- Email: [email protected]
Crosstalk between cytosolic pH and intracellular calcium in human lymphocytes:
Cellular Signalling 12 (2000) 573–581 http://www.elsevier.com/locate/cellsig
Crosstalk between cytosolic pH and intracellular calcium in human lymphocytes Effect of 4-aminopyridin, ammoniun chloride and ionomycin A.G. Cabadoa, A. Alfonsob, M.R. Vieytesa, M. Gonza´lezc, M.A. Botanad, L.M. Botanab,* Departamento de Fisiologı´aa y Farmacologı´ab 1, Facultad de Veterinaria, Servicio de Hematologı´ac, Hospital Xeral, Seccio´n de Endocrinologı´ad, E-27002 Lugo, Universidad Santiago de Compostela, Spain Received 26 March 2000; accepted 22 May 2000
Abstract Stimulation of lymphocytes by specific antigens is followed by the activation of different signal transduction mechanisms, such as alterations in the cytoplasmic levels of Ca2⫹, H⫹ and variations in membrane potential. To study interrelationships among these parameters, changes in pHi and Ca2⫹ were measured with the fluorescent probes BCECF and Fura-2 in freshly isolated blood human lymphocytes. Moreover, membrane potential qualitative alterations were recorded with the fluorescent dye bisoxonol. In a bicarbonate-free medium, cell alkalinization with NH4Cl slightly decreased intracellular Ca2⫹ concentration ([Ca2⫹]i) due to efflux of Ca2⫹ from the cell. In contrast, an elevation of pHi induced with 4-AP increased [Ca2⫹]i, either in the presence or absence of external Ca2⫹. The increase in Ca2⫹-free medium is likely to be due to Ca2⫹ release from thapsigargin and caffeineindependent intracellular stores. Both 4-AP or NH4Cl induced a plasma membrane depolarisation, although with different kinetics. The ionosphere ionomycin increased pHi, Ca2⫹ levels and also induced membrane depolarisation. Together, these observations demonstrate a lack of correlation between the magnitude of changes in pHi and Ca2⫹. 2000 Elsevier Science Inc. All rights reserved. Keywords: Lymphocytes; pHi; Na⫹/H⫹ exchanger; Ca2⫹; 4-AP; NH4Cl; Thapsigargin; Caffeine; Ionomycin
1. Introduction It is a well established paradigm of cell activation that calcium is necessary at the initial steps in almost any cellular model. In lymphocytes, whose activation and proliferation forms the basis of the mechanism of immune defence, binding of antigen to surface receptors leads to intracellular generation of IP3, which releases Ca2⫹ from intracellular stores, thus raising Ca2⫹ transiently. Influx of Ca2⫹ from the external compartment is, however, necessary to maintain the plateau of such response, a prerequisite for the final mitogenic effect. In non-excitable cells such as lymphocytes, the capacitative sustained Ca2⫹ signal depends on plasma membrane Ca2⫹ channels activated by depletion of Ca2⫹ intracellular stores [1]. On the other hand, protons may function as a second messenger in a manner similar to Ca2⫹ [2]. In many * Corresponding author. Tel.: 34-982-252-242; fax: 34-982-252-242. E-mail address: ffbotana @lugo.usc.es (L.M. Botana)
cells, Na⫹/H⫹ exchanger is the most important among the mechanisms involved in pHi homeostasis and some studies have questioned the physiological relevance of Na⫹/H⫹ exchanger activation in the induction of cell proliferation [3–5]. At least for some cellular types, there is a correlation between pHi and intracellular Ca2⫹, since alkalinization, before cytosolic calcium increase, is a sufficient signal to trigger cell activation [2]. In some cell types it was proposed that alkalinization induced by mitogens may contribute to the length and magnitude of the Ca2⫹ signal. Although some studies found no pH response in Jurkat T cells stimulated with the mitogen OKT3 [6], in resting T lymphocytes, stimulation with OKT3 induces pHi rise via Na⫹/H⫹ exchanger activation. This mechanism is dependent on external Ca2⫹ and can be blocked by La3⫹. Also, OKT3 induces a rapid membrane potential-sensitive Ca2⫹ influx in T lymphocyte cell lines [7,8]. In addition, lectins that stimulate T cells raise intracellular Ca2⫹ accompanied by hyperpolarization of the membrane potential, apparently due to a Ca2⫹-acti-
0898-6568/00/$ – see front matter 2000 Elsevier Science Inc. All rights reserved. PII: S0898-6568(00)00101-7
574
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
vated K⫹ conductance [9]. Therefore, alterations in membrane potential are also important in cell signalling and they can be correlated with Ca2⫹ and protons changes. Mechanisms through which changes in pHi could affect Ca2⫹ include not only effects of plasma membrane potential and hence gating of voltage-sensitive Ca2⫹ channels, but also mobilisation of Ca2⫹ from intracellular stores [10]. Several possible interactions between pHi and Ca2⫹ have been described: sharing of common intracellular buffer sites by H⫹ and Ca2⫹; stimulation of metabolic acid production by increases of Ca2⫹; modulation of the membrane transports responsible for the regulation of pHi and Ca2⫹ by changes of Ca2⫹ and pHi, respectively; sensitisation of the IP3 receptor by increasing pHi and consequently stimulation of Ca2⫹ release from internal stores by changes of pHi [11]. Different reports in muscle, fibroblasts and nerve cells have also tried to relate Ca2⫹ responses to pHi changes. Depending on the cell type, Ca2⫹ increased, fell or did not change [12–15]. The link between protons and Ca2⫹ ions is complex because it is dependent on the cell model and the procedure used to induce the cellular change. To evaluate the relationship between intracellular Ca2⫹ and protons in human lymphocytes we used the fluorescent dyes Fura-2 and BCECF to examine changes in intracellular Ca2⫹ and cytosolic pH imposed by the weak base NH4Cl, 4-AP (a K⫹ channel blocker), and ionomycin (a calcium ionophore). 2. Methods 2.1. Materials Bis-(1,3-diethylthiobarbirturic acid) trimethine oxonol (bis-oxonol), 2⬘,7⬘-bis(carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) and Fura-2/AM were purchased from Molecular Probes (Eugene, OR). Ionomycin, gramicidin, 4-aminopyridine (4-AP), caffeine, amiloride and thapsigargin were purchased from Sigma Chemical (St. Louis, MO). Ficoll was from Pharmacia (Sweden). All other chemicals were from standard commercial sources and reagent grade or the highest purity. Physiological saline solution (Umbreit) was composed of (in mM): Na⫹, 142.3; K⫹, 5.94; Ca2⫹, 1; Mg2⫹, 1.2; Cl⫺, 126.1; CO3⫺, 22.85; PO4H2, 1.2 and SO24⫺,1.2 giving a final osmotic pressure of 300 ⫾ 5 mOsm/Kg H2O. BSA and glucose (1 mg/ml) were added to the solution; the pH was adjusted to 7.4. Na⫹-containing solution consisted of (in mM): KCl, 14; MgCl2, 1; CaCl2, 1; NaCl, 140; Hepes, 10 and glucose, 10. To make Na⫹free medium, NaCl was substituted by N-methyl-D-glucamine. K⫹ solution for pH calibration was composed of (in mM): KCl, 141; MgCl, 1; Ca Cl2, 1; Hepes, 20 and glucose, 10.
2.2. Lymphocyte isolation and purification Peripheral blood lymphocytes were isolated from freshly drawn blood from healthy donors. Purification of cells was carried out by means of Ficoll-Hypaque centrifugation. In order to remove platelets, blood was diluted and washed (3⫻) with Umbreit solution. Washed cells were placed on Ficoll and centrifugation (1000 g max) was carried out at room temperature for 30 min. After centrifugation lymphocytes, appearing in the interface, were washed (2⫻) and resuspended in Umbreit. Cells were counted using a model System 9000 Coulter Counter (Menarini Diagno´stico, S.A.). The isolated lymphocyte populations were maintained at 4⬚C in Umbreit solution containing heparin and buffered to pH 7.4. 2.3. Determination of changes in membrane potential Plasma membrane potential changes were recorded with the fluorescent probe negatively charged bis-oxonol. 0.5 ⫻ 106/ml lymphocytes were used for each experiment. Dye loading and determination of changes in the membrane potential were performed as previously reported [16]. 2.4. Determination of [Ca2⫹]i Purified lymphocytes were incubated in Na⫹-containing solution with 2.5 M fura-2/AM for 20 min at 37⬚C. After dye loading, cells were then washed twice and resuspended at 2.5 ⫻ 106/ml in a quartz cuvette with continuous stirring. Experiments were performed at 37⬚C. Calibration and calculation of [Ca2⫹]i was performed as described [17]. 2.5. Cell labelling and pHi measurement Lymphocytes (1 ⫻ 106) were loaded with 1 M BCECF and incubated for 20 min at 37⬚C. Cells were pulsed with 50 mM NH4Cl during the last 15 min of the incubation with the BCECF as described [18]. After incubation, cells were washed twice and placed in a thermostated quartz cuvette with continuous stirring. The calibration of fluorescence vs. pHi was made using nigericin in K⫹ solution as per [19]. Briefly, a calibration curve was obtained with four known values of pH, measuring the fluorescence ratio obtained in the presence of nigericin for each pH value. 2.6. Fluorescence measurements The fluorescence of fura-2, BCECF and bis-oxonol was measured with a Shimadzu RF-5000 spectrofluorometer. Bis-oxonol fluorescence was recorded at excitation and emission wavelengths of 540 nm and 560 nm, respectively, as described [16]. For fura-2-loaded cells, the ratio of fluorescence was determined by switching two excitation wavelengths, 340 and 380 nm (bandwidth 5 nm) and recording the emission at 500 nm (bandwidth
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 1. (a) Recovery of pHi in acid-loaded human lymphocytes vs. basal pHi. Lymphocytes were incubated with BCECF as described in Methods. The data are representative of five experiments. (b) Recovery of pHi in acid-loaded lymphocytes incubated with BCECF as described in Methods. Effect of 100 M amiloride and absence of external Na⫹. Data are means ⫾ S.E.M. of four experiments.
10 nm). BCECF fluorescence was recorded by alternating the excitation at 490 and 440 nm and measuring emitted fluorescence at 530 nm. The ratio of the two excitation wavelengths allowed the calculation of pHi without interference from differences in dye loading, dye bleaching, or cell number [20]. Drugs were added from concentrated solutions with a pipette directly into the cuvette through a small hole on top of the cuvette lid. 3. Results Fig. 1a shows basal pHi in human lymphocytes incubated 20 min with BCECF (open circles). After incubation with the dye cells were washed (2⫻) and resuspended in a bicarbonate-free medium containing 140 mM external Na⫹. When the cells were added to the cuvette the temperature was 4⬚C and in 2 min the temperature reached 37⬚C. The slight decrease of pHi is due to the increase of temperature. Resting pHi in human lymphocytes decreased from 7.347 ⫾ 0.027 at 4⬚C to 7.275 ⫾ 0.03 (n ⫽ 17) at 37⬚C when extracellular pH is 7.4. Fig. 1a also shows recovery of pHi in lymphocytes
575
after an acid load imposed by 50 mM NH4Cl (closed circles). Cells acidified to pHi 6.79 followed by a slower alkalinization (⬇10 min) toward the resting pHi (7.06). However, cells never recovered their basal pHi of 7.27, indicating the existence of a bicarbonate-dependent component of the recovery (unpublished observation). To know if part of the mechanism involved in the recovery of pHi after an acid load is a Na⫹/H⫹ exchanger we carried out experiments in the absence of external Na⫹ and in the presence of 100 M amiloride, an inhibitor of this transporter. Fig. 1b shows recovery of pHi in lymphocytes after an acid load imposed by 50 mM NH4Cl in a medium containing 140 mM external Na⫹ (closed circles), in Na⫹-free medium (triangles) and in the presence of 100 M amiloride (open circles). Cells recovered from the acid load only in the presence of external Na⫹ (closed circles), strongly suggesting that the alkalinization is due to Na⫹/H⫹ exchanger since pHi recovery is inhibited in the absence of external Na⫹ and in the presence of amiloride. To examine if basal pHi influences [Ca2⫹]i we carried out experiments monitoring lymphocytes after incubation with Fura-2. Cells were acidified with NH4Cl and then allowed to recover their pHi, or introduced into the cuvette at their resting pHi. After 10 min the [Ca2⫹]i was similar (⬇115 nM) either in normal or in pre-acidified lymphocytes, independent of the [H⫹]i (data not shown). Because we could not observe any effect of pHi on intracellular Ca2⫹ levels when cells were acidified with respect to normal cells, we added NH4Cl in order to increase the pHi in lymphocytes and try to correlate it with Ca2⫹. Fig. 2a shows that addition of NH4Cl (50 mM) to cells first acidified and then recovered from an acid load induces a rapid rise of pHi from 7.1 to 7.5 (closed circles). In this case the initial pHi is lower because cells never recovered their basal pHi in the absence of bicarbonate (see Fig. 1a). In non-acidified cells, NH4Cl also induces an increase of pHi from 7.27 to 7.5 (open circles). In both cases, cells tend to recover their basal pHi after NH4Cl addition. Fig. 2b illustrates an experiment to check if NH4Cl induces intracellular Ca2⫹ changes after cells equilibrated their [Ca2⫹]i. Either in cells recovered from an acid load (closed circles), or in cells with their normal pHi (open circles), addition of NH4Cl (100 mM) slightly decreased [Ca2⫹]i. The decrease in Ca2⫹ levels activated by NH4Cl could be due to the exit of the cation from the cell. To test this hypothesis, experiments were done in the presence of La⫹3. First, we wanted to see if La⫹3 treatment could affect the rise of pHi caused by NH4Cl in lymphocytes. As shown in Fig. 3a, addition of La⫹3 (triangles) did not modify the NH4Cl-induced pHi increase produced in its absence (closed circles). However, as shown in Fig. 3b, intracellular Ca2⫹ levels are increased from 115 nM to ⬇170 nM after application of La⫹3 (triangles), which is likely to be due to inhibition of the Ca2⫹ efflux mediated
576
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 2. (a) Effect of addition of NH4Cl (50 mM) on pHi of pre-acidified or control lymphocytes loaded with BCECF. Data are means ⫾ S.E.M. of four experiments. (b) Effect of addition of NH4Cl (100 mM) on [Ca2⫹]i in control or pre-acidified Fura-2-loaded lymphocytes. Data are means ⫾ S.E.M. of seven experiments.
by the Ca2⫹-ATPase. Further treatment with NH4Cl caused no change in comparison with the decrease in [Ca2⫹]i that occurs in the absence of La⫹3 (closed circles), suggesting that NH4Cl activated Ca2⫹ exit from the cell. To get more information about the relationship between pHi and Ca2⫹ we used another stimulus, the unspecific K⫹-channel blocker, 4-AP, which is a dihydropyridine. This compound is known to cause not only increase of intracellular Ca2⫹, but also a rise in pHi in Jurkat T-lymphocytes [6]. As Fig. 4a shows, 4-AP (5 mM) induced cellular alkalinization from a basal pHi of 7.25 to ⵑ7.8 in human lymphocytes, either in the presence (closed circles) or the absence (open circles) of extracellular Ca2⫹. In contrast to results with NH4Cl, cells did not recover their pHi after 4-AP treatment. To correlate pHi variations with changes in [Ca2⫹]iwe monitored lymphocytes after incubation with Fura-2 and stimulation with 4-AP. Results show that in the presence of extracellular Ca2⫹ (closed circles), 4-AP induces an increase of Ca2⫹ from 110 nM to 550 nM (Fig. 4b). In a Ca2⫹-free medium (open circles), 4-AP rises intracellular Ca2⫹ from 100 to 160 nM suggesting that this increase is due to the release of Ca2⫹ from intracellular stores.
Fig. 3. Effect of La⫹3 (10 M) on the increase of pHi caused by NH4Cl (50 mM) in BCECF-loaded human lymphocytes. The data are representative of four experiments. (b) Effect of La⫹3 (10 M) on the reduction of [Ca2⫹]i stimulated by NH4Cl (100 mM) in Fura-2-loaded lymphocytes. The data are representative of four experiments.
To analyse the nature of these intracellular Ca2⫹ pools we used thapsigargin, a very potent inhibitor of intracellular Ca2⫹-ATPases. Fig. 5a shows that in the absence of external Ca2⫹, pretreatment with thapsigargin (2 M) 8 min before addition of 4-AP does not preclude the elevation of Ca2⫹ induced by 4-AP. Thapsigargin stimulates depletion of Ca2⫹ from intracellular pools. However further addition of 4-AP elevates [Ca2⫹]i (closed circles), reaching levels similar to those obtained only with 4-AP (open circles). To know if the concentration of thapsigargin is not enough to deplete thapsigargin-sensitive intracellular Ca2⫹ pools, we added a second dose of the inhibitor before addition of 4-AP. A second dose of thapsigargin does not prevent the increase of Ca2⫹ induced by 4-AP (data not shown). These results suggest that 4-AP-sensitive intracellular pools of Ca2⫹ are different from those dependent on thapsigargin. Next we used caffeine, a methylxanthine commonly used to deplete intracellular Ca2⫹ pools [21]. In a Ca2⫹free medium, caffeine (10 mM) induces Ca2⫹ release from intracellular pools (Fig. 5b, closed triangles). Pretreatment with caffeine before addition of 4-AP does
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 4. (a) Effect of addition of 4-AP (5 mM) on pHi in BCECFloaded lymphocytes in the presence or absence of external Ca2⫹. Data are means ⫾ S.E.M. of four experiments. (b) Effect of addition of 4-AP (5 mM) on [Ca2⫹]i in Fura-2-loaded lymphocytes in the presence or absence of external Ca2⫹. The data are representative of six experiments.
not inhibit the increase of Ca2⫹ caused by 4-AP (open circles). On the contrary, the increase of Ca2⫹ is higher than the level reached with only 4-AP suggesting the existence of a calcium-induced calcium release mechanism activated by caffeine in lymphocytes. The role of ionomycin (a Ca2⫹ ionophore) on pHi and Ca2⫹ was also studied. As shown in Fig. 6a, 5 M ionomycin induces a small increase of cytosolic pH either in the presence or absence of extracellular Ca2⫹ (closed vs. open circles), although pHi elevation is lower than in the presence of Ca2⫹. Fig. 6b shows the dual effect of this compound on intracellular Ca2⫹, with external Ca2⫹ influx and release from intracellular pools (closed vs. open circles). Complementary information was obtained by studying the relationship between pH and Ca2⫹ changes with the changes on lymphocyte membrane potential. Fig. 7a shows changes in membrane potential induced by NH4Cl (50 mM) in the presence (closed circles) or absence (open circles) of external Ca2⫹. NH4Cl produced a fast depolarisation in human lymphocytes with a kinetic similar to pH and Ca2⫹ changes (see previous figures).
577
Fig. 5. (a) Effect of thapsigargin (2 M) on the [Ca2⫹]i response induced by 4-AP (5 mM) in Fura-2-loaded lymphocytes in the absence of external Ca2⫹. The data are representative of six experiments. (b) Effect of caffeine (10 mM) on the [Ca2⫹]i signal induced by 4-AP (5 mM) in Fura-2-loaded lymphocytes in the absence of external Ca2⫹. The data are representative of five experiments.
4-AP also induced depolarisation of lymphocytes as illustrated in Fig. 7b, either in the presence (closed circles) or absence (open circles) of external Ca2⫹. However, the time-course of membrane potential change is slower than that induced by NH4Cl. Fig. 7c illustrates the change in membrane potential induced by ionomycin (2M). As for NH4Cl and 4-AP, the ionophore also depolarises the plasma membrane, although only in the presence of external Ca2⫹ (closed circles). No change takes place in a Ca2⫹-free medium (open circles). An original recording is illustrated in Fig. 8, which shows the effect of NH4Cl and 4-AP on the kinetics the membrane potential changes in lymphocytes. As a control, gramicidin was added at the end of each experiment to induce complete membrane depolarisation as shown in the figure. 4. Discussion The functional relationships or crosstalk between calcium and pH receive in general little attention, and almost none regarding human cells. This article reports data which intends to provide information with regard to this crosstalk in human lymphocytes. Binding of a mitogen to its receptor triggers a number of events including elevation of cytosolic Ca2⫹ and intra-
578
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 6. (a) Effect of ionomycin (5 M) on pHi in BCECF-loaded lymphocytes in the presence or absence of external Ca2⫹. Data are means ⫾ S.E.M. of four experiments. (b) Effect of ionomycin (2 M) on [Ca2⫹]i in Fura-2-loaded human lymphocytes in the presence or absence of external Ca2⫹. The data are representative of three experiments.
cellular alkalinization. An alkalinization mediated by the electroneutral Na⫹/H⫹ antiport occurs in a variety of cells following growth factors activation [22]. The increase in cytoplasmic [Ca2⫹] is due to release of Ca2⫹ from intracellular stores and is sustained by influx of external Ca2⫹. In lymphocytes, evidence is accumulating in support of the theory that electrogenic plasma membrane Ca2⫹ influx is mainly regulated by the content of intracellular pools (store-regulated Ca2⫹ uptake, SRCU) [23,24]. A Na⫹/H⫹ exchanger was already described in the plasma membrane of thymocytes and B lymphocytes [25,26]. By using different procedures our results indicate that freshly isolated human lymphocytes have a Na⫹/H⫹ exchanger involved in the recovery of pHi after acid-loading. According to those authors, this exchanger is inhibited by amiloride and absence of external Na⫹. The concentration of amiloride necessary to inhibit the recovery of pHi is 100 M, consistent with the existence of the housekeeping isoform NHE1 in human lymphocytes [27]. However, lymphocytes have a bicarbonatedependent mechanism because in bicarbonate-free medium cells never recovered their basal pHi [28]. Several studies have been carried out to correlate cytosolic pH and Ca2⫹ in different cell types. However, interrelationships between pH and Ca2⫹ are rather com-
Fig. 7. (a) Changes in membrane potential induced by NH4Cl (50 mM) in lymphocytes with or without external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in methods. Data are representative of four experiments. (b) Changes in membrane potential induced by 4-AP (5 mM) in lymphocytes with or without external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in Methods. Data are representative of four experiments. (c) Changes in membrane potential induced by ionomycin (2 M) in human lymphocytes in the presence or absence of external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in Methods. Data are representative of four experiments.
plex and depend on the cell model [2]. We could not observe a clear relationship between the steady state level of pHi and intracellular Ca2⫹ in human lymphocytes, which agrees with [12] and is the contrary to [15]. For this reason we used different agents known to induce pHi and intracellular Ca2⫹ changes in different cells. 4.1. Effect of NH4Cl on Ca2⫹ and pHi Alkalinization of vascular smooth muscle cells by exposure to the weak base NH4Cl resulted in a rise of intracellular Ca2⫹ [14,15]. Also, in glomerular epithelial cells and bovine lactotrophs, NH4Cl loading produced alkalinization and a concurrent rise in Ca2⫹. However, in this study when NH4Cl was removed and the cells became acidic, a second rise in Ca2⫹ was recorded, and
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 8. Effect of NH4Cl (50 mM) and 4-AP (5 mM) on membrane potential of human lymphocytes in the presence or absence of external Ca2⫹. The figure is an original recording representative of four experiments. Gramicidin was added at the end of each experiment and at different times (two arrows) for both agents.
both changes were dependent on intracellular stores [12,13]. In contrast, Ca2⫹ is not mobilized in response to either intracellular alkalinization or acidification in Syrian hamster embryo cells [29]. In rat thymic lymphocytes, increasing pHi with monensin or NH3 did increase [Ca2⫹]i [30]. In Jurkat T-lymphocytes, NH4Cl induced increase in cytosolic pH but no detectable Ca2⫹ signal [6]. Thus, the effect of NH4Cl on Ca2⫹ signal has been contradictory in some cases. In our experiments with human lymphocytes using NH4Cl, we were able to record a rapid alkalinization and a small decrease of intracellular Ca2⫹. Besides different cell models, rat thymic lymphocytes or Jurkat vs. human lymphocytes, the effect found with NH4Cl on the Ca2⫹ level is probably due to the higher concentration used by us in these experiments, since 50 mM NH4Cl does not induce a Ca2⫹ response (data not shown). It was necessary to use 100 mM NH4Cl to observe a decrease of intracellular Ca2⫹. Using a different procedure, a report in collecting duct cells shows that cell alkalinization also produced a lowering of Ca2⫹ [31]. One explanation for our results is that NH4Cl could activate sequestration mechanisms for Ca2⫹. However, La3⫹ obliterates the fall in [Ca2⫹]i indicating that the decrease of Ca2⫹ is due to exit from the cell. Further studies are needed to explain how NH4Cl could stimulate Ca2⫹ efflux from the cell. 4.2. Effect of 4-AP on Ca2⫹ and pHi 4-AP, a K⫹ channel blocker, was used as a tool to increase pHi in Jurkat T cells and to induce a Ca2⫹ signal similar to that observed in response to physiological stimulation [6]. When we used 4-AP in human lymphocytes, an increase of pHi and Ca2⫹ was observed. 4-APinduced alkalinization preceded the Ca2⫹ signal providing further evidence that alkalinization is the trigger for the Ca2⫹ response, as it was reported in Jurkat T cells [6]. Raising pHi in human lymphocytes with 4-AP leads to a rise in intracellular Ca2⫹, even in the absence of external Ca2⫹, indicating the presence of pH-sensitive
579
intracellular Ca2⫹ stores emptied by 4-AP. However, contrary to our results, those authors found that in Ca2⫹free medium previous addition of thapsigargin abolished any effect of 4-AP, suggesting that 4-AP induced Ca2⫹ release from the agonist/IP3-sensitive pool of Jurkat T cells. In our experiments 4-AP increased the level of Ca2⫹ released by either caffeine or thapsigargin suggesting that 4-AP is mobilizing a different pool or the same pool with different kinetics. In fact, the existence of four different pools of Ca2⫹ was shown in Jurkat T-lymphocytes [32]. IP3 and Ca2⫹-induced Ca2⫹ release (CICR) mechanisms are closely related and coexpressed in muscle and nerve cells, both being sensitive to pH [11]. Caffeine, an agonist of the ryanodine receptor CICR, might be involved in the propagation of Ca2⫹ signal in human lymphocytes as it was suggested in other non-excitable cells [32,33]. The use of Jurkat vs. human lymphocytes (culture cells vs. fresh isolated lymphocytes) and the distinct Ca2⫹ pools sensitivity, could explain the different effect found with 4-AP in the two cell models. One alternative to explain the effect of 4-AP on Ca2⫹ signal is the sensitisation of the IP3 R [6], although the alkalinizing effect of NH4Cl does not induce the same Ca2⫹ response. The basis for the opposite findings is unclear and awaits further studies. 4.3. Effect of ionomycin on Ca2⫹ and pHi Ca2⫹ ionophores, such as ionomycin, are used to determine the role of Ca2⫹ in cellular regulation. However, studies concerning the effect of ionomycin on pHi are contradictory. In some cells the dual effect of ionomycin on pHi and Ca2⫹ has been mentioned [34,35]. In collecting duct cells, ionomycin induced a biphasic change in pH; a transient acidification followed by alkalinization. This cell alkalinization was mediated by depolarisation [31]. In human B lymphocytes ionomycin induced increase of Ca2⫹ as well as a transient acidification [36]. However, in some cells ionomycin induces influx of Ca2⫹ but not alteration in cytosolic pH [13]. In human lymphocytes, ionomycin increases pHi and Ca2⫹ levels, similar to that found in hepatocytes [34]. Because ionomycin is a Ca2⫹/H⫹ exchanger the increase of pHi could be a consequence of Ca2⫹/H⫹ activation as suggested in that report. However, we should preclude an effect of ionomycin on the Ca2⫹/H⫹ exchanger since this is an electroneutral exchanger and our results show depolarisation of plasma membrane potential in lymphocytes activated with ionomycin (see next paragraph). In human lymphocytes increase of Ca2⫹ takes place either in the presence or absence of external Ca2⫹, although in the absence of external Ca2⫹ the increase of pHi is almost undetectable. The magnitude of alkalinization is proportional to the increase of Ca2⫹ levels. Although thapsigargin, which also increases [Ca2⫹]i,
580
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
failed to increase pHi in human lymphocytes (data not shown) as occurs in hepatocytes [34]. 4.4. Effects on membrane potential Concerning membrane potential changes the situation is even more complex and several reports have tried to establish a correlation with Ca2⫹ and pHi. Intracellular acidification with organic acids, such as lactate or acetate, caused an increase in intracellular Ca2⫹ and depolarisation in an insulinoma cell line. Alkalinization caused no apparent change in Ca2⫹. However, NH4Cl application and withdrawal were both accompanied by transient increases in Ca2⫹ and no depolarisation occurred [10]. In our study, all three stimuli induced depolarisation of human lymphocytes. The change in membrane potential induced by NH4Cl and 4-AP does not depend on the presence of external Ca2⫹ despite the opposite effect caused by these agents on [Ca2⫹]i. However, the distinct kinetic patterns induced by both agents on membrane potential should be pointed out. Depolarisation activated with NH4Cl takes place very fast, in less than one minute, in a similar fashion to the NH4Cl-induced alkalinization and reduction of intracellular Ca2⫹ levels. Changes produced by NH4Cl on membrane potential, [Ca2⫹]i and pHi are synchronised in lymphocytes. Thus, alkalinization is related to reduction of Ca2⫹ levels and membrane depolarisation. Kinetically, 4-AP-induced Ca2⫹ change occurs more slowly than plasma membrane depolarisation and alkalinization is faster. Then, as described by other studies, an alkalinization due to 4-AP induced an elevation of [Ca2⫹]i [6] and depolarisation in human lymphocytes. In rat thymocytes and in human lymphocytes, Grinstein et al. found that ionomycin induced membrane potential hyperpolarization [37,38]. However, contradictory results have been found even in different populations of the same cell type. Ionomycin hyperpolarises mouse T lymphocytes while depolarising B lymphocytes [39]. Our results show that a high concentration of ionomycin induces membrane depolarisation although this effect does not occur in Ca2⫹-free medium. In our view this is due to a high influx of Ca2⫹ stimulated by a higher concentration of ionophore (unpublished results). The elevated intracellular Ca2⫹ level attained by the ionophore could by itself cause the depolarisation. Additionally, the electrogenic uptake described by Grinstein and Smith [38] is expected to decline dramatically with increasing concentrations of ionomycin. In fact, it has been shown that ionophores are able to transport both electroneutrally and electrogenically [40]. In this study we demonstrate that in human lymphocytes, raising pHi with 4-AP leads to a rise in [Ca2⫹]i, even in the absence of external Ca2⫹. On the contrary, alkalinization with NH4Cl activates the efflux of Ca2⫹ from the cell. The mechanism by which alkalinization
with 4-AP induces a rise and NH4Cl a decrease in intracellular Ca2⫹ is unclear. It is not known if NH4Cl and 4-AP affect Ca2⫹ directly or by some other mechanisms. We found a lack of correlation between pHi and Ca2⫹ since alkalinization with three alkalinising agents does not lead to the same change in Ca2⫹. In summary, intracellular alkalinization can, under certain circumstances, cause a rise or a decrease in [Ca2⫹]i. The three alkalinising agents caused plasma membrane depolarisation in human lymphocytes, independent of external Ca2⫹ for NH4Cl and 4-AP, but completely dependent on the cation for the ionophore. The present report demonstrates that the effect of these agents in human lymphocytes is not limited to the change in [Ca2⫹]i. They induce substantial alterations in the overall cellular ionic homeostasis, including changes in the pHi and in the membrane potential.
Acknowledgments This work was funded with grants FEDER-CICYT1FD97-0153, Xunta Galicia PGIDT99INN26101, and FISS 96/0378. This work could not have been done without the invaluable collaboration of the staff at the General Hospital of Lugo from the Centro de Transfusio´n de Galicia, (Dr. Marcelo Alvarenga and Dr. Jose´ A. Gonza´lez, Ms. Sara Go´mez, Ana Pe´rez, Dolores Sa´nchez, Mr. J.M. Rodrı´guez, and Ms. Dolores Atanes and Ms. Noelia Lauda).
References [1] Putney JW. Cell Calcium 1990;11:611–24. [2] Alfonso A, Cabado AG, Vieytes MR, Botana LM. Br J Pharmacol 2000;130:1809–16. [3] Helmle-Kolb C, Counillon L, Roux D, Pouyssegur J, Mrkic B, Murer H. Pflugers Arch 1993;425:34–40. [4] van Haelst-Pisani C, Cragoe EJ, Rothstein TL. Exp Cell Research 1989;183:251–6. [5] Gore J, Besson P, Hoinard C, Bougnoux P. Am J Physiol 1994; 266:C110–20. [6] Guse AH, Roth E, Emmrich F. Biochem J 1994;301:83–8. [7] Oettgen HC, Terhorst C, Cantley LC, Rosoff PM. Cell 1985;40: 583–90. [8] Rosoff PM, Terres G. J Cell Biol 1986;103:457–63. [9] Tsien RY, Pozzan T, Rink TJ. Nature 1982;295:68–71. [10] Best L, Ammala C, Rorsman P, Tomlinson S. Biochimica et Biophysica Acta 1994;1192:107–11. [11] Tsukioka M, Iino M, Endo M. J Physiol (Lond) 1994;475:369–75. [12] Zorec R, Hoyland J, Mason WT. Eur J Physiol 1993;423:41–50. [13] Wiegmann TB, Welling LW, Beatty DM, Howard DE, Vamos S, Morris SJ. Am J Physiol 1993;34:C1184–90. [14] Siskind MS, McCoy CE, Chobanian A, Schwartz JH. Am J Physiol 1989;256:C234–40. [15] Putnam RW, Douglas PB. Mol Cell Biochem 1990;99:89–95. [16] Cabado AG, Vieyres MR, Botana LM. J Cell Physiol 1994;158: 309–16. [17] Grynkiewicz G, Poenie M, Tsien RY. J Biol Chem 1985;260: 3440–50.
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581 [18] Cabado AG, Yu FH, Kapus A, Lukacs G, Grinstein S, Orlowski J. J Biol Chem 1996;271:3590–9. [19] Thomas JA, Buchsbaum RN, Zimniak A, Racker E. Biochemistry 1979;18:2210–8. [20] Bright GR, Fisher GW, Rogowska J, Taylor DL. J Cell Biol 1987;104:1019–33. [21] Ehrlich BE, Kaftan E, Bezprozvannaya S, Bezprozvanny I. TIPS 1994;15:145–9. [22] Sardet C, Franchi A, Pouyssegur J. Cell 1989;56:271–80. [23] Mason MJ, Mahaut SMP, Grinstein S. J Biol Chem 1991;266: 10872–9. [24] Mason M, Grinstein S. Biochem J 1993;296:33–9. [25] Grinstein S, Cohen S, Goetz JD, Rothstein A, Gelfand EW. Proc Natl Acad Sci USA 1985;82:1429–33. [26] Gaidano G, Chigo D, Schena M, Bergui L, Treves S, Turrini F, et al. J Immunol 1989;142:913–8. [27] Siczkowski M, Davies JE, Ng LL. Am J Physiol Cell Physiol 1994;267:C745–52. [28] De la Rosa LA, Cabado AG, Botana MA, Vieytes MR, Vidal JI, Botana LM. Pflugers Arch 1999;437:935–43.
581
[29] Isfort R, Cody DB, Asquith N, Ridder GM, Stuard SB, Leboeuf A. Eur J Biochem 1993;213:349–57. [30] Grinstein S, Goetz JD. Biochim Biophys Acta 1985;819:267–70. [31] Slotki I, Schwartz JH, Alexander EA. Am J Physiol 1993;265: C432–8. [32] Guse AH, Roth E, Emmrich F. Biochem J 1993;291:447–51. [33] Bennett DL, Cheek TR, Berridge MJ, Smedt HD, Parys JB, Missiaen L, et al. J Biol Chem 1996;271:6356–62. [34] Anwer MS. Hepatology 1993;18:433–9. [35] Hendey B, Mamrack MD, Putnam RW. J Biol Chem 1989;264: 19540–7. [36] Schoyen H, Iversen JG, Smeland EB, Heikkila R. Acta Physiol Scand 1990;138:221–7. [37] Dixon SJ, Stewart D, Grinstein S, Spiegel S. J Cell Biol 1987; 105:1153–61. [38] Grinstein S, Smith JD. J Gen Physiol 1990;95:97–120. [39] Ishuda Y, Chused TM. J Exp Med 1988;168:839–952. [40] Fasolato C, Pozzan T. J Biol Chem 1989;264:19630–6.
Crosstalk between cytosolic pH and intracellular calcium in human lymphocytes Effect of 4-aminopyridin, ammoniun chloride and ionomycin A.G. Cabadoa, A. Alfonsob, M.R. Vieytesa, M. Gonza´lezc, M.A. Botanad, L.M. Botanab,* Departamento de Fisiologı´aa y Farmacologı´ab 1, Facultad de Veterinaria, Servicio de Hematologı´ac, Hospital Xeral, Seccio´n de Endocrinologı´ad, E-27002 Lugo, Universidad Santiago de Compostela, Spain Received 26 March 2000; accepted 22 May 2000
Abstract Stimulation of lymphocytes by specific antigens is followed by the activation of different signal transduction mechanisms, such as alterations in the cytoplasmic levels of Ca2⫹, H⫹ and variations in membrane potential. To study interrelationships among these parameters, changes in pHi and Ca2⫹ were measured with the fluorescent probes BCECF and Fura-2 in freshly isolated blood human lymphocytes. Moreover, membrane potential qualitative alterations were recorded with the fluorescent dye bisoxonol. In a bicarbonate-free medium, cell alkalinization with NH4Cl slightly decreased intracellular Ca2⫹ concentration ([Ca2⫹]i) due to efflux of Ca2⫹ from the cell. In contrast, an elevation of pHi induced with 4-AP increased [Ca2⫹]i, either in the presence or absence of external Ca2⫹. The increase in Ca2⫹-free medium is likely to be due to Ca2⫹ release from thapsigargin and caffeineindependent intracellular stores. Both 4-AP or NH4Cl induced a plasma membrane depolarisation, although with different kinetics. The ionosphere ionomycin increased pHi, Ca2⫹ levels and also induced membrane depolarisation. Together, these observations demonstrate a lack of correlation between the magnitude of changes in pHi and Ca2⫹. 2000 Elsevier Science Inc. All rights reserved. Keywords: Lymphocytes; pHi; Na⫹/H⫹ exchanger; Ca2⫹; 4-AP; NH4Cl; Thapsigargin; Caffeine; Ionomycin
1. Introduction It is a well established paradigm of cell activation that calcium is necessary at the initial steps in almost any cellular model. In lymphocytes, whose activation and proliferation forms the basis of the mechanism of immune defence, binding of antigen to surface receptors leads to intracellular generation of IP3, which releases Ca2⫹ from intracellular stores, thus raising Ca2⫹ transiently. Influx of Ca2⫹ from the external compartment is, however, necessary to maintain the plateau of such response, a prerequisite for the final mitogenic effect. In non-excitable cells such as lymphocytes, the capacitative sustained Ca2⫹ signal depends on plasma membrane Ca2⫹ channels activated by depletion of Ca2⫹ intracellular stores [1]. On the other hand, protons may function as a second messenger in a manner similar to Ca2⫹ [2]. In many * Corresponding author. Tel.: 34-982-252-242; fax: 34-982-252-242. E-mail address: ffbotana @lugo.usc.es (L.M. Botana)
cells, Na⫹/H⫹ exchanger is the most important among the mechanisms involved in pHi homeostasis and some studies have questioned the physiological relevance of Na⫹/H⫹ exchanger activation in the induction of cell proliferation [3–5]. At least for some cellular types, there is a correlation between pHi and intracellular Ca2⫹, since alkalinization, before cytosolic calcium increase, is a sufficient signal to trigger cell activation [2]. In some cell types it was proposed that alkalinization induced by mitogens may contribute to the length and magnitude of the Ca2⫹ signal. Although some studies found no pH response in Jurkat T cells stimulated with the mitogen OKT3 [6], in resting T lymphocytes, stimulation with OKT3 induces pHi rise via Na⫹/H⫹ exchanger activation. This mechanism is dependent on external Ca2⫹ and can be blocked by La3⫹. Also, OKT3 induces a rapid membrane potential-sensitive Ca2⫹ influx in T lymphocyte cell lines [7,8]. In addition, lectins that stimulate T cells raise intracellular Ca2⫹ accompanied by hyperpolarization of the membrane potential, apparently due to a Ca2⫹-acti-
0898-6568/00/$ – see front matter 2000 Elsevier Science Inc. All rights reserved. PII: S0898-6568(00)00101-7
574
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
vated K⫹ conductance [9]. Therefore, alterations in membrane potential are also important in cell signalling and they can be correlated with Ca2⫹ and protons changes. Mechanisms through which changes in pHi could affect Ca2⫹ include not only effects of plasma membrane potential and hence gating of voltage-sensitive Ca2⫹ channels, but also mobilisation of Ca2⫹ from intracellular stores [10]. Several possible interactions between pHi and Ca2⫹ have been described: sharing of common intracellular buffer sites by H⫹ and Ca2⫹; stimulation of metabolic acid production by increases of Ca2⫹; modulation of the membrane transports responsible for the regulation of pHi and Ca2⫹ by changes of Ca2⫹ and pHi, respectively; sensitisation of the IP3 receptor by increasing pHi and consequently stimulation of Ca2⫹ release from internal stores by changes of pHi [11]. Different reports in muscle, fibroblasts and nerve cells have also tried to relate Ca2⫹ responses to pHi changes. Depending on the cell type, Ca2⫹ increased, fell or did not change [12–15]. The link between protons and Ca2⫹ ions is complex because it is dependent on the cell model and the procedure used to induce the cellular change. To evaluate the relationship between intracellular Ca2⫹ and protons in human lymphocytes we used the fluorescent dyes Fura-2 and BCECF to examine changes in intracellular Ca2⫹ and cytosolic pH imposed by the weak base NH4Cl, 4-AP (a K⫹ channel blocker), and ionomycin (a calcium ionophore). 2. Methods 2.1. Materials Bis-(1,3-diethylthiobarbirturic acid) trimethine oxonol (bis-oxonol), 2⬘,7⬘-bis(carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) and Fura-2/AM were purchased from Molecular Probes (Eugene, OR). Ionomycin, gramicidin, 4-aminopyridine (4-AP), caffeine, amiloride and thapsigargin were purchased from Sigma Chemical (St. Louis, MO). Ficoll was from Pharmacia (Sweden). All other chemicals were from standard commercial sources and reagent grade or the highest purity. Physiological saline solution (Umbreit) was composed of (in mM): Na⫹, 142.3; K⫹, 5.94; Ca2⫹, 1; Mg2⫹, 1.2; Cl⫺, 126.1; CO3⫺, 22.85; PO4H2, 1.2 and SO24⫺,1.2 giving a final osmotic pressure of 300 ⫾ 5 mOsm/Kg H2O. BSA and glucose (1 mg/ml) were added to the solution; the pH was adjusted to 7.4. Na⫹-containing solution consisted of (in mM): KCl, 14; MgCl2, 1; CaCl2, 1; NaCl, 140; Hepes, 10 and glucose, 10. To make Na⫹free medium, NaCl was substituted by N-methyl-D-glucamine. K⫹ solution for pH calibration was composed of (in mM): KCl, 141; MgCl, 1; Ca Cl2, 1; Hepes, 20 and glucose, 10.
2.2. Lymphocyte isolation and purification Peripheral blood lymphocytes were isolated from freshly drawn blood from healthy donors. Purification of cells was carried out by means of Ficoll-Hypaque centrifugation. In order to remove platelets, blood was diluted and washed (3⫻) with Umbreit solution. Washed cells were placed on Ficoll and centrifugation (1000 g max) was carried out at room temperature for 30 min. After centrifugation lymphocytes, appearing in the interface, were washed (2⫻) and resuspended in Umbreit. Cells were counted using a model System 9000 Coulter Counter (Menarini Diagno´stico, S.A.). The isolated lymphocyte populations were maintained at 4⬚C in Umbreit solution containing heparin and buffered to pH 7.4. 2.3. Determination of changes in membrane potential Plasma membrane potential changes were recorded with the fluorescent probe negatively charged bis-oxonol. 0.5 ⫻ 106/ml lymphocytes were used for each experiment. Dye loading and determination of changes in the membrane potential were performed as previously reported [16]. 2.4. Determination of [Ca2⫹]i Purified lymphocytes were incubated in Na⫹-containing solution with 2.5 M fura-2/AM for 20 min at 37⬚C. After dye loading, cells were then washed twice and resuspended at 2.5 ⫻ 106/ml in a quartz cuvette with continuous stirring. Experiments were performed at 37⬚C. Calibration and calculation of [Ca2⫹]i was performed as described [17]. 2.5. Cell labelling and pHi measurement Lymphocytes (1 ⫻ 106) were loaded with 1 M BCECF and incubated for 20 min at 37⬚C. Cells were pulsed with 50 mM NH4Cl during the last 15 min of the incubation with the BCECF as described [18]. After incubation, cells were washed twice and placed in a thermostated quartz cuvette with continuous stirring. The calibration of fluorescence vs. pHi was made using nigericin in K⫹ solution as per [19]. Briefly, a calibration curve was obtained with four known values of pH, measuring the fluorescence ratio obtained in the presence of nigericin for each pH value. 2.6. Fluorescence measurements The fluorescence of fura-2, BCECF and bis-oxonol was measured with a Shimadzu RF-5000 spectrofluorometer. Bis-oxonol fluorescence was recorded at excitation and emission wavelengths of 540 nm and 560 nm, respectively, as described [16]. For fura-2-loaded cells, the ratio of fluorescence was determined by switching two excitation wavelengths, 340 and 380 nm (bandwidth 5 nm) and recording the emission at 500 nm (bandwidth
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 1. (a) Recovery of pHi in acid-loaded human lymphocytes vs. basal pHi. Lymphocytes were incubated with BCECF as described in Methods. The data are representative of five experiments. (b) Recovery of pHi in acid-loaded lymphocytes incubated with BCECF as described in Methods. Effect of 100 M amiloride and absence of external Na⫹. Data are means ⫾ S.E.M. of four experiments.
10 nm). BCECF fluorescence was recorded by alternating the excitation at 490 and 440 nm and measuring emitted fluorescence at 530 nm. The ratio of the two excitation wavelengths allowed the calculation of pHi without interference from differences in dye loading, dye bleaching, or cell number [20]. Drugs were added from concentrated solutions with a pipette directly into the cuvette through a small hole on top of the cuvette lid. 3. Results Fig. 1a shows basal pHi in human lymphocytes incubated 20 min with BCECF (open circles). After incubation with the dye cells were washed (2⫻) and resuspended in a bicarbonate-free medium containing 140 mM external Na⫹. When the cells were added to the cuvette the temperature was 4⬚C and in 2 min the temperature reached 37⬚C. The slight decrease of pHi is due to the increase of temperature. Resting pHi in human lymphocytes decreased from 7.347 ⫾ 0.027 at 4⬚C to 7.275 ⫾ 0.03 (n ⫽ 17) at 37⬚C when extracellular pH is 7.4. Fig. 1a also shows recovery of pHi in lymphocytes
575
after an acid load imposed by 50 mM NH4Cl (closed circles). Cells acidified to pHi 6.79 followed by a slower alkalinization (⬇10 min) toward the resting pHi (7.06). However, cells never recovered their basal pHi of 7.27, indicating the existence of a bicarbonate-dependent component of the recovery (unpublished observation). To know if part of the mechanism involved in the recovery of pHi after an acid load is a Na⫹/H⫹ exchanger we carried out experiments in the absence of external Na⫹ and in the presence of 100 M amiloride, an inhibitor of this transporter. Fig. 1b shows recovery of pHi in lymphocytes after an acid load imposed by 50 mM NH4Cl in a medium containing 140 mM external Na⫹ (closed circles), in Na⫹-free medium (triangles) and in the presence of 100 M amiloride (open circles). Cells recovered from the acid load only in the presence of external Na⫹ (closed circles), strongly suggesting that the alkalinization is due to Na⫹/H⫹ exchanger since pHi recovery is inhibited in the absence of external Na⫹ and in the presence of amiloride. To examine if basal pHi influences [Ca2⫹]i we carried out experiments monitoring lymphocytes after incubation with Fura-2. Cells were acidified with NH4Cl and then allowed to recover their pHi, or introduced into the cuvette at their resting pHi. After 10 min the [Ca2⫹]i was similar (⬇115 nM) either in normal or in pre-acidified lymphocytes, independent of the [H⫹]i (data not shown). Because we could not observe any effect of pHi on intracellular Ca2⫹ levels when cells were acidified with respect to normal cells, we added NH4Cl in order to increase the pHi in lymphocytes and try to correlate it with Ca2⫹. Fig. 2a shows that addition of NH4Cl (50 mM) to cells first acidified and then recovered from an acid load induces a rapid rise of pHi from 7.1 to 7.5 (closed circles). In this case the initial pHi is lower because cells never recovered their basal pHi in the absence of bicarbonate (see Fig. 1a). In non-acidified cells, NH4Cl also induces an increase of pHi from 7.27 to 7.5 (open circles). In both cases, cells tend to recover their basal pHi after NH4Cl addition. Fig. 2b illustrates an experiment to check if NH4Cl induces intracellular Ca2⫹ changes after cells equilibrated their [Ca2⫹]i. Either in cells recovered from an acid load (closed circles), or in cells with their normal pHi (open circles), addition of NH4Cl (100 mM) slightly decreased [Ca2⫹]i. The decrease in Ca2⫹ levels activated by NH4Cl could be due to the exit of the cation from the cell. To test this hypothesis, experiments were done in the presence of La⫹3. First, we wanted to see if La⫹3 treatment could affect the rise of pHi caused by NH4Cl in lymphocytes. As shown in Fig. 3a, addition of La⫹3 (triangles) did not modify the NH4Cl-induced pHi increase produced in its absence (closed circles). However, as shown in Fig. 3b, intracellular Ca2⫹ levels are increased from 115 nM to ⬇170 nM after application of La⫹3 (triangles), which is likely to be due to inhibition of the Ca2⫹ efflux mediated
576
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 2. (a) Effect of addition of NH4Cl (50 mM) on pHi of pre-acidified or control lymphocytes loaded with BCECF. Data are means ⫾ S.E.M. of four experiments. (b) Effect of addition of NH4Cl (100 mM) on [Ca2⫹]i in control or pre-acidified Fura-2-loaded lymphocytes. Data are means ⫾ S.E.M. of seven experiments.
by the Ca2⫹-ATPase. Further treatment with NH4Cl caused no change in comparison with the decrease in [Ca2⫹]i that occurs in the absence of La⫹3 (closed circles), suggesting that NH4Cl activated Ca2⫹ exit from the cell. To get more information about the relationship between pHi and Ca2⫹ we used another stimulus, the unspecific K⫹-channel blocker, 4-AP, which is a dihydropyridine. This compound is known to cause not only increase of intracellular Ca2⫹, but also a rise in pHi in Jurkat T-lymphocytes [6]. As Fig. 4a shows, 4-AP (5 mM) induced cellular alkalinization from a basal pHi of 7.25 to ⵑ7.8 in human lymphocytes, either in the presence (closed circles) or the absence (open circles) of extracellular Ca2⫹. In contrast to results with NH4Cl, cells did not recover their pHi after 4-AP treatment. To correlate pHi variations with changes in [Ca2⫹]iwe monitored lymphocytes after incubation with Fura-2 and stimulation with 4-AP. Results show that in the presence of extracellular Ca2⫹ (closed circles), 4-AP induces an increase of Ca2⫹ from 110 nM to 550 nM (Fig. 4b). In a Ca2⫹-free medium (open circles), 4-AP rises intracellular Ca2⫹ from 100 to 160 nM suggesting that this increase is due to the release of Ca2⫹ from intracellular stores.
Fig. 3. Effect of La⫹3 (10 M) on the increase of pHi caused by NH4Cl (50 mM) in BCECF-loaded human lymphocytes. The data are representative of four experiments. (b) Effect of La⫹3 (10 M) on the reduction of [Ca2⫹]i stimulated by NH4Cl (100 mM) in Fura-2-loaded lymphocytes. The data are representative of four experiments.
To analyse the nature of these intracellular Ca2⫹ pools we used thapsigargin, a very potent inhibitor of intracellular Ca2⫹-ATPases. Fig. 5a shows that in the absence of external Ca2⫹, pretreatment with thapsigargin (2 M) 8 min before addition of 4-AP does not preclude the elevation of Ca2⫹ induced by 4-AP. Thapsigargin stimulates depletion of Ca2⫹ from intracellular pools. However further addition of 4-AP elevates [Ca2⫹]i (closed circles), reaching levels similar to those obtained only with 4-AP (open circles). To know if the concentration of thapsigargin is not enough to deplete thapsigargin-sensitive intracellular Ca2⫹ pools, we added a second dose of the inhibitor before addition of 4-AP. A second dose of thapsigargin does not prevent the increase of Ca2⫹ induced by 4-AP (data not shown). These results suggest that 4-AP-sensitive intracellular pools of Ca2⫹ are different from those dependent on thapsigargin. Next we used caffeine, a methylxanthine commonly used to deplete intracellular Ca2⫹ pools [21]. In a Ca2⫹free medium, caffeine (10 mM) induces Ca2⫹ release from intracellular pools (Fig. 5b, closed triangles). Pretreatment with caffeine before addition of 4-AP does
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 4. (a) Effect of addition of 4-AP (5 mM) on pHi in BCECFloaded lymphocytes in the presence or absence of external Ca2⫹. Data are means ⫾ S.E.M. of four experiments. (b) Effect of addition of 4-AP (5 mM) on [Ca2⫹]i in Fura-2-loaded lymphocytes in the presence or absence of external Ca2⫹. The data are representative of six experiments.
not inhibit the increase of Ca2⫹ caused by 4-AP (open circles). On the contrary, the increase of Ca2⫹ is higher than the level reached with only 4-AP suggesting the existence of a calcium-induced calcium release mechanism activated by caffeine in lymphocytes. The role of ionomycin (a Ca2⫹ ionophore) on pHi and Ca2⫹ was also studied. As shown in Fig. 6a, 5 M ionomycin induces a small increase of cytosolic pH either in the presence or absence of extracellular Ca2⫹ (closed vs. open circles), although pHi elevation is lower than in the presence of Ca2⫹. Fig. 6b shows the dual effect of this compound on intracellular Ca2⫹, with external Ca2⫹ influx and release from intracellular pools (closed vs. open circles). Complementary information was obtained by studying the relationship between pH and Ca2⫹ changes with the changes on lymphocyte membrane potential. Fig. 7a shows changes in membrane potential induced by NH4Cl (50 mM) in the presence (closed circles) or absence (open circles) of external Ca2⫹. NH4Cl produced a fast depolarisation in human lymphocytes with a kinetic similar to pH and Ca2⫹ changes (see previous figures).
577
Fig. 5. (a) Effect of thapsigargin (2 M) on the [Ca2⫹]i response induced by 4-AP (5 mM) in Fura-2-loaded lymphocytes in the absence of external Ca2⫹. The data are representative of six experiments. (b) Effect of caffeine (10 mM) on the [Ca2⫹]i signal induced by 4-AP (5 mM) in Fura-2-loaded lymphocytes in the absence of external Ca2⫹. The data are representative of five experiments.
4-AP also induced depolarisation of lymphocytes as illustrated in Fig. 7b, either in the presence (closed circles) or absence (open circles) of external Ca2⫹. However, the time-course of membrane potential change is slower than that induced by NH4Cl. Fig. 7c illustrates the change in membrane potential induced by ionomycin (2M). As for NH4Cl and 4-AP, the ionophore also depolarises the plasma membrane, although only in the presence of external Ca2⫹ (closed circles). No change takes place in a Ca2⫹-free medium (open circles). An original recording is illustrated in Fig. 8, which shows the effect of NH4Cl and 4-AP on the kinetics the membrane potential changes in lymphocytes. As a control, gramicidin was added at the end of each experiment to induce complete membrane depolarisation as shown in the figure. 4. Discussion The functional relationships or crosstalk between calcium and pH receive in general little attention, and almost none regarding human cells. This article reports data which intends to provide information with regard to this crosstalk in human lymphocytes. Binding of a mitogen to its receptor triggers a number of events including elevation of cytosolic Ca2⫹ and intra-
578
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 6. (a) Effect of ionomycin (5 M) on pHi in BCECF-loaded lymphocytes in the presence or absence of external Ca2⫹. Data are means ⫾ S.E.M. of four experiments. (b) Effect of ionomycin (2 M) on [Ca2⫹]i in Fura-2-loaded human lymphocytes in the presence or absence of external Ca2⫹. The data are representative of three experiments.
cellular alkalinization. An alkalinization mediated by the electroneutral Na⫹/H⫹ antiport occurs in a variety of cells following growth factors activation [22]. The increase in cytoplasmic [Ca2⫹] is due to release of Ca2⫹ from intracellular stores and is sustained by influx of external Ca2⫹. In lymphocytes, evidence is accumulating in support of the theory that electrogenic plasma membrane Ca2⫹ influx is mainly regulated by the content of intracellular pools (store-regulated Ca2⫹ uptake, SRCU) [23,24]. A Na⫹/H⫹ exchanger was already described in the plasma membrane of thymocytes and B lymphocytes [25,26]. By using different procedures our results indicate that freshly isolated human lymphocytes have a Na⫹/H⫹ exchanger involved in the recovery of pHi after acid-loading. According to those authors, this exchanger is inhibited by amiloride and absence of external Na⫹. The concentration of amiloride necessary to inhibit the recovery of pHi is 100 M, consistent with the existence of the housekeeping isoform NHE1 in human lymphocytes [27]. However, lymphocytes have a bicarbonatedependent mechanism because in bicarbonate-free medium cells never recovered their basal pHi [28]. Several studies have been carried out to correlate cytosolic pH and Ca2⫹ in different cell types. However, interrelationships between pH and Ca2⫹ are rather com-
Fig. 7. (a) Changes in membrane potential induced by NH4Cl (50 mM) in lymphocytes with or without external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in methods. Data are representative of four experiments. (b) Changes in membrane potential induced by 4-AP (5 mM) in lymphocytes with or without external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in Methods. Data are representative of four experiments. (c) Changes in membrane potential induced by ionomycin (2 M) in human lymphocytes in the presence or absence of external Ca2⫹. Recordings were done in the presence of Bis-oxonol as described in Methods. Data are representative of four experiments.
plex and depend on the cell model [2]. We could not observe a clear relationship between the steady state level of pHi and intracellular Ca2⫹ in human lymphocytes, which agrees with [12] and is the contrary to [15]. For this reason we used different agents known to induce pHi and intracellular Ca2⫹ changes in different cells. 4.1. Effect of NH4Cl on Ca2⫹ and pHi Alkalinization of vascular smooth muscle cells by exposure to the weak base NH4Cl resulted in a rise of intracellular Ca2⫹ [14,15]. Also, in glomerular epithelial cells and bovine lactotrophs, NH4Cl loading produced alkalinization and a concurrent rise in Ca2⫹. However, in this study when NH4Cl was removed and the cells became acidic, a second rise in Ca2⫹ was recorded, and
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
Fig. 8. Effect of NH4Cl (50 mM) and 4-AP (5 mM) on membrane potential of human lymphocytes in the presence or absence of external Ca2⫹. The figure is an original recording representative of four experiments. Gramicidin was added at the end of each experiment and at different times (two arrows) for both agents.
both changes were dependent on intracellular stores [12,13]. In contrast, Ca2⫹ is not mobilized in response to either intracellular alkalinization or acidification in Syrian hamster embryo cells [29]. In rat thymic lymphocytes, increasing pHi with monensin or NH3 did increase [Ca2⫹]i [30]. In Jurkat T-lymphocytes, NH4Cl induced increase in cytosolic pH but no detectable Ca2⫹ signal [6]. Thus, the effect of NH4Cl on Ca2⫹ signal has been contradictory in some cases. In our experiments with human lymphocytes using NH4Cl, we were able to record a rapid alkalinization and a small decrease of intracellular Ca2⫹. Besides different cell models, rat thymic lymphocytes or Jurkat vs. human lymphocytes, the effect found with NH4Cl on the Ca2⫹ level is probably due to the higher concentration used by us in these experiments, since 50 mM NH4Cl does not induce a Ca2⫹ response (data not shown). It was necessary to use 100 mM NH4Cl to observe a decrease of intracellular Ca2⫹. Using a different procedure, a report in collecting duct cells shows that cell alkalinization also produced a lowering of Ca2⫹ [31]. One explanation for our results is that NH4Cl could activate sequestration mechanisms for Ca2⫹. However, La3⫹ obliterates the fall in [Ca2⫹]i indicating that the decrease of Ca2⫹ is due to exit from the cell. Further studies are needed to explain how NH4Cl could stimulate Ca2⫹ efflux from the cell. 4.2. Effect of 4-AP on Ca2⫹ and pHi 4-AP, a K⫹ channel blocker, was used as a tool to increase pHi in Jurkat T cells and to induce a Ca2⫹ signal similar to that observed in response to physiological stimulation [6]. When we used 4-AP in human lymphocytes, an increase of pHi and Ca2⫹ was observed. 4-APinduced alkalinization preceded the Ca2⫹ signal providing further evidence that alkalinization is the trigger for the Ca2⫹ response, as it was reported in Jurkat T cells [6]. Raising pHi in human lymphocytes with 4-AP leads to a rise in intracellular Ca2⫹, even in the absence of external Ca2⫹, indicating the presence of pH-sensitive
579
intracellular Ca2⫹ stores emptied by 4-AP. However, contrary to our results, those authors found that in Ca2⫹free medium previous addition of thapsigargin abolished any effect of 4-AP, suggesting that 4-AP induced Ca2⫹ release from the agonist/IP3-sensitive pool of Jurkat T cells. In our experiments 4-AP increased the level of Ca2⫹ released by either caffeine or thapsigargin suggesting that 4-AP is mobilizing a different pool or the same pool with different kinetics. In fact, the existence of four different pools of Ca2⫹ was shown in Jurkat T-lymphocytes [32]. IP3 and Ca2⫹-induced Ca2⫹ release (CICR) mechanisms are closely related and coexpressed in muscle and nerve cells, both being sensitive to pH [11]. Caffeine, an agonist of the ryanodine receptor CICR, might be involved in the propagation of Ca2⫹ signal in human lymphocytes as it was suggested in other non-excitable cells [32,33]. The use of Jurkat vs. human lymphocytes (culture cells vs. fresh isolated lymphocytes) and the distinct Ca2⫹ pools sensitivity, could explain the different effect found with 4-AP in the two cell models. One alternative to explain the effect of 4-AP on Ca2⫹ signal is the sensitisation of the IP3 R [6], although the alkalinizing effect of NH4Cl does not induce the same Ca2⫹ response. The basis for the opposite findings is unclear and awaits further studies. 4.3. Effect of ionomycin on Ca2⫹ and pHi Ca2⫹ ionophores, such as ionomycin, are used to determine the role of Ca2⫹ in cellular regulation. However, studies concerning the effect of ionomycin on pHi are contradictory. In some cells the dual effect of ionomycin on pHi and Ca2⫹ has been mentioned [34,35]. In collecting duct cells, ionomycin induced a biphasic change in pH; a transient acidification followed by alkalinization. This cell alkalinization was mediated by depolarisation [31]. In human B lymphocytes ionomycin induced increase of Ca2⫹ as well as a transient acidification [36]. However, in some cells ionomycin induces influx of Ca2⫹ but not alteration in cytosolic pH [13]. In human lymphocytes, ionomycin increases pHi and Ca2⫹ levels, similar to that found in hepatocytes [34]. Because ionomycin is a Ca2⫹/H⫹ exchanger the increase of pHi could be a consequence of Ca2⫹/H⫹ activation as suggested in that report. However, we should preclude an effect of ionomycin on the Ca2⫹/H⫹ exchanger since this is an electroneutral exchanger and our results show depolarisation of plasma membrane potential in lymphocytes activated with ionomycin (see next paragraph). In human lymphocytes increase of Ca2⫹ takes place either in the presence or absence of external Ca2⫹, although in the absence of external Ca2⫹ the increase of pHi is almost undetectable. The magnitude of alkalinization is proportional to the increase of Ca2⫹ levels. Although thapsigargin, which also increases [Ca2⫹]i,
580
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581
failed to increase pHi in human lymphocytes (data not shown) as occurs in hepatocytes [34]. 4.4. Effects on membrane potential Concerning membrane potential changes the situation is even more complex and several reports have tried to establish a correlation with Ca2⫹ and pHi. Intracellular acidification with organic acids, such as lactate or acetate, caused an increase in intracellular Ca2⫹ and depolarisation in an insulinoma cell line. Alkalinization caused no apparent change in Ca2⫹. However, NH4Cl application and withdrawal were both accompanied by transient increases in Ca2⫹ and no depolarisation occurred [10]. In our study, all three stimuli induced depolarisation of human lymphocytes. The change in membrane potential induced by NH4Cl and 4-AP does not depend on the presence of external Ca2⫹ despite the opposite effect caused by these agents on [Ca2⫹]i. However, the distinct kinetic patterns induced by both agents on membrane potential should be pointed out. Depolarisation activated with NH4Cl takes place very fast, in less than one minute, in a similar fashion to the NH4Cl-induced alkalinization and reduction of intracellular Ca2⫹ levels. Changes produced by NH4Cl on membrane potential, [Ca2⫹]i and pHi are synchronised in lymphocytes. Thus, alkalinization is related to reduction of Ca2⫹ levels and membrane depolarisation. Kinetically, 4-AP-induced Ca2⫹ change occurs more slowly than plasma membrane depolarisation and alkalinization is faster. Then, as described by other studies, an alkalinization due to 4-AP induced an elevation of [Ca2⫹]i [6] and depolarisation in human lymphocytes. In rat thymocytes and in human lymphocytes, Grinstein et al. found that ionomycin induced membrane potential hyperpolarization [37,38]. However, contradictory results have been found even in different populations of the same cell type. Ionomycin hyperpolarises mouse T lymphocytes while depolarising B lymphocytes [39]. Our results show that a high concentration of ionomycin induces membrane depolarisation although this effect does not occur in Ca2⫹-free medium. In our view this is due to a high influx of Ca2⫹ stimulated by a higher concentration of ionophore (unpublished results). The elevated intracellular Ca2⫹ level attained by the ionophore could by itself cause the depolarisation. Additionally, the electrogenic uptake described by Grinstein and Smith [38] is expected to decline dramatically with increasing concentrations of ionomycin. In fact, it has been shown that ionophores are able to transport both electroneutrally and electrogenically [40]. In this study we demonstrate that in human lymphocytes, raising pHi with 4-AP leads to a rise in [Ca2⫹]i, even in the absence of external Ca2⫹. On the contrary, alkalinization with NH4Cl activates the efflux of Ca2⫹ from the cell. The mechanism by which alkalinization
with 4-AP induces a rise and NH4Cl a decrease in intracellular Ca2⫹ is unclear. It is not known if NH4Cl and 4-AP affect Ca2⫹ directly or by some other mechanisms. We found a lack of correlation between pHi and Ca2⫹ since alkalinization with three alkalinising agents does not lead to the same change in Ca2⫹. In summary, intracellular alkalinization can, under certain circumstances, cause a rise or a decrease in [Ca2⫹]i. The three alkalinising agents caused plasma membrane depolarisation in human lymphocytes, independent of external Ca2⫹ for NH4Cl and 4-AP, but completely dependent on the cation for the ionophore. The present report demonstrates that the effect of these agents in human lymphocytes is not limited to the change in [Ca2⫹]i. They induce substantial alterations in the overall cellular ionic homeostasis, including changes in the pHi and in the membrane potential.
Acknowledgments This work was funded with grants FEDER-CICYT1FD97-0153, Xunta Galicia PGIDT99INN26101, and FISS 96/0378. This work could not have been done without the invaluable collaboration of the staff at the General Hospital of Lugo from the Centro de Transfusio´n de Galicia, (Dr. Marcelo Alvarenga and Dr. Jose´ A. Gonza´lez, Ms. Sara Go´mez, Ana Pe´rez, Dolores Sa´nchez, Mr. J.M. Rodrı´guez, and Ms. Dolores Atanes and Ms. Noelia Lauda).
References [1] Putney JW. Cell Calcium 1990;11:611–24. [2] Alfonso A, Cabado AG, Vieytes MR, Botana LM. Br J Pharmacol 2000;130:1809–16. [3] Helmle-Kolb C, Counillon L, Roux D, Pouyssegur J, Mrkic B, Murer H. Pflugers Arch 1993;425:34–40. [4] van Haelst-Pisani C, Cragoe EJ, Rothstein TL. Exp Cell Research 1989;183:251–6. [5] Gore J, Besson P, Hoinard C, Bougnoux P. Am J Physiol 1994; 266:C110–20. [6] Guse AH, Roth E, Emmrich F. Biochem J 1994;301:83–8. [7] Oettgen HC, Terhorst C, Cantley LC, Rosoff PM. Cell 1985;40: 583–90. [8] Rosoff PM, Terres G. J Cell Biol 1986;103:457–63. [9] Tsien RY, Pozzan T, Rink TJ. Nature 1982;295:68–71. [10] Best L, Ammala C, Rorsman P, Tomlinson S. Biochimica et Biophysica Acta 1994;1192:107–11. [11] Tsukioka M, Iino M, Endo M. J Physiol (Lond) 1994;475:369–75. [12] Zorec R, Hoyland J, Mason WT. Eur J Physiol 1993;423:41–50. [13] Wiegmann TB, Welling LW, Beatty DM, Howard DE, Vamos S, Morris SJ. Am J Physiol 1993;34:C1184–90. [14] Siskind MS, McCoy CE, Chobanian A, Schwartz JH. Am J Physiol 1989;256:C234–40. [15] Putnam RW, Douglas PB. Mol Cell Biochem 1990;99:89–95. [16] Cabado AG, Vieyres MR, Botana LM. J Cell Physiol 1994;158: 309–16. [17] Grynkiewicz G, Poenie M, Tsien RY. J Biol Chem 1985;260: 3440–50.
A.G. Cabado et al. / Cellular Signalling 12 (2000) 573–581 [18] Cabado AG, Yu FH, Kapus A, Lukacs G, Grinstein S, Orlowski J. J Biol Chem 1996;271:3590–9. [19] Thomas JA, Buchsbaum RN, Zimniak A, Racker E. Biochemistry 1979;18:2210–8. [20] Bright GR, Fisher GW, Rogowska J, Taylor DL. J Cell Biol 1987;104:1019–33. [21] Ehrlich BE, Kaftan E, Bezprozvannaya S, Bezprozvanny I. TIPS 1994;15:145–9. [22] Sardet C, Franchi A, Pouyssegur J. Cell 1989;56:271–80. [23] Mason MJ, Mahaut SMP, Grinstein S. J Biol Chem 1991;266: 10872–9. [24] Mason M, Grinstein S. Biochem J 1993;296:33–9. [25] Grinstein S, Cohen S, Goetz JD, Rothstein A, Gelfand EW. Proc Natl Acad Sci USA 1985;82:1429–33. [26] Gaidano G, Chigo D, Schena M, Bergui L, Treves S, Turrini F, et al. J Immunol 1989;142:913–8. [27] Siczkowski M, Davies JE, Ng LL. Am J Physiol Cell Physiol 1994;267:C745–52. [28] De la Rosa LA, Cabado AG, Botana MA, Vieytes MR, Vidal JI, Botana LM. Pflugers Arch 1999;437:935–43.
581
[29] Isfort R, Cody DB, Asquith N, Ridder GM, Stuard SB, Leboeuf A. Eur J Biochem 1993;213:349–57. [30] Grinstein S, Goetz JD. Biochim Biophys Acta 1985;819:267–70. [31] Slotki I, Schwartz JH, Alexander EA. Am J Physiol 1993;265: C432–8. [32] Guse AH, Roth E, Emmrich F. Biochem J 1993;291:447–51. [33] Bennett DL, Cheek TR, Berridge MJ, Smedt HD, Parys JB, Missiaen L, et al. J Biol Chem 1996;271:6356–62. [34] Anwer MS. Hepatology 1993;18:433–9. [35] Hendey B, Mamrack MD, Putnam RW. J Biol Chem 1989;264: 19540–7. [36] Schoyen H, Iversen JG, Smeland EB, Heikkila R. Acta Physiol Scand 1990;138:221–7. [37] Dixon SJ, Stewart D, Grinstein S, Spiegel S. J Cell Biol 1987; 105:1153–61. [38] Grinstein S, Smith JD. J Gen Physiol 1990;95:97–120. [39] Ishuda Y, Chused TM. J Exp Med 1988;168:839–952. [40] Fasolato C, Pozzan T. J Biol Chem 1989;264:19630–6.