The Regulation of Store-dependent Ca2+ Influx in HL-60 Granulocytes Involves GTP-sensitive Elements*

In granulocytes, emptying of intracellular Ca2+ stores activates Ca2+ influx across the plasma membrane. To study the putative role of GTP-binding proteins in this process, we have introduced non-hydrolyzable guano- sine phosphate analogues into the cytosol of non-perme-abilized HL-60 granulocytes using an endocytosis- hypoosmotic shock procedure. At the cytosolic concentrations obtained (100-500 w), neither guanosine 5'-3- 0-(thi0)triphosphate (GTPyS) nor guanosine 5'-3-0-(thio)diphosphate (GDPPS) affected basal [Ca2+Ii. Ca2+ release in response to the receptor agonist Met-Leu-Phe, the Ca2+-ATPase inhibitor thapsigargin, or the Ca2+ iono- phore ionomycin r,as also unaffected by GTPyS or GDPPS. In contrast, the activation of the Ca2+ influx pathway by Met-Leu-Phe or by thapsigargin was blocked by GTPyS but not by GDPPS. The GTP+ effect was mimicked by NaF. The GTP@ and NaF effects were independent of protein kinase C activation and actin polymerization. Our results demonstrate that a GTP- sensitive element is involved in the signaling between intracellular Ca2+ stores and plasma membrane Ca2+ channels. The identical effects of GTP+ and NaF suggest that the GTP-sensitive element is


G-protein.
Ca2+ influx in human granulocytes does not appear to involve voltage-operated, receptor-operated, or second messenger-operated Ca2+ channels. The only cellular parameter that is consistently correlated with the activity of the plasma membrane Ca2+ influx pathway is the filling state of intracellular Ca2+ stores; any procedure that empties intracellular Ca2+ stores increases the divalent cation permeability of granulocytes (for review see Ref. 1). This store-dependent Ca2+ influx, often also referred to as capacitative Ca2+ influx, is found in many cell types and may represent the most widely distributed mechanism of Ca2+ influx found in eukaryotic cells (2,3). Storedependent Ca2+ influx occurs through highly selective Ca2+ channels with a very small single channel conductance (4, 5). Terwindt Foundation, Geneva.
Very little is known about the biochemical basis of the signaling between intracellular Ca2+ stores and plasma membrane Ca2+ channels. The involvement of soluble messengers (6-81, protein phosphatases (6,9), or cytoskeletal elements (lo), or a direct interaction between proteins from Ca2+ stores and the plasma membrane (11) has been suggested. GTP-binding proteins (G-proteins)' appear to be ideal candidates to participate in the regulation of store-dependent Ca2+ influx. They have been implicated in the regulation of ion channels and also in the regulation of communication between intracellular organelles (for recent reviews see Refs. This study was designed to test the involvement of G-proteins in the mediation of store-dependent Ca2+ influx in granulocytes. Our results demonstrate that cytosolic GTPyS but not GDPOS blocks the activation of store-dependent Ca2+ influx. The GTPyS effect is mimicked by fluoride. Our results suggest the involvement of GTP-sensitive elements, most likely a trimeric G-protein, in the regulation of store-dependent Ca2+ influx. EXPERIMENTAL PROCEDURES Materials and Buffers-fMLP, thapsigargin, Me2S0, HEPES, and DTPA were purchased from Sigma and fura-WAM from Molecular Probes (Eugene, OR). RPMI 1640 culture medium and fetal calf serum were purchased from Gibco (Paisley, Scotland, UK). Other chemicals were of analytical grade and were obtained from Fluka (Buchs, Switzerland) or Sigma. When drugs were added as Me2S0 solutions, the final concentration of Me2S0 in the recording medium did not exceed 0.25%. Experiments were performed in a medium containing (in m): NaCl, 138; KC1, 6; MgCl,, 1; glucose, 20; HEPES, 20; pH 7.4.
Culture of HL-60 Cells-HL-60 promyelocytes were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum and induced to differentiate to granulocytes by addition of 1.3% Me2S0 (v/v) for 7 days to the culture medium, as described previously (19).
Endocytosis-Osmotic Shock Procedure-The procedure applied in this study is a slight modification of the original procedure described by Rechsteiner (20). 40 x loe HL-60 granulocytes were incubated in 250 pl of a buffer containing 143 n m NaCl, 6 m KCl, 1 m MgSO,, 20 m HEPES pH 7.4,0.1% glucose, 375 m sucrose, 7.5% polyethylene glycol 1000, 7.5% fetal calf serum. Where indicated, the solutions also contained 50 m GTPyS or 50 m GDPPS. The cells were incubated for 15 min at 25 "C to allow fluid-phase endocytosis of extracellular material. To induce hypoosmotic lysis of endosomes, 4 ml of Hz0 were added, and cells were incubated under hypoosmotic conditions for 60 s. Isoosmolarity was restored by addition of 3.5 ml of 1.8% NaCl. To quantify the efficacy of the endocytosis-hypoosmotic shock procedure, we introduced lucifer yellow (10 mglml) or [3Hlmannose (10 x lo6 cpdml) into the cytosol by the same method. After three washes, the cytosolic content of lucifer yellow or [3Hlmannose was measured as the amount of the respective compound that could be released by 20 p m digitonin (21). Assuming a cytosolic volume of 0.5 puce11 (22), the cytosolic concentrations were estimated as 2 * 1 pglml for lucifer yellow and 0.1 * 0.02 x lo6 cpdml for [3Hlmannose, corresponding to 0.2 * 0.1 and 1.0 * 0.2%, respectively, of the extracellular concentration present during the endocytosis-hypoosmotic shock procedure. Inspection of lucifer yellowloaded granulocytes by fluorescence microscopy showed that approximately  Fluorescence Measurements-Details of the procedure have been described previously (19). HL-60 granulocytes were loaded with 2 p~ fura-WAM (37 "C for 45 mid. Experiments were performed on a Perkin-Elmer fluorimeter (LS3, Perkin-Elmer Cetus), thermostated at 37 "C. Fluorescence emission was set at 505 nm, and fluorescence excitation at 360 nm (Mn2+ influx measurements) or at 340 nm ([Ca2+l, measurements). Mn2+ influx through the stimulated Ca2+ influx pathway was measured using the fura-2 quenching technique, as previously described (19). To standardize our measurement procedure and to avoid the use of an arbitrary selected "initial slope," we measured the fluorescence decrease within the first 60 s after Mn2+ addition (= total fura-2 quenching) and subtracted from this value the fluorescence increase after addition of the heavy metal chelator DTPA (= "unquenching" of extracellular fura-2).
Data Presentation and Statistics-Data are shown either as typical traces or as mean 2 S.E. of 3-9 determinations of at least three independent experiments. When indicated, statistical significance was tested by a paired t test.

RESULTS
To study a putative role of GTP-binding proteins in the activation of store-dependent Ca2+ influx, we have introduced nonhydrolyzable guanosine phosphate analogues into the cytosol of HL-60 granulocytes using an endocytosis-osmotic shock procedure. This technique allows the introduction of macromolecules into the cytoplasm of cells without disruption of the plasma membrane (20). It has been recently applied to granulocytes and did not interfere with complex motile functions, indicating that cellular integrity is preserved (23). In our studies, the endocytosis-osmotic shock procedure by itself (ie. in the absence of added GTP analogues) was not cytotoxic and did not interfere with basal and stimulated cellular Ca2+ homeostasis. The cytosolic concentrations of the compound of interest obtained with the method were between 0.2 and 1% of the concentrations present in the extracellular solution during the procedure. The GTPyS and GDPPS concentrations used during the endocytosis-osmotic shock procedure were 50 m~ and therefore yielded intracellular concentrations of approximately 100-500 p~ (see "Experimental Procedures" for details).
We first investigated the effect of non-hydrolyzable guanosine phosphate analogues on granulocyte Ca2+ homeostasis in the absence of extracellular Ca2+. At the obtained cytosolic concentrations, neither GTPyS nor GDPPS affected basal Ca2+ levels or Ca2+ release in response to the receptor agonist fMLP, the Ca2+-ATPase inhibitor thapsigargin, or the Ca2+ ionophore ionomycin (Fig. 1, A 4 and G I , and Table I). The absence of a n effect on basal [Ca2+Ii levels suggested that, at submillimolar concentrations, GTPyS did not induce a sustained activation of phospholipase C. The normal fMLP-induced Ca2+ release suggested that the achieved cytosolic GDPpS concentrations were not sufficient to block fMLP-induced phospholipase C activation. As the Ca2+ release in response to the Ca2+-ATPase inhibitor thapsigargin is thought to reflect the basal permeability of Ca2+ stores, the normal thapsigargin-induced Ca2+ release suggested that neither GTPyS nor GDPPS affected the permeability of intracellular Ca2+ stores. The normal ionomycin-induced Ca2+ release indicated that GTPyS and GDPPS had no effect on the total content of intracellular Ca2+ stores.
We next investigated the effect of non-hydrolyzable guanosine phosphate analogues on the activation of the Ca2+ influx pathway (Fig. 1, D-F and J-L, and Fig. 2). We measured the activity of the Ca2+ influx pathway as quenching of cytosolic  I thapsigargin-or ionomycin-induced Ca2+ release from intracellular Ca2+ stores GTPyS and GDPpS were introduced into the cytosol of HL-60 cells using the endocytosis-osmotic shock procedure, as described under "Experimental Procedures." Control cells were subjected to the same procedure in the absence of non-hydrolyzable nucleotide analogues. Cells were loaded with fura-2, and fluorimetric Ca2+ measurements were performed in a nominally Ca2+-free medium. Basal [Ca2+li was defined as the [Ca2+Ii measured before addition of a stimulus; fMLP-, thapsigargin-, or ionomycin-induced Ca2+ release was defined as the peak increase in [Ca2+]; after addition of the respective compound. No statistically significant ( p > 0.05) differences between GTPyS-loaded, GDPpS-loaded, and control cells were found.

Control GTPyS GDPRS
Basal  (19,24). The rate of fura-2 quenching in unstimulated cells was unaffected by GTPyS and was slightly, but not significantly, increased by GDPPS. The rate of fura-2 quenching increased approximately %fold after stimulation with either the receptor agonist or with the Ca2+-ATPase inhibitor thapsigargin (19,24). In GTPyS-loaded cells the stimulated increase in Mn2+ influx was inhibited by approximately 70%. The extent of inhibition was comparable for fMLP-induced and for thapsigar-  2. GTP+ blocks the activation of Mu2* entry by depleted Ca2* stores. GTP$ and GDPPS were introduced into the cytosol of HL60 cells using an endocytosis-osmotic shock procedure. Control cells were subjected to the same procedure in the absence of non-hydrolyzable nucleotide analogues. Cells were loaded with fura-2, and fluorimetric Ca2+ measurements were performed in a nominally Ca2+-free medium. Basal Mn2+ influx (empty columns), and MnZ+ influx in response to 100 m fMLP (gray columns) and 50 m thapsigargin (black columns) was assessed. fura-2 quenching (= Mn2* entry) is expressed as percent of control cells. Basal fura-2 quenching in control cells was 3.7 0.5 fluorescence uniW60 s; fMLPand thapsigargin (TG)-stimulated values were 6.3 0.7 and 8.2 0.8% (after subtraction of basal). The cytosolic GTPyS and GDPPS concentrations were estimated to be between 100 and 500 (see "Experimental Procedures"). The slightly increased Mn2* influx in GDPPS-loaded cells was statistically not different from control ( p > 0.05).
gin-induced Mn2+ influx. GDPPS-loaded cells showed a small but statistically not significant increase of stimulated Mn2+ influx.
Thus, our results demonstrate an exquisite sensitivity of the store-regulated Ca2+ influx to inhibition by GTPyS. Two families of G-proteins are thought to mediate GTPyS effects: heterotrimeric G-proteins and small G-proteins. To distinguish between heterotrimeric and small G-proteins, we have used fluoride, which can mimic the effect of GTPyS on large Gproteins (25,26) but is not an activator of small G-proteins (27). For these experiments, cells were preincubated for 10 min with different NaF concentrations (5,10, and 20 mM) and stimulated with thapsigargin for 5 min; Mn2+ was then added and fura-2 quenching was measured. NaF blocked thapsigargin-induced Mn2+ influx in a dose-dependent manner (Fig. 3, triangzes). To investigate whether NaF affected the filling state of Ca2+ stores, we estimated the amount of Ca2+ remaining in intracellular stores by measuring the increase of [Ca2+Ii in response to the Ca2+ ionophore ionomycin. Independently of the preincubation with NaF, about 20% of total ionomycin-releasable Ca2+ was left in the cells after thapsigargin stimulation (Fig. 3,  circles). Thus, NaF did not inhibit Ca2+ influx by preventing thapsigargin-induced emptying of Ca2+ stores. When NaCl instead of NaF was added to the cells no inhibition of Mn2+ influx was observed (see legend of Fig. 3), demonstrating that the inhibition by NaF is not due to the hyperosmolarity of the extracellular solution. Taken together, these results suggest that NaF, similar to GTPyS, inhibited the Ca2+ store to plasma membrane signaling.
To investigate whether GTPyS ions might act through activation of protein kinase C, we have compared the effects of GTPyS, PMA, and fMLP on [Ca2+Ii homeostasis in HL-60 granulocytes. Under our experimental conditions (ie. in a nominally Ca2+-free medium), fMLP did not inhibit, and PMA only weakly2 inhibited Mn2+ influx in HL-60 granulocytes (104.9 24.5%, n = 4, and 79.2 2 5.2%, n = 7, of control, respectively, mean S.E.; Fig. 4A). This is in accordance with influx by PMA might be accounted for simply by the diminished driving The quantitatively minor inhibition of thapsigargin-induced Mn2* force for divalent cation influx secondary to the profound and long lasting PMA depolarization of HL60 granulocytes (31, 32). previous studies in human neutrophils showing no inhibition of store-dependent Ca2+ influx by fMLP and PMA in a nominally Ca2+-free medium (28). To study whether the cytosolic GTPyS activated protein kinase C a t all, we compared the PMA and the GTPyS effect on fMLP-induced Ca2+ mobilization. It has been previously shown that in HL-60 granulocytes activation of protein kinase C inhibits phospholipase C activation (29). In accordance with this observation, we found an inhibition of fMLPinduced Ca2+ release by preincubation of cells with PMA, contrasting with the entirely normal fMLP-induced Ca2+ release in GTPyS-treated cells (Fig. 4B). Taken together these results demonstrate that activation of protein kinase C does not mimic the GTPyS inhibition of Ca2+ influx and that the cytosolic GTPyS concentrations obtained in our experiments did not activate protein kinase C.

-
GTPyS and fluoride are known to induce actin polymerization in granulocytes (30). Actin polymerization might inhibit Ca2+ store to plasma membrane signaling, for example by steric hindrance of a putative direct interaction of Ca2+ stores with the plasma membrane. We therefore tested the effect of cytochalasin B, which efficiently disrupts the F-actin network. As shown in Fig. 4C, cytochalasin B did not affect the activation of Ca2+ influx by thapsigargin (the initial rate of Mn2+-induced fura-2 quenching was 9.  FIG. 4. The GTP+ effect is not mediated by protein kinase C activation or actin polymerization. GTPyS was introduced into the cytosol of HL-60 cells using an endocytosis-osmotic shock procedure. Control cells were subjected to the same procedure in the absence of non-hydrolyzable nucleotide analogues. Cells were loaded with fura-2, and fluorimetric Ca2+ measurements were performed in a nominally Ca2+-free medium. Panel A, Mn2+-induced fura-2 quenching 5 min after control, dotted line), in cells exposed to 50 m PMA 2 min prior to addition of 50 m thapsigargin (TG) was recorded in control cells (= thapsigargin (= PMA), in cells exposed to 100 m fMLP 30 s prior to thapsigargin (= fMLP), or in GTPyS-loaded cells (= GTPyS, dotted line). Panel B, Ca2+ release in response to 100 m fMLP was recorded in a Ca2+-free medium in control cells (= control), cells exposed to 50 m PMA prior to fMLP (= fMLP), or in GTPyS-loaded cells (= GTPyS, dotted line). Panel C, Mn2+-induced fura-2 quenching was recorded in control cells and GTPyS-loaded cells in the presence and absence of 2.5 p g / d cytochalasin B (Cyto B ) . The cytosolic GTPyS concentration was estimated to be between 100 and 500 (see "Experimental Procedures").

DISCUSSION
Our results demonstrate that GTP-sensitive elements are involved in the regulation of store-dependent Ca2+ influx in HL-60 granulocytes; GTPyS and fluoride block the activation of the influx pathway by depleted Ca2+ stores in intact cells.
These observations are novel and appear to represent a new step toward understanding of the Ca2+ store to plasma membrane ~i g n a l i n g .~ The GTPyS and fluoride inhibition of Ca2+ influx are clearly not due to an effect on Ca2+ release or on the Ca2+ content of Ca2+ stores. The presence of GTPyS or GDPpS in the cytosol did not change the amount of ionomycin-releasable Ca2+ and did not influence the fMLP and thapsigargin-induced Ca2+ release. Similarly, the Ca2+ content of intracellular Ca2+ stores after thapsigargin stimulation of cells was not modified by preincubation with fluoride. As GTPyS was equally efficient in inhibiting fMLP-and thapsigargin-stimulated Mn2+ influx and fMLP did not block thapsigargin-induced Ca2+ influx, the GTPyS effect is probably not mediated by a fMLP receptorcoupled inhibitory G-protein. Our results also demonstrate that the GTPyS effect is not mediated by protein kinase C activation or by actin polymerization. Thus, the effects of GTPyS and fluoride observed in this study are best explained Two articles published after submission of this manuscript show the block of store-dependent Ca2+ influx by GTPyS in hepatocytes (33) and mast cells (34). As opposed to our results in HL-60 granulocytes, the authors suggest the involvement of small rather than trimeric G-proteins.
by postulating the involvement of a G-protein in the signaling between Ca2+ stores and Ca2+ channels.
What type of G-protein might be involved in the regulation of store-dependent Ca2+ influx? Although our data do not allow us to exclude a role for small G-proteins, the identical effects of GTPyS and of fluoride clearly suggest the involvement of a trimeric G-protein. In previous studies, we did not find evidence for an enhancement of basal or stimulated Ca2+ influx by pertussis toxin (311, suggesting that the GTPyS effect is not mediated by a pertussis toxin substrate; similarly, preliminary studies in our laboratory did not find any marked effect of cholera toxin on basal or stimulated Ca2+ influx.4 Thus, GTPyS and Fmost likely act either through the activation of an a subunit of a pertussis toxin-and cholera toxin-insensitive Gprotein or through an increase in free Ply subunits.
Taken together, our results would be best compatible with a tonic inhibition of the store-dependent Ca2+ channel by a heterotrimeric G-protein. The activity of this G-protein might determine the sensitivity of the Ca2+ channel to the depletion of Ca2+ stores; alternatively, however, inhibition of the activity of this G-protein might be involved in the mechanism of activation of store-dependent Ca2+ influx.
for critical reading of the manuscript. We thank Drs. G. St. J. Bird and