Lowering Extracellular pH Evokes Inositol Polyphosphate Formation and Calcium Mobilization*

Changing extracellular pH (pH,) from 7.4 to 6.1 increased [3H]inositol bis- and trisphosphates -10- and &fold, respectively, in 15 s in human fibroblasts. [‘HI Inositol phosphate increased less rapidly than the pol- yphosphates. Bradykinin similarly increased 13H]ino-sitol phosphates. Shifting pH, from 7.4 to 6.0 evoked a large spike in cytosolic free Ca2+ ([Ca2’li) which was primarily caused by the release of stored Ca2+. Chang- ing pH. from 7.4 to 6.0 decreased cytoplasmic pH to -7.0. Moderate decreases in intracellular pH had no effect on [Ca2+Ii or 4aCa2+ efflux. Decreasing pH, strik- ingly increased *‘Ca2+ efflux and decreased total cell Ca2+ similarly to bradykinin. Changing pH, from 7.4 to -6.4 produced half-maximal effects on [Ca2+]i, 4aCa2+ efflux, and total Ca2+. Cycling pH, between 7.4 and 6.0 produced repetitive decreases and increases in total Ca2+. Bradykinin released the Ca2+ which was reaccumulated after an acid pulse indicating that Ca2+ had returned to the hormone-sensitive pool. Decreasing pH, also released stored Ca2+ from coronary endothelial, neuroblastoma, and umbilical artery muscle cells, but not from rat aortic smooth muscle or human epidermoid carcinoma (A431) cells. We suggest that lowering pH, stimulates a

mobilizing metals was the same (Cd" > Co'+ > Ni2+ > Fez+ > Mn2+) in all of the cell types (Smith et al., 1989b). The divalent metals appear to stimulate an extracellular site that is reversibly blocked by Zn2+ and which may be considered a unusual Ca2+ mobilizing stimuli, we observed that decreasing pH, induces cell Ca2+ mobilization.' Here we present evidence that the protonation of a critical functional group, possibly imidazolium, in a cell surface protein induces inositol polyphosphate production and mobilizes cell Ca2+. We suggest that the three diverse Ca2+-mobilizing stimuli, divalent metals, or decreasing [Na+], or pH,, act on the same cell surface "receptor." UCd2+ receptor" (Smith et al., 1989b). In studying these two

EXPERIMENTAL PROCEDURES
Cell Culture-The methods for culturing the fibroblasts from human forearm skin (Smith et al., 1989a), endothelial cells from dog coronary arteries (Smith et al., 1989a), and smooth muscle cells from rat aorta  were the same as reported previously. Human neuroblastoma cells (SK-N-SH) and epidermoid carcinoma (A431) cells were obtained from the American Type Culture Collection (Rockville, MD) and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Smith et al., 1989b).
Cell pH-Fibroblasts were grown on cover glasses (12 X 18 mm) as described (Smith et al., 1989a) and incubated with 3 I~M BCECF acetoxymethyl ester (Molecular Probes, Inc., Eugene, OR) in 1 ml of DME for 30 min at 37 "C. Then the cover glass was rinsed and incubated with 2 ml of PSS with 10 mM glucose for 30 min before placing the cover glass on the diagonal face of a triangular cuvette. Fluorescence emission at 527 nm minus the autofluorescence of a cover glass that was not loaded with BCECF was recorded at two excitation wavelengths, 500 and 440 nm, with a Deltascan dual wavelength fluorometer (Photon Technology International Inc., Princeton, NJ). Autofluorescence was not greater than 5% dye fluorescence. Calibration of cytoplasmic dye fluorescence was done with a high K+ PSS (KC1 substituted for NaCl) (pH 8.0 to 6.0) containing 10 p M nigericin as described (Moolenaar et al., 1984).

Effect of
Decreasing pH, on [Ca2+],-Changing pH, from 7.4 to 6.0 evoked a large spike in [Ca2+], which was very similar to that produced by bradykinin (Fig. 1). [Ca2+Ii increased rapidly in response to both stimuli to a peak (-1 HM) and then fell to nearly the basal level (Fig. 1). Smaller de-' These findings were presented at the 73rd Annual Meeting of the Federation of American Societies for Experimental Biology (Smith et al., 1989~).

Decreasing External p H Triggers the Release of Stored
Effect of decreasing pH, on cytosolic free Ca" and pH. A, pH, was shifted at the A p H arrow by removing pH 7.4 PIPES-PSS containing 10 mM glucose and replacing it with PIPES-PSSglucose at the indicated pH. A separate coverslip was used for each pH value. B, bradykinin (100 nM) was added at the BK arrow, and pH, was decreased to 6.1 at the A pH arrow. C, the extracellular medium was removed and replaced each time the tracing of the ratio of BCECF fluorescence at 500 to 440 nm was interrupted. At the first interruption PSS was removed and replaced; at the second interruption PSS was replaced with pH 6.0 PIPES-PSS. From the point indicated by pHi = p H , the extracellular medium was high K'-PSS containing 10 Fg/ml nigericin at the pH indicated. Fluorescence at 440 nm was 200,000 counts/s and changed by 4 0 % . The tracings are representative of three independent experiments on duplicate cover glasses. creases in pH, evoked smaller increases in [Ca2+];. Changing pH, to 6.9 increased [Ca2+]; to about twice the basal level ( Fig.  1). Decreasing pH, increased [Ca"]i in the absence of extracellular Ca2+. Changing pH from 7.4 to 6.0 increased [Ca2+]; from 134 f 9 nM (n = 9) to 978 f 52 (n = 5) in the presence of external Ca2+ or to 952 f 70 nM (n = 4) in the absence of Ca2+. For this experiment PIPES-PSS was removed and replaced with PIPES-PSS (pH 6.0) or PIPES-PSS with 0.1 mM EGTA and no added Ca2+. These data indicate that the release of stored Ca2+ is largely responsible for the increase in [Ca2+], produced by lowering external pH.
Previously, we (Smith et ai., 1989) reported that bradykinin increases [Ca2+]; primarily by releasing stored ea2+ in human skin fibroblasts. Prior stimulation of the cells with the hormone abolished the [Ca2+]; response to decreasing pH, (Fig.  1B). Therefore, it appears that lowering pH, releases Ca2+ from the same Ips-sensitive pool as bradykinin.
Decreases in Intracellular p H Have No Effect on [Ca2'l; or Ca2+ Efflux-Changing pH, from 7.4 to 6.0 decreases intracellular pH from -7.3 to -7.0 ( Fig. 1). Moderate decreases in intracellular pH are readily produced by exposing cells to a weak acid at constant extracellular pH as described previously of the initial portion of efflux were determined by exponential curve fitting with a Hewlett-Packard 11C calculator. The pH, that halfmaximally increased efflux was determined from a log-log plot of the data that was done with commercially available software (Dose-Effect Analysis with Microcomputers by J. Chou and T.-C. Chou, Elsevier BIOSOFT, Amsterdam). Fa/Fu (Chou, 1976) is fraction affected/ fraction unaffected with 100% affected equal to the difference between the first-order rate coefficients at pH, 5.7 and 7.4. The regression coefficient of the line was 0.9963.
for fibroblasts (Moolenaar et al., 1984) and other cell types (Roos and Boron, 1981). The addition of 5, 10, 20, or 40 mM sodium propionate (pK 4.87) decreases cell pH by 0.1-0.4 units (Moolenaar et al., 1984 and Footnote 3). The addition of these concentrations of sodium propionate had no effect on [Ca"]; or 46Ca2" efflux. Because moderate decreases in intracellular pH at constant pH, failed to evoke a [Ca2+], spike, the protonation of an extracellular site probably triggers Ca2+ mobilization. Effect of pH, on 45Ca2+ Efflux-Lowering pH, strikingly increased 45Ca2+ efflux (Fig. 2). The maximal rate of efflux evoked by decreasing pH, is similar to that produced by bradykinin (Smith et al., 1989a). Changing pH, from 7.4 to 6.4 caused the half-maximal increase in the first-order rate coefficient of efflux (Fig. 2B). The effect of pH, was independent of the buffer used. Changing pH, from 7.4 to 6.0 produced similar increases in 45Ca2+ efflux when phosphate J. B. Smith, unpublished data. or imidazole was used to buffer PSS instead of PIPES.
The cells slowly regained much of the lost Ca2+ even when pH, was kept at 6.0 (Fig. 4A). Changing pH, back to 7.4 markedly increased the rate of recovery of total cell Ca2+ (Fig.   4A). After cell Ca2+ had returned to the basal level, a second 2-min acid pulse decreased total Ca2+ similarly to the first one (Fig. 4). Next, we cycled pH, between 7.4 and 6.0 to find out if total Ca2+ would repeatedly rise and fall in response to the pH, changes. Five successive changes in pH, repeatedly increased and decreased total cell Ca2+ by substantial amounts (Fig. 4B). Furthermore, bradykinin decreased cell Ca2+ after a partial recovery from an acid pulse (Fig. 4B), indicating that Ca2+ had been reaccumulated by the hormone-sensitive organelle.
Decreasing pH, had no effect on total cell K+ (p = 0.615, Student's t test). Cell K+ was 1.21 f 0.06 pmol/mg protein (n = 8) after 10 min at pH, 6.0 compared to 1.25 +-0.04 nmol/ mg (n = 12) in control cultures incubated at pH, 7.4. Changing pH, to 6.0 for 10 min significantly increased cell Na+ from 0.118 f 0.002 ( n = 12) to 0.155 k 0.006 pmol/mg protein (n = 8, p < 0.001). The cultures were incubated for 1 h in PSS containing glucose prior to changing pH, as described for "Ca2' efflux. The lack of an effect of pH, on cell K+ indicates The cultures were labeled overnight in 1 ml of DME containing 2% fetal bovine serum and 10 pCi of 4'Ca2+. The medium was aspirated and replaced with 1 ml of PIPES-PSS containing 10 mM glucose, 1.8 mM CaC12, and 10 pCi of "Cas+. At zero time the medium was aspirated and replaced with 1 ml of PIPES-PSS containing 1.8 mM CaCL and 10 pCi of "Cas+ at the indicated pH. The pH which halfmaximally decreased total Cas+ was determined from a log-log plot of the data that was done as indicated in the legend to Fig. 2 with change in total Ca2+ at pH 5.7 as 100% affected. The regression coefficient of the line was 0.9698. The cultures were labeled overnight in 1 ml of DME containing 2% fetal bovine serum and 10 pCi of 4sCa2+. A, thirty pl of 1 N HCl was added to the medium to decrease pH, to 6.0. After 2 min the medium was aspirated and replaced with 1 ml of DME containing 2% fetal bovine serum and 10 pCi of ' ' Cas+ and incubated for the indicated time. Values are mean k S.E. (n = 2-9) for two experiments. B, the medium was removed and replaced with freshly prepared medium of the same composition 1 h before starting the experiment. The medium was acidified at the indicated times. After 2 min the medium was aspirated and replaced with 1 ml of DME containing 2% fetal bovine serum and 10 pCi of %a2+ as in A. Values are mean f S.E. (n = 6-19) for three experiments. Bradykinin was added (20 pl of 1 p~) to some cultures (triangles) at 0 or 12 min.
that pH, selectively affects cell Ca2+ regulation and does not generally alter permeability of the cells to cations. Exposing the cells to pH 6 for 2-10 min had no immediate effect on cell morphology, and there was no decrease in cell viability for a t least 3 days after the acid treatment as judged by phase contrast microscopy or plating efficiency after detachment with trypsin.

Effect of Lowering pH, and Bradykinin on Cellular pH]
Inositol Phosphates-Changing pH, from 7.4 to 6.1 for 15 s increased [3H]IPz and 13H]IP3 by -10-and -5-fold, respectively (Fig. 5). [3H]IP increased less rapidly than the polyphosphates (Fig. 5). [3H]IP4 increased by -90% 30 s after shifting pH, (Fig. 5). Bradykinin produced similar changes in [ 3 H ] I P~ as lowering pH, (Fig. 5). Neither the hormone nor the change in pH, affected [3H]glycerophosphoinositol (Fig.  5). pH Triggers the Release of Stored ea2+  Decreasing p H , Mobilizes Cell Ca2+ in Endothelial and Neuroblastoma Cells-Changing pH, from 7.4 to 6.0 transiently increased [Ca2+]i in endothelial cells cultured from dog coronary arteries (Fig. 6). [Ca2+], increased rapidly from 187 f 11 nM (n = 12) to 527 f 48 nM (n = 7) and returned to the basal level about 2 min after lowering pH, (Fig. 6). Removing external Ca2+ from the low pH buffer and adding 0.  Fig. 2. Data are representative of two experiments at pH 5.7 and 6.0-6.9 and at least six experiments at pH 7.4 and 6.0. B, efflux from SK-N-SH cells was assayed in PSS containing 20 mM maleate adjusted to pH 5.7-6.7 with Tris or in PSS containing 20 mM Tris adjusted to pH 6.9-7.4 with maleate. Three additional experiments gave similar results. The pH, that half-maximally increased efflux was determined as indicated in the legend to Fig. 28. The regression coefficients of the lines in endothelial cells and SK-N-SH cells were 0.9796 and 0.9988, respectively. creased to 489 +-93 nM (n = 5) in response to the pH 6 containing EGTA and no added Ca2+ (Fig. 6).
Lowering pH, strongly stimulated 46Ca2+ efflux in endothelial and neuroblastoma cells (Fig. 7). Changing pH, to 6.4 or 6.3 half-maximally increased efflux rate coefficient in the endothelial and neuroblastoma cells, respectively. Therefore, the pH, dependence of Ca2+ mobilization was almost the same in the three cell types.

DISCUSSION
A variety of external stimuli trigger phosphatidylinositol 4,5-bisphosphate hydrolysis, including hormones, neurotransmitters, spermatozoa, photons, antigens, nucleotides, and mitogens (Berridge and Irvine, 1984). Our findings indicate that a decrease in pH, triggers cell Caz+ mobilization in fibroblasts, endothelial, smooth muscle, and neuroblastoma cells (Table  I). Decreasing intracellular pH failed to mobilize cell Ca2+. Therefore, acidifying the extracellular medium apparently triggers Ca2+ mobilization by protonating a functional group, possibly imidizolium, in a cell surface protein. The imidizole  group (pK. 6-7) of histidine is the most common protein functional group with a pKa near the pH, (6.4) which halfmaximally induced Ca2+ mobilization. Acidifying the external medium produced a large and rapid increase in IPS (Fig. 5). Because IP3 releases Ca2+ from a nonmitochondrial pool in a variety of cells (Berridge, 1987), it is likely that low pH,evoked IPS production causes cell Ca2+ mobilization. Hormone-receptor binding triggers IPS production by activating a phosphoinositidase which is regulated by a G protein (Berridge, 1987;Gilman, 1987). It is unlikely that changing pH, directly affects either the G protein or phosphoinositidase, because neither of these proteins have membrane-spanning domains (Gilman, 1987;Katan et al., 1988).
Cell surface receptors that mediate the endocytosis of specific macromolecules cycle continuously between the plasma membrane and intracellular organelles (Goldstein et al., 1985). After internalization the receptors encounter mild acidity (pH 5.0-6.5) in endocytic vesicles and lysosomes (Yamashiro et al., 1984). The low pH of the endocytic compartments usually causes the macromolecule to dissociate from the receptor, which is essential for receptor sorting (Brown et al., 1983). DiPaola and Maxfield (1984) observed that mild acidity induces conformational changes in the receptor for epidermal growth factor in A431 cells and in the purified asialoglycoprotein receptor reconstituted in liposomes. Turkewitz et al. (1988) showed recently that a soluble fragment of the transferrin receptor that contains 95% of its external domain undergoes a reversible conformational transition and selfassociation below pH 6. Lowering pH, may trigger inositol polyphosphate production by inducing a conformational change in the ectodomain of a cell surface protein that normally encounters low pH only after endocytosis.
In spite of the vigorous investigation over the past decade of the influences of stimuli that alter cell Ca2+ on intracellular pH (Busa and Nuccitelli, 1984;Moody, 1984), there have been relatively few studies of the effects of pH, on cell Ca2+ regulation. Kim and Smith (1987) reported that changing pH, from 7.4 to 6.0 decreased 45Ca2+ uptake by 29% and increased 45Ca2+ efflux by 17% from cultured chick embryo ventricular cells. The authors suggested that the shift in pH, may affect Na+/Ca2+ antiport activity and Ca2+ binding on the cell surface. Iijima et al. (1986) examined the effects of external pH on gating and permeation in Ca2+ channels with the whole cell configuration of the patch clamp technique. They concluded that protonation reduces the amplitude of the negative surface potential which is sensed by the gating mechanism.
Drapeau and Nachshen (1988) examined the effects of lowering internal and/or external pH on Ca2+ regulation in synaptosomes. They found that changing internal pH to 5.8 or external pH to 5.5, which decreased internal pH to 6.4 in 30 s, had no effect on [Ca2+Ii which was measured with fura-2. Changes in intracellular pH produce relatively small changes in [Ca2+Ii which are inconsistent with respect to the direction of the [Ca2+]; change among different cell types (Moody, 1984). We found that changing pH, to 6 decreased cell pH to only about 7 and that moderate decreases in intracellular pH had no effect on [Ca2+], or 45Ca2+ efflux in the fibroblasts.
The response of the fibroblasts to changing pH, to 6.0 is remarkably similar to stimulating the bradykinin receptor in these cells. First, both stimuli caused similar changes in cellular [3H]inositol phosphates, suggesting that both stimuli activate phosphoinositidase rather than another phospholipase (Fig. 5). Second, both stimuli cause a large spike in [Ca2+Ii which follows a very similar time course (Fig. 1, Smith et al., 1989a). Third, both stimuli evoke a similar increases in 45Ca2+ efflux (Fig. 2, Smith et al., 1989a). Fourth, both stimuli provoke a rapid and reversible depletion of total cell Ca", which amounts to about 5 nmol/mg protein Ca2+ being expelled in 60 s (Figs. 3 and 4, Smith et al., 1989a). About 200,000 Ca2+ pumps/cell would be required to expel Ca2+ at this rate (Smith et al., 1989a). The rapidity of the recovery of total cell Ca2+ after changingpH, from 6.0 to 7.4 is noteworthy ( Fig. 4) and suggests that IPS production abruptly stops and that IPS is rapidly metabolized to compounds that do not activate the intracellular Ca2+ release channel. Comparable studies with bradykinin have not been possible because of the lack of an effective receptor antagonist.
Previously, we (Smith et al., 1989a(Smith et al., , 1989b observed that removing extracellular Na+ and certain divalent metals trigger IP3 production and mobilize cell Ca2+. Table I shows the responsiveness of various cell types to the three stimuli, decreasing pH,, decreasing [Na+Ia, or the addition of Cd". Four different cell types, including human neuroblastoma, dog coronary endothelial, human umbilical artery muscle, and human lung and skin fibroblasts, respond to all three stimuli. Two cell types, rat embryo fibroblasts and aortic muscle cells, did not respond to any of the stimuli. One exceptional cell type, A431 cells, responded to [Na'], removal, but not to Cd2+ or decreasing pH,. The mechanism of cell Ca2+ mobilization by [Na+], removal in the case of A431 cells is clearly different from that in the other cell types. In the skin fibroblasts, endothelial, and neuroblastoma cells, Ca2+ mobilization is unaffected by raising intracellular Na+ (Smith, 1989 and Footnote 3). In contrast, raising cell Na+ with ouabain completely abolishes the stimulation of 46Ca2+ efflux in A431 cells. Therefore, decreasing intracellular rather than extracellular Na+ appears to provoke the release of stored Ca2+ in A431 cells unlike the other cell types we studied. The observation that diverse cell types respond to all three stimuli or to none of them suggests that a single receptor may be involved. Recently, we found that photooxidation strongly inhibits cell Ca2+ mobilization in response to all three stimuli without affecting the response to bradykinin! To prove that a single receptor confers responsiveness to all three stimuli will require receptor purification and reconstitution or gene cloning and expression.
Bacteria and taste cells have chemosensors for acid. Acid is a repellent for Escherichia coli (Tso and Adler, 1974). The pH, dependence of negative chemotaxis in E. coli is similar to that of Ca2+ mobilization in the cell types we studied, but it is unclear whether the pH sensor is an intracellular or extracellular component of the bacterium. The taste intensity of acids appears to be primarily determined by proton concentration, although the anion influences the response to acid (Biedler, 1971). Kurihara et al. (1986) observed that decreasing pH, depolarizes mouse neuroblastoma (N-18) cells; however, it is unclear whether the depolarization of N-18 cells by acid is related to chemoreception by taste cells. Recently Akabas et al. (1988) showed that a bitter substance, denatonium, transiently increases [Ca2+]i in a subpopulation of rat taste cells. Taste reception appears to trigger the release of stored Ca2+ because denatonium increased [Ca2+]i in the absence of extracellular Ca2+. The cell surface receptor that triggers inositol polyphosphate formation and cell Ca2+ mobilization in response to decreases in pH, as shown here may be structurally related to an acid taste sensor.