Ryanodine Receptor Type III (Ry3R) Identification In Mouse Parotid Acini

Immunoblot analysis and [3H]ryanodine binding were used to characterize and identify ryanodine receptors (RyRs) in nonexcitable mouse parotid acini. Western analysis revealed ryanodine receptor type III (Ry3R) to be the only detectable isoform in parotid microsomal membranes. Binding of [3H]ryanodine to microsomal fractions was dependent on Ca2+, salt, pH, and temperature. At 23 °C, and in the presence of 0.5 m KCl and 100 μm Ca2+, [3H]ryanodine bound specifically to membranes with high affinity (K d = 6 nm); maximum binding capacity (B max) was 275 fmol/mg protein. Mg2+ and ruthenium red inhibited [3H]ryanodine binding (IC50 = 1.4 mm and 0.5 μm, respectively). 4-Chloro-3-ethylphenol enhanced the binding of [3H]ryanodine 2.5-fold; whereas ATP and caffeine were much less efficacious toward activating Ry3R (56% and 18% maximal enhancement, respectively). Bastadin, a novel modulator of the 12-kDa FK506 binding protein·RyR complex, increased [3H]ryanodine binding 3–4-fold by enhancingK d . The immunosuppressant FK506 enhanced [3H]ryanodine receptor occupancy at >100 μm and antagonized the action of bastadin, suggesting that an immunophilin modulates Ry3R in parotid acini. These results suggest that Ry3R may play an important role in Ca2+ homeostasis in mouse parotid acini.

Increases in cellular Ca 2ϩ occur in response to agonist stimulation in many cell types and are important in regulating a number of cellular functions. Two distinct classes of channels that mediate the release of Ca 2ϩ from intracellular stores have been identified: one is sensitive to the second messenger inositol 1,4,5-trisphosphate and the second is sensitive to ryanodine (1). Ryanodine receptors (RyRs) 1 were first identified in skele-tal and cardiac muscle junctional sarcoplasmic reticulum (SR) using radiolabeled ryanodine binding analysis (2,3) where they function as Ca 2ϩ release channels during excitation-contraction coupling (4,5). More recently, the availability of antibodies/antisera selective for the skeletal (Ry 1 R), cardiac (Ry 2 R), and brain (Ry 3 R) isoforms, in conjunction with molecular approaches, have clearly demonstrated expression of RyRs in a variety of nonmuscle tissues (6 -8). In addition to expression in skeletal muscle, Ry 1 R has been shown to be expressed at low levels in cerebral cortex and hippocampus and exhibits especially high abundance in cerebellar Purkinje cells (9). Ry 2 R is expressed in cardiac tissue and throughout most of the brain, with relatively high levels of expression in the olfactory nerve layer, layer VI of the cerebral cortex, the dentate gyrus, cerebellar granule cells, the motor trigeminal nucleus, and the facial nucleus. The recently identified Ry 3 R isoform appears to be the major isoform in smooth muscle (10 -12) and has also been found to be expressed in several nonexcitable cells, including HeLa and LLC-PKI cells, mink lung cells, and submandibular cells (7,8). Interestingly, Ry 3 R was recently found to be generally expressed in mammalian brain at very low levels, raising questions about its precise function in Ca 2ϩ signaling (6).
[ 3 H]Ryanodine binding to receptors isolated from muscle SR and brain microsomes indicates that the RyRs are modulated by a number of physiologic and pharmacologic agents. Activators include micromolar Ca 2ϩ , millimolar caffeine and adenine nucleotides, and nanomolar ryanodine. Inhibitors include micromolar ryanodine, nanomolar ruthenium red, and millimolar Mg 2ϩ or Ca 2ϩ (5). In addition, the immunosuppressive agents FK506 and rapamycin, which bind to immunophilins such as FKBP-12 (12-kDa FK506 binding protein) and inhibit the immune response, have been implicated in the regulation of gating properties of the RyR in skeletal muscle (13)(14)(15), cardiac muscle (16), and neurons (17). Associated with the Ry 1 R, FKBP-12 has been shown to stabilize the closed conformation of the Ca 2ϩ release channel. FK506 was found to promote dissociation of FKBP-12 from the Ry 1 R complex (14) and to alter SR Ca 2ϩ transport and single channel gating behavior (15). The ability of FK506 to dissociate FKBP-12 from SR membranes was shown to be enhanced by macrocyclic bastadins isolated from the marine sponge Ianthella basta (18). Like FK506, bastadins also appear to target FKBP-12. Unlike FK506, bastadins themselves do not dissociate FKBP-12 from the SR calcium release channel complex but instead facilitate FK506-induced dissociation (18).
Although RyRs have been well characterized in skeletal and heart muscle, less is known about RyRs in nonexcitable cells. The inability to observe caffeine-induced Ca 2ϩ release from intracellular stores has been misinterpreted to mean that RyRs are absent from some cell types. Results with caffeine are especially confusing since cells known to express the Ry 3 R isoform fail to respond to caffeine (10,19), and various laboratories have reported differences in caffeine sensitivity in the same cell type (19,20). Much of the evidence supporting the presence of RyRs in nonexcitable cells, including exocrine cells, has been based on functional studies (7,(21)(22)(23) with little information concerning expression of RyR protein (7,11,24). Furthermore, few studies have either demonstrated or characterized specific ryanodine binding in microsomal fractions from nonexcitable cells (7,(25)(26)(27).
The focus of the present study is to identify the expression of RyR protein(s) in nonexcitable mouse parotid acini and to characterize its ability to bind [ 3 H]ryanodine. Western blot analysis of microsomes isolated from primary acini reveals the expression of Ry 3 R protein without detectable levels of Ry 1 R or Ry 2 R protein. The data presented show that high-affinity binding of [ 3 H]ryanodine to Ry 3 R from parotid acini is modulated in a similar fashion to Ry 1 R and Ry 2 R from muscle SR with important differences with respect to modulation by caffeine and adenine nucleotides.
Acinar Cell and Tissue Preparation-Acinar cells were isolated as small aggregates from the parotid glands of male Swiss Webster mice as described by Watson et al. (28). Briefly, parotid glands from male Swiss Webster mice (27-30 g) were removed quickly, trimmed, and minced in a siliconized dish in Krebs-Henseleit bicarbonate solution (KHB), pH 7.4, containing 0.9 mM Mg 2ϩ and 1.28 mM Ca 2ϩ , 30 mM Hepes, 90 units/ml collagenase (CLS2), and 1 mg/ml hyaluronidase. Enzyme digestion was conducted in a rotary water bath at 37°C for 60 min under continuous CO 2 /0 2 (5/95%) gassing. After the first 40 min of digestion, the suspension was pipetted up and down twelve times with a 10-ml plastic pipette. This was repeated two more times at approximately 5-min intervals. The pH during the dispersion was maintained at 7.2-7.4. Following digestion, the cells were centrifuged at 50 ϫ g for 2 min, washed with buffer (KHB minus enzymes with 4% BSA, pH 7.4), filtered through two layers of nylon, and washed two additional times. Cells were suspended in the same KHB minus enzyme buffer containing 1% BSA to rest for 30 min at 37°C with continuous gassing.
Skeletal muscle (tibia anterialis) and whole brain were dissected from the same mice for use as control tissue in some experiments. Both acinar cells and control tissues were processed immediately to obtain membrane fractions.
Microsomal Membrane Preparation-Microsomal membranes were isolated at 4°C from parotid acinar cells and brain by fractionation of a 10% (wet w/v) homogenate of the cells or tissue using differential centrifugation in isomolar sucrose as described below. Cells were suspended in 250 mM sucrose, 10 mM HEPES, 1 mM EDTA, buffer (pH 7.4) containing 1 mM DTT, and the protease inhibitors of benzamidine (1 mM), leupeptin (1 g/ml), pepstatin A (0.7 g/ml), and PMSF (0.1 mM), and homogenized in a glass mortar using a motor-driven Teflon pestle and 10 complete passes. The homogenate was centrifuged for 5 min at 250 ϫ g. Homogenization of the pellet was repeated, and the pooled 250 ϫ g supernatants were centrifuged for 20 min at 10,000 ϫ g. The 10,000 ϫ g supernatant was centrifuged for 1 h at 100,000 ϫ g, and the resulting microsomal fraction (pellet) was separated from the soluble fraction (supernatant), held overnight submerged in suspension buffer (200 mM sucrose, 50 mM KCl, 20 mM MOPS buffer (pH 6.8) containing the protease inhibitors) at 4°C, and resuspended the next day at a protein concentration of 5-11 mg/ml by homogenization in the same buffer for immediate use or storage at Ϫ80°C. A total membrane fraction of skeletal muscle was prepared in homogenate medium without CHAPS according to Damiani and Margreth (29). EGTA and EDTA, each at 2 mM concentration, were substituted for 1 mM EDTA in homogenization buffer for the preparation of microsomes in some experiments.
SR membranes used in Western blots as positive controls were isolated from rabbit fast skeletal muscle according to Saito,et. al. (30), rat cardiac muscle according to Feher and Davis (31), and avian pectoralis muscle according to Airey,et. al. (32). and by passage of sample through a Whatman GF/F glass fiber filter followed immediately by 3 ϫ 4 ml washes of filter with the same buffer; all procedures were completed within 1 min. The filters were dried overnight and placed in vials containing scintillant, and the bound [ 3 H]ryanodine was measured by liquid scintillation counting using a Packard Tri-Carb 2200CA analyzer. Specific bound [ 3 H]ryanodine was calculated by subtracting nonspecific binding, measured for parallel assays in the presence of 10 M unlabeled ryanodine, from the total bound. For acinar cell microsomal membranes, nonspecific binding averaged 81 and 57% at 0.3 and 20 nM, respectively, and was linear with the concentration of [ 3 H]ryanodine. Free Ca 2ϩ effects were obtained by titrating CaCl 2 in binding media containing 1 mM EGTA using theoretical estimates derived from the computer program BAD3 (33). [ 3 H]Ryanodine binding to proteins in the soluble fraction of acinar cells was determined by the PEG precipitation method as described by Shoshan-Barmatz and Zarka (34). Protein was determined by the Hartree (35) modified method of Lowry (36).

Association and Dissociation Kinetics for [ 3 H]Ryanodine
Binding-For association kinetics, acinar cell microsomal membrane (0.4 mg of protein/ml) was incubated with 5 nM [ 3 H]ryanodine in binding media with and without 10 M cold ryanodine at 23 and 37°C. At time intervals ranging between 10 min and 22 h, the reaction was terminated by rapid filtration as described above.
For dissociation kinetics, acinar cell microsomal membrane (0.4 mg of protein/ml) was incubated at 23°C with 5 nM [ 3 H]ryanodine in binding media with and without 10 M cold ryanodine for 16 h until association was completed. Dissociation was initiated by diluting duplicate aliquots of each treatment group by 50-fold using binding media without ryanodine. Dissociation was terminated by filtering aliquots (250 l) at the indicated intervals of time up to 20 h, and the bound radioactivity was assayed as above.
Analysis of Binding Data-EC 50 and IC 50 values for activators and inhibitors of [ 3 H]ryanodine binding, respectively, were determined by linear analysis of log-logit transformation of concentration response curves. The equilibrium binding constants, K d and B max , for [ 3 H]ryanodine with and without effectors were derived by curvilinear analysis using the computer program RADLIG sub-routine EBDA (Elsevier-BIOSOFT) and depicted graphically using a Rosenthal (Scatchard) plot and linear analysis. Kinetic association and dissociation rate constants were calculated using sub-routine KINETIC of RADLIG. Mean treat-ment differences of effectors from control were tested for significance using the student t statistic for paired observations at p Ͻ 0.05. Values reported represent the mean Ϯ S.E. of (N) number of independent experiments performed in duplicate unless otherwise noted.
Electrophoreses and Western Blot Analysis-Constituent proteins from membrane preparations were resolved on 3-10% gradient gels by the method of Laemmli (37). Proteins were electroblotted overnight at 30 V followed by a 60-min fast transfer at 100 V. Nonspecific antibody binding was blocked by incubating blots for 1 h at 37°C in TTBS solution (20 mM Tris-HCl, 0.5 M NaCl, 0.5% Tween-20, pH 7.5) with the addition of 5% nonfat dry milk. Specific binding of primary antibody was performed by incubating blots for 1 h at 37°C in TTBS buffer in the presence of 1% BSA and antibody. Resulting immunoblots were labeled with horseradish peroxidase-conjugated goat anti-mouse or donkey anti-rabbit secondary antibody for 1 h at 37°C and then visualized using an enhanced chemiluminescence technique. Nonspecific binding of each secondary antibody to each membrane preparation was minimized by performing a dilution series in the absence of primary antibody.

RESULTS
For the results presented herein, microsomal membranes were prepared from isolated mouse parotid acini. The enzymatic digestion method of isolation, which has been widely used for the preparation of rat and mouse salivary cells, yielded a preparation containing 95% acinar and 5% ductal cells. The percent composition of ductal cells is consistent with earlier quantitative estimations of ductal cell volume observed in parotid tissue sections, i.e. 7.7% (38), and with visual estimates in isolated parotid cells (39 -41).
Western blot analysis of microsomal membranes isolated from rat parotid acinar cells revealed no detectable levels of specific immunoreactive skeletal (Ry 1 R) or cardiac (Ry 2 R) protein even when the blots were visualized with the highly sensitive chemiluminescent technique. Two antibodies selective for the "brain" (Ry 3 R) isoform of the ryanodine receptor detected a high molecular weight protein in several microsomal preparations (n ϭ 4 determinations; Fig. 1). The immunoreactive band migrated at a slightly smaller apparent molecular weight than the Ry 1 R protomer found in rabbit skeletal muscle.
Assay conditions of Ca 2ϩ and KCl concentrations, pH, and Equilbrium binding of 6 nM [ 3 H]ryanodine was also biphasic with respect to pH. When assay pH was buffered with HEPES between 7 and 8, ryanodine binding was between 70 and 90%, respectively, of its maximum at pH 7.6 ( Fig. 2C). Test of pH effects below 7 (MOPS) and above 8 (TRIS) showed binding to decrease steeply to 20% of maximum at pH 6.2 and to be relatively unaffected in alkaline solution, with pH 8.8 giving 87% of maximum binding (data not shown).
Assay temperature dramatically influenced the time course and magnitude of [ 3 H]ryanodine binding (Fig. 2D). At 37°C in binding media (pH 7.4) containing 100 M Ca 2ϩ and 0.5 M KCl, maximum binding of 5 nM [ 3 H]ryanodine occurred within 2 h with an association half-time of 45 min. Thereafter, total binding declined to approximately 50% by 4 h and to negligible levels by 22 h. At 23°C, the rate of [ 3 H]ryanodine (5 nM) association was significantly slower (half-time ϭ 100 min; k obs ϭ 0.0069 min Ϫ1 ) but reached a 2-fold higher level of occupancy that remained stable for at least 22 h (Fig. 2D). After equilibrating the membranes for 16 h and subsequent to a 50-fold dilution with binding media lacking ryanodine, the radioligand dissociated with a half-time ϭ 595 min (k -1 ϭ 0.0012 min Ϫ1 ). The time courses of association (Fig. 2D) and dissociation (Fig.  3) of 5 nM [ 3 H]ryanodine at 23°C were both best described by a mono-exponential equation. The association rate constant calculated according to the relationship k ϩ1 ϭ k obs Ϫ k -1 /[L] was 1.16 ϫ 10 6 min Ϫ1 M Ϫ1 . The equilibrium dissociation constant calculated based on K d ϭ k -1 /k ϩ1 was 1 nM.
A standard assay condition that includes binding media (pH 7.4) with 100 M Ca 2ϩ , 0.5 M KCl and temperature at 23°C was used to characterize equilibrium binding of [ 3 H]ryanodine to mouse parotid acinar cell membranes. The distribution of specific binding of 6 nM [ 3 H]ryanodine among acinar subcellular fractions was compared for two experiments. [ 3 H]Ryanodine binding to the 10,000 -100,000 ϫ g particulate (microsomal) fraction comprised 26% of the total acinar cell specific binding. Percent distribution of [ 3 H]ryanodine binding to the 250 ϫ g and 250 -10,000 ϫ g particulate (putative nuclear and mitochondrial) fractions averaged 17 and 57% of the total, respectively. The ratio of specific binding activity for these fractions relative to the microsomal fraction (relative specific binding FIG. 1. Parotid acini express Ry 3 R proteins by Western blot analysis. Constituent proteins from rabbit junctional sarcoplasmic reticulum, rat cardiac sarcoplasmic reticulum, avian pectoralis sarcoplasmic reticulum, or crude acini membranes were resolved by SDS-PAGE using 3-10% gradient gels and transferred overnight onto polyvinylidene difluoride membranes. Secondary antibody was visualized using chemiluminescent methods. Shown are the Ry 1 R Western blot skm lane: 1 g of rabbit skeletal junctional SR, and acr lane: 30 g of crude acini membrane proteins; the Ry 2 R Western blot crd lane: 5 g rat cardiac SR, and acr lane: 30 g of crude acini membrane proteins; and the Ry 3 R Western blot avi lane: 5 g of avian pectoralis SR, and acr lane, 20 g of crude acini membrane proteins (arrow marks the relative mobility of Ry 1 R protein). Ry 1 R and Ry 2 R antibodies were labeled using a goat anti-mouse secondary antibody (Sigma). Ry 3 R antibody was labeled using a donkey anti-rabbit antibody (Amersham Corp.). Western analysis with each antibody was repeated at least three times on two independent preparations with the same results. activity ϭ 1) averaged 0.92 and 0.86. No binding was detectable in the 100,000 ϫ g soluble (cytosolic) fraction.
Overall, [ 3 H]ryanodine equilibrium binding in the standard assay media at 22 h averaged 132 Ϯ 10 fmol/mg of protein for 27 preparations of microsomal membrane; the concentration of added [ 3 H]ryanodine averaged 6.2 Ϯ 0.06 nM. [ 3 H]Ryanodine (6 nM) binding was a linear function of protein concentration from 0.1 to 1.0 mg of protein/ml (Fig. 4); assay concentrations of membrane used in this study were within this range. Since this is the first report of RyR in parotid acini, we chose to illustrate that the amount of specific binding is directly proportional to the amount of acini receptor in the assay.
As shown in Fig. 5A for three independent preparations of microsomes assayed in standard media at 23°C, specific binding of [ 3 H]ryanodine approached saturation near 20 nM. Nonspecific binding was linear with increasing concentration of radionuclide but approached 45% of total binding at the higher concentrations of [ 3 H]ryanodine. Nonlinear regression analysis of equilibrium binding curves reveals the data is best fit statistically by a mathematical model describing [ 3 H]ryanodine binding to a single site and results in a linear Scatchard plot (Fig. 5B). The apparent K d was 6.2 Ϯ 0.55 nM (n ϭ 3), and maximum occupancy (B max ) was 275 Ϯ 22 fmol/mg of protein (n ϭ 3).
In Table I, [ 3 H]ryanodine saturation binding to mouse parotid acinar microsomal membranes is compared with binding parameters measured with membrane preparations from mouse fast twitch skeletal muscle and whole brain tissues. Brain and parotid preparations differed from those derived from skeletal muscle in having a lower percentage of specific bound radioligand and greater than 10-fold less binding sites/mg of membrane protein. Brain had the highest affinity for [ 3 H]ryanodine, having approximately 10-fold higher binding affinity than either parotid or muscle membranes.
In addition to Ca 2ϩ , other compounds reported to modulate the ryanodine-sensitive release of Ca 2ϩ from muscle SR (caffeine, adenine nucleotide, ruthenium red, and MgCl 2 ), and from the endoplasmic reticulum of nonexcitable cells (4-CEP) were tested for their influence on [ 3 H]ryanodine binding to acinar cell microsomes. Typical concentration-response curves of activators and inhibitors of specific [ 3 H]ryanodine binding determined at equilibrium with a non-saturating concentration of 6 nM [ 3 H]ryanodine are shown in Fig. 6, A and B, respectively. Concentration estimates of effectors producing halfmaximal activation (EC 50 ) or inhibition (IC 50 ) are summarized in Table II. In the presence of optimal Ca 2ϩ , AMP-PCP (2 mM) maximally enhanced [ 3 H]ryanodine occupancy by 40% and exhibited an EC 50 of 1 mM. Caffeine (3-10 mM) enhanced occupancy by 25% although the activation was too small to calculate an EC 50 (Fig. 6A). The activating action of 4-CEP on receptor binding was sharply biphasic with 1 mM producing a 24% reduction of control occupancy. Ruthenium red was 50-fold less potent than unlabeled ryanodine toward competing with 6 nM [ 3 H]ryanodine (IC 50 of 10 and 500 nM, respectively); whereas Mg 2ϩ was inhibitory at a concentration above physiological relevance (Ͼ1 mM). Table II summarizes EC 50 and IC 50 values for modulators of RyR in microsomes isolated from acinar cells.
The effects of caffeine and AMP-PCP in the absence and presence of an inhibitory concentration of MgCl 2 are given in Table III. In the presence of 100 M Ca 2ϩ , increases in [ 3 H]ryanodine binding of 18 and 56% above controls were measured at 10 mM caffeine and 1 mM AMP-PCP, respectively. In the presence of 100 M Ca 2ϩ , 1 mM Mg 2ϩ reduced [ 3 H]ryanodine binding by 40% and 10 mM caffeine could not fully restore occupancy to control levels. In contrast, 1 mM AMP-PCP in the presence of equimolar Mg 2ϩ nearly restored occupancy to that seen with 1 mM AMP-PCP alone, revealing that the Mg⅐AMP-PCP complex maintains activating properties toward RyR.
Equilibrium saturation binding of [ 3 H]ryanodine to acinar microsomes was examined in the presence and absence of modulators in binding media containing 100 M Ca 2ϩ and 0.5 M KCl, which was found to be optimal for the binding of [ 3 H]ryanodine to acinar microsomes (Table IV, Fig. 7). At a concentration shown to enhance occupancy by 50%, 4-CEP decreased the value of K d (i.e. enhanced the apparent affinity of [ 3 H]ryanodine for its binding site) nearly 4-fold and produced a small but significant reduction in B max . Interestingly, the macrocyclic bromotyrosine bastadins (20 M) significantly enhanced [ 3 H]ryanodine binding site affinity (4-fold) without a significant change in B max . Ruthenium red and Mg 2ϩ significantly reduced [ 3 H]ryanodine binding affinity (K d from 6 to 8.5 and 10 nM, respectively) and lowered B max by 42% and 18%, respectively. The effects of Mg 2ϩ on K d were fully restored to the control value by caffeine, and the Mg⅐AMP-PCP complex enhanced K d nearly 3-fold (Table IV). Although caffeine did not significantly reverse the actions of Mg 2ϩ on B max , the Mg⅐AMP-PCP complex enhanced B max by 41%.
The immunosuppressive macrolactam, FK506, and bastadins have been shown to interact with the FKBP-12⅐Ry 1 receptor complex found in skeletal muscle (18). These compounds were used to probe the possible involvement of immunophilins toward modulating [ 3 H]ryanodine binding sites in acinar cell microsomes. The effect of dimethyl sulfoxide, used as a vehicle in studies with FK506 and bastadins, was examined and found to inhibit the equilibrium binding of 6 nM [ 3 H]ryanodine by 9% and 37% at final concentrations of 1% and 5% (v/v), respectively. Bastadin enhanced [ 3 H]ryanodine receptor occupancy in a concentration-dependent manner with maximum stimulation of Ͼ4-fold over dimethyl sulfoxide control observed between 100 -200 M when equilibrium binding was determined at the nonsaturating concentration of 6 nM [ 3 H]ryanodine; whereas, FK506 was much less efficacious, producing a 2.0-fold increase in occupancy at 300 M (Fig. 8). The EC 50 for bastadin was 20 M. Interestingly, FK506 negated the remarkable activity of bastadin toward activating [ 3 H]ryanodine occupancy, resulting in a concentration-response relationship very similar to that seen with FK506 alone (Fig. 9). DISCUSSION In the present study, RyRs were characterized and identified in mouse parotid acinar cells by analysis of [ 3 H]ryanodine binding and immunoblot analysis using antibodies selective toward the three known isoforms. Western analysis revealed that the major, perhaps only, RyR isoform expressed in mouse parotid acini was the so-called brain isoform, Ry 3 R. Data show that cells expressing Ry 3 R possess very low levels of Ry 3 R protein. Data also showed saturable, high affinity [ 3 H]ryanodine binding in subcellular parotid membrane fractions that was Ca 2ϩ -, salt-, pH-, and temperature-dependent. Both kinetic and equilibrium binding studies with the microsomal membrane fraction revealed that [ 3 H]ryanodine interacts with a single population of high affinity binding sites with K d calculated between 1 and 7 nM in equilibrium and kinetic measurements in the standard assay media employed. Attempts to demonstrate different allosteric binding conformations of low and high binding affinities by using higher concentrations of [ 3 H]ryanodine (Ͼ30 nM) produced unacceptably high nonspecific binding (42). Given that parotid cells are composed of approximately 95% acinar cells, it is not likely that the specific [ 3 H]ryanodine-binding sites or the immunoreactive protein toward Ry 3 R antibody originates from a contaminating cell other than acini. The contaminating cells would have to express very high levels of protein to account for the measurements made.
In general, the effects of temperature (42)(43)(44)(45), alkaline pH (46 -49), ions (species specific, Ca 2ϩ ) (2,44,50,51), and monovalent ions (46 -48, 51, 52) on [ 3 H]ryanodine binding reported here using parotid acinar cells were similar to those observed in excitable cells. [ 3 H]Ryanodine binding showed a near absolute requirement for certain monovalent ions (K ϩ and Na ϩ ), even in the presence of optimum Ca 2ϩ . Since NaCl, at least partially, substituted for KCl, but sucrose, at a concentration close to physiologic molar equivalents, did not, the increase in [ 3 H]ryanodine binding at 25°C appears to be dependent on ionic strength as reported for smooth muscle (47), and not on osmolarity of the binding media, as reported for bullfrog skeletal muscle (53). There appears to be a temperature-and concentration-dependent monovalent ion effect on [ 3 H]ryanodine binding with tissue-specific response. Concentrations of KCl and/or NaCl above 0.25 M are activating at 25°C in rabbit skeletal muscle (51), bullfrog skeletal muscle (53), smooth muscle (47), and, as reported here, in acinar cells. At 37°C, concentrations of KCl and/or NaCl are activating in rat heart muscle, but inhibitory in rabbit skeletal muscle sarcoplasmic reticulum (52). The higher sensitivity of the receptor for K ϩ over Na ϩ in parotid acinar cells, observed with less and varying magnitude in other cell types (47,52), suggests that the larger size of the K ϩ cation may be a factor distinguishing the activation efficiency of monovalent ions.
The pharmacology of 1) inhibition by competing cold ryanodine (47,49,54,55), ruthenium red (49,51,52), and MgCl 2 (46,49,51,54), and 2) activation by AMP-PCP (45,46,49,52,54) and caffeine in excitable cells (47,49,52,54,56) was qualitatively similar to that found in mouse parotid acinar cells. Marked differences in binding assay conditions used by investigators preclude direct quantitative comparisons of the magnitude and potency of each effector in the parotid acinar cell with other cell types. The very low efficacy of caffeine and AMP-PCP toward enhancing [ 3 H]ryanodine occupancy may reflect an inherently low sensitivity of the Ry 3 R complex to these agents compared with Ry 1 R and Ry 2 R isoforms expressed in striated muscle and is consistent with a lack of caffeine responsiveness of Ry 3 R in other cell types (10,19). Deletion of Ry 1 R expression using gene targeting has revealed a small response to caffeine (25 mM) and is maintained in dyspedic myotubes, which has been attributed to expression of Ry 3 R (57, 58). However, this interpretation is based solely on the presence of Ry 3 R mRNA in selected dyspedic myotubes (57). Since the presence of Ry 3 R protein has not been positively identified in dyspedic  muscle, this does not preclude the possibility that in the absence of Ry 1 R, other effects of caffeine on the Ca 2ϩ permeability of plasma membrane and mitochondria become apparent. However, the small effects of caffeine and AMP-PCP seen in acinar microsomes to partially and completely overcome MgCl 2 inhibition of [ 3 H]ryanodine binding, respectively, are consistent with the proposed roles of Ca 2ϩ , ATP, and Mg 2ϩ to modulate Ca 2ϩ -induced Ca 2ϩ release via the ryanodine-sensitive calcium channel (59, 60). 4-Chloro-phenolic compounds have been shown to release Ca 2ϩ from ruthenium red-sensitive intracellular stores of both excitable (61,62) and nonexcitable cells (63). We sought to verify that 4-CEP specifically interacts with the ryanodinesensitive calcium release channel of nonexcitable cells (63)  activation was monophasic with no inhibition of [ 3 H]ryanodine binding at the highest concentration tested (l mM). They also reported 4-Cl-3-methylphenol to increase receptor affinity for [ 3 H]ryanodine without altering binding site number. Differences observed between muscle and acinar cell in potency (112 versus 54 M, respectively) and inhibitory properties of these 4-chloro-phenolic compounds may reflect differences in phenol group substitution, sensitivity of RyR isoforms (1 versus 3), or in binding assay conditions (62).
The known association of rabbit FKBP-12 with skeletal muscle Ry 1 R (13,14) and its modulator role in binding to the Ry 1 R (14,15), coupled with the potential for expression of FKBP-12 in mouse parotid acinar cells (64), prompted us to test for immunophilin modulation of RyR function in mouse parotid acinar cells. FKBP-12 is the primary target for the immunosuppressant FK506 (65,66) that, in binding to FKBP-12, dissociates the immunophilin from skeletal muscle (14). We found that FK506 produced a concentration-dependent increase in [ 3 H]ryanodine binding to acinar cell membranes, suggesting that a change in the open state of the RyR had occurred, presumably through an FK506-induced release of a FKBP-l2. Bastadin, which enhances FK506-induced release of FKBP-12 from the ryanodine membrane complex (18), was also found to increase [ 3 H]ryanodine binding in acinar cell membranes in a concentration-dependent manner. The EC 50 value determined in acinar cell membranes for the bastadin mixture was 20 M. Mack et al. (18) found, for skeletal muscle, that EC 50 values for bastadin 5 (2.2 M) and bastadin 7 (6.3 M), which are part of the bastadin mixture, were approximate1y 10-and 3-fold more potent, respectively, than the bastadin used in the present study. Saturating levels of bastadin mixture produced a 2-4fold higher binding level in acinar cell membranes compared with a 40 -50-fold higher binding in skeletal muscle sarcoplasmic reticulum. Differences in the amount of KCl (140 versus 500 mM) in the assay medium may underlie the apparent differences in efficacy of the two preparations. In parotid acinar cells, and as shown by Mack et al. (18) in skeletal muscle, bastadin was more effective in enhancing [ 3 H]ryanodine binding than FK506. Equilibrium [ 3 H]ryanodine saturation binding studies showed bastadin to increase RyR affinity for [ 3 H]ryanodine in parotid acinar cell membranes comparable to bastadininduced increases (approximately 3-fold) in ryanodine affinity in skeletal muscle (18). However, unlike skeletal muscle, an increase in maximum binding capacity was not detected in mouse parotid acinar cells. This difference also appears to be related to the assay conditions (high versus physiological salt levels and 23°C versus 37°C. 2 FK506 inhibited bastadin-augmented [ 3 H]ryanodine binding in acinar cells analogous to that observed in skeletal muscle,  (51,67,68), are nonphysiological. Lowering the salt concentration from 1 M to near physiological levels (200 mM) has been found to result in marked changes in the extent to which MgCl 2 , caffeine, and AMP-PCP affect binding properties of receptor for [ 3 H]ryanodine (44,54). The sensitivity of the RyR to activation by Ca 2ϩ has also been reported to be influenced by salt concentration of binding media (6,54).
The association of [ 3 H]ryanodine with its high affinity site is highly dependent on temperature (Q 10 Ͼ 3) (42,45). By lowering the assay temperature from 37 to 23°C, as reported here for acinar cell microsomes or by Carroll et al. (42) in skeletal muscle sarcoplasmic reticulum, the binding occupancy of [ 3 H]ryanodine at equilibrium was increased, implying a temperature-dependent conformational change of the receptor occurred that favored [ 3 H]ryanodine binding with a decrease in  temperature. The observation that bastadin (mixture) did not increase B max in parotid acinar cells, but did in skeletal muscle (18) under similar binding conditions but different assay temperatures, suggests that the reported effect of bastadin to stabilize the high affinity binding site conformation at 37°C may be mimicked by lowering the binding assay temperature.
A significant finding of the present study was the expression of the Ry 3 R isoform in mouse parotid cells. Ry 3 R has also been identified in other nonexcitable cells including HeLa, LLC-PK1, and mink lung epithelial cells (7,10). However, this is the first biochemical characterization of Ry 3 R in the absence of detectable levels of isoforms Ry 1 R and Ry 2 R. Thus far, the function of RyRs in nonexcitable cells is unclear. Demonstration of a Ca 2ϩ -induced Ca 2ϩ release process (CICR) has been hampered by the fact that caffeine, commonly employed as an activator of CICR, has been found to produce little or no effect on [Ca i ] in some cells, including mouse parotid acini. 3 In cell types expressing Ry 3 R, the response to caffeine has been controversial. Caffeine did not release Ca 2ϩ from intracellular stores of transforming growth factor ␤-treated mink epithelial cells or Jurkat T-cells (10,19). The reason(s) for the lack of or small effects of caffeine on [Ca i ] in cells known to express RyRs is unclear. One explanation given was that caffeine insensitivity is due to the expression of low levels of RyR. This may be the case for HeLa cells where [ 3 H]ryanodine binding was 17-fold less than in rabbit brain (7). In mouse parotid cells, however, [ 3 H]ryanodine binding capacity was comparable with mouse brain (54,69). Another possibility is that lack of caffeine sensitivity may be related to the species of animal used. Caffeine has been shown to increase [Ca i ] in rat parotid and pancreatic acinar cells (21,22) but not in mouse parotid cells. Similarly, endothelial cells from bovine and rabbit were also found to differ in their responses to caffeine. Caffeine did not produce a marked release of Ca 2ϩ from intracellular stores in bovine endothelial cells but was effective in rabbit aortic endothelial cells (70).
In summary, the expression and characterization of Ry 3 R in mouse parotid acini supports the conclusion that this receptor subtype contributes to the changes in intracellular Ca 2ϩ following cell stimulation. Although in mutant mice lacking Ry 3 R where results do not suggest an involvement of Ca 2ϩ mobilization via Ry 3 R in lymphocyte proliferation, mutant mice did show increased locomotor activity that may indicate abnormal Ca 2ϩ signaling of certain neurons (71). In nonexcitable mouse parotid cells, Ry 3 R may be involved in the cross-talk that occurs between the Ca 2ϩ and cAMP pathways, i.e. Ca 2ϩ regulation of cAMP synthesis 3 suggests that an RyR plays a role in muscarinic augmentation of stimulated cAMP accumulation reported previously by Watson et al. (28).