Aberrant Localization of Intracellular Organelles, Ca2+ Signaling, and Exocytosis in Mist1 Null Mice*

Ca2+ signaling and exocytosis are highly polarized functions of pancreatic acinar cells. The role of cellular architecture in these activities and the capacity of animals to tolerate aberrant acinar cell function are not known. A key regulator of acinar cell polarity is Mist1, a basic helix-loop-helix transcription factor. Ca2+ signaling and amylase release were examined in pancreatic acini of wild type and Mist1 null mice to gain insight into the importance of cellular architecture for Ca2+ signaling and regulated exocytosis. Mist1–/– acinar cells exhibited dramatically altered Ca2+ signaling with up-regulation of the cholecystokinin receptor but minimal effect upon expression of the M3 receptor. However, stimulation of inositol 1,4,5-trisphosphate production by cholecystokinin and carbachol was inefficient in Mist1–/– cells. Although agonist stimulation of Mist1–/– cells evoked a Ca2+ signal, often the Ca2+ increase was not in the form of typical Ca2+ oscillations but rather in the form of a peak/plateau-type response. Mist1–/– cells also displayed distorted apical-to-basal Ca2+ waves. The aberrant Ca2+ signaling was associated with mislocalization and reduced Ca2+ uptake by the mitochondria of stimulated Mist1–/– cells. Deletion of Mist1 also led to mislocalization of the Golgi apparatus and markedly reduced digestive enzyme content. The combination of aberrant Ca2+ signaling and reduced digestive enzyme content resulted in poor secretion of digestive enzymes. Yet, food consumption and growth of Mist1–/– mice were normal for at least 32 weeks. These findings reveal that Mist1 is critical to normal organelle localization in exocrine cells and highlight the critical importance of maintaining cellular architecture and polarized localization of cellular organelles in generating a propagating apical-to-basal Ca2+ wave. The studies also reveal the spare capacity of the exocrine pancreas that allows normal growth and development in the face of compromised exocrine pancreatic function.

Ca 2ϩ signaling regulates virtually all cell functions, including long term functions such as transcription and translation and short term functions such as neurotransmission and exocytosis (1,2). In the pancreas, activation of Ca 2ϩ signaling by G protein-coupled receptors (GPCRs) 1 plays a central role in digestive enzyme secretion (3). Abnormal enzyme secretion can lead to numerous diseases, including malnutrition in cystic fibrosis (4) and, more commonly, acute pancreatitis (5). Several studies have shown a strong association between aberrant Ca 2ϩ signaling and pancreatitis (6 -8). The polarized function of exocrine secretory cells requires polarized Ca 2ϩ signaling. Indeed, stimulation of pancreatic acini GPCRs with physiological agonist concentrations triggers repetitive [Ca 2ϩ ] i oscillations in the form of Ca 2ϩ waves that initiate at the apical pole and propagate to the basal pole (9 -13).
The apical-to-basal Ca 2ϩ waves are achieved by clustering Ca 2ϩ signaling complexes at the apical pole (14 -16). This leads to an apical-to-basal gradient of responsiveness, with the most responsive GPCR Ca 2ϩ signaling complexes at the apical pole (16). In fact, functional mapping of Ca 2ϩ signaling complexes has revealed that physiological Ca 2ϩ signals are mostly triggered by the stimulation of Ca 2ϩ signaling complexes at the apical pole (16). Once launched, the shape of the Ca 2ϩ signals are regulated by many processes that affect the biochemical component of the Ca 2ϩ signal that generates inositol 1,4,5trisphosphate (IP 3 ) as well as the biophysical component that includes the Ca 2ϩ release and influx channels and the sarco/ endoplasmic reticulum ATPase (SERCA2b) and plasma membrane Ca 2ϩ ATPase (PMCA) pumps (16).
Other important regulators of Ca 2ϩ signaling in all cells (17), including pancreatic acinar (18,19) and other secretory cells (20), are mitochondria. The seminal work of Rizzuto, Pozzan and co-workers showed that the mitochondria are in close proximity to the ER (21), thereby aligning the mitochondrial Ca 2ϩ uptake pathway with the IP 3 Rs to incorporate a large portion of the Ca 2ϩ released from the ER (22,23). Subsequently, mitochondria were shown to communicate with I crac channels to regulate Ca 2ϩ influx across the plasma membrane (24,25). Mitochondrial localization in pancreatic acini is even more intricate. In these cells, three populations of energized mitochondria have been identified, namely a belt capping the secretory granules, a ring surrounding the nucleus, and a string lining the plasma membrane (19). The belt capping the secre-tory granules functions to confine Ca 2ϩ signals to the apical pole at very low stimulus intensity (18), whereas the ring surrounding the nucleus isolates the nucleus from cytosolic Ca 2ϩ tides (15,19).
The intricate localization of Ca 2ϩ signaling complexes and intracellular organelles is likely to be critical for the precise operation of the Ca 2ϩ signaling apparatus and for regulated exocytosis in acinar cells. These assumptions can be directly examined only by testing the effect of perturbation of the cellular architecture on pancreatic acinar cell function. This has been a difficult problem to address but has now become possible with the availability of Mist1 null (Mist1 Ϫ/Ϫ ) mice (26). Mist1 is a basic helix-loop-helix transcription factor that is essential for the normal development of serous acinar cells in various secretory glands, including the pancreas and salivary glands (26 -30). Deletion of the Mist1 gene (26,30) or inhibition of Mist1 function (28,29) leads to severe distortion of acinar cellular architecture, including loss of gap junctions and intercellular communication (28,30), disorder of secretory granules (26 -30), and acinar-to-ductal metaplasia (29). Additional changes in Mist1 Ϫ/Ϫ cells include up-regulation of mRNA coding for the cholecystokinin (CCK) receptors and down-regulation of expression of IP 3 R3 (26). However, the effect of distorted cellular architecture on Ca 2ϩ signaling, pancreatic function, and animal growth and development has not been examined. In the present work we show that the Ca 2ϩ signaling machinery in Mist1 Ϫ/Ϫ cells is dramatically altered, resulting in aberrant agonist-evoked Ca 2ϩ oscillations and Ca 2ϩ waves. The altered Ca 2ϩ response is not due to mislocalization of the remaining IP 3 receptors but rather is a consequence of mislocalization and aberrant Ca 2ϩ uptake into mitochondria. The Golgi apparatus of Mist1 Ϫ/Ϫ cells is diffuse and fragmented, which may explain the markedly decreased digestive enzyme content in the secretory granules. Exocytosis by Mist1 Ϫ/Ϫ acinar clusters is undetectable at physiological agonist concentrations and is greatly reduced at pharmacological agonist concentrations. Remarkably, food consumption and weight gain remain similar for WT and Mist1 Ϫ/Ϫ mice. These findings provide experimental evidence for the importance of cellular architecture in the generation and propagation of Ca 2ϩ signals and illustrate the spare capacity of the exocrine pancreas to allow normal growth at reduced exocytosis.
Experimental Animals, Body Weights, and Food Intake-Mist1 Ϫ/Ϫ mice were generated as described previously (26). All animals were housed in the animal care facility, and all protocols were approved by the University of Texas Medical Center at Dallas Animal Care Committee and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. WT and Mist1 Ϫ/Ϫ mice were housed individually. Food intake was measured three times per week using an Ohaus portable electronic scale with a sensitivity of 0.1 g. The body weight of litter mates was measured on the day of birth and then at 1-week intervals.
Preparation of Pancreatic Acini-Pancreatic acini and small acinar clusters were prepared by collagenase digestion as detailed previously (31). After isolation, the cells were suspended in solution A (140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM HEPES (pH 7.4 with NaOH), 10 mM glucose, 0.1% mM bovine serum albumin, and 0.02% soybean trypsin inhibitor) and kept on ice until used.
[Ca 2ϩ ]i Imaging-Pancreatic acinar cells were loaded with Fura2, and [Ca 2ϩ ] i was imaged as described (32). In brief, to measure the dose response for agonists the Fura2 fluorescence ratio was measured at excitation wavelengths of 350 and 380 nm, and the ratio was calibrated to obtain [Ca 2ϩ ] i . Ca 2ϩ waves were recorded using a single excitation wavelength of 380 nm. The image of resting cells was acquired and taken as the fluorescence signal at time 0 (F 0 ). All subsequent images were divided by this image, and the traces and images are the calculated F t /F 0 , where F t is the fluorescence at time t.
Simultaneous Measurement of Cytosolic and Mitochondrial Ca 2ϩ -Rhod-2 and Fluo-3 were used for the measurement of mitochondrial and cytosolic Ca 2ϩ signals, respectively. Loading with Rhod-2 was achieved by incubating acini is solution A with 8 M Rhod-2/AM at 4°C for 15 min. The acini were washed by centrifugation for 2 min at 30 ϫ g to remove excess dye and then resuspended in solution A. The acini were incubated at room temperature for 30 min to hydrolyze the trapped Rhod-2. During this incubation, the acinar suspension was supplemented with 4 M Fluo-3/AM. Finally, the acini were washed with solution A and kept on ice until use. Dye-loaded cells were transferred to a perfusion chamber, and Fluo-3 and Rhod-2 fluorescence was measured with a confocal laser-scanning system from Bio-Rad (MRC-1024) using the 488-and 568-nm lines, respectively. Laser intensity was reduced to 1-3% with neutral density filters to reduce photo bleaching. Images were recorded at a frequency of 1 Hz. The images were analyzed using MetaMorph software.
Mitochondrial Membrane Potential-Mitochondrial membrane potential was estimated from the incorporation of the potential-sensitive dye TMRM. Preliminary experiments showed that when monitored continuously, even at 1 M the extent of TMRM incorporation was proportional to the mitochondrial membrane potential. TMRM uptake was measured by continuously perfusing acinar cells with a solution containing 1 M TMRM. Once a plateau was reached, the acini were perfused with dye-free medium and then with medium containing 10 M carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone. TMRM fluorescence was recorded with the 488-nm line of the Bio-Rad MRC-1024 confocal microscope.
Measurement of IP 3 -IP 3 levels were measured by a radioligand assay as described elsewhere (33). Acini suspended in solution A and incubated at 37°C were stimulated with the indicated CCK or carbachol concentrations for 5-15 s, depending on agonist concentration. The reactions were stopped by the addition of perchloric acid, vigorous mixing, and incubation on ice for at least 10 min to allow precipitation of proteins. The supernatants were collected and transferred to clean tubes. Standards of IP 3 were prepared in the same manner. The perchloric acid was removed and IP 3 extracted by the addition of 0.15 ml of Freon and 0.15 ml of tri-n-octylamine. IP 3 content in the aqueous phase was measured by displacement of [ 3 H]IP 3 using microsomes prepared from bovine brain cerebella.
RT-PCR Analysis of CCK and M3 Receptors Expression-Acinar cell digests were placed in a Petri dish, and small clusters consisting of 3-5 cells were collected with a Pasteur pipette under microscopic examination to ensure lack of contamination with other cell types. The RNA was extracted from the acinar cells and brains of WT and Mist1 Ϫ/Ϫ mice with TRIzol reagent (Invitrogen) and dissolved in diethyl pyrocarbonate-treated water. RT-PCR was performed using the same amount of RNA isolated from three WT and three Mist1 Ϫ/Ϫ mice. The RT reaction was performed with the SuperScript TM II RT kit (Invitrogen) in a 20-l reaction volume as suggested by the manufacturer. PCR primers were designed using Primer 3 version 0.2 and alignment by NCBI BLAST software. The primer sequences used were as follows: CCKA receptor sense, 5Ј-TCAGTGACCTCATGCTTTGC-3Ј, and CCKA receptor antisense, 5Ј-ATGAGTCCGTAAGCCACCAC-3Ј (size of PCR product, 442 bp); muscarinic receptor 3 sense, 5Ј-TGCTGGTGATCAGCTTGAC-3Ј, and muscarinic receptor 3 antisense, 5Ј-TTCGTGCCTTGCTGTTG-TAG-3Ј (size of PCR product, 427 bp); ␤-actin-sense, 5Ј-TGTTAC-CAACTGGGACGACA-3Ј, and ␤-actin antisense, 5Ј-TCTCAGCTGTG-GTGAAG-3Ј (size of PCR product, 392 bp). The PCR reaction was initiated by a 5-min hot start at 94°C followed by 35 amplification cycles that consisted of 50 s of incubation at 94°C, 1 min at 55°C, and 1 min at 72°C, extended by 10 min at 72°C, and terminated by incubation at 4°C. Preliminary experiments using different amounts of cDNA were used to determine the optimal conditions, and amplification of actin mRNA was used to calibrate between samples.
Immunoblotting-Brain microsomes were prepared by homogenizing brain tissue from WT and Mist1 Ϫ/Ϫ mice in a buffer containing (pH 7.6 with KOH) 100 mM KCl, 20 mM Tris-base, 1 mM EDTA, 1 mM benzamidine, and 1 mM phenylmethylsulfonyl fluoride. The homogenate was centrifuged at 1000 ϫ g for 10 min at 4°C. The supernatant was collected and centrifuged at 40,000 ϫ g for 30 min. The pellet was resuspended in homogenization buffer, and the microsomes were stored at Ϫ80°C until use. The microsomes were extracted by a 1-h incubation on ice with a buffer containing 50 mM Tris (pH 6.8 with HCl), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA and 0.5% Triton X-100 supplemented with protease inhibitors (0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 15 g/ml aprotinin, and 1 mM benzamidine). Released proteins were separated by SDS-PAGE and transferred to 0.2-m polyvinylidene difluoride membranes, and the membranes were then blocked by a 1-h incubation at room temperature in 5% nonfat dry milk in TTBS solution containing 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.05% Tween 20. The IP 3 R1, IP 3 R2, IP 3 R3, PMCA, SERCA2b, and M3 receptors were detected by a 1-2-h incubation of individual membranes with the respective antibodies diluted in TTBS.
Immunocytochemistry-Immunostaning was performed as detailed previously (31,32). Frozen pancreatic sections were fixed and permeabilized with 0.5 ml of cold methanol for 10 min at Ϫ20°C. The sections were washed with phosphate-buffered saline alone and phosphatebuffered saline containing 50 mM glycine, and the nonspecific sites were blocked by a 1-h incubation with PBS containing 5% goat serum, 1% bovine serum albumin, and 0.1% gelatin. The medium was aspirated and replaced with 50 l of blocking medium containing control serum or a 1:10 dilution of antibodies against the M3 muscarinic receptor (M3R), IP 3 R2, and IP 3 R3 and a 1:1000 dilution of antibodies against mannosidase II. After incubation with the primary antibodies overnight at 4°C and three washes with the incubation buffer, the antibodies were detected with goat anti-rabbit or anti-mouse IgG tagged with fluorescein or rhodamine. Images were collected with a Bio-Rad MRC 1024 confocal microscope.
Measurement of Amylase, Lipase, and Trypsin-Freshly isolated acini from one mouse were suspended in 50 ml (for amylase) or 15 ml (for lipase and trypsin) of solution A. To measure total enzyme activity, 1 ml of cell suspension was lysed by the addition of 1% Triton X-100 and centrifuged for 2 min at 2000 ϫ g at 4°C, and the supernatant was collected. A second sample was used to measure protein content. Enzyme activity was measured in parallel from WT and Mist1 Ϫ/Ϫ mice. To reduce variation due to feeding schedule, the mice were fasted for 24 h before the preparation of acini, and the results are expressed as the ratio between the activities in WT/Mist1 Ϫ/Ϫ cells. For the measurement of exocytosis, portions of the 1.5-ml cell suspension were transferred to vials containing agonists to give the desired final concentrations. After 30 min of incubation at 37°C, samples were transferred to Eppendorf tubes, the supernatants were separated from the acini by centrifugation, and the amylase released to the medium was measured. In each experiment, samples of cells were lysed to measure the total amylase content, and exocytotic amylase release was calculated as a fraction of total amylase content.
Amylase activity was measured with a Phadebas kit (Pharmacia and Upjohn Diagnostics, catalog number 10-5380-33) as described previously (34). In brief, 10-l samples were diluted into 200 l of buffer containing 20 mM NaH 2 PO 4, 20 mM Na 2 HPO 4 , 50 mM NaCl, and 0.02% NaN 3, pH 7.0. 20 l of the diluted samples were mixed with 1 ml of blue starch (10 mg/ml) and incubated for 10 min at 37°C with gentle shaking. The reaction was stopped by the addition of 250 l of 2 M NaOH, the supernatant was cleared by 5 min of centrifugation at 14,000 rpm, and the absorbance was measured at a wavelength of 595 nm.
Lipase activity was measured using para-nitrophenyl palmitate as a substrate (35). Samples of 0.1 ml were added to 2.4 ml of a freshly prepared para-nitrophenyl palmitate solution (30 mg of para-nitrophenyl palmitate in 100 ml of 100 mM Tris buffer, pH 8.5, 207 mg of Na ϩ -deoxycholate, and 100 mg of gum arabic) and incubated for 1 h at 37°C with gentle shaking. The reactions were terminated by the addition of 0.2 ml of 100 mM CaCl 2 , the samples were centrifuged at 14,000 rpm for 2 min, and the absorbance was measured at a wavelength of 410 nm.
Trypsin activity was determined with the substrate N-␣-benzoyl-arginine-p-nitroanilide as described previously (36) with a slight modification. Samples of 100 l were added into 2.4 ml of 1 mM substrate N-␣-benzoylarginine-p-nitroanilide dissolved in a solution containing 100 mM Tris, pH 9.0, and 10 mM CaCl 2 and incubated for 1 h at 37°C. The reactions were stopped by addition of 5 l of 0.5 mg/ml soybean trypsin inhibitor, and the change in absorbance at 410 nm was monitored.
Statistics-When appropriate, results are presented as the mean Ϯ S.E. of the indicated number of experiments. Statistical significance was evaluated by a two-way analysis of variance. All immunostaining experiments were repeated at least five times with similar results.

Impaired Stimulation of Ca 2ϩ Signaling in Mist1 Ϫ/Ϫ Cells-
The up-regulation of the CCK receptor mRNA in Mist1 Ϫ/Ϫ mouse pancreatic acinar cells (26) was confirmed by RT-PCR in pancreatic acinar cells (Fig. 1A). However, this increase was not observed for all GPCRs. Initial analysis by RT-PCR showed that expression of M3R mRNA in the pancreas of Mist1 Ϫ/Ϫ mice was unchanged or slightly up-regulated (Fig. 1A). An attempt to quantitate the extent of protein expression of M3Rs by Western blot in pancreatic extracts failed because of poor signal/noise. However, the antibodies gave a reasonable signal in immunostaining. Fig. 1, B and C, show enrichment of M3R expression at the apical pool of cells from WT and Mist1 Ϫ/Ϫ mice and comparable staining intensity. To further analyze receptor expression, we examined the expression of mRNA and protein in the brains of the mice. Fig. 1D shows that, similar to findings in the pancreas, brain mRNA for CCK receptors was up-regulated, and for M3Rs it was unchanged. Western blot analysis of extracts prepared from five brains of WT and five brains of Mist1 Ϫ/Ϫ mice showed a small reduction in the level of M3Rs protein in Mist1 Ϫ/Ϫ mice (18 Ϯ 5%, n ϭ 5) (Fig. 1E).
The aberrant localization of secretory granules, the reduction in expression of IP 3 R3, and the up-regulation of CCK receptor mRNA in the Mist1 Ϫ/Ϫ mouse pancreatic acini (26) raised the question of how Ca 2ϩ signaling is affected in these cells. Measurement of IP 3 production showed that signaling by all GPCRs was impaired in Mist1 Ϫ/Ϫ cells. Loss of the Mist1 protein reduced the EC 50 for CCK stimulation from ϳ0.83 to 0.032 nM and reduced the maximal production of IP 3 by ϳ50% ( Fig. 2A). Mist1 Ϫ/Ϫ acinar cells showed an increased EC 50 for carbachol from ϳ3.7 to 46 M and a reduced maximal production of IP 3 of ϳ30% (Fig. 2B). The increased apparent affinity to CCK may relate to the increased mRNA levels of the CCK receptors (26) (Fig. 1A). The modest change in M3R mRNA and protein is consistent with this interpretation. However, the reduction in maximal IP 3 production indicates a generally impaired activation of G proteins by the M3R and an impaired activation of phospholipase C␤ by GPCRs.
The consequence of impaired IP 3 production on the pattern of Ca 2ϩ signaling is shown in Figs. 3 and 4. Panels A and B in each figure show example traces from individual experiments, whereas the panels C (Figs. 3 and 4) summarize results from multiple experiments. CCK stimulated Ca 2ϩ signaling with a higher affinity in Mist1 Ϫ/Ϫ cells (Fig. 3C), whereas the affinity for carbachol to trigger a Ca 2ϩ signal was lower in Mist1 Ϫ/Ϫ cells (Fig. 4C), as was expected from their effects on IP 3 production. However, maximal concentrations of both agonists increased [Ca 2ϩ ] i to similar levels in WT and Mist1 Ϫ/Ϫ cells (CCK, 723 Ϯ 81 nM in WT and 744 Ϯ 79 nM in Mist1 Ϫ/Ϫ ; carbachol, 726 Ϯ 88 nM in WT and 711 Ϯ 63 nM in Mist1 Ϫ/Ϫ ; n was Ͼ35 acini from 11 Mist1 Ϫ/Ϫ and 11 age-matched WT mice, ages 2-11 months). A notable difference between WT and Mist1 Ϫ/Ϫ cells was that often agonists did not induce [Ca 2ϩ ] i oscillations in the Mist1 Ϫ/Ϫ cells but rather a single Ca 2ϩ transient with a subsequent plateau (Figs. 3, A and B and 4,  A and B).
The Ca 2ϩ signal in pancreatic acinar cells occurs as a propagated Ca 2ϩ wave (9 -13). Therefore, we measured Ca 2ϩ waves in WT and Mist1 Ϫ/Ϫ cells. Stimulation of WT acini with low agonist concentrations such as 1 M carbachol (not shown) or 10 pM CCK (Fig. 5A) resulted in a slowly propagating Ca 2ϩ wave that was initiated at the apical pole. On the other hand, in multiple attempts it was almost impossible to resolve a Ca 2ϩ wave at low agonist concentrations in Mist1 Ϫ/Ϫ cells (for example, Fig. 5B). However, we noted that [Ca 2ϩ ] i increased faster in Mist1 Ϫ/Ϫ cells. The Ca 2ϩ waves in Mist1 Ϫ/Ϫ cells could sometimes be resolved at intermediate agonist concentrations. Stimulation of heterozygous Mist1 ϩ/Ϫ cells (as well as WT cells) with 5 M carbachol generated Ca 2ϩ waves that propagated through the cell at a rate of 16.7 Ϯ 0.5 m/sec (Fig. 5C). Interestingly, although Mist1 Ϫ/Ϫ cells also initiated a Ca 2ϩ signal at the apical pole, even at the intermediate agonist concentrations it was not always possible to observe a clear Ca 2ϩ wave for two main reasons (Fig. 5B). First, the Ca 2ϩ wave in Mist1 Ϫ/Ϫ cells propagated very rapidly. When it could be resolved, the Ca 2ϩ wave propagated at a rate of 43.6 Ϯ 5.8 (n ϭ 14) m/s. Second, the Ca 2ϩ wave in Mist1 Ϫ/Ϫ cells propagated along the cell periphery and only then proceeded concentrically toward the cell center (Fig. 5D). These results suggest that the proteins and organelles that control the propagation of the Ca 2ϩ wave do not function properly in the Mist1 Ϫ/Ϫ cells. This can be due to mislocalization of IP 3 Rs that are concentrated at the apical pole of pancreatic acini (14,15), mislocalization of the mitochondria that determine the pattern of the Ca 2ϩ wave (18 -20), or both.
Expression and Localization of Ca 2ϩ Signaling Proteins-The Ca 2ϩ waves in pancreatic acini are regulated by localization of IP 3 receptors (14,15) and other Ca 2ϩ signaling proteins (13,16) and by Ca 2ϩ uptake into the mitochondria (18 -20). Our previous work has shown that the level of IP 3 R3 is downregulated in Mist1 Ϫ/Ϫ cells (26). To test if the down-regulation of IP 3 R3 is a more general phenomenon and whether deletion of the Mist1 gene affects localization of Ca 2ϩ signaling complexes, we analyzed the expression of individual IP 3 Rs in brain extracts and the localization of IP 3 R2 and IP 3 R3 in pancreatic acini. The Western blot analysis in Fig. 6A shows that deletion of Mist1 reduced expression of IP 3 R3 in the brain by ϳ65 Ϯ 11% (n ϭ 5). On the other hand, expression of IP 3 R1, IP 3 R2, SERCA2b, and PMCA was not affected. This indicates that down-regulation of IP 3 R3 expression in Mist1 Ϫ/Ϫ mice is not specific to acinar cells, raising the possibility that loss of Mist1 may effect IP 3 R3 expression in other cell types. The immunolocalization in Fig. 6B revealed a lack of IP 3 R3 and completely normal localization of IP 3 R2 in pancreatic acini of Mist1 Ϫ/Ϫ cells.
The results in Fig. 6 indicate that the distorted Ca 2ϩ signaling and Ca 2ϩ waves observed in Mist1 Ϫ/Ϫ cells are not due to a general aberrant expression of Ca 2ϩ signaling proteins or to mislocalization of the remaining IP 3 Rs. Furthermore these findings and those in Figs. 3 and 4 indicate that the function of IP 3 R3, and perhaps other IP 3 Rs, is cell-specific. A recent work provides strong evidence in cell lines that IP 3 R1 is essential for receptor-triggered Ca 2ϩ oscillations, whereas IP 3 R3 functions as an anti-Ca 2ϩ oscillatory unit (38). That is, knock-down of IP 3 R3 by small interfering RNA enhanced Ca 2ϩ oscillations. On the other hand, Ca 2ϩ oscillations were rare in Mist1 Ϫ/Ϫ cells, where the IP 3 R3 is markedly down-regulated (Figs. 3  and 4). This would suggest a cell-specific function of IP 3 R3.

Mitochondrial Localization and Function in Mist1
Ϫ/Ϫ Cells-The Ca 2ϩ waves in pancreatic acini and Ca 2ϩ signaling in many cells are prominently regulated by Ca 2ϩ uptake into the mitochondria (18 -25). Therefore, we next compared localization and function of the mitochondria in WT and Mist1 Ϫ/Ϫ cells. Localization of mitochondria was followed using Mito-Tracker Green detection, and mitochondrial and cytosolic Ca 2ϩ levels were measured with Rhod-2 and Fluo-3, respectively. As expected, the energized mitochondria in WT cells were clustered around the secretory granules and the nucleus and next to the plasma membrane, but they were completely excluded from the secretory granule area in pancreatic acini (Fig. 7, A, B, and E) (19). By contrast, mitochondria in Mist1 Ϫ/Ϫ acinar cells were highly disorganized (Fig. 7, C, D, and F). Energized mitochondria were found at all regions of the basal pole. In addition, although mostly excluded from the apical pole in Mist1 Ϫ/Ϫ cells, some mitochondria could also be found at the periphery of the apical pole or sometimes within the apical pole itself.
Disorganization of the mitochondria in Mist1 Ϫ/Ϫ cells resulted in the inhibition of Ca 2ϩ uptake into the mitochondria when the cells were stimulated with low agonist concentrations. In WT cells, [Ca 2ϩ ] i and mitochondrial Ca 2ϩ increases suggests that mislocalization of mitochondria rather than loss of mitochondrial membrane potential was the major reason for reduced Ca 2ϩ uptake into the mitochondria at a low agonist concentration. This was verified directly by measuring the mitochondrial membrane potential with TMRM. Fig. 7, I and J show the similar TMRM uptake in WT and Mist1 Ϫ/Ϫ cells, respectively.
Mitochondrial localization has a critical role in Ca 2ϩ signaling. Previous work has emphasized the importance of the close apposition of the mitochondria to the ER for Ca 2ϩ uptake into the mitochondria (22,23), regulation of the I crac Ca 2ϩ influx channel (24,25), and regulation of the Ca 2ϩ wave in secretory cells (18 -20). Furthermore, Ca 2ϩ uptake into the mitochondria is essential for mitochondrial and cellular energy metabolism (39,40). The disorganization of the mitochondria in Mist1 Ϫ/Ϫ cells can explain the lack of Ca 2ϩ uptake by the mitochondria when the cells are stimulated at a low agonist concentration. Such an uptake requires close communication between the mitochondria and the ER (21,23). Because Ca 2ϩ uptake by the mitochondria plays an important role in controlling the Ca 2ϩ wave (18 -20), the disorganization of mitochondria is likely responsible for the rapid spreading of Ca 2ϩ from the apical to the basal pole and the concentric Ca 2ϩ signal observed in Mist1 Ϫ/Ϫ cells (Fig. 5).
Digestive Enzyme Content in Mist1 Ϫ/Ϫ Cells-To determine the effect of deleting the Mist1 gene on pancreatic exocrine function, it was necessary to first measure how the absence of the Mist1 protein affects digestive enzyme content. Examination of the ER and the Golgi apparatus, which govern protein synthesis, revealed that the overall ER structure remained normal in Mist1 Ϫ/Ϫ cells (29). On the other hand, deletion of Mist1 modified the structure of the Golgi apparatus. Staining the Golgi with mannosidase II revealed a diffused Golgi organization in Mist1 Ϫ/Ϫ cells (Fig. 8A). As expected, the diffused Golgi resulted in reduced digestive enzyme content in pancreatic acini. Mist1 Ϫ/Ϫ cells contained 2.5-3-fold less amylase, trypsin, and lipase than WT cells (Fig. 8B). Amylase content measured in mice as young as 1 month and as old as 15 months were found to be similarly reduced, suggesting that reduction in digestive enzyme content does not develop with time but is a relatively early defect in Mist1 Ϫ/Ϫ acinar cells. This observation is consistent with Mist1 controlling the transcriptional regulation of genes involved in the localization of several organelles, including the secretory granules, the mitochondria, and the Golgi apparatus in pancreatic acini, and possibly functioning as a master regulator of intracellular organelle localization. Further studies will be needed to establish this intriguing possibility.  (3), and aberrant Ca 2ϩ signaling is intimately associated with pancreatitis (6 -8). In addition, digestive enzyme content is low in Mist1 Ϫ/Ϫ cells. Therefore, we expected that digestive enzyme secretion would be modified in Mist1 Ϫ/Ϫ cells and that the Mist1 Ϫ/Ϫ mice would show retarded growth or higher food consumption. Gross inspection of the pancreases of Mist1 Ϫ/Ϫ mice did not reveal any major differences in the size or shape of the organ. However, the pancreas of Mist1 Ϫ/Ϫ mice tended to be more fibrotic than that of WT mice, as suggested by the need for longer digestion by collagenase to liberate Mist1 Ϫ/Ϫ acini (not shown). Large acinar clusters comprise of 8 -30 cells prepared from WT and Mist1 Ϫ/Ϫ mice were used to measure stimulated amylase secretion as a measure of exocytosis. In the large clusters, the structure of the acini and the stimulated enzyme secretion is preserved and faithfully reflects enzyme secretion in vivo (41). As shown in Fig. 9, exocytosis in response to both carbachol and CCK was markedly impaired in Mist1 Ϫ/Ϫ acini. In fact, no exocytosis could be measured in Mist1 Ϫ/Ϫ acini stimulated with physiological agonist concentrations. Low exocytosis in Mist1 Ϫ/Ϫ acini was observed only at the very high agonist concentrations that are pathological in WT animals. Furthermore, after correction for total amylase content, the peak amylase release by Mist1 Ϫ/Ϫ acini detected at the higher agonist concentrations was only 50% of that measured in WT acini. Because high agonist concentrations increase [Ca 2ϩ ] i to similar levels in WT and Mist1 Ϫ/Ϫ cells (Figs. 3 and 4), this finding suggests that mislocalization of secretory granules contributed to the aberrant exocytosis. This was tested directly by measuring the response to increasing [Ca 2ϩ ] i with ionomycin and stimulating protein kinase C with phorbol 12-myristate 13-acetate. Fig. 9C shows that deletion of Mist1 impaired exocytosis in response to both ionomycin and phorbol 12-myristate 13-acetate stimulation by ϳ40 -50%. However, it is important to note that exocytosis stimulated by physiological agonist concentrations was impaired by 80 -95%, indicating that impaired Ca 2ϩ signaling was the major cause for the poor exocytosis in Mist1 Ϫ/Ϫ cells.
An important implication of the findings in Fig. 9 is that the precise pattern of Ca 2ϩ oscillations and waves are critical for exocytosis. Thus, CCK between 1 and 100 pM and carbachol at 0.5 and 1 M evoked robust Ca 2ϩ increases in Mist1 Ϫ/Ϫ cells but did not stimulate any exocytosis. On the other hand, the same or lower agonist concentrations evoked a smaller Ca 2ϩ signal but markedly stimulated exocytosis in WT acini. The only difference in the Ca 2ϩ signals in WT and Mist1 Ϫ/Ϫ cells is that in WT cells the agonists induced repetitive Ca 2ϩ oscillations and propagated Ca 2ϩ waves.
The impaired stimulated exocytosis was expected to lead to malnutrition, retarded growth, and/or increased food consumption in the Mist1 Ϫ/Ϫ mice. Remarkably, this was not the case for male or female Mist1 Ϫ/Ϫ mice (Fig. 10, A and B). Food consumption and growth were similar in WT and Mist1 Ϫ/Ϫ mice for the 32 weeks that the data were recorded. This was the case during both the rapid (first 10 weeks) and slow growth (weeks 11-32) phases of the animals development. These results suggest that either secretion in vivo was affected less than what was observed with isolated acinar clusters or that the residual pancreatic function in Mist1 Ϫ/Ϫ mice is sufficient to support the food digestion and nutritional requirement of these animals.
In conclusion, the findings of the present work reveal several new roles for Mist1. We have extended previous studies to show that Mist1 controls the localization and perhaps the integrity of the mitochondria and the Golgi apparatus in addition to regulating the localization of secretory granules. The disordered localization of intracellular organelles leads to aberrant Ca 2ϩ signaling that prevents Ca 2ϩ oscillations and distorts the apical-to-basal Ca 2ϩ waves. These findings provide evidence for the importance of cellular architecture in the generation and propagation of the Ca 2ϩ signals. The aberrant Ca 2ϩ signaling and cell architecture also result in greatly reduced exocytosis, further demonstrating the importance of Ca 2ϩ signaling in exocrine pancreatic function. The unexpected finding is the lack of any effect of the compromised acinar pancreatic function on food consumption and animal growth. Whether this is the result of a less compromised pancreatic function in vivo or the large spare capacity of the pancreatic function remains to be determined. However, given the known spare capacity of the pancreas and the fact that secretion by acinar clusters accurately reflects secretion in vivo (41), it is likely that the spare pancreatic function protected the Mist1 Ϫ/Ϫ animals to allow normal growth with no need for increased food consumption.