Na+-dependent release of Mg2+ from an intracellular pool in rat sublingual mucous acini.

Muscarinic stimulation induces release of Mg2+ from an intracellular pool in rat sublingual mucous acini (Zhang, G. H., and Melvin, J. E. (1992) J. Biol. Chem. 267, 20721-20727). In the present study we examined the interdependence of Mg2+ mobilization on intracellular Na+ and Ca2+ by monitoring the intracellular free concentrations of Na+ ([Na+]i), Mg2+ ([Mg2+]i), and Ca2+ ([Ca2+]i) using ion-sensitive fluorescent indicators. Gramicidin increased the intracellular concentrations of all three ions. Comparable to agonist-stimulated mobilization of Mg2+, the gramicidin-induced [Mg2+]i increase was independent of extracellular Mg2+ indicating release of Mg2+ from an intracellular pool. Clamping the [Ca2+]i near 30 nM with the Ca2+-selective chelator BAPTA failed to alter the [Na+]i or [Mg2+]i increases generated by gramicidin. In contrast, depletion of intracellular Na+ markedly suppressed the muscarinic-stimulated [Mg2+]i increase, whereas the [Ca2+]i increase was similar to that seen in physiological extracellular Na+. These results revealed that intracellular Mg2+ mobilization did not directly relate to the [Ca2+]i, but required an increase in [Na+]i. Consistent with this hypothesis, increasing [Na+]i by activating Na+ influx via the Na+/H+ exchanger also increased the [Mg2+]i. The Na+/Mg2+ exchange inhibitor quinidine suppressed both the gramicidin- and muscarinic-induced discharge of internal Mg2+. These results suggest that release of Mg2+ from an intracellular pool is mediated by a Na+-dependent Mg2+ transport mechanism in salivary acinar cells.

Magnesium ions are essential cofactors for many cell functions including enzymatic reactions and transmembrane ion movements (1,2). In this latter role, Mg 2ϩ is important for regulating Na ϩ and Ca ϩ pump activity. The inward-directed chemical gradients created by these pumps are especially critical in the modulation of fluid secretion by salivary acinar cells. These physiological processes require that the [Mg 2ϩ ] i 1 be con-trolled within a narrow range. However, the mechanism(s) by which the intracellular free Mg 2ϩ concentration is regulated is still poorly understood. Mg 2ϩ is not passively distributed, the [Mg 2ϩ ] i can be more than 100 times lower than the concentration predicted from its electrochemical equilibrium. Acutely raising or lowering the external [Mg 2ϩ ] has little effect on the total magnesium content of most cells suggesting that the plasma membrane has a very low Mg 2ϩ permeability (3,4). In cells artificially loaded with an elevated [Mg 2ϩ ] i , a plasma membrane Na ϩ /Mg 2ϩ exchanger is activated that shuts down when the [Mg 2ϩ ] i returns to resting levels (5,6). These results suggest that the primary function of Na ϩ /Mg 2ϩ exchange is to maintain the resting [Mg 2ϩ ] i . Consistent with the plasma membrane having a low Mg 2ϩ permeability, the [Mg 2ϩ ] i changes observed during stimulation in many cell types does not reflect increased Mg 2ϩ movement across the plasma membrane, but results from mobilization of Mg 2ϩ from an intracellular pool (7,8).
The total intracellular magnesium content consists of cytosolic free Mg 2ϩ , cytosolic bound magnesium, and magnesium stored within organelles. More than 90% of the cellular magnesium is in the bound form (9,10). Cytosolic free Mg 2ϩ accounts for ϳ6% of the total cytosolic magnesium content in hepatocytes (11) and ϳ3% in murine S49 lymphoma cells (12). It has been suggested that ATP and RNA play a key role in the Mg 2ϩ buffering system (1,10). Muscarinic stimulation of salivary acinar cells greatly enhances ATP consumption (13). However, inhibition of Na ϩ ,K ϩ -ATPase does not influence the muscarinic-stimulated increase in [Mg 2ϩ ] i in sublingual acinar cells (7) suggesting that this increase in [Mg 2ϩ ] i does not result from the liberation of Mg 2ϩ during ATP consumption. Therefore, release of Mg 2ϩ from an intracellular organelle is apparently responsible for the agonist-induced [Mg 2ϩ ] i increase seen in salivary acinar cells.
Muscarinic stimulation induces a marked increase in [Ca 2ϩ ] i that subsequently triggers an increase in [Na ϩ ] i in sublingual acini (14,15). The agonist-induced increases in [Na ϩ ] i and [Ca 2ϩ ] i occur simultaneously with the mobilization of Mg 2ϩ from an intracellular pool (7). When the increase in either [Na ϩ ] i or [Ca 2ϩ ] i is prevented, the stimulated increase in [Mg 2ϩ ] i is blunted as well (7). These results suggest that muscarinic-induced Mg 2ϩ mobilization is both Na ϩ -and Ca 2ϩdependent; however, the nature of this Na ϩ and Ca 2ϩ dependence is unknown. In the present study, the role of Na ϩ and Ca 2ϩ in regulating intracellular Mg 2ϩ movement in rat sublingual acini was examined. Although both muscarinic receptor activation and Ca 2ϩ mobilization were sufficient for stimulating the release of Mg 2ϩ from the intracellular pool, neither was required. In contrast, we found that an increase in [Na ϩ ] i is not only sufficient but also necessary to mobilize Mg 2ϩ from a quinidine-sensitive pool.
Preparation of Sublingual Mucous Acini-Sublingual mucous acini were prepared from male, Wistar strain rats (150 -250 gm, Charles River, Kingston facility, NY) as described previously (14). Rats were killed by exsanguination after exposure to CO 2 and sublingual glands removed and placed in ice-cold digestion medium which consisted of Earle's minimal essential medium containing 1% bovine serum albumin, 50 units/ml collagenase, and 0.02 mg/ml hyaluronidase. The glands were finely minced in 2 ml of the digestion medium and then placed in 10 ml of the same medium, incubating at 37°C in a Dubnoff shaker with continuous gassing with 95% O 2 , 5% CO 2 (humidified), and agitation (80 cycles/min). The mince was dispersed by gently pipetting 10 times with a 10 ml plastic pipette at 15 min intervals. After 45 min of digestion, the preparation was centrifuged at 400 ϫ g for 30 s, the supernatant was discarded and replaced with fresh digestion medium. After a total of 1.5 h of digestion, the preparation was washed three times with a physiological salt solution (PSS) containing 0.01% bovine serum albumin and resuspended in the same medium. The PSS consisted of (mM): 110 NaCl, 25 NaHCO 3 , 20 HEPES, 10 glucose, 5.4 KCl, 1.2 CaCl 2 , 0.8 MgSO 4 , 0.4 KH 2 PO 4 , 0.33 NaH 2 PO 4 , adjusted to pH 7.4 with NaOH. For the nominally Ca 2ϩ -free solution, CaCl 2 was omitted. For the Na ϩ -free solution, Na ϩ was replaced with N-methyl-D-glucamine.
Determination of [Ca 2ϩ ] i -The intracellular free Ca 2ϩ concentration was determined by using the Ca 2ϩ -sensitive fluorescent indicator fura-2 as described previously (14). Briefly, dispersed sublingual acini were loaded with fura-2 by incubating in 2 M fura-2/AM for 30 min at 23°C, rinsed twice with PSS containing 0.01% bovine serum albumin, and resuspended in the same medium. Acini were attached to a coverslip mounted to the bottom of a perfusion chamber on the stage of a Nikon inverted microscope interfaced to a SPEX AR-CM fluorometer (Edison, NJ). A Nikon fluor X40 1.3-NA oil immersion objective was used to isolate five to eight acinar cells using a pinhole turret. Fluorescence ratios, obtained by exciting the dye at 340 and 380 nm and collecting the 505-nm emission, were converted to [Ca 2ϩ ] i by in situ calibration. The [Ca 2ϩ ] i was calculated according to Grynkiewicz et al. (16) using 224 nM as the K d of fura-2 for Ca 2ϩ . In some studies, the intracellular [Ca 2ϩ ] i was clamped near 30 nM by incubating acini with 50 M BAPTA/AM for 60 min at room temperature. BAPTA-loaded acini were switched to the nominally Ca 2ϩ -free solution just prior to stimulation.
Determination of [Na ϩ ] i -[Na ϩ ] i was determined as described previously (15). A stock solution of SBFI/AM (2 mM) and the nonionic detergent Pluronic F127 (25%, w/v) were prepared in dimethyl sulfoxide and mixed in equal volumes before loading. This mixture was added to the sublingual acinar suspension at a final concentration of 1 M SBFI/AM. The acini were then incubated with the indicator for 60 min at 23°C. Fluorescence ratio determinations were as described above for [Ca 2ϩ ] i . In situ calibration of the SBFI fluorescence was performed according to the method described by In some experiments [Na ϩ ] i was estimated using sodium green (see Fig. 9) because quinidine interfered with the fluorescence of SBFI. A stock solution of sodium green/tetraacetate (5 mM) and the nonionic detergent Pluronic F127 (25%, w/v) were prepared in dimethyl sulfoxide and mixed in equal volumes just prior to cell loading. This mixture was added to the sublingual acinar suspension at a final concentration of 2.5 M sodium green/tetraacetate. The acini were then incubated with the indicator for 120 min at 23°C. It should be noted that calculation of [Na ϩ ] i using single-wavelength dyes is inherently less accurate than using dual-wavelength ratio dyes such as SBFI. Therefore, fluorescence, obtained by exciting the dye at 500 nm and collecting the 530 nm emission, was normalized to the initial fluorescence in unstimulated acini and presented as arbitrary units.
Statistics-All results are presented as means Ϯ S.E. Traces are shown as the representative response of experiments from at least four separate cell preparations. Comparisons were made between different treatments using the unpaired Student's t test. Differences were considered significant at p Ͻ 0.05.   (20). Fig. 1 (ϩBAPTA)  Muscarinic stimulation increases the [Mg 2ϩ ] i by mobilizing Mg 2ϩ from an intracellular pool (7). To examine whether gramicidin increases [Mg 2ϩ ] i by a similar mechanism, acini were exposed to gramicidin in a Mg 2ϩ -free medium to eliminate Mg 2ϩ influx. Fig. 2 shows that gramicidin increased the [Mg 2ϩ ] i 57 Ϯ 14% in a Mg 2ϩ -free solution (n ϭ 5), comparable to the magnitude of the Mg 2ϩ increase seen in a Mg 2ϩ -containing medium (55 Ϯ 3%, n ϭ 5). This indicates that the gramicidin-induced increase in [Mg 2ϩ ] i is due to intracellular Mg 2ϩ mobilization.

The Na
Gramicidin causes depolarization of the plasma membrane. We exposed acini to high extracellular K ϩ to test the possibility that depolarization stimulated the increase in [Mg 2ϩ ] i in acini treated with gramicidin. Depolarization by this maneuver did not significantly alter the [Mg 2ϩ ] i (n ϭ 5; Fig. 3). Furthermore, this cytosolic-like, low Na ϩ medium (126 mM K ϩ and 15 mM Na ϩ ) abolished the Na ϩ gradient and eliminated the gramicidin-induced increase in [Mg 2ϩ ] i (Fig. 3). Thus, under experimental conditions which prevented the ionophore-induced increase in Na ϩ (data not shown), gramicidin failed to mobilize intracellular Mg 2ϩ . These results indicate that a Na ϩ -dependent mechanism is involved.
We further explored the Na ϩ dependence of the intracellular Mg 2ϩ response by activating Na ϩ /H ϩ exchange to increase the [Na ϩ ] i . Acid loading salivary acinar cells increases Na ϩ /H ϩ exchange activity and increases the [Na ϩ ] i (15). Fig. 4 (ϪBAPTA) shows that acid-loading sublingual acini by incu-

FIG. 5. Intracellular Na ؉ depletion inhibits Mg 2؉ mobilization.
Mag-fura-2-loaded acini were depleted of intracellular Na ϩ for either 1 (n ϭ 6) or 7 (n ϭ 9) min by superfusing with Na ϩ -free PSS. Ten M CCh was added at the time indicated by the arrow. Each trace is a representative response. further depletion of [Na ϩ ] i produced by exposing acini to the Na ϩ -free medium for 7 min suppressed the CCh-induced increase in [Mg 2ϩ ] i Ͼ75% (n ϭ 9; p Ͻ 0.001). Fig. 6 displays the association between Na ϩ depletion and the stimulated increases in [Ca 2ϩ ] i , [Na ϩ ] i , and [Mg 2ϩ ] i . Here, depletion of intracellular Na ϩ did not reduce the CCh-stimulated increase in [Ca 2ϩ ] i (n ϭ 5) but significantly inhibited the increases in both [Na ϩ ] i and [Mg 2ϩ ] i (n ϭ 5). These results indicate that agonist-stimulated mobilization of intracellular Mg 2ϩ does not directly involve an increase in [Ca 2ϩ ] i , whereas the increase in the [Na ϩ ] i is required.
The sustained increase in [Mg 2ϩ ] i induced by CCh is also contingent upon the [Na ϩ ] i . Fig. 7 shows that after three minutes CCh stimulation [Ca 2ϩ ] i increased 423% (n ϭ 5), [Na ϩ ] i increased 126% (n ϭ 6), and [Mg 2ϩ ] i increased approximately 45% (n ϭ 7). Removal of Na ϩ after 3 min stimulation did not alter the sustained increase in [Ca 2ϩ ] i . Nonetheless, [Mg 2ϩ ] i decreased rapidly in parallel with changes in [Na ϩ ] i . These results are in accord with the hypothesis that the muscarinicinduced mobilization of the intracellular Mg 2ϩ pool is mediated by a Na ϩ -dependent transport mechanism.
Inhibition of Mg 2ϩ Release by Quinidine-The Na ϩ dependence of both the CCh-and the gramicidin-induced discharge of Mg 2ϩ from the intracellular pool suggests that a Na ϩ /Mg 2ϩ exchange mechanism may be involved. The Na ϩ /Mg 2ϩ exchangers in several cell types are sensitive to quinidine (5,6,21). Fig. 8 shows that both the CCh-induced and gramicidininduced increases in [Mg 2ϩ ] i were inhibited by quinidine. Acini were pretreated with 0.5 mM quinidine for approximately 1 min to permit the inhibitor to enter the cell (22). Acini were then stimulated with either gramicidin or CCh in the continued presence of quinidine. The [Mg 2ϩ ] i responses were dramatically inhibited by quinidine, whereas Fig. 9 demonstrates that CCh-stimulated [Ca 2ϩ ] i (control, n ϭ 5; ϩ quinidine, n ϭ 5) and [Na ϩ ] i (control, n ϭ 9; ϩ quinidine, n ϭ 7) increases were essentially unchanged suggesting that quinidine blocked Na ϩdependent release of Mg 2ϩ from an intracellular pool. DISCUSSION We previously observed in rat sublingual acini that muscarinic stimulation induces Mg 2ϩ mobilization from an intracellular pool and this increase in [Mg 2ϩ ] is both Na ϩ -and Ca 2ϩdependent (7). The [Na ϩ ] i is tightly coupled to the [Ca 2ϩ ] i in salivary acinar cells (14,15,23). In the present study we examined the interdependence of [Mg 2ϩ ] i on Na ϩ and Ca 2ϩ . Our results clearly demonstrate that Mg 2ϩ release is mediated by a Na ϩ -dependent ion transport mechanism, and the receptor-stimulated rise in [Ca 2ϩ ] i activates this mechanism indirectly by increasing the [Na ϩ ] i .
The muscarinic-stimulated increase in [Ca 2ϩ ] i is sufficient but not necessary for Mg 2ϩ mobilization. In fact, it appears that liberation of the intracellular Ca 2ϩ pool does not release Mg 2ϩ but induces uptake of Mg 2ϩ by the Ca 2ϩ pool, most likely to maintain the charge balance of this pool (24). The results displayed in Fig. 1  The plasma membrane of numerous cell types, including sublingual acinar cells, contains Na ϩ /Mg 2ϩ exchangers (5,6,21,25,26) which utilize the Na ϩ gradient generated by Na ϩ ,K ϩ -ATPase as the energy source for extrusion of Mg 2ϩ . When the Na ϩ gradient was reduced about 4-fold by gramici-FIG. 6. Effects of extracellular Na ؉ removal on the cytosolic free Ca 2؉ , Na ؉ , and Mg 2؉ concentrations. Rat sublingual acini were loaded with ion-sensitive fluorescent indicators (fura-2, SBFI, or magfura-2). Dye-loaded acini were superfused for 7 min with Na ϩ -free PSS to deplete intracellular Na ϩ as indicated by the bar. Ten din-induced Na ϩ influx, the Na ϩ /Mg 2ϩ exchanger should mediate little if any Mg 2ϩ influx. Indeed, Fig. 2 shows that gramicidin induced a comparable increase in [Mg 2ϩ ] i in a Mg 2ϩ -free medium to that seen in Mg 2ϩ -containing solution ruling out this possibility, and clearly demonstrated the release of Mg 2ϩ from an intracellular pool. Moreover, superfusing acinar cells in a cytosolic-like [Na ϩ ] and [K ϩ ] medium to prevent net movement of these ions upon exposure to gramicidin abolished the increase in [Mg 2ϩ ] i , inferring that Mg 2ϩ release requires an increase in the [Na ϩ ] i . In agreement with this observation, increasing [Na ϩ ] i by treating acini with sodium propionate produced an increase in [Mg 2ϩ ] i similar to that seen with gramicidin (Fig. 4). The effects of gramicidin on the increase in [Mg 2ϩ ] i were not likely due to increased permeability of cations across the membranes of intracellular organelles considering gramicidin primarily affects the plasma membrane (27).
The muscarinic-stimulated mobilization of Mg 2ϩ is Na ϩdependent (7). Depletion of cytosolic Na ϩ blunted the CChstimulated [Mg 2ϩ ] i increase without significantly influencing the CCh-induced increase in [Ca 2ϩ ] i (Figs. 6 and 7). Furthermore, suppressing Na ϩ influx mediated by the Na ϩ /K ϩ /2Cl Ϫ cotransporter by replacing Cl Ϫ (28) with either gluconate or SCN Ϫ blunted the CCh-stimulated increase in [Mg 2ϩ ] i (data not shown), clearly indicating that the muscarinic-stimulated Mg 2ϩ release, like the gramicidin-induced increase in [Mg 2ϩ ] i , required an increase in the [Na ϩ ] i . Thus, it appears likely that the increases in [Mg 2ϩ ] i induced by both gramicidin and CCh were from the same intracellular pool.
In summary, our data indicated that increasing the intracellular [Na ϩ ] i was necessary for the mobilization of Mg 2ϩ from an intracellular pool by Ca 2ϩ -mobilizing agonists. Bypassing receptors, with either gramicidin or Na-propionate, raised the intracellular [Na ϩ ] and increased the [Mg 2ϩ ] i . These results show that neither receptor activation nor Ca 2ϩ mobilization is required to discharge the intracellular Mg 2ϩ pool. Thus, the agonist-stimulated increase in [Mg 2ϩ ] i involves a cascade of events including an initial increase in [Ca 2ϩ ] i which activates Na ϩ influx. The resultant increase in [Na ϩ ] i then triggers the release of Mg 2ϩ from the intracellular pool. Mg 2ϩ is important for regulating Na ϩ and Ca ϩ pump activity and therefore is critical in the modulation of fluid secretion by salivary acinar cells (14). Quinidine, an inhibitor of Na ϩ /Mg 2ϩ exchangers (5,6,21), effectively blocked Mg 2ϩ mobilization without altering the [Na ϩ ] i increase. Consequently, the release of Mg 2ϩ was apparently not due to Na ϩ competing with Mg 2ϩ for binding sites on an intracellular Mg 2ϩ buffering system; it is hard to visualize how gramicidin could disrupt such an interaction. Thus, the simplest interpretation of our data is that an increase in the [Na ϩ ] i activates Na ϩ /Mg 2ϩ exchangers located in the membrane of an intracellular pool.