Secretogranin III Directs Secretory Vesicle Biogenesis in Mast Cells in a Manner Dependent upon Interaction with Chromogranin A¹ Free

The granin family (Chromogranin A (CgA),⁴ Chromogranin B (CgB), Secretogranin (Sg) II, and the less well studied Secretogranins III-VII) comprises a group of acidic proteins that are present in the secretory granules of a wide variety of endocrine and neuro-endocrine cells (1, 2, 3, 4). The molecular functions of the granins are poorly understood. There is evidence that granins may be the precursors of biologically active peptides (3, 5, 6, 7, 8, 9). However, granins cannot be considered simply as precursors of granule cargo. They may act as helper proteins in the packaging of peptide hormones and neuro-peptides (3, 5, 6, 7, 8, 9). The EF-hand containing CgA and CgB may also act as high capacity, low affinity, calcium binding proteins that are closely associated with calcium-regulated intracellular ion channels (10, 11, 12, 13). Moreover, emerging evidence suggests that these proteins play an active role in directing the biosynthesis of secretory granules. This biological role for granins has been proposed on the basis of work in neuroendocrine cells, where several granin proteins are expressed and have been shown to cooperatively contribute to secretory vesicle biogenesis (14, 15). This cooperation is exemplified by the interaction between SgIII and CgA (15, 16). In this study, an intermolecular interaction between the two granins is needed for their targeting to secretory granules. Courel and coauthors (15) have described a “granulogenic domain” in CgA that promotes the biogenesis of secretory granules probably through exerting control over cholesterol sequestration and the dynamics of intracellular vesicle budding/fusion. The potential contribution of granins to the granularity of secretory immunocytes has not been addressed.

Mast cells are densely granular, secretory leukocytes that reside in tissues (17, 18). In response to a variety of stimuli, including immunological challenge, exposure to secretagogues and exposure to physical stimuli, mast cells release preformed, and de novo synthesized, inflammatory mediators into their surrounding tissues. The preformed mediators are housed within membrane-delimited secretory granules. The contents of mast cell secretory granules include inflammatory mediators such as histamine, serotonin, platelet activating factor, some cytokines (TNF-α), and matrix-active proteases of the chymase and tryptase families (17, 18). Together, these mediators exert diverse proinflammatory effects upon tissue. These include the promotion of vasodilation, induction of inflammatory pain, and recruitment of other immunocytes to the inflammatory site. The central role of these granules in inflammation has led to an intense interest in the signaling mechanisms that control granule exocytosis, because targets for anti-inflammatory therapeutics could lie within this system. Moreover, the mechanisms that control the biogenesis and abundance of these granules are also of therapeutic interest, but remain poorly studied.

Mast cell granules are structurally, and functionally, related to the dense core vesicles of neuro-endocrine cells and neurons. There has been no published analysis of the expression, or function, of granins in the mast cell system, although their involvement in secretory vesicle biogenesis in other neuroendocrine cells and neurons has been established (6, 7, 14, 19, 20, 21, 22, 23). In the current report, we tested the idea that SgIII drives the production of mast cell secretory granules. Moreover, we propose that its interaction with CgA is critical to this biological role. Although SgIII and CgA proteins are abundant in various mast cells, we show that SgIII is depleted from the cytosol after stimulation of mast cells via either pharmacological or immunological methods. Moreover, SgIII levels in mast cells recover during the poststimulation refractory phase during which mast cells replenish their granule population. Over-expression of SgIII in mast cells causes an increase in the number of mast cell granules, and thus when SgIII-expressing mast cells are stimulated we note an increase in the release of proinflammatory mediators. Our results also confirm that CgA interacts with SgIII via a specific granulogenic domain and that the granulogenic activity of SgIII in mast cells depends upon its ability to form a functional interaction with CgA. Although the enigmatic granin proteins should therefore be considered as components of the secretory apparatus in mast cells, our data also suggest potentially wider functions as regulators of organellar ion channels.

RBL2H3, HEK293, and PC12 cells were maintained in DMEM (Mediatech), 10% FBS (Mediatech), and 2 mM glutamine (Invitrogen) in humidified 5% CO₂ at 37°C. LAD2 cells were a generous gift from Dr. A. Kirshenbaum (National Institute for Allergic and Infections Diseases, Bethesda, MD; Ref. 24). LAD2 cells were maintained in DMEM with 2% FBS, 2 mM Gln and rSCF at 100 ng/ml (PeproTech). Growth medium for stably transfected cells were supplemented with appropriate antibiotics; 200 μg/ml zeocin (Invivogen), 5 μg/ml blasticidin (Invitrogen), 200 μg/ml hygromycin (Invivogen).

Full-length human Secretogranin III cDNA (Open Biosystems) was obtained by PCR and cloned into the pcDNA4/TO and pcDNA5/TO GFP (A. Stokes, unpublished observations) vectors using NotI/KpnI or EcoRV/NotI digestion respectively. A V5 epitope tag was incorporated into the carboxyl terminus of SgIII during this PCR. A truncated SgIII cDNA (residues 1–210) was obtained by PCR and subcloned into pcDNA5/TO empty vector and pcDNA5/TO GFP vector (A. Stokes, unpublished observations). All constructs were verified by sequencing. The pcDNA5/TO TRPA1 construct was generated as described previously (25).

For transient transfection, HEK293 cells were seeded and grown until 50% confluent. All cDNA was purified using a QIAfilter Plasmid Maxi Kit (Qiagen). The cDNA quality was evaluated using a spectrophotometer and was used in transfection only if O.D. 260/280 >1.7. HEK293 cells were transiently transfected using reagent LT1 (Mirus). Serum-free DMEM (100 μl) and LT1 reagent (10 μl) were added together and vortexed for 2 min. The mixture was then left at room temperature for 15 min. After 15 min, 7.5 μg of V5-SgIII cDNA was added to the mixture and mixed gently. After 15 min at room temperature (RT), the mixture was added to the cells in a drop-wise manner while swirling the plate. Cells were incubated for 48–72 h and then harvested and lysed for Western blots.

For stable transfection, the RBL2H3 and PC12 cell lines were electroporated with V5-SgIII, V5-SgIII-GFP, V5-ΔSgIII, and V5-ΔSgIII-GFP constructs. Cells were electroporated with 15 μg of DNA in 500 μl of 1× DMEM, 10% FBS, and 2 mM glutamine at 280 V and 950 μF in a 4 mm path length cuvette (BioRad). Clonal cell lines were selected by limiting dilution in 96-well plates. Selection was conducted in zeocin (V5-SgIII construct) or hygromycin (V5-SgIII-GFP, V5-ΔSgIII, and V5-ΔSgIII-GFP constructs).

For production of TRex HEK293 cells with inducible expression of SgIII constructs, parental cells were electroporated with 15 μg of plasmid DNA in a 500 μl volume of DMEM/10% FBS/2 mM glutamine. Electroporations were performed at 280 V/950 μF in a 4-mm path length cuvette (BioRad). Clonal cell lines were selected by limiting dilution in 400 μg/ml zeocin and 5 μg/ml blasticidin, and screened by immunoprecipitation and Western blot. Transfected gene expression was induced using 1 μg/ml tetracycline for 16 h at 37°C.

Whenever stable cell lines were produced, steps were taken to avoid the possibility that experiments would be performed on progeny cells derived from an original parental clone. After electroporation, a brief (12–14 h) recovery period for the cells was used such that no discernable cell division took place. Recovered cells were trypsinized to a single cell suspension to disperse any potential daughter cell pairs. The trypsinized cell suspension was subjected to a limiting dilution (plated in a 96-well plate at less than one cell per well, (i.e., 0.5 cells per 250 μl medium)). Each well is therefore unlikely to have more than one cell, but of course across the 960 wells plated there will occasionally be wells where more than one cell is dispensed. After 3 wk, typically 5–15 wells on a given 96-well plate had visible colonies (i.e., colonies at the 100–200 cell level). Before these colonies were harvested and transferred to a larger culture vessel, the wells were inspected microscopically and any wells in which more than one colony appears were discarded. After harvesting, great care was taken to avoid cross-contamination of the clonal cell lines. Depending on the efficiency of the transfection, we observed between five to fifteen clonal colonies per 96-well plate. Between five and ten individual clones were selected at random, expanded, and cryopreserved. These clones were screened for expression of the transfected gene by immunoprecipitation and Western blot. Experiments were performed on three independent, randomly selected clones, but results from one clonal cell line are presented throughout this article.

Affinity-purified goat polyclonal anti-Secretogranin III (Santa Cruz Biotechnology) was used to study SgIII protein expression. Rabbit polyclonal anti-Secretogranin I (Abcam) and affinity purified goat polyclonal anti-Secretogranin II (Santa Cruz Biotechnology) were used to perform comparative studies. Rabbit monoclonal anti-serotonin (Sigma-Aldrich) and mouse monoclonal anti-V5 (Serotec) were used for immunostaining and immunoblotting. Mouse anti-CgA (Spring Bioscience) was used to study the interaction of SgIII and CgA. Anti-FLAG M2 was from Sigma-Aldrich. The anti-TRPA1 has been described previously (26). IgE anti-DNP, DNP-BSA, PMA, and ionomycin were all from Sigma-Aldrich.

Cells were pelleted (2000 × g, 2 min) and washed once in PBS at room temperature. Approximately 10⁷ cells were lysed on ice for 20 min in 500 μl of lysis buffer (100 mM HEPES (pH 7.4), 150 mM NaCl, 80 mM NaF, 20 mM iodoacetamide, 50 mM PMSF (phenyl methyl sulfonyl fluoride), 500 mg/ml aprotinin, 1.0 mg/ml leupeptin, and 2.0 mg/ml chymostatin, 1% (weight-to-volume ratio, w/v) IGEPAL) dissolved in distilled water. Lysates were then clarified (10000 × g, 5 min at 4°C) to remove nuclei. For preparation of total protein, lysates were acetone precipitated and again centrifuged at 10000 × g for 5 min at 4°C. For immunoprecipitation, supernatants were tumbled (4°C/2 h) with the indicated Ab, followed by capture of immunocomplexes using ∼50 μg Protein A conjugated to agarose beads (Sigma-Aldrich). Excess liquid was aspirated with a Hamilton syringe. Samples were then boiled (8 min) in reducing sample buffer (20% (v/v) glycerol, 62.5 mM Tris-HCl (pH 6.8), 0.05% (w/v) bromophenol blue, 2 mM 2-mercapto-ethanol) and then resolved by 10% SDS-PAGE for 16 h.

After being resolved on SDS-PAGE, the proteins were electro-transferred to Polyvinylidene fluoride membrane (1.4 A for 180 min in 25 mM Tris base-HCl, 192 mM glycine (pH 8.63)). Membranes were blocked using 5% (w/v) nonfat milk (1 h at RT). Membranes were incubated with primary Abs (Ab dissolved in TBS, 0.05% Tween, 0.05% NaN₃, and 0.5% w/v BSA) for 16 h at 4°C. Membranes were washed with TTBS (4 × 5 min) and incubated with secondary Abs (HRP-conjugated IgG) for 1 h. The membranes were washed with TTBS (4 × 5 min) and then incubated with ECL Plus Western Blotting Detection Reagent (Amersham Biosciences). Emitted light was then detected using autoradiography film (Kodak Scientific Imaging).

Protein samples were matched for total protein levels on the basis of a colorimetric protein determination assay, the D_c Protein Determination Kit (BioRad) according to the manufacturer’s instructions. Color development was read at 710 nm in a Benchmark Plus (BioRad) plate reader.

Adherent RBL2H3 and RBL-SgIII (1.25 × 10⁵ cells/cm²) were incubated with 1 μg/ml IgE–anti DNP and 0.5 μCi/ml [³H]serotonin for 16 h at 37°C. Monolayers were then washed with normal medium at 37°C (2×) and cells were incubated with 10 μl of indicated stimuli (1 μM PMA, 10 μM ionomycin, or 10% Triton X-100) or vehicle (PBS) in 190 μl/cm² medium for 1 h at 37°C. Reactions were stopped on ice. The supernatant was centrifuged (10,000 × g for 30 s) and 100 μl of supernatant were counted in liquid scintillation mixture (Scintiverse BD mixture, Fisher Scientific). Release in response to PMA/ionomycin was between 60 and 80% of the Triton X-100 releasable signal. Signals from FcεRI stimulation approached 30–50% of the Triton X-100 releasable [³H] serotonin.

Cells were seeded onto glass coverslips. The cells were fixed with pure methanol at −20°C for 5 min. They were then washed with PBS, and then blocked with 0.7% (w/v). Fish skin gelatin (prepared in PBS) for 20 min. The cells were incubated with primary Ab diluted in the blocking reagent (0.7% (w/v) fish skin gelatin) for 1 h. The cells were washed four times with PBS and then incubated with Alexa fluorophore conjugated Abs (Invitrogen) prepared in the blocking reagent for 20 min. Cells were washed four times with PBS and then finally the cover-slips were dipped in water and mounted on microscope slides using Crystal mount (Biomedia). Imaging was performed with an Olympus IX70 fluorescence inverted microscope with quadruple dichroic filter block and excitation filter set 88000 (Chroma), connected to an F-view monochrome CCD camera.

Cells were seeded onto glass cover-slips. The cells were first fixed with pure methanol at −20°C for 5 min. They were then washed with PBS, followed by incubation for 30 min with 0.5% (w/v) Toluidine blue (J. T. Baker, Phillipsburg, NJ) dissolved in distilled water. Coverslips were washed 4 times with PBS and mounted using Crystal mount (Biomedia).

Total RNA was extracted from adherent RBL2H3 cells that were either left untreated or exposed to IgE (1 μg/ml for 16 h followed by two washes) followed by 250 ng/ml DNP-BSA for the time indicated. Total RNA was purified using a Nucleospin RNA II kit according to the manufacturer’s instructions (BD Biosciences). Before hybridization on the BD Atlas Rat 1.2 Arrays, total RNA samples were labeled with [α-³³P]dATP using the Atlas Pure Total RNA Labeling System (BD Biosciences). A set of matched BD Atlas Rat 1.2 Array blots were hybridized (68°C, 16 h). Membranes were washed four times in 2 × SSC/1% SDS (68°C, 15 min), then once in 0.1 × SSC/0.5% SDS (68°C, 15 min) and once in 2 × SSC (RT, 5 min). Blots were exposed to storage phosphor screens (Packard Biosciences) for 8 days. Data were captured using a Cyclone System (Packard Biosciences). BD Atlas Image 2.7 (BD Biosciences) software was used to assign genes to specific coordinates on the array membranes, to correct for background counts, to establish the threshold for positive gene expression (2 × background level counts + 10%), and to normalize between-blot hybridization intensity differences based on the median-signal intensity. Per gene differences in mRNA levels were expressed as a fold-change relative to the corresponding unstimulated controls.

Fura-2 AM ester (Molecular Probes) loading of intact cells was performed by incubating cells for 45 min in a standard modified Ringer’s solution of the following composition (in mM): NaCl 145, KCl 2.8, CsCl 10, CaCl₂10, MgCl₂2, glucose 10, HEPES · NaOH 10 (pH 7.4), supplemented with 5 μM fura-2-AM (Molecular Probes). Cytosolic calcium concentration was monitored at a rate of 5 Hz with a photomultiplier-based system using a monochromatic light source that excited fura-2 at 360 and 390 nm for 20 ms each. Emission was detected at 450–550 nm with a photomultiplier whose analog signals were sampled and processed by the X-Chart software package (HEKA Electronics). Fluorescence ratios were translated into free intracellular calcium concentration based on calibration parameters derived from patch-clamp experiments with calibrated calcium concentrations. Osmolarity was maintained at 300 mOsm. Local perfusion of individual cells was achieved through a wide-tipped, pressure-controlled application pipette (3-μm diameter).

Experiments are n = 3 unless otherwise indicated. Results are shown as the mean ± SD. Statistical significance was determined based on a two-way ANOVA (Student’s t test). Adjacent to data points in the respective graphs, significant differences were recorded as follows: *, p < 0.05; **, p < 0.01; ***, p < 0.001; no symbol, p > 0.05.

SgIII and CgA are considered to be components of the secretory vesicle biogenesis system in neuroendocrine cells. Because the secretory vesicle system in mast cells shares multiple features with the neuroendocrine system we asked whether these proteins are expressed in mast cells. We assayed the presence of SgIII and CgA by Western blot in rodent and human mast cells. The data in Figs. 1, A and B show an expression survey for Secretogranin III and CgA. IGEPAL lysates were prepared from the following mast cell lines RBL2H3 (rat basophilic leukemia), P815 (mouse mastocytoma), mouse bone marrow derived mast cell, and LAD2 (human mast cell leukemia). LAD2 cells exhibit a high MW (>75 kDa) form of SgIII. This variation in apparent MW of SgIII may reflect an, as yet undefined, posttranslational modification. The size variance is unlikely to be attributatble to species differences because HEK are derived from humans, as are the LAD2. The neuroendocrine PC12 cells were used as positive controls for SgIII expression (Fig. 1,C). SgIII protein is present in each of the mast cell lines tested. SGIII was also detected in CAD catecholaminergic neurons, RIN5MF pancreatic β cells and the Jurkat T lymphocyte line (data not shown). We also noted that in some cell lines SgIII migrated as a single species on SDS-PAGE (Fig. 1 A), supporting the data of Ottiger and coauthors (25) who suggested that SgIII, unlike other granin family members, is not inevitably proteolysed to form bioactive peptides. A high MW (75–80 kDa), uncleaved, version of SgIII that binds to CgA has recently been reported by the Hosaka laboratory (27).

We hypothesized that SgIII is a component of the mast cell secretory granule. We reasoned that if so, then SgIII levels in the mast cells may be altered after stimulation of mast cells using agents that evoke secretory responses. We assayed the levels of SgIII protein in mast cells treated with secretion-evoking pharmacological stimuli or antigenic stimuli. The presence of SgIII was assessed by Western blot of matched cytosolic extracts from unstimulated and stimulated RBL2H3. Our results show that resting cells express SgIII protein, which is markedly depleted at 1–2 h after stimulation with PMA/Ionomycin (Fig. 2,A) or via antigenic cross-linking of the FcεRI (Fig. 2,B). Control experiments suggest that this is not a protein degradation event as secreted SgIII can be detected by Western blot in the extra-cellular milieu of resting mast cells (Fig. 2,C). We have also detected full-length endogenous SgIII in the supernatants of resting RBL2H3 but at very low levels (data not shown). SgIII protein starts to reappear at 4 h after stimulation and this expression stabilizes by 24 h after stimulation, and re-attains resting levels (Fig. 2, A and B). The depletion in SgIII levels is consistent with the idea that this protein is a component of the mast cell secretory granule. Other granule components such as chymase and tryptase show a distinctive pattern of depletion and then replenishment in activated mast cell populations (28). SgIII appears to behave in a similar fashion. Moreover, the recovery in SgIII protein is apparently due to a transient transcriptional up-regulation of the SGIII gene (Fig. 2 D) following mast cell stimulation.

The granulogenic properties of granins have been established in neuroendocrine cells on the basis that over-expression of these proteins is necessary and sufficient to cause the appearance of a novel secretory vesicle compartment. We evaluated the secretory compartment of RBL2H3 mast cells with stable over-expression of SgIII using several methods. First, we used an immunofluorescence approach in RBL2H3 mast cells, HEK cells, and PC12 neuroendocrine cells. We compared each of these cell types with a matched clonal cell line in which V5-SgIII is over-expressed (Fig. 3,A). Interestingly, we observed that V5-SgIII over-expression in each cell type was accompanied by the appearance of a prominent vesicle population that could be immunodecorated with anti-V5 (Fig. 3 A). This vesicular organelle was most prominent in the secretory RBL2H3 and PC12 cells, while the SgIII-associated structures in HEK293 were relatively smaller and less numerous.

The vesicular structures associated with SgIII over-expression in RBL2H3 cells did not immunodecorate with Abs toward markers of the Golgi apparatus (anti-Golgi-58K), lysosomes (anti-LAMP2), endoplasmic reticulum (anti-calreticulin), or peroxisomes (anti-PMP70) (data not shown). We therefore hypothesized that these structures may represent secretory granules in mast cells. Because serotonin is a marker for secretory granules (29, 30) we asked whether SgIII-associated vesicles could be counterstained with an anti-serotonin Ab. We noted that these vesicles contained significant serotonin, which colocalized with V5-SgIII, indicating that they may be bona fide granules capable of packaging inflammatory mediators in mast cells (Fig. 3,B). We also detected some subplasma membrane staining with anti-serotonin, which may reflect its presence in a subpopulation of small, endogenous, exocytotic vesicles. Further, we performed staining on mast cells with toluidine blue, a metachromatic dye that detects the numerous sulfated groups in mast cell granules. RBL2H3 with over-expression of SgIII displayed a large number of toluidine blue-positive granules that were not observed in control cells (Fig. 3 C), again suggesting that SgIII over-expression leads to an increase in the secretory vesicle load of mast cells.

With the idea that V5-SgIII over-expressing granules could package inflammatory mediators such as serotonin, we hypothesized that these V5-SgIII over-expressing mast cells may exhibit enhanced secretion of proinflammatory mediators. In V5-SgIII over-expressing RBL2H3 resting cells, there is a significant increase in basal efflux of [³H]serotonin (Fig. 3,D). Together with the fact that the detergent-releasable amount of [³H]serotonin is also enhanced in V5-SgIII over-expressing cells (Fig. 3,E), these data suggest that there may be an expansion of the compartment storing [³H]serotonin in these cells. Indeed, the V5-SgIII cells exhibit enhanced secretory responses to pharmacological stimulation (Fig. 3 F), suggesting that the vesicles arising from V5-SgIII over-expression may be functionally equivalent to mast cell granules. The potential for V5-SgIII over-expression to induce apparent granulogenesis in mast cells may speak to a role for endogenous SgIII in directing this process physiologically.

Taken together, our data indicate similarities in the role of SgIII in two major physiological secretory systems: the neuroendocrine cells and mast cells. In neuroendocrine cells, over-expression of CgA has been shown to be granulogenic, as is over-expression of SgIII. Moreover, in neuroendocrine cells the granulogenic properties of CgA are critically dependent upon its interaction with SgIII. Hosaka and coauthors (16) showed that the SgIII (214–373) and CgA (48–111) domains mediate the functional interaction between CgA and SgIII in mouse corticotrope-derived AtT-20 cells. We asked whether SgIII and CgA interact functionally in the mast cell system, and we produced a truncated form of SgIII that lacks the putative CgA-interacting (and granulogenic) domain (Fig. 4,A). We were able to coimmunoprecipitate CgA and SgIII in mast cells (Fig. 4,B). Our data indicate that removal of residues 211–468 of SgIII (Fig. 4,A) resulted in a loss of interaction with CgA (Fig. 4 B).

We asked whether full-length and the SgIII (1–210) truncation mutant were equivalent in their ability to induce the formation of a novel granule population in mast cells. RBL2H3 were transfected with either V5-SgIII or V5-SgIII (1–210). The presence of SgIII-associated vesicle population in the cytosol of these cells was assessed by anti-V5 immunofluorescence. These data (Fig. 5) show that the over-expression of the truncated V5-SgIII (1–210) did not result in the appearance of a prominent vesicular compartment. Diffuse, faint staining of the truncated V5-SgIII was observed, which quenched rapidly. Moreover, a GFP-tagged version of SgIII (1–210) with truncation of CgA binding domain showed a lack of vesicular organization (data not shown). Thus, the granulogenic potential of SgIII in mast cells appears to be dependent upon an intact SgIII C terminus (residues 211–468), which is the region that mediates interaction between SgIII and CgA.

The data presented above suggest that SgIII in mast cells participates in an interaction with CgA, and that the CgA-SgIII complex plays an, as yet poorly understood, role in directing granule biosynthesis. However, there are precedents for the interactions of granin proteins with nongranin binding partners, including the inositol (1, 4, 5) trisphosphate receptor, and intracellular calcium release channel (10, 11, 12, 13, 31). We performed an interaction trap analysis (a Sos-Recruitment Screen yeast two-hybrid (Stratagene) performed using the amino-terminal cytoplasmic tail of rat TRPA1 as bait and a rat brain cDNA library a target, data not shown), which suggested that the ion channel TRPA1 may form a protein complex with SgIII. Because the work of several groups does suggest that ion channels in intracellular membranes may interact with certain granins, we asked whether this is a physiological interaction in mast cells. Fig. 6, A and B show that TRPA1 and SgIII can be coimmunoprecipitated in both an HEK over-expression system and in RBL2H3 mast cells. It is clear that a small proportion of the over-expressed, or endogenous, TRPA1 can be immunoprecipitated in a complex with SgIII. Similarly, we find that of the multiple mobility forms of SgIII that are present in cells over-expressing V5-SgIII, only the lowest apparent m.w. (perhaps corresponding to the least processed, immature form of SgIII) species forms a complex with TRPA1. The functional significance of this remains unclear, but we note that TRPA1 predominantly localizes to vesicular structures in resting mast cells (Fig. 6,C), although these immuno-decorate only weakly with anti-SgIII (data not shown). The interaction between TRPA1 and SgIII clearly has functional implications for the channel however, because intracellular calcium release responses evoked by the TRPA1 ligand icilin are compromised in cells over-expressing SgIII (Fig. 6 D). These data present the first hint of a wider role for SgIII in mast cells, and their implications will be discussed below.

Granin family members are present in the secretory granules of a wide variety of endocrine and neuroendocrine cells (1, 3, 4, 32). Granins are involved in granulogenesis in neuroendocrine cells (6, 7) but there has been no published analysis of expression or function of granins in the mast cell system. Our data demonstrate the expression of granins in mast cells, and imply an important role for these proteins in this secretory immunocyte.

The biological roles of granin proteins are not well understood. Several ideas, which are not mutually exclusive, have been proposed for the molecular functions of granins. It is clear that CgA and CgB act as cargo precursors that are proteolytically cleaved to form bioactive peptides. In addition to being sources of granule content in the form of proteolytically derived peptides, Chromogranins have also been described to chaperone secretory granule components (1, 3, 9, 33, 34). Chromogranins have also been described as cholesterol binding proteins, and a model for their role in the cholesterol sequestration events that participate in granulogenesis has been derived from several studies (35). One common observation for several granins suggests that their over-expression result in expansion of the granule compartment (6, 15, 20, 22). Both their proposed roles as precursors of granule cargo, and as regulators of granule membrane dynamics, would be consistent with these observations of granulogenic capacity. In a further, intriguing, set of publications, the CgA and CgB proteins have been described as low affinity, high capacity, calcium buffering proteins that associate with intracellular calcium channels in the ER and Golgi apparatus. In this context, the calcium reservoir that is provided by CgA appears to play a feedback role in regulating the open status of these calcium permeant channels (10, 11, 12, 13, 31).

Secretogranin III has not been extensively investigated, and is only weakly similar to CgA/CgB at the sequence level. Our data, and that from other laboratories, allow us to make a number of statements about SgIII. First, it is heavily sulfated, which is consistent with it being a component of the mast cell secretory granule (Ref. 36 and Prerna Prasad, unpublished observation). Second, SgIII is not extensively proteolyzed to form peptides in resting mast cells (this article and Ref. 25). These data suggest that its major physiological role in this system is not to act as a precursor for bioactive peptides. Although Holthuis and coauthors (36) have reported proteolytic fragments that derive from SgIII, no biological activity has been assigned to any of the resulting peptides, in contrast with the situation for the CgA and CgB (37, 38, 39). It remains to be seen whether any SgIII-derived fragments act as signaling peptides in neuroendocrine cells, and we have not seen any evidence that pharmacological or immunological activation of mast cells can induce proteolysis of SgIII. In the mast cell system at least, we can argue that the granulogenic capacity of transfected SgIII does not simply reflect an adaptive response to the sudden need to package large quantities of SgIII-derived bioactive peptides.

Our third observation concerning SgIII is that transfection of SgIII cDNA is necessary and sufficient to drive expansion of a vesicular compartment in mast cells, neuroendocrine cells, and even in the nonsecretory HEK cell line. This expansion is accompanied by a concomitant increase in the proinflammatory capacity of the transfected mast cells. The granulogenic potential of SgIII depends on the presence of an intact carboxyl terminus, which in turn is the domain of SgIII that interacts with the amino terminus of CgA. Thus, in mast cells, as in sympathoadrenal AtT-20 cells (16, 35), a functional interaction between SgIII and CgA seems to be required for their granulogenic capacity. Hosaka and coauthors (34, 35) have suggested that SgIII binds to the amino terminus of CgA thereby initiating the secretory granule biogenesis in the AtT-20 cells. This SgIII-CgA complex binds to the cholesterol moieties within the secretory granule membrane, but the exact mechanism by which this binding promotes granulogenesis is not yet fully understood. Several groups have determined that cholesterol depletion blocks the formation of constitutive secretory vesicles from the Golgi apparatus (30), and the formation of microdomains of the “cone-shaped” cholesterol molecules may contribute to the negative curvature of the lumenal leaflet of the vesicle membrane. One key open question regards the localization of endogenous SgIII in untransfected mast cells. We have not been able to produce any convincing immunofluorescence analysis for endogenous SgIII in mast cells, because the available anti-SgIII gives only weak, diffuse vesicular staining that may either indicate a vesicular localization or may represent the type of background staining that is typically associated with low affinity Abs in this application.

The data from our SgIII over-expression experiments in mast cells suggest that SgIII drives granule biogenesis. The data concerning an expansion of the granule content that can be released in response to antigenic or pharmacological treatment of mast cells could also reflect an alternate explanation, such as an action of SgIII in promoting expression, or activity, of the transporters that load serotonin into preexisting granules. Similarly, we cannot exclude that SgIII over-expression enhances the activity of serotonin efflux transporters, a caveat to the secretion assays presented in this study. However, our microscopy data are more consistent with the fact that SgIII over-expression simply results in the presence of more mast cell granules. Knockdown or knockout of SgIII would reveal whether this protein is needed for the biosynthesis of endogenous vesicles in a native expression system. We have found SgIII knockdown in mast cells to be associated with suppression in secretory capacity, but these experiments are difficult to interpret in the light of the confounding effects of decreased viability, possibly because this protein may plays central role in membrane trafficking (data not shown). A SgIII knockout mouse has been described by Dopazo and coauthors (Ref. 40 ; unpublished observations) but no endocrine or immunological phenotype was described and the mice are no longer extant. However, a recent paper by Hendy and coauthors (23) suggests that knockout of granins may not be very informative, due to a compensatory up-regulation in the expression levels of other granin genes.

The data summarized above suggest that the function of SgIII in mast cells closely parallels that described in the neuroendocrine secretory system, where SgIII is involved in regulating the membrane dynamics of secretory vesicles via an interaction with CgA. Intriguingly, our interaction trap and coimmunoprecipitation data suggest that CgA is not the only binding partner for SgIII in mast cells. SgIII clearly interacts with the ligand-gated cation channel TRPA1. These data are strongly reminiscent of the observations of the Ehrlich and Yoo laboratories (10, 11, 13, 31), where CgA and CgB bind to and regulate the inositol (1, 4, 5) trisphosphate receptors, which are intracellular calcium channels. We had initially considered the possibility that interaction between SgIII and TRPA1 might reflect the fact that an SgIII-derived peptide could be a ligand for this channel. However, the clear interaction between TRPA1 and full-length SgIII, and the data from studies of the granin/InsP3R interaction would seem to argue against this model. TRPA1 seems, like TRPV1, to function as an intracellular calcium release channel when supplied with an appropriate membrane-permeant (lipophilic) ligand (41), and it is clearly localized to vesicular intracellular structures. Moreover, channels of the TRP family have been reported to be associated with cholesterol-rich microdomains, at least in the plasma membrane (42). Thus, the association between TRPA1 and SgIII may simply reflect the localization of the latter to lipid membrane microdomains. However, there is also a clear precedent for a regulatory role of granins that associate with certain intracellular calcium release channels, the InsP3R.

Choe and coauthors (10) demonstrated that CgA and CgB associate with the InsP3R, and that provision of competing concentrations of a peptide derived from the interacting portion of the granin had a dominant inhibitory effect upon Ins(1,4,5)P3-evoked calcium release. Similarly, our over-expression of SgIII seems to suppress calcium release via TRPA1. In the case of the Chromogranin/InsP3R interaction, the calcium-binding activity of the granin, through conserved EF-hand motifs, has been proposed to play a role in shaping the channel’s open probability (10, 43). In the case of SgIII, there is no obvious EF-hand motif (CgA/CgB and SgIII are only weakly similar at the sequence level; Ref. 4) and no calcium-buffering activity has been ascribed to this protein. At this time, it seems likely that the loss of function in TRPA1 that is associated with SgIII over-expression is likely to reflect a disruption of its localization to cholesterol-rich microdomains. Future work will determine whether the activity of TRPA1 in intracellular membranes absolutely depends upon correct localization to SgIII-containing lipid rafts.

In summary, the present report suggests that granins in mast cells recapitulate the varied functional contributions made by these proteins to the neuroendocrine secretory system. Granins such as SgIII can obviously contribute to multiple facets of cell biology, and the reaching of a consensus as to their molecular roles is clearly a high priority. In the mast cell, granins such as SgIII may be critical determinants of granule load, and hence the intensity and duration of inflammatory responses.

We gratefully acknowledge the practical assistance of Clay Wakano and Lori Shimoda, and the constructive advice from Dr. Kimberly del Carmen. DNA sequencing services were provided by the Center for Genomics, Proteomics, and Bioinformatics (CGBP) at the University of Hawaii, with the kind assistance of Dr. Shaobin Hu.

The authors have no financial conflict of interest.