Calnuc plays a role in dynamic distribution of Galphai but not Gbeta subunits and modulates ACTH secretion in AtT-20 neuroendocrine secretory cells.

In AtT-20 cells ACTH secretion is regulated by both Ca2+ and G proteins. We previously demonstrated that calnuc, an EF-hand Ca2+ binding protein which regulates Alzheimer's beta-amyloid precursor protein (APP) biogenesis, binds both Ca2+ as well as Galpha subunits. Here we investigate calnuc's role in G protein-mediated regulation of ACTH secretion in AtT-20 neuroendocrine secretory cells stably overexpressing calnuc-GFP. Similar to endogenous calnuc, calnuc-GFP is mainly found in the Golgi, on the plasma membrane (PM), and associated with regulated secretion granules (RSG). By deconvolution immunofluorescence, calnuc-GFP partially colocalizes with Galphai1/2 and Galphai3 at the PM and on RSG. Cytosolic calnuc(DeltaSS)-CFP with the signal sequence deleted also partially colocalizes with RSG and partially cosediments with Galphai1/2 in fractions enriched in RSG. Overexpression of calnuc-GFP specifically increases the distribution of Galphai1/2 on the PM whereas the distribution of Gbeta subunits and synaptobrevin 2 (Vamp 2) is unchanged. Overexpression of calnuc-GFP or cytosolic calnuc(DeltaSS)-CFP enhances ACTH secretion two-fold triggered by mastoparan or GTPgammaS but does not significantly affect glycosaminoglycan (GAG) chain secretion along the constitutive pathway or basal secretion of ACTH. Calnuc's facilitating effects on ACTH secretion are decreased after introducing anti-Galphai1/2, Galphai3, Gbeta or calnuc IgG into permeabilized cells but not when Galpha12 or preimmune IgG is introduced. The results suggest that calnuc binds to Galpha subunits on the Golgi and on RSG and that overexpression of calnuc causes redistribution of Galphai subunits to the PM and RSG, indicating that calnuc plays a role in dynamic distribution of only Galpha but not Gbeta subunits. Thus calnuc may connect G protein signaling and calcium signaling during regulated secretion.

Transport along the regulated secretory pathway and exocytosis of secretion granules involves vesicular trafficking, fusion of secretory granules with the plasma membrane (PM), followed by release of granule contents. Regulated secretion is stimulated by Ca 2+ [9] and heterotrimeric G proteins, including several Gα and Gβγ subunits [10][11][12]. Among these, Gαi3 was found to facilitate histamine release from mast cells [13], noradrenaline release from adrenal chromaffin cells [14], and adrenocorticotropic hormone (ACTH) secretion from AtT-20 cells [15]. The recent discovery that corticotrophin releasing hormone (CRH) and vasopressin (VP) regulate ACTH secretion via binding to the Type 1 CRH receptor and the V1b receptor, which are G protein coupled receptors (GPCRs), verifies the regulation of ACTH secretion by G proteins [16]. Moreover, several G proteins have been found on intracellular membranes as well as on the PM. Gαi3 is associated with Golgi membranes as well as at the PM [17,18], and Gαi1/2 is found on secretory vesicles [14,19,20].
We have previously reported that calnuc is associated with regulated secretion granules (RSG) [21] and binds to Gαi3 in the Golgi [8]. In addition, we hypothesized that calnuc might modulate regulated secretion by virtue of its ability to bind Gαi3 and Ca 2+ . To obtain direct evidence for the role of calnuc in the regulation of G protein mediated ACTH secretion we overexpressed calnuc-GFP in AtT-20 cells. We report here that overexpressed calnuc-green fluorescent protein (GFP) partially codistributes with Gαi1/2 as well as Gαi3 on the cytoplasmic surface of regulated secretory granules (RSG), facilitates ACTH secretion triggered by the G protein activators GTPγS or mastoparan and causes redistribution of Gαi subunits by increasing Gαi1/2 on the PM and Gαi3 on RSG. Thus calnuc, the only protein demonstrated to bind both Ca 2+ and Gα subunits [3], appears to play an important role in regulation of G protein and Ca 2+ -related signaling events in endocrine cells.

Distribution of Endogenous Calnuc in AtT-20 Cells and in Cells Stably Overexpressing Calnuc-GFP or Calnuc (ΔSS)-CFP
We have previously reported [21] that most of the endogenous calnuc is concentrated in the Golgi region in AtT-20 cells (Fig. 1A) with some also associated with the PM and immature secretory granules. Calnuc-GFP directly visualized in live AtT-20 cells (Fig. 1B) or seen by deconvolution analysis of immunostained fixed sections (Fig. 1E) is similarly concentrated in the Golgi region, but it is also associated with regulated secretion granules (RSG) which contain ACTH located at the tips of the cell processes ( Fig.  1D-F). In addition to its presence in the Golgi and RSG, some calnuc-GFP is also found at the PM (Fig. 1B) [8]. Our previous morphological and biochemical data established that calnuc is secreted by the constitutive-like pathway, as it is packaged into immature secretory granules in the Golgi, sorted out of RSG during their maturation, and secreted by the constitutive pathway [21].
When calnuc(ΔSS)-CFP with the signal sequence deleted which is located in the cytoplasm [8] is expressed and viewed by live cell imaging, it is seen to be distributed throughout the cytoplasm (Fig. 1C). However, when fixed cells are permeabilized before fixation (to release cytosolic calnuc), immunostained with an anti-GFP IgG, and examined by immunofluorescence and deconvolution analysis, calnuc(ΔSS)-CFP is also seen to be associated with RSG based on colocalization with ACTH ( Fig.  1G-I). The findings with this mutant suggest that some of the cytosolic calnuc binds to the cytoplasmic surface of RSG. Thus calnuc appears to be located both inside RSG as well as bound to the cytoplasmic surface of RSG membranes.

Distribution of Gαi Subunits in Parental (NT) AtT-20 Cells and Those Stably Expressing Calnuc-GFP and Calnuc(ΔSS)-CFP
Next we investigated the distribution of Gα subunits in AtT-20 cells. In parental AtT-20 cells, Gαi3 is found on both the PM and the Golgi ( Fig. 2A) as previously reported [21], and Gαi1/2 is associated mainly with RSG located in both the cell bodies and tips of the AtT-20 cell processes (Figs. 2C and 3A-C). In AtT-20 cells stably expressing calnuc-GFP, more Gαi3 appears to be associated with RSG (Figs. 2B and 3D-F), and both Gαi1/2 (Figs. 2D and 3A-C) and Gαi3 (Fig 2B and 3D-F) are observed on the PM as well as on RSG.
Calnuc-GFP partially colocalizes with Gαi3 in the Golgi (Fig. 3G-I) and partially co-distributes with Gαi1/2 ( Fig.  3J-L) and Gαi3 (Fig. 3G-I) on RSG and the PM. In cells stably expressing cytosolic calnuc(ΔSS)-cyan fluorescent protein (CFP), the majority of the Gαi1/2 and membrane associated calnuc colocalize on RSG (Fig. 3M-O). Both Gαi1/2 and calnuc(ΔSS)-CFP are concentrated on the cytoplasmic surface of RSG and the PM. Thus, based on deconvolution analysis of our immunofluorescence results, it is evident that 1) both Gαi1/2 and Gαi3 are found on RSG where they partially colocalize with ACTH, and 2) in cells expressing calnuc-GFP the distribution of Gαi subunits along the PM and RSG is enhanced.

Distribution of ACTH, Calnuc and Its Mutants as Well as G Protein Subunits in Membrane vs Cytosolic Fractions from Parental and Stably Transfected AtT-20 Cells
To further check the distribution and expression of calnuc and G proteins in AtT-20 cells we analyzed their distribution in membrane (100,000 × g pellet) vs cytosolic (100,000 × g supernatant) fractions. Figure 4A shows that endogenous calnuc (63 kD) and calnuc-GFP (91 kD) are found in both cytosolic and membrane fractions. The amount found in cytosolic fractions was greater for calnuc-GFP (80%) than endogenous calnuc (50%). Most (>90%) of the calnuc(ΔSS)-CFP (89 kD) was found in the cytosolic fraction, but ~10% pelleted with membranes in keeping with the immunofluorescence results demonstrating its association with RSG. The distribution of proopiomelanocortin (POMC)-derived products including POMC, POMC intermediate, glycosylated ACTH (gACTH) and mature ACTH was also checked and was similar in parental AtT-20 cells and those overexpressing calnuc-GFP (Fig. 4B). Synaptobrevin 2 (Vamp 2) which localizes on both RSG and synaptic vesicles was found exclusively in membrane fractions (Fig. 4B). 80% of the Gαi1/2, Gαi3 and Gβ subunits were associated with membrane fractions in parental AtT-20 cells and 90% in Characterization of AtT-20 cells stably overexpressing calnuc-GFP or Calnuc(ΔSS)-CFP Figure 1 Characterization of AtT-20 cells stably overexpressing calnuc-GFP or Calnuc(ΔSS)-CFP. (A). Immunofluorescence showing that most of the endogenous calnuc is found in the Golgi region (G) with some found on the PM (arrow) (B) Stably expressed calnuc-GFP in live AtT-20 cells is also concentrated in the Golgi with some found on the PM, but it is also associated with RSG stored at the tips of cell processes. . An additional Gβ band was also found in the calnuc-GFP cells. We conclude that overexpression of calnuc-GFP does not significantly change the expression level or localization ratio (membrane vs cytosolic fractions) of Gαi subunits, ACTH or its derivatives.

Overexpression of Calnuc-GFP Does Not Affect Constitutive Secretion of GAG Chains or Basal Secretion of gACTH and Its Precursors in AtT-20 Cells
Next we examined the effects of overexpression of calnuc-GFP on secretion of glycosaminoglycan (GAG) chains, commonly used as a marker of constitutive secretion from AtT-20 cells [22] and on basal secretion of ACTH and its precursors. No change was detected in GAG secretion, as similar amounts of GAG chains were secreted from both parental AtT-20 cells and those expressing calnuc-GFP over a period of 15 min to 2 h (Fig. 5A). The intracellular and basal secreted metabolically labeled POMC, intermediate, glycosylated ACTH (gACTH) was also similar in parental (NT) and in calnuc-GFP (CNG) cells after a 2 h chase. However, certain increased secreted (Lane 2) and intracellular mature ACTH (Lane 4) were observed following overexpression of calnuc (Fig. 5B).

Overexpression of Calnuc-GFP Increases ACTH Secretion from AtT-20 Cells Stimulated by Mastoparan or GTPγS and the Effect Is G protein-dependent
To investigate if calnuc plays a role in regulated secretion we used permeabilized, parental AtT-20 cells and those overexpressing calnuc-GFP or calnuc(ΔSS)-CFP. Cells were permeabilized with digitonin, and ACTH secretion was triggered with the wasp venom peptide mastoparan which mimics receptor activation of G proteins by binding to the C-terminus of Gαi and Gαq [23]. Mastoparan has been used to study exocytosis in AtT-20 by others [15]. We found that when cells were stimulated with mastoparan, 30% of the total ACTH was secreted by parental cells whereas the amount secreted by cells expressing calnuc-GFP was ~60% of the total or double that found in parental cells (Fig. 6A). The amount secreted by cells expressing calnuc(ΔSS)-CFP was also increased (~55% of total). ACTH secretion was similarly increased in calnuc-GFP or calnuc(ΔSS)-CFP cells after stimulation with GTPγS (another G protein activator), where 70-75% of the total ACTH was secreted vs 40% in parental cells or those overexpressing GFP alone (Fig. 6A). The result indicates that calnuc enhances regulated secretion of ACTH triggered by the G protein activators mastoparan and GTPγS.
Next we investigated whether the effects of calnuc on ACTH secretion is G protein dependent by introducing affinity purified antibodies against Gαi, Gβ or calnuc into the cytoplasm of calnuc-GFP cells. ACTH secretion triggered by mastoparan from cells pre-treated with anti-calnuc, anti-Gα1/2, anti-Gαi3 and anti-GβIgG decreased by 42%, 33%, 50%, and 42% respectively, compared to nontreated AtT-20 cells or those treated with preimmune IgG (Fig. 6B). No significant change in ACTH secretion was seen from non-stimulated calnuc-GFP cells or those pretreated with either anti-Gα12 as a negative control or calnuc preimmune IgG. Together the results obtained with mastoparan treatment and antibody inhibition support the conclusion that calnuc as well as G protein subunits stimulate regulated secretion of ACTH in AtT-20 cells.

Calnuc Does Not Regulate G protein Activity
We next investigated if calnuc's effects on ACTH secretion triggered by mastoparan and GTPγS are due to regulation of G protein activity. Calnuc was previously demonstrated by us to bind to the same region of the C-terminal α5helix on Gαi3 as mastoparan [7]. We therefore investigated if purified, recombinant calnuc behaves similarly to mastoparan (Fig. 7). Mastoparan was found to increase GTPγS binding to Gαi3 over a 5 to 60 min period. 5 min after the reaction was started, mastoparan increased binding of GTPγS to Gαi3 4.15 ± 1.33 fold, which is similar to that described for Gαi1 [23]. However, in contrast to mastoparan or AGS3 (Activator of G protein signaling 3) [24], no effect was observed when calnuc was added in the same assay (Fig 7). This result indicates that although it binds Gαi in vivo [8], it most likely does not regulate G protein activity. In addition, calnuc was not found to possess GAP activity (data not shown). We conclude that although calnuc binds Gα subunits, it most likely does not act as a GDI, GEF or GAP for these G proteins.

Calnuc-GFP Localizes in Both Light and Heavy Membrane Fractions
Our findings that calnuc increases ACTH secretion and anti-G protein antibodies partially block this effect, suggests that calnuc's effects on ACTH secretion occur via G proteins. Because our immunofluorescence results (Fig. 2) suggested a shift in the distribution of Gα subunits occurs following overexpression of calnuc-GFP, we analyzed the effects of overexpressing calnuc on G protein distribution by subcellular fractionation. Initially we used a sucrose gradient centrifugation protocol [25] aimed at separating Golgi and PM (light fractions) from ER and RSG (heavy fractions). In keeping with the immunofluorescence results, we found that the majority of the endogenous calnuc (>90%) found in parental AtT-20 (NT) cells is concentrated in light fractions (9)(10)(11)(12) containing Golgi and PM (Fig. 8) where it cosediments with the Golgi marker αmannosidase II (Man II). Overexpressed calnuc-GFP is also concentrated mostly in light fractions (>90%), but some (<10%) is also found in heavy fractions (3-8) containing ER and granules where it cosediments with calnexin, an ER marker, and ACTH and synaptobrevin 2 (Vamp 2), markers for RSG. This is in keeping with our finding that calnuc-GFP colocalizes with RSG. The distribution in the gradient and level of expression of Vamp 2 and ACTH are not significantly changed following overexpression of calnuc-GFP. The results indicate that unlike endogenous calnuc mainly localize in the Golgi, overexpressed calnuc localzed in both heavy (ER and granules) and light (plasma membrane and Golgi) fraction. However, overexpression of calnuc doesn't change ACTH and Vamp 2 distribution.

Overexpression of Calnuc Affects Distribution of Gαi Subunits
We next examined whether the distribution of G proteins changes upon overexpression of calnuc-GFP using the same sucrose gradient protocol. Figure 9 shows that in parental AtT-20 cells (NT), the majority of the Gαi1/2 (70%) is in heavy fractions (3)(4)(5)(6)(7)(8) containing ER and RSG, and the majority of the endogenous Gαi3 (72%) sediments in the light fractions (10-12) containing Golgi and PM. In cells overexpressing calnuc-GFP, the total Gαi1/2 in the heavy fractions is reduced to 55% with a corre-

Overexpression of Calnuc-GFP Increases Distribution of Gαi1/2 on the PM in AtT-20 Cells
To further distinguish whether the increased Gαi1/2 found in light fractions is associated with Golgi membranes or the PM, we used a sucrose gradient flotation method [26] designed to separate light fractions (5-8) enriched in PM from heavy fractions (1-3) containing Golgi and ER. The % of the total Gαi1/2 in PM fractions (fractions 5-8) increased from 38% to 63% following overexpression of calnuc-GFP (Fig. 10). No striking change in the distribution of Gαi3 in fractions 6-8 (95% in calnuc-GFP cells vs. 90% in parental cells) was observed except that a broader spread throughout the gradient (fractions 1-8) was seen, similar to that of calnuc-GFP. In parental cells, the Golgi marker (β-COP) and Calnuc-GFP or Calnuc(ΔSS)-CFP increases ACTH secretion triggered by mastoparan or GTPγS and the effect is G pro-tein-dependent  . (B). Introduction of anti-calnuc or anti-G protein antibodies decreases ACTH secretion triggered by mastoparan. In control cells (no-treatment) or those treated with preimmune or anti-Gα12 IgG, 55-60% of the total ACTH was secreted. After pre-treatment with anti-calnuc, anti-Gαi1/2, anti-Gαi3 and anti-GβIgG, ACTH secretion triggered by mastoparan was decreased to 37%, 43%, 32% and 37%, respectively. Results (mean ± SD) represent the average of values obtained in two separate experiments performed in duplicate. a n t i -G α i 1 / 2 a n t i -C a l n u c a n t i -G α i 3 a n t i -G Overexpressed calnuc-GFP is concentrated in light membrane fractions containing plasma membranes and Golgi membranes Figure 8 Overexpressed calnuc-GFP is concentrated in light membrane fractions containing plasma membranes and Golgi membranes. (A) Majority (90%) of the endogenous calnuc (NT) is in light fractions 10-12 containing Golgi and PM, with the remaining 10% found in heavy fractions (3)(4)(5)(6)(7)(8). Calnuc-GFP is similarly concentrated in light fractions (75%), but some (25%) is also found in heavy fractions 3-8 containing ER and RSG, where it cosediments with synaptobrevin 2 (Vamp 2) and ACTH, a marker for RSG. No major change in the distribution of Vamp 2 and ACTH is observed following overexpression of calnuc-GFP. The ER marker calnexin and the Golgi marker α-mannosidase II (Man II) are found in heavy fractions (3)(4)(5)(6)(7)(8) and light fractions (9-12), respectively. Overexpression of calnuc-GFP was not found to significantly change the distribution of ACTH and Vamp in the gradients. The percent of the total for a given protein in each quantified fraction was plotted in (B). The amount of each protein found in fractions 3-8 and 10-12 were pooled and plotted in the bar graph as the percent of total as shown in (C). Similar distribution change of indicated proteins was observed in 3 separate experiments depicted. Redistribution of Gαi subunits in AtT-20 cells overexpressing calnuc-GFP. In parental AtT-20 cells (NT), the majority of the endogenous Gαi1/2 (70%) is found in heavy fractions (3)(4)(5)(6)(7)(8) containing ER and secretory granules, whereas most of the endogenous Gαi3 (72%) is present in light fractions (10)(11)(12) containing Golgi and PM. In cells overexpressing calnuc-GFP (CNG) the % of the total Gαi1/2 found in heavy fractions is reduced (55%), that in light fractions is correspondingly increased (41%), and Gαi3 is more broadly distributed in fractions 3-12. No obvious change in the distribution of Gβ is observed following calnuc-GFP overexpression. Similar distribution change of indicated proteins was observed in 3 separate experiments depicted. The quantitative analysis shown by bar graph which was calculate based on the dot curve in left panel represent the result from the same illustrated experiment as indicated above. The experimental protocol is the same as for Figure 6. "B" and "C" were calculated and plotted as in Figure 8A.

Discussion
Calnuc, the first protein identified that binds to both Ca 2+ and heterotrimeric G proteins, was previously shown to be localized both in the Golgi and in the cytoplasm [3]. Our previous studies established that cytoplasmic calnuc specifically interacts with several Gα subunits in vivo as shown by both co-immunoprecipitation [3,4] and FRET analysis [8] and that calnuc binds the C-terminal α5-helix region of Gαi3 through its EF-hand Ca 2+ -binding region [7].
In this study we focused on defining the role of calnuc in G protein dependent activation of ACTH secretion and used GTPγS or mastoparan [15,23], a receptor mimetic activator of Gi/o subunits, to trigger ACTH secretion. We found that overexpression of calnuc increased nearly twofold ACTH secretion induced by mastoparan or GTPγS compared to non-transfected cells. The fact that no differences were seen between calnuc-GFP (located both within organelles along the secretory pathway and in the cytoplasm) and calnuc lacking its signal sequence (located exclusively in the cytoplasm) together with our finding that ACTH secretion is reduced when anti-calnuc antibodies are introduced into permeabilized cells demonstrate that it is the cytosolic pool that is responsible for calnuc's effects on secretion. By immunofluorescence we found that some of the calnuc(ΔSS)-CFP is associated with the cytoplasmic side of RSG, indicating that this association must occur by protein-protein interaction between calnuc and a binding protein found on the cytoplasmic surface of the RSG membrane.
Using antibodies against Gαi2 or Gαi3 we were able to similarly impair ACTH secretion, but antibodies against Gαi12 were without effect because mastoparan does not activate Gα12. We found that mastoparan increases the initial rate of GTPγS binding to Gαi3 four-fold compared to Gαi3 alone or in the presence of calnuc. Mastoparan does so by unwinding the α5 helix of Gαi1 which is highly homologous to Gαi3 [23]. Although mastoparan and calnuc have the same binding site in the α5 helix of Gαi3, it appears that mastoparan and calnuc have different mechanisms regarding regulation of ACTH secretion because unlike mastoparan, calnuc can not regulate G protein activity but affects G protein distribution.
We also investigated the effects of overexpression of calnuc-GFP on G protein distribution. Our immunofluorescence results demonstrate that in AtT-20 cells stably expressing calnuc-GFP, calnuc-GFP partially colocalizes with Gαi1/2 as well as Gαi3 on the PM and Golgi membranes which is similar to what we reported earlier in EcR-CHO cells overexpressing calnuc and Gαi3-GFP [7] or in COS-7 cells overexpressing calnuc-GFP and Gαi3-YFP visualized by FRET [8]. Here we show (see Fig. 2C) that in non-transfected AtT-20 cells Gαi1/2 is found mainly on secretory granules (RSG) but not the PM which is consistent with reports by others on chromaffin cells [14,19] and rat melanotrophs [20]. However, in AtT-20 cells overexpressing calnuc the distribution of Gαi1/2 is shifted in that more is found on the PM and on RSG based on results obtained by both immunofluorescence (Fig. 2D) and cell fractionation (Figs. 9, 10). When calnucΔSS-GFP lacking the calnuc signal sequence is overexpressed in the cytoplasm, the majority of Gαi1/2 was found to colocalize with cytosolic calnuc on RSG. These results suggest that there is a dynamic distribution of Gαi between the cytoplasm, the PM, and membranes of subcellular compartments such as the Golgi.
Based on our results we propose that cytosplasmic calnuc may play a role in recruiting Gαi3 onto the granules and Gαi1/2 onto the PM either from its cytosolic pool and/or from its Golgi pool (Fig. 11). However, it is possible that increasing the Ca 2+ concentration in the Golgi by calnuc overexpression [5] may also contribute to regulation of membrane trafficking [25,[27][28][29] which further results in redistribution of G proteins.
Calnuc resembles several other proteins that are also localized in two pools [30,31]. We established that calnuc is distributed both in the Golgi lumen and in the cytoplasm in numerous cell lines [3,7]. Recently, NEFA [32,33] a protein that has high homology to calnuc and also has a signal sequence, has also been found both within the Golgi lumen and in the cytoplasm [34]. Several mechanisms have been proposed to explain the dual localization of the same protein [30,31,35,36]. It has been suggested that under some conditions the signal sequence may be masked by binding proteins other than SRP [35,36]. The detailed mechanism for generating two pools of calnuc remains to be elucidated.
Regulated exocytosis has been extensively investigated to date in AtT-20 and other cells [37][38][39][40]. It has become clear that it requires multiple steps that need to be controlled in time and space. Ca 2+ and heterotrimeric G proteins have been shown to affect exocytosis either separately or in synergy depending on the cell type. In AtT-20 cells exocytosis is dependent on both Ca 2+ and heterotrimeric G proteins [12,41]. We found that calnuc binds both Ca 2+ and heterotrimeric G proteins and is found on secretory vesicles which places it in a strategic location to serve as a vesicle bound controller of regulated secretion.
In summary, in this study we demonstrate that overexpression of calnuc, a Ca 2+ and heterotrimeric G protein Total Gαi1/2 associated with PM is increased following overexpression of calnuc-GFP

Materials
Polyclonal rabbit IgG (F-5059) against recombinant full length calnuc was generated and affinity purified as previously described [3]. Affinity purified rabbit anti-calnuc IgG raised against the C-terminal 14 amino acids of rat/mouse calnuc (EQPPVLPQLDSQHL) or human calnuc (LLERLPE-
For immunofluorescence, cells were fixed in 2% paraformaldehyde in phosphate buffer and permeabilized as Diagram of dynamic distribution of calnuc Figure 11 Diagram of dynamic distribution of calnuc. There are two pools for caluc: cytosolic and the Golgi membrane associated. Cytosolic calnuc interacts with Gαi subunits on the Golgi and vesicles budded from the Golgi, whereas luminal calnuc retains in the Golgi lumen for several hours, followed by secretion along the constitutive-like pathway. Calnuc binding the Golgi lumen is subjected to post-translational modification, and as the major calcium binding protein, plays a key role in constitution of a calcium pool in the Golgi. Cytosolic calnuc involves distribution of Gαi subunits via binding to them, and further participates regulation of G protein related intracellular pathways.
For deconvolution analysis of immunofluorescence results, images were collected with an Applied Precision DeltaVision imaging system (Issaquah, WA) coupled to a Zeiss S100 fluorescence microscope (Carl Zeiss; Thornwood, NY). Cross-sectional images of cells were obtained with 150-nm step width to optimize reconstruction of the center plane image. Deconvolution was done on a Silicon Graphics Octane ® visual workstation (SGI, Mountain View, CA) equipped with Delta Vision reconstruction software.

Assessment of Constitutive Secretion of Glycosaminoglycans (GAG) from AtT-20 Cells
Non-transfected AtT-20 cells or those transfected with calnuc-GFP were pretreated with 0.5 mM xyloside at 37°C for 30 min and subsequently pulse-labeled with [ 35 S] sulfate (150 μCi/ml) (ICN Biomedicals) for 5 min as described [22]. Labeled GAG chains secreted into the medium at selected intervals from 15 min to 2 h were precipitated with cetylpyridinium chloride (CPC). Samples were collected by vacuum filtration and counted by liquid scintillation as previously described [22].

Immunoprecipitation of Metabolically Labeled ACTH and its Precursors
Parental AtT-20 cells or cells stably expressing calnuc-GFP were pulse-labeled [42] with [ 35 S]Met (0.5 mCi/ml) (NEN ® Life Science Products, Boston, MA) for 20 min at 37°C and subsequently chased in unlabeled DMEM for 2 h. ACTH released into the medium was immunoprecipitated with anti-ACTH IgG and protein A beads (Cytelligen Corp., San Diego, CA), followed by separation on 10-20% Tris-Tricine gels (Bio-Rad) and autoradiography with Kodak film [21].

ACTH Release from Permeabilized AtT-20 Cells
The permeabilization protocol used followed that described previously by others [13]. In brief, stably transfected or non-transfected AtT-20 cells were plated on tissue culture plates (5 × 10 4 /well) for 48 h, and subsequently rinsed with 0.1% BSA in DMEM, followed by permeabilization buffer (20 mM digitonin, 137 mM NaCl, 2.7 mM KCl, 5.6 mM glucose, 1 mg/ml BSA, 20 mM Hepes, pH 7.2), plus either 10 μM mastoparan (Neosystem Laboratoire, Strasbourg, France) or 100 μM GTPγS (Roche Molecular Biochemicals, Indianapolis, IN) at 37°C for 15 min [13,15,43]. To introduce antibodies into cells, affinity purified anti-calnuc (F-5059), anti-Gαi1/2 (AS), anti-Gαi3 (EC) or anti-GβIgG (30 μg/ml) were added and incubated with cells in permeabilization buffer at 37°C for 10 min according to a published protocol [44], followed by addition of 10 μM mastoparan for 15 min. Media were collected, and cells were lysed in 0.5% Triton X-100 at 4°C for 30 min, followed by centrifugation (14,000 × g for 5 min). ACTH in supernatants of both the medium and cell lysate were assessed by enzymelinked immunosorbent assay (ELISA) using ACTH (Rat) EIAH kits (Peninsula Laboratories, San Carlos, CA) with a Vmax Kinetic Microplate Reader (λ = 450 nm) (Molecular Devices, Sunnyvale, CA). ACTH secretion was plotted as percent of total ACTH (secreted ACTH in supernatants vs secreted + intracellular ACTH). The results from each experiment were subjected to statistic analysis, and the final plotted results (mean ± SD) represent the average of values obtained in indicated separate experiments performed in either duplicate or triplicate as shown in figure legend.

Subcellular Fractionation
Membrane (100,000 g pellets) and cytosolic fractions (100,000 g supernatants) were prepared by centrifugation of postnuclear supernatants from AtT-20 cells and analyzed by immunoblotting and ECL [3].
Light membranes containing PM and Golgi membranes were separated from heavy fractions containing ER and granules by centrifugation on discontinuous sucrose gradients using protocols similar to those previously published with minor modifications [3,5]. Briefly, postnuclear supernatants (PNS) prepared from AtT-20 cells were loaded on the top of sucrose step gradient containing 0.2, 0.4, 0.6, 1.0, 1.4, and 1.8 M in 1 mM Tris-HCl,