Characterization of calcium translocation across the plasma membrane of primary osteoblasts using a lipophilic calcium-sensitive fluorescent dye, calcium green C18.

The synthesis of Calcium Green C18, a lipophilic fluorescent calcium-sensitive dye, and its use as a monitor of Ca2+ efflux from cells is described. This indicator consists of a Calcium Green-1 molecule conjugated to a lipophilic 18-carbon alkyl chain which will intercalate into cell membranes. The Kd of the indicator for Ca2+ in aqueous solution (pH 7.2, 22°C, ionic strength 0.1 M) is 0.23 ± 0.04 μM and in the presence of liposomes is 0.062 ± 0.007 μM. Due to its high negativity, the calcium chelating fluorophore faces the cell exterior, when loaded under a defined set of conditions. The dye was found largely on the surface of the cells when loaded at a concentration of 5 μM for 10 min at 37°C. Five minutes after introduction of EGTA, 83-95% fluorescence disappeared, indicating that most of the fluorophore was on the cell surface. Photobleaching was minimal (3-13%). A confocal laser scanning microscope was used to detect and quantify fluorescence. Internalized dye was apparent in cells loaded for longer times (30-60 min) and in membrane-impaired cells, as shown by uptake of propidium iodide. Under defined confocal laser scanning microscope settings, a transient fluorescence at the periphery of ∼30% of the cells was observed following 10−8M parathyroid hormone treatment, indicating the presence of outwardly directed calcium transport across the plasma membrane. Calcium efflux usually lasted 7-10 min, peaking at around 2-3 min. Changes in cell shape were also observed. Calcium efflux was shown to be sensitive to (a) 10 μM quercetin and 10 μM vanadate, partially specific inhibitors of plasma membrane Ca2+-ATPase, to (b) 0.1 mM trifluoperazine, an agent which renders calmodulin ineffective, and to (c) 10 mM neomycin sulfate, which blocks release of Ca2+ from intracellular stores. Thapsigargin (5 μM), an inhibitor of Ca2+-ATPase of the endoplasmic reticulum, prolonged fluorescence. These observations indicate that cell surface fluorescence was due to the capture of Ca2+ by Calcium Green C18 after Ca2+ had been translocated across osteoblast plasma membranes. Involvement of the plasma membrane Ca2+-ATPase, known to be present in osteoblasts in substantial amounts, is implicated.

Osteoblasts line bone surfaces at sites of bone formation and secrete a protein matrix that has the capacity to calcify. The precise manner by which Ca 2ϩ passes through the layer of osteoblasts to sites of calcification is not known. The demonstration of substantial amounts of a plasma membrane Ca 2ϩ -ATPase in osteoblasts (Akisaka et al., 1988) raised the interesting possibility that the enzyme may be involved in movement of Ca 2ϩ to calcifying sites in bone. More recently, a plasma membrane Ca 2ϩ -ATPase has been cloned and sequenced from human osteoblast cell lines (SAOS-2; HOBIT) and shown to bear high sequence homology to other plasma membrane Ca 2ϩ -ATPases (Kumar et al., 1993). In a preparation of vesicles derived from osteoblast plasma membranes, the magnitude and direction of calcium translocation has been determined (Gay and Lloyd, 1995). Sealed inside-out vesicles took up Ca 2ϩ , implying that the direction of pumping in cells was outward, as has been found for numerous other cell types (Carafoli, 1987;Garrahan and Rega, 1990). The uptake rate by osteoblast plasma membrane vesicles was 9.9 Ϯ 2.3 nmol/mg of protein/min, which falls in the range reported for several other vesicle studies, including red blood cells (Caride et al., 1983), parotid gland (Takuma et al., 1985), neutrophils (Ochs and Reed, 1983), and liver (Kraus-Friedmann et al., 1982). Na ϩ / Ca 2ϩ exchange also exists in osteoblasts (Krieger, 1992;Short et al., 1994) and may contribute in a minor way to the Ca 2ϩ efflux observed, as subsequently discussed.
The present study was undertaken to establish characteristics of the fluorophore, Calcium Green C 18 , when applied to living cells and to utilize the dye to evaluate calcium efflux in intact osteoblasts. Parathyroid hormone was employed to stimulate an increase in [Ca 2ϩ ] i . 1 This occurs as a consequence of Ca 2ϩ release from intracellular stores as has been shown in other studies of osteoblasts (e.g. Reid et al. (1987), Civitelli et al. (1989)). It was anticipated that the return of [Ca 2ϩ ] i to prestimulatory levels would involve, in part, efflux by plasma membrane Ca 2ϩ -ATPase activity, since this enzyme system is a major pathway utilized by cells to restore [Ca 2ϩ ] i to basal levels (Carafoli et al., 1990). The lipophilic calcium-sensitive fluorescent dye, Calcium Green C 18 , was selected as a likely candidate for monitoring Ca 2ϩ efflux. Since the fluorophore is hydrophilic and is negatively charged, it is excluded from the lipid bilayer; the lipophilic hydrocarbon chain lodges in the plasma membrane, holding the dye in close proximity to the membrane. We reasoned that Ca 2ϩ , upon emerging from the cell, would be trapped by the fluorophore, which would then fluoresce. A laser scanning confocal imaging system was used to visualize sites of calcium efflux and to measure both the intensity and duration * This work was supported by National Institutes of Health Grant DE09459. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Synthesis of Calcium Green C 18 , Tetrapotassium Salt-Calcium
Green-1 is a fluorescent indicator of free Ca 2ϩ ions. It consists of a dichlorofluorescein conjugated to an aromatic Ca 2ϩ binding site (BAPTA) through an amide spacer (Haugland, 1992). This indicator can be excited at 488 nm and gives an increase in emission intensity at 530 nm on reversibly binding Ca 2ϩ , with a K d for Ca 2ϩ of 0.189 M. The increase in emission intensity for free Calcium Green-1 in solution is approximately 12-fold. Dichlorofluorescein was chosen as the fluorophore for the synthesis of Calcium Green because it is less pH-sensitive than fluorescein itself and still retains a high quantum yield (0.75 at saturating Ca 2ϩ ) when excited at 488 nm.
Based on previous experience with the synthesis of a lipophilic Fura-2 (Etter et al., 1994), we synthesized a lipophilic derivative of Calcium Green-1. The synthesis starts with catalytic reduction of 5,5dinitro-BAPTA tetramethyl ester to the diamine (Pethig et al., 1989), which is then reacted with 1 eq of stearoyl chloride to give the monosubstituted lipophilic BAPTA tetramethyl ester. This is conjugated to a protected dichlorofluorescein, and the protecting groups on the BAPTA and dichlorofluorescein are hydrolyzed with base. The resulting soapy fluorescent product is purified by reverse-phase chromatography to give pure Calcium Green C 18 tetrapotassium salt. The synthesis of Calcium Green C 18 tetrapotassium salt requires synthesis of four intermediates, shown in Fig. 1.
Synthesis of Compound I: 5,5Ј-bis-Amino-BAPTA Tetramethyl Ester-A solution of 11 g (17.7 mmol) of 5,5Ј-dinitro-BAPTA tetramethyl ester as described by Pethig et al. (1989) is dissolved in 200 ml of DMF and shaken under hydrogen at 40 p.s.i. for 3 h in the presence of 0.8 g of 10% palladium on charcoal. TLC in ethyl acetate shows the formation of a single new product that is colorless and gives a brown/red product on reaction with ninhydrin. The catalyst is removed by filtration through diatomaceous earth, and the colorless filtrate is concentrated under vacuum to a gray oil, which is stirred with 200 ml of methanol for 1 h to give a gray solid. The solid is filtered and dried to give 7.0 g (88% yield), pure to TLC in ethyl acetate or 5% methanol in chloroform.
Synthesis of Compound II: N-Stearoyl-5,5Ј-bis-amino-BAPTA Tetramethyl Ester-A solution of stearoyl chloride (1.6 g; 5.4 mmol) in 20 ml of dry DMF is added dropwise over 2 h to a stirring solution of compound I (3.0 g, 5.4 mmol) in 70 ml of dry DMF. TLC analysis in ethyl acetate shows the formation of two UV quenching products. The predominant product runs with an R f of 0.4 and reacts with ninhydrin to give a ruddy brown product. The reaction is added to 150 ml of ethyl acetate and washed three times with saturated NaCl and once with pure water. The organic layer is dried over Na 2 SO 4 and evaporated to a tan oil. This is purified over 150 ml of silica gel (40 -60 m) eluted with 20% ethyl acetate, 80% chloroform. Purified fractions are pooled and evaporated to a semicrystalline solid that is recrystallized from methanol to give 1.4 g of pure material (40% yield). NMR in CDCl 3 shows 0.95 ppm of 3 H, t; 1.4 ppm of 27H, s(b); 1.7 ppm of 2H, m; 2.4 ppm of 2H, m; 3.6 ppm of 12H, s; 4.1 ppm of 4H, s; 4.15 ppm of 4H, s; 4.2-4.4 ppm of 4H, d; 6.8 ppm of 2H, d; 7.05 ppm of 2H, d; 7.3 ppm of 2H, s.
Synthesis of Compound III: 5-Carboxy-2,7-dichlorofluorescein Diacetate, Isobutyl Anhydride-The 5-isomer of carboxy-2,7-dichlorofluorescein diacetate is purified from mixed 5 and 6 isomers (Molecular Probes) by repeated crystallization from ethanol. The resulting colorless solid is dissolved in 20 ml of CH 2 Cl 2 at 0°C with isobutyl chloroformate (4.67 mmol) added. The reaction is stirred for 1 h and is evaporated to a colorless semisolid; ethyl acetate (50 ml) is added, and the insoluble salts are removed by filtration. The filtrate is evaporated and dried for 5 h under high vacuum to give the isobutyl anhydride as a colorless foam (2.2 g; 3.5 mmol, 90% yield). NMR in CDCl 3 shows 1.55 ppm of 6H, s; 2.35 ppm of 6H, s; 3.95 ppm of 1H, s; 4.4 ppm of 2H, q; 6.85 ppm of 2H, s; 7.2 ppm of 2H, s; 7.85 ppm of 1H, s; 8.15 ppm of 1H, d; 8.35 ppm, m.
Synthesis of Compound IV: Calcium Green C 18 Diacetate, Tetramethyl Ester-A solution of Compound III (0.23 g; 0.36 mmol) in 1.5 ml of dry DMF is added dropwise over 0.5 h to a stirring solution of Compound II (0.3 g, 0.36 mmol) in 1.5 ml of dry DMF at 0°C. After stirring at this temperature for 2 h, TLC in 5% methanol, 95% chloroform shows conversion of the two starting materials to a new product of intermediate R f . On exposure to ammonia fumes, the product turns dark red and is only weakly fluorescent. The reaction is evaporated at 15°C to a tan oil, which is purified on silica gel (40 -60 m) eluted with pure ethyl acetate. The pure fractions are combined and evaporated to a tan oil with a mass of 250 mg.
Synthesis of Compound V: Calcium Green C 18 , Hexapotassium Salt-Compound IV (0.25 g, 0.19 mmol) is dissolved in 1 ml of dioxane and diluted with 2 ml of methanol. Potassium hydroxide (0.11 g, 1.9 mmol) is dissolved in 0.5 ml of pure water and added to the stirring ester over 10 min. The solution immediately turns red as the acetates are cleaved from the dichlorofluorescein. The reaction is stirred at room temperature for 20 h until TLC in 1:1 CHCl 3 :CH 3 OH shows a single fluorescent product with an R f of 0.5. Cleavage products are also visible as minor impurities formed in the hydrolysis. The reaction is evaporated, diluted with 5 ml of pure water, and purified twice on 300 ml wet bed volume of lipophilic Sephadex LH-20 eluted with pure water adjusted to pH 8 with the addition of dilute KOH. Pure fractions are combined, frozen, and lyophilized to give 210 mg of red powder. NMR in D 2 O shows 0.7 ppm of 3H, t; 1.0 -1.5 ppm of 28H, s(b); 1.6 ppm of 2H, m(b); 2.4 ppm of 2H, s(b); 3.9 ppm of 4H, s; 4.0 ppm of 4H, s; 4.5 ppm of 4H, s; 6.75 ppm of 3H, s; 6.85 ppm of 1H, d; 6.9 ppm of 1H, d; 7.05 ppm of 1H, d; 7.15 ppm of 1H, s; 7.3-7.4 ppm of 2H, m; 7.4 ppm of 2H, s; 7.45 ppm of 1H, d; 8.15 ppm of 1H, d; 8.45 ppm of 1H, s.
Spectral Properties of Calcium Green C 18 -Dissociation constants for Ca 2ϩ were determined in EGTA/CaEGTA-buffered Ca 2ϩ solutions with free Ca 2ϩ concentrations varying between 0 and 1.35 M in 100 mM KCl and 10 mM MOPS (pH 7.2, 22°C) using the Calcium Calibration Kit I (Molecular Probes) and an MPF-44B fluorescence spectrophotometer (Perkin-Elmer); excitation, 506 nm; emission, 526 nm; slit widths, 5 nm. Calcium buffers were made by the "pHimetric" method of Tsien and Pozzan (1989). Calcium Green C 18 was tested for response to free Ca 2ϩ both in aqueous solution and in the presence of DOPC liposomes at a dye concentration of 1 M. To create liposomes, 100 l of 40 mM DOPC (in ethanol) was injected slowly (over 1 min) into vortexing buffers, one that contained 10 mM CaEGTA and one that contained 10 mM EGTA, using a Hamilton Syringe and 23-gauge needle. The buffered DOPC solutions were stirred slowly for 15 min. Calcium Green C 18 in anhydrous dimethyl sulfoxide (1 mM) was then added to both DOPC solutions to achieve a dye to DOPC ratio of 1:400. These solutions were stirred for 45 min to ensure an even distribution of the dye and DOPC, then mixed in graded proportions according to Tsien and Pozzan (1989). The response and affinity of the indicator is dependent on the environment, with an increase in emission intensity of 9-fold in aqueous solution and 3.5-fold in the presence of liposomes. A plot of log(F Ϫ F min / Defining Conditions for Loading Calcium Green C 18 into Osteoblasts-Primary osteoblasts were isolated as described previously (Gay et al., 1994). Briefly, cells were scraped from periosteal surfaces of 2-3-week-old chick tibias, after mild enzyme treatment, and cultured on 25-mm glass coverslips in Dulbecco's modified Eagle's medium (Sigma) plus 10% heat-inactivated fetal bovine serum for 6 to 8 days.
Introduction of Calcium Green C 18 into the plasma membranes involved first rinsing osteoblasts adherent to coverslips three times in a reduced calcium balanced salt solution (RCBSS) which contained 127 mM NaCl, 3.8 mM KCl, 1.2 mM KH 2 PO 4 , 0.8 mM MgCl 2 , 5 mM glucose, and 10 mM HEPES buffer at pH 7.3. The cells were then exposed to 1, 2, 5, or 10 M Calcium Green C 18 , initially dissolved in anhydrous dimethyl sulfoxide, then diluted in RCBSS for 1, 2, 5, 10, 30, or 60 min at room temperature. Cells were rinsed again three times with RCBSS to remove excess indicator. The coverslip was inverted and placed in a Dvorak-Stotler Controlled-Environment Culture Chamber (DSC200, Nicholson Precision Instrument, Gaithersburg, MD) that was preloaded with RCBSS. The 5 M concentration (with a final concentration of dimethyl sulfoxide at 0.5%) and 10-min loading time were found to be optimal. Propidium iodide (2 g/ml) was present in the RCBSS in some experiments as a test of cell viability.
Determination of [Ca 2ϩ ] in the Reduced Calcium Balanced Salt Solution-Fluo-3 (1 mM, Molecular Probes) was dissolved in anhydrous dimethyl sulfoxide and further diluted to 0.1 mM with RCBSS or with distilled, deionized H 2 O. Fluo-3 (10 l, 0.1 mM) was added to 500 l of RCBSS in a 0.7-ml microcell cuvette (type 4, NSG Precision Cell, Inc., Farmingdale, NY), and fluorescence was measured (MPF-44B, Perkin-Elmer; excitation, 506 nm; emission, 526 nm; slit widths, 5 nm). The calcium concentration was determined using a standard Ca 2ϩ concentration curve (0 to 39.8 M Ca 2ϩ ) generated from the Calcium Calibration Buffer Kit II (Molecular Probes). This analysis revealed that the concentration of Ca 2ϩ in RCBSS was 0.22 M.
Determining Specificity of Membrane Localization and Extent of Photobleaching-In order to demonstrate that the dye was localized specifically in the plasma membrane, Calcium Green C 18 -loaded cells were exposed to 1.2 mM CaC1 2 added to RCBSS which was added through a port in the controlled environment chamber. Patterns of fluorescence were observed in optical sections using the 488 nm laser line, zoom 1.0, 10% neutral density filter, aperture opening 2.7 mm, gain 8, and black level 2 settings on the confocal microscope (Bio-Rad MRC 600). For a control, cells were treated with 50 mM EGTA for 5 min in order to chelate added calcium.
In order to assess the extent of photobleaching, cells were loaded with 5 M Calcium Green C 18 for 10 min, then placed in buffer that contained 1.2 mM CaCl 2 , and exposed to the laser beam every 5 s for 5 min ( Table I).
Characterization of Ca 2ϩ Efflux in Osteoblasts Using Calcium Green C 18 As an Ion Trap-In order to characterize Ca 2ϩ efflux by osteoblasts, cells loaded with Calcium Green C 18 were placed in the Dvorak-Stotler Chamber in RCBSS, and a bolus (20 l) of PTH (final concentration 10 Ϫ8 M) was injected through a port into the chamber as a means of increasing [Ca 2ϩ ] i . Optical sections, 1.0 micron in thickness, in the x-y plane through cell centers were recorded every 5 s for the first 200 s, then every 60 s thereafter, depending on the duration of the fluorescence due to Ca 2ϩ efflux. Under the chosen settings for the confocal system, the background fluorescence caused by the trace amount of Ca 2ϩ present in RCBSS was negligible. Controls included: thapsigargin (5 M), an inhibitor of endoplasmic reticulum Ca 2ϩ -ATPase; trifluoperazine (0.1 mM), an inactivator of calmodulin; quercetin (10 M) and sodium vanadate (10 mM), inhibitors of plasma membrane Ca 2ϩ -ATPase; and neomycin sulfate (10 mM), an inhibitor of phospholipase C and therefore of generation of IP 3 , the intracellular signal that causes Ca 2ϩ release from intracellular stores. Cells were pre-incubated with the inhibitors in RCBSS for 10 min at room temperature. RCBSS in the culture chamber also contained the same concentration of each inhibitor. Fluorescence images were recorded, and image analysis was carried out using COMOS commands in the Bio-Rad software package. The image space was divided into pixels (0.17 m 2 when using the 40ϫ objective lens) and assigned a gray level (on a scale of 1-256) as a measure of fluorescence intensity.

Synthesis and Fluorescence Excitation Spectra of Calcium
Green C 18 -The structures of the compounds synthesized, then utilized in the synthesis of Calcium Green C 18 , described under "Experimental Procedures," are diagramed in Fig. 1. Fig. 2 shows the excitation spectra of 1 M Calcium Green C 18 in the presence of increasing concentrations of Ca 2ϩ in aqueous solution. As [Ca 2ϩ ] increased, the emission intensity (at 530 nm) increased with little shift in wavelength.
Localization and Characterization of the Lipophilic Calcium Green C 18 in Cells-Cells loaded with Calcium Green C 18 that were exposed to 1.2 mM CaCl 2 appeared as rings in optical sections produced by confocal microscopy (Fig. 3A). A low level of cytoplasmic staining was also evident in many cells. This was, however, a minor contribution to total cell fluorescence. Occasionally, cell cytoplasm was intensely fluorescent; this phenomenon coincided with cell death, as these cells were found to stain with propidium iodide. Following the addition of EGTA to the observation chamber, fluorescence of cell peripheries were markedly reduced ( Fig. 3B; Table II). Controls indi-cated that photobleaching contributed only to a minor extent to the disappearance of the fluorescent images; as shown in Table  I, loss of fluorescence due to photobleaching after a total of 60 scans over a time frame of 5 min ranged between 3 and 13%. Serial optical sections along the z axis of dye-loaded cells further revealed that Calcium Green C 18 was evenly located along peripheries of cells (Fig. 4).
Characterization of Calcium Efflux-The plasma membrane of approximately 30% of the 8-day-old cultured osteoblasts were found to fluoresce following PTH stimulation. A rapid and transient increase in fluorescence was observed on the surface of the responding osteoblasts when PTH was added to the cell chamber (Fig. 5). The initial efflux was usually detectable 10 s after the addition of PTH. Fluorescence intensity peaked around 2-3 min and persisted for about 7-10 min. Cell shape changes, including the appearance of cell processes, were observed in some cases after PTH stimulation (Fig. 5).
Quercetin, sodium vanadate, trifluoperazine, and neomycin sulfate substantially impaired Ca 2ϩ efflux, both in magnitude and duration (Table III). Thapsigargin caused the fluorescence to to persist for 15-17 min (Table III), nearly twice the time found for untreated controls.

DISCUSSION
In this paper we have described the synthesis and spectral properties of a lipophilic long-chain Ca 2ϩ -sensitive fluorescent dye, Calcium Green C 18 , that localizes preferentially in the plasma membrane of cells. Our goal was to utilize this fluorescent Ca 2ϩ -specific probe to monitor Ca 2ϩ translocation out of osteoblasts. The Ca 2ϩ -trapping head group of Calcium Green C 18 facing the external medium served as a reporter of Ca 2ϩ efflux. Conjugation of Calcium Green-1 with a lipophilic tail, 18 carbons long, as shown in Fig. 1, provided a means of situating the fluorescent probe along a lipid-based membrane while preserving the capacity of the indicator to fluoresce. It is likely that the method of loading the dye achieved placement of most of the dye into the plasma membrane with the fluorophore facing the outward direction since added external EGTA abolished fluorescence. Furthermore, the observation that the plasma membrane Ca 2ϩ -ATPase inhibitors, quercetin, vanadate and trifluoperazine, greatly reduced membrane fluorescence (Table III) indicates that most of the Calcium Green C 18 was loaded in the preferred orientation. If substantial amounts of Calcium Green C 18 had been inwardly oriented, the dye would have fluoresced when [Ca 2ϩ ] i increased whether Ca 2ϩ efflux was inhibited or not. While some fluorescence was observed in the cytosol, it was minor compared to the cell periphery under short (10 min) loading times. At the longer loading times, internal fluorescence was common. Using three-dimensional imaging techniques Calcium Green C 18 has been found useful for demonstrating Ca 2ϩ efflux from primary osteoblasts stimulated with parathyroid hormone. It was also possible to monitor the duration of the response. This is the first study to demonstrate that Calcium Green C 18 can be used to detect Ca 2ϩ efflux from cells.
The spectral response of Calcium Green C 18 to increasing concentrations of Ca 2ϩ in aqueous solution, where the K d was 0.23 M, shows a linear increase in emission intensity in the submicromolar range (Fig. 2). In this state, the lipid tail of the dye is not embedded in a membrane, so the response is expected to differ slightly from that in cells. In the presence of DOPC liposomes (spectra not shown), a linear increase in fluorescence with increasing concentrations of Ca 2ϩ was also observed, but the affinity of the dye for Ca 2ϩ was higher (K d ϭ 0.062 M). The water solubility of this probe is better than for C 18 -Fura-2, as Calcium Green-1 has an extra negative charge. This should facilitate loading the dye into cell membranes from x ϭ Ϫ8.2 Ϯ 0.8 an aqueous solution. Further, once in place in the membrane, dye-dye interactions would be reduced due to the negativity of the head group. In cultured osteoblasts, Calcium Green C 18 appears to be localized along the entire plasma membrane, except for regions associated with the coverslip. Ca 2ϩ specificity is indicated since the calcium greens have been shown to preferentially bind calcium (Girard et al., 1992). Further, little fluorescence was observed under the chosen confocal settings when cells were bathed with RCBSS, which intrinsically contained trace amounts of Ca 2ϩ . Fluorescence was induced when Ca 2ϩ was added to the RCBSS (Fig. 3A) and then largely disappeared when the Ca 2ϩ -specific chelating agent EGTA was added externally (Fig. 3B). Because of the special orientation of Calcium Green C 18 and its relatively low K d , it is critical that trace amounts of calcium present in the RCBSS be small. At a Ca 2ϩ concentration present in the RCBSS of 0.22 M, ϳ50% of the indicator would be free to bind calcium emerging from the cell. The K d measured in aqueous solution, rather than in liposomes, is considered to better represent the K d of the dye intercalated into the plasma membrane since the fluorescent head group is entirely extracellular and is bathed in aqueous solution.
Another near-membrane Ca 2ϩ concentration indicator has been described recently that consists of Fura-2 conjugated to a lipophilic C 18 alkyl chain (Etter et al., 1994). C 18 -Fura-2 orients in the plasma membrane so that the fluorophore faces from the side to which it was applied. This indicator was found to detect rapid, localized changes in [Ca 2ϩ ] i which are undetectable by water-soluble bulk cytosolic fluorescent Ca 2ϩ indicators. Calcium Green C 18 , on the other hand, has a higher negative charge, which would assist in maintaining the outward orientation of the dye in the plasma membrane due to the negative membrane potential of intact cells. In addition, the negativity of Calcium Green C 18 would result in fewer dye-dye interactions once the dye is positioned in the cell membrane. The calcium greens are useful for kinetic analysis due to increased quantum yields on binding Ca 2ϩ (Haugland, 1993). However, since the calcium greens are not ratiometric dyes, they are not useful for measuring ion concentrations accurately. The lower K d of Calcium Green C 18 in liposomes than for C 18 -Fura-2 (0.062 M versus 0.15 M) indicates that Calcium Green C 18 can be used to detect lower levels of Ca 2ϩ and therefore is more sensitive. In this study we did not use an EGTA/CaEGTA buffer in the assay medium surrounding the cells because of concerns with cell membrane instability. Use of such a buffering system, needs to be explored, however, for detection of very low levels of Ca 2ϩ . The improved sensitivity of the calcium greens has been discussed by Eberhard and Erne (1991). In addition, the calcium greens show increased specificity, speed of response, and higher spatial and temporal resolution (Haugland, 1992). Advantages of Calcium Green C 18 being excitable at longer wavelengths over that of C 18 -Fura-2 include less cellular photodamage and decreased autofluorescence. In addition, Fura-2 may bind to cellular proteins and could lead to alterations in the response of the indicator to Ca 2ϩ (Blatter and Wier, 1990).
The present study is the first direct demonstration of Ca 2ϩ efflux from osteoblasts. The source of Ca 2ϩ is believed to be from intracellular stores since cytosolic calcium in osteoblast cell lines has been shown to increase following PTH treatment  (Reid et al., 1987;van Leeuwen et al., 1988;Bidwell et al., 1991;Yamaguchi et al., 1991). The pathways by which increased [Ca 2ϩ ] i is restored to basal levels are incompletely understood but are believed to involve generation of IP 3 (Ferris and Snyder, 1992). In the present study, we have shown that a portion of increased [Ca 2ϩ ] i is handled by a plasma membrane Ca 2ϩ -ATPase-dependent efflux mechanism and that re-entry into intracellular stores formed by the endoplasmic reticulum also occurs. By blocking the latter process with thapsigargin (Thastrup et al., 1987), more cytosolic calcium was available for translocation to the cell exterior as shown by a prolonged fluorescence at the cell surface. An alternative explanation for the prolonged effect could be that ATP-dependent processes in the plasma membrane were partially inhibited by thapsigargin. This possibility is contraindicated by studies which show the specificity of thapsigargin for Ca 2ϩ -ATPase of the endoplasmic reticulum (Lytton et al., 1991). It has been shown that once PTH binds to receptors on the cell membrane, phospholipase C is activated (Lowik et al., 1985). This in turn generates IP 3 which will then bind to the endoplasmic reticulum membrane derived Ca 2ϩ storage compartment, causing Ca 2ϩ to be released. Our experiments show that a substantially reduced amount of Ca 2ϩ efflux occurred in the presence of neomycin sulfate, an inhibitor of phospholipase C and, therefore, IP 3 generation (Bidwell et al., 1991). This supports the view that the Ca 2ϩ that is secreted in response to PTH is derived mainly from intracellular stores, through the IP 3 pathway.
Two major Ca 2ϩ efflux systems are known to exist in the plasma membrane of cells, the ubiquitous Ca 2ϩ -ATPase (Carafoli et al., 1990), and the less widespread Na ϩ /Ca 2ϩ exchanger (Carafoli, 1987). The plasma membrane Ca 2ϩ -ATPase has been shown to be present in osteoblasts in substantial amounts (Akisaka et al., 1988;Kumar et al., 1993;Gay and Lloyd, 1995). Na ϩ /Ca 2ϩ exchange has also been demonstrated in osteoblasts, but at an unknown level of participation (Krieger, 1992;Short et al., 1994). Typically, excitable tissues have notably high levels of the exchanger, whereas nonexcitable tissues have less or even no Na ϩ /Ca 2ϩ exchange (Philipson and Nicoll, 1993). The extent to which the Ca 2ϩ -ATPase and Na ϩ /Ca 2ϩ exchange systems may have contributed to this study can be partially assessed from the data in Table III. The three inhibitors used, trifluoperazine, vanadate, and quercetin, all markedly affect the Ca 2ϩ -ATPase. Na ϩ /K ϩ -ATPase, the enzyme which drives Na ϩ /Ca 2ϩ exchange by creating a sodium gradient, is also affected by these inhibitors. However, while trifluoperazine has been shown to inhibit Na ϩ /K ϩ -ATPase activity in red blood cell hemolysates, at the concentration we used (0.1 mM), trifluoperazine is more effective (2-5ϫ) in reducing Ca 2ϩ -ATPase activity (Luthra, 1982). In ameloblasts, cells which also form a calcifiable matrix, trifluoperazine has no effect on Na ϩ /K ϩ -ATPase activity (Sasaki and Garant, 1987). Quercetin, at the concentration we used (10 M), has almost no effect on eel Na ϩ /K ϩ -ATPase (Kuriki and Racker, 1976), but has a K i ϭ 4 -6 M for Ca 2ϩ -ATPase of erythocytes (Wü thrich and Schatzmann, 1980). Vanadate is also notably effective for red blood cell Ca 2ϩ -ATPase having a K i ϭ 5 M (Niggli et al., 1981). Under some conditions, however, vanadate has been shown to more effective for Na ϩ /K ϩ -ATPase than for Ca 2ϩ -ATPase (Bond and Hudgins, 1980). Our finding of similar degrees of inhibition of Ca 2ϩ efflux with all three inhibitors, which have widely differing degrees of effectiveness on Na ϩ /K ϩ -ATPase, suggests that Na ϩ /Ca 2ϩ exchange may have made little to no contribution to the Ca 2ϩ -efflux monitored in this study. Further, PTH may have inhibited Na ϩ /K ϩ -ATPase activity as has been shown in the renal tubule (Ribeiro and Mandel, 1992). However, manipulation of external Na ϩ and treatment with ouabain, a specific Na ϩ /K ϩ -ATPase inhibitor, are needed for critical assessment of the participation of Na ϩ /Ca 2ϩ exchange.
It has been reported that parathyroid hormone alters the shape of osteoblasts, usually from a spreadout to a stellate morphology (Egan et al., 1991;Ali et al., 1990). Our observations of cell shape change induced by PTH corroborate the earlier reports.
In summary, this study has provided evidence that Calcium Green C 18 can be employed to monitor Ca 2ϩ efflux by cells. Plasma membrane Ca 2ϩ -ATPase of primary osteoblasts is implicated as having a major role in actively transporting intracellular calcium to the extracellular space. The enzyme activity is calmodulin-dependent and can be partially blocked by Ca 2ϩ -ATPase inhibitors. Also shown is that calcium is released from intracellular stores which utilize the IP 3 pathway upon receiving signals from PTH-occupied receptors; this process is substantially impaired by interfering with IP 3 production. Apparently, when thapsigargin blocks re-entry of intracellular calcium into the endoplasmic reticulum, more calcium is transported through the plasma membrane to the exterior.