Activation of the Inositol 1,4,5-Trisphosphate Receptor by the Calcium Storage Protein Chromogranin A*

Secretory granules of neuroendocrine cells are inositol 1,4,5-trisphosphate (InsP 3 )-sensitive Ca 2 (cid:1) stores in which the Ca 2 (cid:1) storage protein, chromogranin A (CGA), couples with InsP 3 -gated Ca 2 (cid:1) channels (InsP 3 R) located in the granule membrane. The functional aspect of this coupling has been investigated via release studies and planar lipid bilayer experiments in the presence and absence of CGA. CGA drastically increased the release activity of the InsP 3 R by increasing the channel open probability by 9-fold and the mean open time by 12-fold. Our results show that CGA-coupled InsP 3 Rs are more sensitive to activation than uncoupled receptors. This modulation of InsP 3 R channel activity by CGA ap- pears to be an essential component in the control of intracellular Ca 2 (cid:1) concentration by secretory granules and may regulate the rate of vesicle fusion and exocytosis.

CGA 1 is a member of the granin protein family and is stored in high concentrations in the large dense core secretory granules of most endocrine and neuroendocrine cells as well as in many nerve cells in the periphery and brain (1,2). CGA, the first member of the granin family to be discovered (3)(4)(5), has a wide variety of functions, both extracellular and intracellular.
As one of its extracellular functions, CGA acts as a prohormone, a protein that contains numerous sites for proteolytic processing. Following secretion, extracellular proteases cleave CGA, generating several peptide fragments with biological activity, including pancreastatin (6,7), vasostatins I and II (8 -10), parastatin (11), catestatin (12), and chromacin (13). In healthy individuals, CGA and its peptide fragments are present in the circulatory system in low nanomolar quantities. However, in patients suffering from pheochromocytoma and other neuroendocrine tumors, concentrations are significantly higher (14). Elevated plasma levels of CGA are associated with a number of pathological conditions making the protein an ideal marker not only for neuroendocrine tumors but also for chronic heart failure and brain disorders such as Parkinson's and Alzheimer's diseases (15).
Among its intracellular roles, CGA has been shown to interact with ATP, catecholamines, and Ca 2ϩ (16,17), to acidify the intravesicular medium and to sort proteins for the regulated secretory pathway via a range of protein-protein interactions (15). These sorting functions include aggregation with chromogranin B, complexing with dopamine ␤-hydroxylase, t-plasminogen activator, and binding secretory granule membrane constituents such as the InsP 3 R (15).
In recent years, secretory granules of neuroendocrine cells have been identified as inositol (1,4,5)-trisphosphate (InsP 3 )sensitive Ca 2ϩ stores (18 -20). In the granules CGA forms a tetramer and appears to bind four molecules of the intraluminal loop of the InsP 3 R at the intravesicular pH 5.5 (21)(22)(23). In vitro studies show that purified InsP 3 R interact directly with CGA at this pH and dissociate from it at pH 7.5, a pH encountered when exocytosis occurs (24). Co-transfection of InsP 3 R and CGA into COS-7 cells followed by co-immunoprecipitation demonstrates that these two proteins form a complex in vivo (24).
We have investigated the functional aspect of this coupling via InsP 3 -mediated Ca 2ϩ release studies using InsP 3 R-reconstituted liposomes in the presence and absence of CGA. We have further characterized the molecular basis of this phenomenon at the single channel level using planar lipid bilayer studies. In the presence of CGA the open probability and mean open time of the InsP 3 R channel increases significantly. Hence, modulation of InsP 3 R channel activity by CGA appears to be an essential component in the control of intracellular Ca 2ϩ concentration in secretory granules.

Purification of the InsP 3 Receptor
For Flux Studies-The type I InsP 3 receptor was isolated from bovine cerebella as described previously (25). Briefly, bovine cerebella were mixed with 3 volumes of buffer I (50 mM Tris-HCl, pH 7.4, 0.32 M sucrose, 1 mM EDTA, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, 10 M pepstatin), homogenized, and centrifuged at 2000 ϫ g for 10 min at 4°C. The supernatants were re-centrifuged at 105,000 ϫ g for 1 h to precipitate the membrane pellet, which was resuspended in buffer II (50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, 10 M pepstatin) containing 1% Triton X-100, stirred for 1 h, and then centrifuged at 32,000 ϫ g for 1 h at 4°C. The resulting supernatant was mixed with an equal volume of buffer III (20 mM Tris-HCl, pH 7.5, 0.1 M NaCl, 1% Triton X-100, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, 10 M pepstatin), and applied to an InsP 3 R antibody-coupled immunoaffinity column (0.35 ϫ 1 cm) equilibrated with 20 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM CaCl 2 . The protein-loaded column was washed with 20 bed volumes of this buffer, and the InsP 3 R was eluted by 10 ml of elution buffer (0.1 M glycine, pH 2.8, 0.2% Triton X-100, 0.5 M NaCl, 1 mM ␤-mercaptoethanol, 0.1 mM PMSF, 10 M leupeptin, 10 M pepstatin). The eluate was immediately neutralized by adding 1 M Tris-HCl, pH 9.5, and mixed with an equal volume of buffer IV (50 mM Tris-HCl, pH 8.0, 0.2% Triton X-100, 0.5 M NaCl, 1 mM ␤-mercaptoethanol); it was then applied to a benzamidine-Sepharose column equilibrated with Buffer V (20 mM * This work was supported by Grant GM63496 from the National Institutes of Health, an American Heart Association fellowship, Heritage Affiliate (to E. C. T.), and the Creative Research Initiatives Program of the Ministry of Science and Technology, Korea. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. HEPES, pH 7.5, 100 mM NaCl, 1 M KCl, and 3 M urea). InsP 3 R containing flow-through was collected and stored at Ϫ70°C. until use.
For Bilayer Experiments-The type I InsP 3 receptor was solubilized in 1% CHAPS and purified from mouse cerebellum using heparin affinity and concanavalin A-Sepharose column chromatography as described previously (26). The purified InsP 3 R was then incorporated into liposomes by adding 15 g of purified protein to 1 ml of liposome solution (consisting of phosphatidylcholine in bilayer buffer), mixing, and then incubating on ice for 10 min.

Flux Studies
InsP 3 Dose Response for InsP 3 R-InsP 3 R proteoliposomes were formed as described previously (25). Some of these proteoliposomes had CGA encapsulated in them, and the remainder was used for control experiments. Ca 2ϩ efflux from the proteoliposomes was measured by observing changes in indo-1 fluorescence. Fluorometric measurements were carried out at 35°C using a Shimadzu RF-5301 PC spectrofluorometer equipped with a temperature-controlled cuvette holder. Fluorescence intensity was measured at the emission wavelength of 393 nm (excitation of 355 nm) with 10 nm of excitation band slit width and 10 nm of emission band slit width. For the kinetic analysis of InsP 3induced Ca 2ϩ release, the data were acquired every 20 ms after each addition of the indicated InsP 3 concentration to 0.5 ml of the proteoliposome solution. The fluorescent intensities of indo-1 were calibrated to free Ca 2ϩ concentrations using a Ca 2ϩ -EGTA buffering system (27).
InsP 3 dose-dependent Ca 2ϩ release was also measured by the intensity of indo-1 fluorescence after each InsP 3 addition and was compared with the fluorescence intensity after the addition of Triton X-100 instead of InsP 3 . In these experiments, 10 M indo-1 was used, which is a high enough concentration to buffer released Ca 2ϩ , thus precluding the possibility of Ca 2ϩ regulation of the InsP 3 -induced release.
Bilayer Experiments-Planar lipid bilayers were formed by painting a solution of phosphatidylethanolamine/phosphatidylserine (3:1; 30 mg/ml in decane) across a 100 m aperture in a Teflon sheet bisecting a Lucite chamber. The hole was pre-painted with phosphatidylcholine/ phosphatidylserine (3:1) prior to membrane formation. The two compartments are defined as cis (corresponding to the cytosol) and trans (corresponding to the lumen of the ER).
The trans (luminal) compartment consisted of 250 mM HEPES, adjusted to pH 5.5 (as purified InsP 3 R was used in these experiments, the pH could be changed using 70 mM HCl), and 53 mM Ba(OH) 2 . Single channel currents were amplified using a bilayer clamp amplifier (Warner Instruments) and recorded on digital tape. Data was filtered with an eight-pole Bessel filter to 500 Hz, digitized to 2 kHz, transferred to a personal computer, and analyzed using the pClamp 6.0 (Axon Instruments) software package. InsP 3 R proteoliposomes were added to the cis compartment and mixed followed by the addition of 2 M InsP 3 to the same compartment. Upon InsP 3 R activation, single channel activity was recorded. CGA (1 g) was added to the trans compartment and mixed. InsP 3 R single channel activity was recorded. The pH inside the trans compartment was changed by adding Tris (final concentration 110 mM) to pH 7.5 (to dissociate CGA from InsP 3 R), and InsP 3 R single channel activity was recorded.
These experiments were repeated (i) in the presence of increasing doses of InsP 3 (over the range of 0.2-2 M) to the cis compartment and (ii) at a fixed InsP 3 concentration of 2 M in the presence of increasing free Ca 2ϩ concentrations (over the range of 0.01-1 M) to the cis compartment. Both steps i and ii were carried out in the presence and absence of 1 g of CGA in the trans compartment, and InsP 3 R single channel activity was recorded.

Effect of CGA on InsP 3 -mediated Ca 2ϩ
Release-The effect of CGA on InsP 3 dose response for type I InsP 3 R from bovine cerebella was investigated initially using Ca 2ϩ release studies. InsP 3 -induced Ca 2ϩ release from InsP 3 R-reconstituted liposomes was monitored both in the presence and absence of CGA (Fig. 1). InsP 3 -induced Ca 2ϩ efflux through the proteoliposomes (300 M Ca 2ϩ inside) was determined by the change of indo-1 fluorescence at 393 nm. The total amount of Ca 2ϩ in the liposomes was determined by adding 1% Triton X-100, and this was the value set at 100%. Given this information, the total amount of InsP 3 -releasable Ca 2ϩ was estimated to be 60%. When CGA was present inside the vesicle at pH 5.5, the pH value at which CGA associates with the InsP 3 R, InsP 3 -induced Ca 2ϩ release was significantly enhanced (see Fig. 1a). A K app value for InsP 3 of 0.2 M was obtained. When the pH was maintained at 7.5, however, the fluorescent changes seen at each InsP 3 dose more closely resembled those seen in the absence of CGA at pH 5.5 (K app values for InsP 3 of 0.8 and 0.9 M, respectively), further supporting the pH dependence of the InsP 3 R/CGA interaction. The presence of CGA at pH 5.5 markedly increased the apparent affinity of the receptor for InsP 3 when compared with the Ca 2ϩ release obtained in the absence of CGA. This result complements the effect of CGA on InsP 3 binding to its receptor (25). Even at InsP 3 concentrations lower than those published previously (starting at 0.05 M) (25), an increase in apparent affinity for InsP 3 was seen (Fig. 1b).
Effect of CGA upon InsP 3 R Channel Activity-The enhancement of Ca 2ϩ release from InsP 3 R-reconstituted liposomes as a result of the pH-dependent interaction of CGA with InsP 3 R, illustrates a functional phenomenon associated with this coupling. To further define the actual mechanism of action, we investigated these effects at the single channel level using

InsP 3 R incorporated into planar lipid bilayers.
Under control conditions, in the absence of luminal CGA and in the presence of cytosolic free Ca 2ϩ (300 nM) and InsP 3 (2 M), mouse InsP 3 R type I, single channel activity was observed (see Fig. 2a, trace i). Single channel currents of ϳ2 pA were seen, and the presence of a pH gradient between the trans and cis compartments (pH 5.5:pH 7.35) did not affect channel activity. Two populations of mean open times were seen with values of 0.864 Ϯ 0.039 and 8.84 Ϯ 0.014 ms (Fig. 3a). The data set is further expanded (Fig. 3b)  Less than 1 min after the addition of 1 g of CGA to the trans compartment, a dramatic change in channel activity was observed (Fig. 2a, trace ii). The magnitude of the single channel current remained unaltered; however, significant differences were apparent in mean open times and P o . Two populations of mean open time were evident, as before, but were greatly increased over control values (Fig. 3c). Values of 2.61 Ϯ 0.024 and 103.5 Ϯ 0.003 ms were obtained, illustrating an approximate 3-fold and 12-fold increase of open time in each respective population. This is displayed even more clearly when the scale from Fig. 3c is expanded to focus on the population of longer open times (Fig. 3d). Furthermore, a large increase in P o over control levels was seen with a value of 33.0 Ϯ 8.5% (S.E.) (n ϭ 4).
By changing the pH of the trans compartment to 7.5, a condition known to cause dissociation of CGA from InsP 3 R, channel activity essentially reverted to control levels (Fig. 2a, trace iii). The two populations of mean open time were reduced to 1.05 Ϯ 0.014 and 7.19 Ϯ 0.013 ms (n ϭ 4, Fig. 3e, and compare with Fig. 3a), and the P o was reduced to 3.0 Ϯ 1.6% (S.E.), a value close to that seen for the control. The addition of heparin, an InsP 3 R-specific antagonist, to the cis compartment inhibited channel activity completely.
A similar study was carried out using microsomes from mouse cerebellum to see whether the effects of CGA would still be seen in native tissue the same as in purified protein. These experiments were complicated by the fact that HCl was present in the trans compartment, a condition necessary for maintaining the pH at 5.5. Thus chloride channels present in the microsomes were activated, making the analysis difficult. Nonetheless, under comparable control conditions to those described for the purified receptor, the open probability was 8%; and upon addition of CGA to the trans compartment this increased to 52%. Altering the pH of the trans compartment lowered the open probability to ϳ1%.
The addition of CGA to the trans compartment in the absence of InsP 3 R had no effect upon the bilayer itself, and CGA did not potentiate any InsP 3 R channel activity in the absence of cytosolic InsP 3 . Furthermore, the addition of 1 g of CGA to the cis compartment in the presence of InsP 3 R and InsP 3 did not affect channel activity.
As CGA is a highly charged protein, control experiments were carried out to exclude the possibility that any charged macromolecule could be responsible for the effects observed in this study. Heparin is one such charged macromolecule and is known for its inhibitory effects on the InsP 3 R when exposed to the cytosolic face of the receptor, although it is not known to bind to its luminal face. The addition of 1 g of heparin to the trans compartment (see Fig. 2b) did not alter channel open probability (5% for the control compared with 4.4% in the presence of luminal heparin), indicating that CGA does have a specific modulatory effect on InsP 3 Rs.
Effect of CGA on InsP 3 Dose Response for InsP 3 R-Single channel activity as a function of InsP 3 concentration, both in the presence and absence of CGA, was characterized next (Fig.  4, a and b). Over a range of InsP 3 concentrations starting at 0.2 M, the open probability was greater in the presence of CGA, with a 14-fold increase observed at 2 M InsP 3 . When the pH of the trans compartment was changed to 7.5, the P o was reduced to that seen in the control experiments. The results obtained from the single channel experiments concur with those seen in the Ca 2ϩ release studies (see Fig. 1) in that, at each InsP 3 concentration and in the presence of CGA, a significant increase in P o is concomitant to an increased amount of released Ca 2ϩ .
Effect of CGA on Ca 2ϩ Dependence for InsP 3 R-The Ca 2ϩdependence of the InsP 3 R was investigated in the absence and presence of CGA (Fig. 5, a and b) at a fixed InsP 3 concentration of 2 M and over a Ca 2ϩ concentration range of 0.01-1.0 M. As the Ca 2ϩ concentration in the cis compartment increased successively from 0.01 to 0.3 M, the P o increased, reaching a maximum value of 4% at 1 M Ca 2ϩ (pCa 6) in the absence of CGA (Fig. 5b). At Ca 2ϩ concentrations higher than 0.3 M (pCa 6.5), no inhibition was seen (see Fig. 5b, expanded section). This lack of inhibition by free Ca 2ϩ has been observed previously for purified receptor (28,29) and contrasts to the inhibition seen with microsomes (30). Repetition of this experiment, with the addition of CGA to the trans compartment, produced dramatic increases in channel activity (Fig. 5a). At a Ca 2ϩ concentration of 0.01 M (pCa 8) the P o is effectively zero in the absence of CGA, in contrast to the P o observed when CGA is present, which expressed as a percentage of total open probability is 34%. Furthermore, the P o effectively remains at this level irrespective despite increasing Ca 2ϩ concentrations, and again no inhibition by Ca 2ϩ is seen. The activating phase of the Ca 2ϩ dependence seen in the absence of CGA is not apparent in its presence. The channel has reached maximal open probability at pCa 8. Thus, the lack of dependence on Ca 2ϩ for activation in the presence of CGA and 2 M InsP 3 is clearly illustrated.

DISCUSSION
Secretory granules of endocrine and neuroendocrine cells have been shown to serve as InsP 3 -sensitive intracellular Ca 2ϩ stores (18,31), and additional evidence from goblet cells has demonstrated their direct participation in the control of cytoplasmic Ca 2ϩ (20). Chromogranins are Ca 2ϩ storage proteins that are found in secretory granules in millimolar concentrations (1-2 mM) (1,32), and CGA binds 32 mol of Ca 2ϩ /mol with a dissociation constant of 2.7 mM at pH 7.5 and 55 mol of Ca 2ϩ /mol with a dissociation constant of 4 mM at pH 5.5 (33). Given the high capacity Ca 2ϩ binding of CGA, most of the 40 mM intravesicular Ca 2ϩ remains bound, thus yielding a total free Ca 2ϩ concentration of ϳ24 M inside the granules (34). As secretory granules occupy about 10% of the total cell volume (at least in bovine chromaffin cells) (32) and have a high storage capacity for Ca 2ϩ , they may play an important role in governing intracellular Ca 2ϩ dynamics.
This hypothesis is further supported by the discovery that chromogranins A and B interact directly with type I InsP 3 Rs located on the secretory granule membrane (24) and that this coupling is functional, at least in terms of CGA (25). The present results obtained from our bilayer studies reemphasize the functional importance of the interaction and provide the first mechanistic insight at the level of a single InsP 3 R. The effects of CGA on InsP 3 R channel activity, at the intravesicular pH of 5. plies a greater release of Ca 2ϩ . When the intravesicular pH is altered to 7.5, the effects of CGA are seen to dissipate almost instantly (Fig. 2, trace iii, and Fig. 3, e and f) and resemble more closely those seen at pH 5.5 in the absence of CGA. The InsP 3 concentration dependence both in bilayer studies and flux studies carried out at comparable concentrations complement one another. At the single channel level the InsP 3 R is seen to have a greater chance of opening when CGA is present, and in the flux studies, the apparent affinity for InsP 3 is greater. Again, a change in pH to 7.5 causes the effect to revert to control levels. As for Ca 2ϩ dependence, the InsP 3 R is already active at maximal levels when CGA and InsP 3 are present even when the level of cytosolic free Ca 2ϩ is only 10 nM (Fig. 5). The data shown in Figs. 4 and 5 indicate that in the presence of coupled CGA and a sufficient concentration of InsP 3 (2 M), the characteristic effects of cytosolic Ca 2ϩ on the InsP 3 R channel disappear. It appears that the conformation of the InsP 3 R is in such a state in the CGA-coupled condition, that high InsP 3 dictates the channel activity regardless of the presence of Ca 2ϩ .
How do these results relate to the physiological situation? In the presence of CGA at intravesicular pH 5.5, the InsP 3 R is primed to respond to low levels of InsP 3 . Hence, when a secretory granule reaches the surface of the cell and docks with the inner surface of the plasma membrane, it is fully loaded with Ca 2ϩ and sensitized for release. Generation of InsP 3 , even in small quantities, will cause a large elevation in local Ca 2ϩ concentration (values as high as 100 M have been seen immediately prior to exocytosis) (23), initiating secretory processes. During exocytosis, the vesicle contents are exposed to the extracellular pH of 7.4, thus causing dissociation of CGA from the InsP 3 R resulting in altered channel properties and hence Ca 2ϩ release. Following these events, vesicular contents dissociate from the vesicle membrane, with secretory cargo moving to the extracellular space and then into the bloodstream.
The chromogranins, particularly CGA, appear to play a role in intracellular Ca 2ϩ dynamics and secretion, which also has significant implications to human disease. For example, patients suffering from pheochromocytoma and other neuroendocrine tumors have significantly higher concentrations of CGA measured in the plasma (14,15). This high plasma CGA indicates that high levels must have been stored in order to be released. Individual cells contain increased CGA, and there are more CGA-containing cells. In various cholestatic liver diseases, a striking increase is seen in the number of bile ductules (35). These reactive bile ductules differ from their normal counterparts in that they display neuroendocrine features, in particular the expression of CGA. Hence, the presence of elevated CGA can lead to abnormal Ca 2ϩ release within these cells, thus activating specific metabolic mechanisms in the cell, which leads to unnecessary growth. The increased release of neuroendocrine substances, including CGA, may also play an autocrine or paracrine regulatory role in ductular metaplasia of hepatocytes or bile ductule growth. Thus, although the role of CGA in disease is largely attributed to extracellular CGA and its derivatives, a contribution may arise also from intragranular CGA with respect to its effects on intracellular Ca 2ϩ release and exocytosis.
In conclusion, our present work is the first electrophysiological study detailing the very important physiological phenomenon arising from the interaction between CGA, a Ca 2ϩ storage protein, and an intracellular Ca 2ϩ release channel, InsP 3 R. We have shown that there is an order of magnitude increase in the open probability and open time of the InsP 3 -gated Ca 2ϩ channel. These effects of the interaction of CGA and InsP 3 R have implications for exocytosis and neuroendocrine cell function in disease.