchased from Santa Cruz Biotechnology. Rabbit anti-ERGIC (ER–Golgi intermediate compartment) 53 antibody was from Abcam, and mouse anti-γ-tubulin was from Thermo Fisher Scientific. Rabbit polyclonal antiserum against the CgA (chromogranin A)-derived peptide EL35 has been described previously . Rabbit polyclonal anti-hGH (human growth hormone) antibody was obtained from Dr A.F. Parlow (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, U.S.A.). ER-Tracker™ Red was purchased from Molecular Probes. Alexa Fluor®-conjugated secondary antibodies for immunofluorescence were purchased from Invitrogen. Horseradish peroxidase-conjugated goat anti-(rabbit IgG) and anti-(mouse IgG) were obtained from Sigma.
Cell lines (PC12, AtT20 and HEK-293 AD cells) were cultured as described previously [23,26]. For BFA (brefeldin A) treatment, cells were incubated with 5 μg/ml BFA (Sigma) for 1 h at 37°C. In washout experiments, BFA was removed by five washes in fresh medium before incubation for a further 1.5 h at 37°C.
Immunofluorescence and confocal microscopy
Isolated cells from rat pituitary were obtained using a dispersion protocol described previously . PC12, AtT20 and pituitary cells on glass coverslips were processed for immunofluorescence and examined by confocal microscopy as described previously . Evaluation of the number and area of GM130-positive objects per cell was assessed using ImageJ 1.36 (http://rsbweb.nih.gov/ij/) as follows: a binary threshold (40–255 grey levels) was applied to z-projection images, and the number and area of immunoreactive objects displaying a size higher than 5 pixels2 was measured in each cell using the Analyze Particles function of this software. The degree of co-localization between NECC1 and intracellular markers was examined by visual inspection of single confocal images and considered as such when fluorescent signals were coincident in the same planes. The co-localization rate was subsequently assessed by Manders' coeffcient using ImageJ. Manders' coefficient varies from 0 to 1, corresponding to non-overlapping images and 100% co-localization between the two images respectively.
SDS/PAGE and immunoblotting
Protein extracts from cultured cells or rat pituitary were prepared by homogenization in RIPA buffer containing protease inhibitors and dissolved in reducing SDS sample buffer. Proteins were separated by SDS/PAGE (7.5% gel) and transferred on to nitrocellulose membranes (BioTrace NT). Membranes were blocked in 5% non-fat dry milk, 0.05% Tween 20, 150 mM NaCl and 25 mM Tris/HCl (pH 7.4) for 1 h and then incubated with the corresponding primary antibody diluted in blocking buffer overnight at 4°C. Blots were washed and incubated for 1 h at room temperature (22°C) with the appropriate horseradish peroxidase-conjugated secondary antibody. The immunoreaction was visualized using ECL (enhanced chemiluminescence) Plus kit (GE Healthcare). Band intensities were quantified using ImageJ 1.36 software, and the data were normalized to the corresponding β-actin values measured on the same blots.
PC12 cells from four 100-mm dishes grown to confluency were scraped into 2 ml of lysis buffer [320 mM sucrose, 10 mM Tris/HCl (pH 7.4), 1 mM EGTA and protease inhibitors] and homogenized by 20 passages through a 25-guage needle. The PNS (post-nuclear supernatant) was obtained by centrifugation of the homogenate at 1000 g for 10 min. The PNS was then centrifuged at 165000 g for 2 h to separate the pellet containing the crude membrane fraction (P2) and the supernatant containing the cytosolic fraction (S2). The P2 fraction was resuspended in 500 μl of lysis buffer and loaded on to a 10 ml continuous sucrose density gradient (0.6–1.8 M) made up in lysis buffer. After ultracentrifugation at 34000 rev./min for 17 h in a Beckman SW-40Ti rotor at 4°C, 13 fractions (0.8 ml) were taken from the top of the gradient. Proteins were precipitated by adding trichloroacetic acid to a final concentration of 15%. After incubation overnight at 4°C, protein precipitates were collected by centrifugation at 16000 g for 15 min, washed twice with acetone at −20°C and resuspended in reducing SDS sample buffer. Equal volumes of each fraction were analysed by immunoblotting.
Membrane protein extraction assay
Aliquots (300 μg) of the crude membrane fraction from PC12 cells were resuspended in 0.6 ml of lysis buffer and extracted with an equal volume of one of the following buffers: lysis buffer (control), 1 M NaCl in lysis buffer, 2% Triton X-100 in lysis buffer, 1% SDS in lysis buffer or 200 mM Na2CO3 (pH 11) in water. After incubation for 1 h at 4°C, the extracted (supernatants) and insoluble (pellets) materials were separated by ultracentrifugation at 35000 rev./min for 2 h. Pellets were resuspended in 1.2 ml of lysis buffer and proteins in supernatants and pellets were precipitated and prepared for immunoblotting.
Plasmid expression vectors and transfection
Constructs encoding the full-length form of human NECC1 fused to the C-terminus of GFP (green fluorescent protein) (GFP–NECC1) or cMyc epitope tag (cMyc–NECC1), and the short isoform of human NECC1 fused to the C-terminus of GFP (GFP–NECC1 S) have been described previously . A vector coding for the short isoform of human NECC1 fused to the N-terminus of GFP (NECC1 S–GFP) was created by inserting the PCR-amplified coding sequence of the short isoform of NECC1 lacking the stop codon into the NotI/SalI sites of the phrGFP-C vector (Stratagene). All plasmid vectors were sequenced. Plasmids coding for hGH (pXGH5) and GFP-tagged human NPY (neuropeptide Y; NPY–GFP) were kindly provided by Dr S. Gasman (University Louis Pasteur, Strasbourg, France) and Dr T. Lang (Max-Planck Institute for Biophysical Chemistry, Gottingen, Germany) respectively. Transient transfections of cells with plasmid vectors were performed using Lipofectamine™ 2000 reagent (Invitrogen) following the manufacturer's instructions. Transfected cells were used for experiments 24–72 h later. Cell viability of transfected cells was assessed using an MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] assay (Sigma) according to the manufacturer's instructions.
RNAi (RNA interference)
NECC1 silencing was induced by infection with a lentiviral vector leading the transcription of a shRNA (short hairpin RNA) for rat NECC1. To produce lentiviral particles, we employed the self-inactivating lentiviral transfer vector pLVTHM, the lentivirus packaging plasmid psPAX2 and the envelope plasmid pMD2G, which were kindly provided by Dr D. Trono (École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland) . The pLVTHM vector also includes the GFP coding sequence, which was used to detect infected cells. Two complementary DNA oligonucleotides were annealed to produce a double-stranded DNA fragment, which was cloned into the MluI and ClaI sites of pLVTHM. The sequences of the oligonucleotides were as follows: 5′-cgcgtccccGAGAAACGATGAACTGATGttcaagagaCATCAGTTCATCGTTTCTCTTtttttggaaat-3′ (sense) and 5′-cgatttccaaaaaAAGAGAAACGATGAACTGATGtctcttgaaCATCAGTTCATCGTTTCTCgggga-3′ (antisense). The sequence in capitals is the NECC1 target sequence corresponding to bases 1284–1302 of rat NECC1 mRNA (GenBank® accession number NM_001107391). As a negative control, we used the empty pLVTHM vector. Lentiviral particles were produced by co-transfection of pLVTHM, psPAX2 and pMD2G plasmids in HEK-293T cells. Recombinant lentiviruses were concentrated by ultracentrifugation (2 h, 24000 rev./min) and the infectious titre was determined by flow-cytometry analysis of GFP-positive HEK-293T cells. PC12 cells were infected 4 h after plating with lentiviral vectors at a MOI (multiplicity of infection) of 2 and then cultured for 4–6 days before analysis.
In addition to the lentiviral vector employed for down-regulation of NECC1, expression of this gene in PC12 cells was also silenced by siRNA (small interfering RNA) transfection. The sequence of the siRNA specific for NECC1 was 5′-GGATGATGACTTGGATGAAAGTCTA-3′ (bases 1662–1686 of rat NECC1 mRNA) (Invitrogen). PC12 cells were transfected with 75 pmol of siRNA duplex and 5 μl of Lipofectamine™ 2000.
hGH secretion assay
PC12 cells were plated at a density of 50000 cells/well in 12-well dishes. To evaluate the effect of NECC1 silencing on hormone release, PC12 cells infected either with the lentiviral vector pLVTHM or pLVTHM NECC1 shRNA were transfected with 400 ng of pXGH5. For overexpression studies, 800 ng of the expression vector cMyc or cMyc–NECC1 was transfected together with 400 ng of the hGH-encoding plasmid pXGH5. hGH release experiments were performed 72 h after transfection with the pXGH5 vector as described previously . hGH secretion is expressed as the percentage of total hGH present in the cells before stimulation.
FM5-95 incorporation analysis
Evaluation of secretory activity in single living PC12 cells with reduced levels of NECC1 expression, which were achieved by infection with lentiviral vectors or by transfection with siRNA, was carried out as described previously . PC12 cells were plated onto poly-L-lysine-coated 25-mm round coverslips. In the case of silencing by siRNA, cells were transfected 3 days after plating with 500 ng of the phrGFP-N1 vector alone or together with siRNA for NECC1. At 3 days after transfection or 6 days after infection either with the lentiviral vector pLVTHM or pLVTHM NECC1 shRNA, cells were rinsed with measurement medium (Dulbecco's modified Eagle's medium supplemented with 20 mM NaHCO3 and 1 mM Hepes) and incubated with 600 μl of medium containing 2 μM of the fluorescent dye FM5-95 (Invitrogen) for 5 min at 37°C. Coverslips bearing the cells were mounted in Sykes–Moore chambers (Belco Glass) and placed on the temperature-controlled stage of a fluorescence microscope fitted with a Plan-Fluor ×40 oil objective (Eclipse TE2000-E, Nikon Instrument). Fields containing transfected or infected cells were selected by GFP fluorescence. Then, cells were epi-illuminated at 540 nm every 1 min for a total of 22 min. After collecting a 10-min baseline, cells were exposed to 59 mM KCl by adding 600 μl of the measurement medium containing 2 μM FM5-95 and 118 mM KCl to the chamber. Image acquisition was controlled using Metafluor PC-software (Universal Imaging) and fluorescence emissions were captured using a CCD (charge-coupled device) camera (ORCA-BT-1024G, Hamamatsu Photonics) running in 1-bit mode. Regions of the same field devoid of cells were selected for continuous monitoring of the background, which was subtracted from the specific signal profile off-line. Before FM5-95 fluorescence monitoring, an image of GFP fluorescence of the regions of interest was captured for reference purposes. Time-course analysis of the intracellular FM5-95 fluorescence was calculated off-line using ImageJ 1.36 software. Secretory activity of each cell was represented by the AUC (area under the curve) of the corresponding kinetic of FM5-95 uptake after stimulation with 59 mM KCl.
NECC1 is localized at the Golgi apparatus in neuroendocrine cells
We have reported previously that NECC1 accumulates in juxtanuclear tubulo-vesicular structures when transfected in HEK-293 AD cells . As shown in Figure 1(A), immunostaining of PC12 cells, AtT20 cells or rat pituitary cells with the antibody raised against NECC1 revealed a juxtanuclear labelling, which is morphologically reminiscent of the appearance of the Golgi. The NECC1 immunosignal was reduced in NECC1 siRNA-transfected PC12 cells (Figure 1A), confirming the specificity of the immunoreaction. In addition, pre-adsorption of the anti-NECC1 antibody with its specific antigen (10−6 M) resulted in no staining (Figure 1A). The ability of the antibody to recognize NECC1 was confirmed in cMyc–NECC1-transfected HEK-293 AD cells, which do not express NECC1 endogenously (Supplementary Figure S1 at http://www.BiochemJ.org/bj/443/bj4430387add.htm).
To investigate the Golgi distribution of NECC1, we performed double-staining immunofluorescence using antibodies against the cis-Golgi marker GM130 or the TGN marker TGN38 and the anti-NECC1 antibody. Examination by confocal microscopy revealed a high degree of overlap of NECC1 with GM130 in both PC12 and AtT20 cells and with TGN38 in PC12 cells (Figure 1B). For PC12 cells, Manders' co-localization coefficients between NECC1 and GM130 and TGN38 were 0.638±0.075 and 0.449±0.025 respectively.
In order to assess the presence of NECC1 in other cellular compartments, NECC1 immunostaining was carried out on PC12 cells transfected with a construct coding for NPY fused to GFP (NPY–GFP), which has been shown previously to be sorted and packaged into secretory granules . Confocal microscopy studies showed co-localization of NECC1 and NPY–GFP in some NPY–GFP-containing vesicles (Manders' coefficient: 0.214±0.048). No significant co-localization was observed between NECC1 and the ER marker, the COPII (coatamer protein II) vesicle-coating component Sec23, or the early endosome marker EEA1 (Manders' coefficients were 0.143±0.054, 0.096±0.022 and 0.098±0.021 respectively) (Figure 1C).
Western blot analysis using the anti-NECC1 antibody revealed an immunoreactive band at 95 kDa in extracts from PC12 cells and rat pituitary (Figure 2A). This apparent molecular mass is close to the expected molecular mass of 96.5 kDa calculated from the full-length cDNA of NECC1 . Specificity of the immunoreactive band was confirmed by abolition of the labelling after pre-adsorption of the antibody with the purified antigen (Figure 2A), as well as after the silencing of NECC1 expression by RNAi (see below). Immunoblot analysis of protein extracts from cMyc–NECC1-transfected HEK-293 AD cells confirmed that the anti-NECC1 antibody reacted strongly with exogenous NECC1 and that the NECC1 cDNA encoded a protein of identical size as the endogenous NECC1 (Supplementary Figure S1).
To characterize NECC1 further, we performed subcellular fractionation studies. Analysis of the PNS, cytosolic fraction and crude membrane fraction from PC12 cells revealed a more intense NECC1 immunolabelling in the membrane fraction, which was also enriched in the Golgi proteins GM130 and TGN38 (Figure 2B), indicating that NECC1 is stably associated with intracellular membranes. The crude membrane preparation was then separated in a continuous sucrose gradient. Immunoblotting of the gradient fractions revealed that NECC1 co-distributes with GM130 in many of the fractions wherein this cis-Golgi protein separates (Figure 2C); indeed, as observed for GM130, NECC1 peaked in fraction 7. We also observed a partial overlap of NECC1 and other markers of the Golgi apparatus, including the TGN marker TGN38 and the ERGIC marker ERGIC53 (Figure 2C). Finally, a second pool of NECC1 partially overlapped with the granule marker CgA (Figure 2C). Together, our immunocytochemical and biochemical data indicate that NECC1 distributes across the Golgi apparatus and, to a lesser extent, in post-Golgi carriers. Unfortunately, the anti-NECC1 antibody was unable to detect the endogenous protein in immunocytochemical electron microscopy studies, thus precluding further analysis of NECC1 distribution at the ultrastructural level.
NECC1 is a peripheral membrane protein and behaves as a Golgi matrix component
To investigate the type of association that NECC1 maintains with Golgi membranes, total membranes from PC12 cells were treated with distinct compounds to differentially extract peripheral and integral membrane proteins. As controls, we examined the distribution of two reference proteins: TGN38, a transmembrane protein that is resistant to both 1 M NaCl and sodium carbonate, pH 11, but is sensitive to Triton X-100 extraction, and GM130 which, as a peripheral membrane protein, was extracted by high pH, but not by high salt conditions (Figure 3A). NECC1 behaved exactly as did GM130 (Figure 3A), indicating that this protein is associated with Golgi membranes as a peripheral protein. In addition, NECC1, as with GM130, is resistant to Triton X-100 extraction, whereas both proteins were extracted by SDS treatment (Figure 3A), indicating that it must be forming tight interactions with other Golgi proteins. Previous studies have proposed that the insolubility of GM130 and other golgins in Triton X-100 is a feature of Golgi matrix proteins [31,32]. It has also been shown that proteins composing the Golgi matrix relocate to punctuate cytoplasmic structures (i.e. remnants) upon treatment of cells with BFA . As shown in Figure 3(B), NECC1 was relocated to cytoplasmic remnants upon BFA treatment, whereas it recovered its juxtanuclear distribution after BFA washout, in a manner similar to that exhibited by GM130. Together, these data suggest that NECC1 is a bona fide component of the Golgi matrix.
Overexpression of NECC1 in PC12 cells
As shown in Figure 4(A), GFP–NECC1 exhibits a juxtanuclear distribution when transfected in PC12 cells. This distribution was found regardless of the amount of construct used for transfection (150 ng or 800 ng) or the time point examined (12, 24 or 72 h after transfection) (results not shown). In contrast with that observed for endogenous NECC1, the ectopically expressed protein showed no co-localization with GM130 or TGN38 (Figure 4A), although the GFP–NECC1 signal was in close proximity to that revealed by the anti-GM130 or the anti-TGN38 antibody. Confocal image rendering revealed that GFP–NECC1 fluorescent signal takes the form of ribbon-like tubules that wrap GM130 immunoreactivity (Supplementary Figure S2 at http://www.BiochemJ.org/bj/443/bj4430387add.htm). This distribution was observed irrespective of the reporter sequence used, GFP (Figure 4A) or cMyc epitope tag (Supplementary Figure S2B), the position of the reporter sequence, either at the N-terminus (Figure 4A and Supplementary Figure S2B) or the C-terminus (Supplementary Figure S2C) of NECC1, or the cell line employed, PC12 cells (Figure 4A), AtT20 cells (Supplementary Figure S2D) or HEK-293 AD cells .
Other Golgi long coiled-coil proteins have also been shown to become mislocalized when overexpressed . In this regard, it has been reported that overexpression of proteins can lead to the formation of aggresomes resulting from the accumulation of misfolded proteins . In particular, it has been shown that the aggresome localizes to the MTOC (microtubule organizing centre) and that nocodazole inhibits this process . In line with this, we observed that exogenous NECC1 accumulated close to γ-tubulin-labelled MTOC and that treatment of NECC1-transfected cells with nocodazole (10 μM for 2.5 h) abrogated the accumulation of NECC1, which appeared as small spots scattered throughout the cytoplasm (Supplementary Figure S2E). Together, our data indicate that overexpression of NECC1 caused the formation of aggresomes, and that this represents an intrinsic feature of this protein.
Given that NECC1 accumulations were located in close proximity to the Golgi apparatus, we investigated whether NECC1 expression could affect Golgi structure. Evaluation of Golgi morphology by GM130 immunofluorescence and image analysis revealed that the number of GM130-labelled structures (Golgi objects) per cell in cMyc–NECC1-transfected PC12 cells (n=35) was almost double that observed in mock-transfected cells (n=66) (9.1±1.3 compared with 5.1±0.3 respectively), although the average area of GM130 labelling was not altered (1905±165 compared with 1968±135 pixels2) (Figure 4B). However, transient transfection of NECC1 had no effect on cell viability of PC12 cells, as demonstrated using an MTT assay (Figure 4C).
In contrast with that found for the Golgi markers, co-localization studies revealed a partial overlap between exogenous NECC1 and the secretory granule markers CgA and exogenous hGH. Thus some CgA- or hGH-containing vesicles exhibited GFP–NECC1 labelling (Figure 4D), indicating that at least some NECC1 is not targeted to the aggresomal pathway. No co-localization was found between exogenous NECC1 and the COPII vesicle marker Sec23 or the early endosome marker EEA1 (Supplementary Figure S3 at http://www.BiochemJ.org/bj/443/bj4430387add.htm).
NECC1 is involved in the control of regulated secretion in neuroendocrine cells
The preferential expression of NECC1 in tissues containing cells with the regulated secretory pathway , together with the localization of the endogenous protein at the Golgi apparatus in neuroendocrine cells, led us to investigate the possible participation of NECC1 in the regulated secretion of cargo. To this end, we analysed the effects of silencing and overexpression of NECC1 on the secretory activity of PC12 cells by quantifying the release of a secretory reporter, the ectopically expressed hGH protein, after stimulation of the cells with a depolarizing pulse of KCl.
To reduce NECC1 expression, PC12 cells were infected with the lentivirus encoding NECC1 shRNA. Transient expression of NECC1 shRNA reduced NECC1 mRNA levels by 85%, as evaluated by real-time RT–PCR (results not shown), and lowered endogenous levels of NECC1 protein by 55%, as measured by quantitative immunoblotting (Figure 5A). As shown in Figure 5(B), a reduction in NECC1 expression significantly increased the release of hGH stimulated by KCl as compared with that observed in mock-infected PC12 cells. On the other hand, decreasing concentrations of NECC1 in PC12 cells did not significantly affect Golgi structure, as monitored by localization of endogenous GM130. Thus GM130 immunostaining and image analysis showed that the number of Golgi objects per cell (5.2±0.3 compared with 4.9±0.2) and the average area of GM130 labelling (1947±121 compared with 1932±149 pixels2) were not altered in NECC1-silenced cells (n=121) as compared with mock-infected cells (n=126).
In addition to the hGH secretion assay, which is based on measurement of hormone release by whole cell populations, we performed real-time analysis of the secretory activity in single living PC12 cells with reduced levels of NECC1. This was enabled by the use of a highly lipophilic probe (FM5-95) which has been shown to internalize in neuroendocrine cells during recycling endocytosis that follows fusion of secretory granules to the plasma membrane, the amount of internalized fluorophore being proportional to the number of secretory granules fused during the exocytotic event . Specifically, this analysis was carried out by using a protocol optimized in our laboratory, which is based on the measurement of the incorporation of the fluorescent probe FM5-95 in PC12 cells after stimulation of secretion by treating cultures with a depolarizing pulse of KCl . In this experimental setting, infection of cells with silencing lentiviral vectors led to a decrease in NECC1 mRNA and protein levels comparable with those obtained in the hGH secretion assays. Cells with reduced NECC1 expression showed a significant increase in FM5-95 uptake after KCl stimulation, the average AUC being 2-fold higher in NECC1-silenced cells than in mock- or non-infected cells (Figure 6A). This observation was confirmed by using a synthetic siRNA for NECC1, which is different in sequence to that produced by the lentiviral vector. Transfection of PC12 cells with NECC1 siRNA evoked a 55% decrease in NECC1 protein levels (Figure 6B). Video-microscopy analysis showed that, after KCl stimulation, dye incorporation was significantly higher in PC12 cells transfected with NECC1 siRNA than in mock- or non-transfected cells (Figure 6B). Accordingly, the AUC of FM5-95 uptake after stimulation in NECC1-silenced cells was increased by 44% with respect to that observed in mock-transfected cells (Figure 6B). In all, the results obtained from the FM5-95 incorporation studies are consistent with those obtained in the hGH secretion assays, thus providing support for the hypothesis that NECC1 acts as a negative player in the regulated secretory pathway in neuroendocrine cells.
Finally, we examined the effect of overexpression of NECC1 on hormone secretion from PC12 cells. To this end, PC12 cultures were co-transfected with the construct coding for the cMyc–NECC1 fusion protein and the plasmid expressing hGH. Immunoblotting of extracts from co-transfected cells confirmed that NECC1 levels were higher in cultures transfected with the cMyc–NECC1 vector than in mock-transfected cultures (Figure 7A). Quantification of hGH release revealed that increased levels of NECC1 induced a significant reduction in KCl-stimulated secretion of hGH (Figure 7B).
In the present study, we have identified the long coiled-coil protein NECC1 as a novel component of the Golgi apparatus in neuroendocrine cells. In addition, our functional studies suggest that NECC1 plays a role in the control of the regulated secretory pathway, providing the first evidence for the participation of a Golgi-associated long coiled-coil protein in this specialized secretory process.
Fluorescence microscopy using a specific anti-NECC1 antibody revealed that NECC1 locates at the Golgi apparatus in neuroendocrine cells. These observations, together with the structural properties of NECC1, indicate that this protein could be considered as a novel member of the golgin family. Nevertheless, whereas golgins normally localize to particular Golgi subdomains , NECC1 was found to distribute across the Golgi apparatus, as shown by our co-localization studies with the cis-Golgi marker GM130 and the TGN marker TGN38. It is plausible that the two distinct isoforms of NECC1, which are both expressed in PC12 cells and are recognized by the anti-NECC1 antiserum when exogenously expressed (results not shown), distribute at different Golgi subdomains and/or that NECC1 transits through the Golgi, which might represent a particular feature of this neuroendocrine-specific long coiled-coil protein. Moreover, NECC1 was also present, although to a minor extent, in post-Golgi carriers containing secretory cargo. In agreement with the microscopic data, our subcellular fractionation studies showed that NECC1 co-fractionated with GM130 as well as with markers of the TGN and secretory granules. Together, these data suggest that NECC1 may function in membrane traffic events that are directed toward the trans face of the Golgi.
Biochemical analysis showed that NECC1 is stably bound to Golgi membranes as a peripheral membrane protein. Nevertheless, our in silico search failed to identify in the NECC1 sequence motifs or domains known to be involved in the targeting of peripheral golgins to Golgi membranes, such as the GRAB (GRIP-related Arf-binding) or the GRIP domains [19,37], the ALPS (amphipathic lipid-packing sensor) motif  or the GRASP65-interacting motif of GM130 . On the other hand, based on the use of bioinformatic tools, we have suggested previously that the 19 amino acid hydrophobic region at the C-terminus of NECC1 could anchor this protein to Golgi membranes . However, our biochemical data argue against the view that this region acts as a transmembrane domain. It is plausible that NECC1, as other peripheral membrane golgins, is recruited from the cytosol to Golgi membranes via binding to a membrane-associated small GTPase of the Rab, ARF or ARL families [7,38,39].
Our studies in PC12 cells have shown that NECC1 has properties of a Golgi matrix protein. In particular, NECC1 was shown to be resistant to Triton X-100 extraction and to relocate to cytoplasmic remnants upon BFA treatment. NECC1 shares these features with other Golgi long coiled-coil proteins, such as GM130, p115, GMAP210, giantin or golgin-160 [3,31,32,40]. It has been proposed that long coiled-coil proteins of the Golgi form a detergent-insoluble proteinaceous structure on the cytosolic side of the Golgi membranes, referred to as the Golgi matrix, which has been suggested to act as a structural scaffold of this organelle [3,9]. In fact, depletion by RNAi of GM130, p115, GMAP210 or golgin-160 induces an extensive fragmentation of the ribbon-like Golgi appearance [14,16,41]. In spite of its similarities with these components of the Golgi matrix, knockdown of NECC1 by RNAi did not alter Golgi morphology when assayed by GM130 immunofluorescence in PC12 cells. These data suggest that NECC1 is not essential for maintaining Golgi structure. Nevertheless, we cannot exclude the possibility that the degree of silencing of NECC1 reached in the present study might not have been sufficient to elicit noticeable changes in Golgi morphology or, alternatively, that NECC1 function could be compensated for by other Golgi components in the silenced cells. On the other hand, overexpression of NECC1 resulted in an increase in the number of GM130-labelled structures. It must be stressed that our transfection studies indicate that exogenous NECC1 accumulates, at least in part, in aggresomes that distribute in close proximity to the Golgi, and that this phenomenon is independent of the cell type, the tag employed or the location of the NECC1 sequence relative to the tag. Similar observations have been reported for the Golgi-associated long coiled-coil protein optineurin, whose overexpression also induces Golgi fragmentation [42,43]. In all, our findings indicate that the presence of high levels of NECC1 alters, either directly or indirectly, Golgi morphology and, therefore, that maintenance of an adequate level of this protein may be important for the organization of the Golgi.
In addition to regulating Golgi structure, golgins have been reported to play major roles in secretory trafficking at the Golgi apparatus [9,10]. The intracellular distribution of NECC1, together with its preferential expression in tissues containing cells equipped with a regulated secretory pathway , has suggested a potential role for NECC1 in the transport of regulated secretory cargoes across the Golgi apparatus in cells specialized in secretion. In support of this proposal are our data on the effects of NECC1 silencing on hormone secretion. Thus K+-induced hGH release by PC12 cells was significantly increased under conditions of decreased NECC1 expression levels, suggesting that NECC1 could limit hormone transport. Further support for an inhibitory role of NECC1 in regulated secretion has been provided by the present study on FM5-95 uptake in single living PC12 cells with diminished NECC1 levels. These functional data are in agreement with our previous findings in primary endocrine cells, i.e. pituitary melanotropes, wherein an inverse relationship between NECC1 expression levels and hormone secretion was observed . These results suggest that NECC1 acts as a negative regulator of secretory cargo trafficking at the Golgi apparatus. Interestingly, NECC1 overexpression reduced the hormone-releasing capacity of PC12 cells in response to K+ stimulation, without affecting cell viability, yet this effect could be secondary to the Golgi structure perturbation caused by overexpression of the protein and/or due to the mislocalization of exogenous NECC1.
The present study is the first to provide experimental evidence supporting an inhibitory function for a Golgi-associated long coiled-coil protein in the regulated transport of proteins. Remarkably, the idea of the existence of long coiled-coil tethering proteins acting as negative regulators of membrane fusion events, by providing transient binding sites for vesicles in order to limit their diffusion or by keeping membranes apart, has already been proposed [3,10]. In this scenario, NECC1 could bind transport vesicles containing secretory cargo at the Golgi to reduce their traffic and/or impair cisternal progression. As described above, NECC1 is expressed in cells displaying an elevated rate of secretory traffic of specific cargoes, endocrine cells and neurons. Hence it is plausible that NECC1 participates in the sorting and/or segregation of cargoes for regulated secretion (i.e. peptidergic hormones) at the Golgi apparatus, thereby controlling their flux through this organelle. Indeed, both cis-Golgi and TGN golgins have been shown to be involved in the sorting and cell-surface delivery of secretory cargo [44–46]. We cannot exclude the possibility that NECC1 may be involved in functions other than vesicle tethering to regulate protein traffic across the Golgi in neuroendocrine cells. Thus golgins have been proposed to act as molecular scaffolds by virtue of their ability to interact with a variety of binding partners . Future identification of the molecular factors interacting with NECC1 in neuroendocrine cells will help to determine its precise role in intracellular protein transport at the Golgi and provide novel insights into the regulation of the secretory pathway in cells specialized in secretion.
In conclusion, our data support the hypothesis that NECC1 plays a role in the control of hormone traffic across the Golgi in neuroendocrine cells. Thus, by reducing the transit of newly made secretory cargo through the Golgi, NECC1 could contribute to finely tune the amount of hormone available for secretion upon stimulation of the regulated secretory pathway.
David Cruz-García, Alberto Díaz-Ruiz, Yoana Rabanal-Ruiz, and Juan Peinado performed the experiments and analysed the data; Marie-Christine Tonon and Hubert Vaudry were responsible for antibody preparation; Maité Montero-Hadjadje prepared the NECC1 constructs and collaborated in the transfection experiments; Youssef Anouar, Hubert Vaudry, Francisco Gracia-Navarro and Justo Castaño provided critical insight into data interpretation; Rafael Vázquez-Martínez and María Malagón conceived and supervised the study, and were primarily responsible for preparing the paper, with the assistance of the other authors.
This work was supported by the Ministry of Science and Innovation/FEDER of Spain [grant numbers BFU2007–60180 and BFU2010–17116 (to M.M.M.)] and Junta de Andalucía/FEDER [grant number CTS-03039 (to M.M.M.)].
We thank all the scientists who generously provided us with plasmids and recombinant proteins (see the Experimental section). We thank Dr. Jérôme Leprince (University of Rouen, France) for kindly providing the NECC1 peptide for rabbit immunization.
Abbreviations: ARF, ADP-ribosylation factor; ARL, ARF-like; AUC, area under the curve; BFA, brefeldin A; CgA, chromogranin A; COPII, coatamer protein II; EEA1, early endosome antigen 1; ER, endoplasmic reticulum; ERGIC, ER–Golgi intermediate compartment; GCC, Golgi coiled-coil; GFP, green fluorescent protein; GM130, 130 kDa Golgi matrix protein; GMAP210, Golgi-associated microtubule-binding protein 210; GRASP, Golgi reassembly stacking protein; GRIP, golgin-97/RanBP2/Imh1p/p230; HEK, human embryonic kidney; hGH, human growth hormone; MTOC, microtubule organizing centre; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide; NECC1, neuroendocrine long coiled-coil protein 1; NPY, neuropeptide Y; PNS, post-nuclear supernatant; RNAi, RNA interference; shRNA, short hairpin RNA; siRNA, small interfering RNA; TGN, trans-Golgi network
- © The Authors Journal compilation © 2012 Biochemical Society