We previously showed that shrinking a barnacle muscle fiber (BMF) in a hypertonic solution (1,600 mosM/kg) stimulates an amiloride-sensitive Na-H exchanger. This activation is mediated by a G protein and requires intracellular Cl−. The purpose of the present study was to determine (a) whether Cl− plays a role in the activation of Na-H exchange under normotonic conditions (975 mosM/kg), (b) the dose dependence of [Cl−]i for activation of the exchanger under both normo- and hypertonic conditions, and (c) the relative order of the Cl−- and G-protein-dependent steps. We acid loaded BMFs by internally dialyzing them with a pH-6.5 dialysis fluid containing no Na+ and 0–194 mM Cl−. The artificial seawater bathing the BMF initially contained no Na+. After dialysis was halted, adding 50 mM Na+ to the artificial seawater caused an amiloride-sensitive pHi increase under both normo- and hypertonic conditions. The computed Na-H exchange flux (JNa-H) increased with increasing [Cl−]i under both normo- and hypertonic conditions, with similar apparent Km values (∼120 mM). However, the maximal JNa-H increased by nearly 90% under hypertonic conditions. Thus, activation of Na-H exchange at low pHi requires Cl− under both normo- and hypertonic conditions, but at any given [Cl−]i, JNa-H is greater under hyper- than normotonic conditions. We conclude that an increase in [Cl−]i is not the primary shrinkage signal, but may act as an auxiliary shrinkage signal. To determine whether the Cl−-dependent step is after the G-protein-dependent step, we predialyzed BMFs to a Cl−-free state, and then attempted to stimulate Na-H exchange by activating a G protein. We found that, even in the absence of Cl−, dialyzing with GTPγS or AlF3, or injecting cholera toxin, stimulates Na-H exchange. Because Na-H exchange activity was absent in control Cl−-depleted fibers, the Cl−-dependent step is at or before the G protein in the shrinkage signal-transduction pathway. The stimulation by AlF3 indicates that the G protein is a heterotrimeric G protein.
Cell swelling generally initiates a rapid sequence of events that results in the efflux of ions and water, a volume-regulatory decrease (VRD) that returns cell volume toward normal (for reviews see Hoffmann and Simonsen, 1989; Hallows and Knauf, 1994; Lang et al., 1995). The ion efflux may be mediated by K/Cl cotransport (Dunham and Ellory, 1981; Jennings and Shulz, 1990; Jennings and Schulz, 1991; Jennings and Al-Rohil, 1990), KCl efflux through parallel K+ and Cl− channels (Knoblauch et al., 1989; Welling and O'Neil, 1990; Banderali and Roy, 1992), or K-H exchange (Cala, 1980, 1983, 1985). Conversely, cell shrinkage often initiates a volume-regulatory increase (VRI), a rapid sequence of events that results in the influx of ions and water (Lang et al., 1995; Hoffman and Simonsen, 1989; Hallows and Knauf, 1994). The influx may be mediated by Na/K/Cl cotransport (Geck et al., 1980; Eveloff and Calamia, 1986) or Na-H exchange augmented by Cl-HCO3 exchange (Kregenow et al., 1985; Grinstein et al., 1983; Jennings et al., 1986). It is believed that, in some cells, the VRD and VRI responses are reciprocal, with cell swelling stimulating VRD and inhibiting VRI, and cell shrinkage having the opposite effects (Lang et al., 1995).
Over a longer time frame, hypertonicity may stimulate osmotic-response elements in some cells, increasing the transcription of enzymes that catalyze the production of intracellular osmolytes. With a delay of ∼48 h, mIMCD-3 (renal medullary collecting duct), PAP-HT25 (rabbit inner medulla), and MDCK cells respond to hypertonicity by inducing aldose reductase (Spring and Siebens, 1988; Garcia-Perez and Burg, 1991; Burg, 1995). This enzyme converts glucose to the relatively impermeant sorbitol, causing osmotic swelling. Other osmolytes, such as inositol, betaine, taurine, and glycerophosphocholine may also be concentrated, and thereby promote re-swelling (Burg, 1995; Garcia-Perez and Burg, 1991).
A major unanswered question in cell physiology is how cells sense cell-volume changes and transduce them to the appropriate changes in ion transport. One important clue may be the observation, made by Parker (1986), that Cl− is necessary for the shrinkage-induced activation of Na-H exchange in dog red blood cells. In earlier work on muscle fibers from the giant barnacle, we confirmed this observation and additionally showed that the shrinkage-induced activation of the Na-H exchanger specifically requires Cl− in the intracellular fluid (Davis et al., 1994). The precise role of Cl− in this process is unclear. However, there is precedent for involvement of Cl− in other biological processes. For example, Cl− increases the affinity of the α subunit of the heterotrimeric G protein Go for GTPγS (Higashijima et al., 1987).
A second important clue into the shrinkage signal-transduction system is that, in barnacle muscle fibers (BMFs),1 the shrinkage-induced activation of the Na-H exchanger appears to be mediated by a G protein (Davis et al., 1992a). Thus, the effect of shrinkage on the exchanger is inhibited by dialyzing the fiber with GDPβS, and mimicked either by dialyzing with GTPγS or by injecting the fibers with activated catalytic subunit of cholera toxin (CTX).
The purpose of the present study was to explore the role of Cl− in the shrinkage-induced activation of Na-H exchange in internally dialyzed barnacle muscle fibers. We used microelectrodes to monitor intracellular pH (pHi) and calculated the Na-H exchange rate ( JNa-H) from the rate of pHi increase and the intracellular buffering power. Because internal Cl− is required for the shrinkage-induced activation of Na-H exchange in BMFs, we hypothesized that the primary signal the cell senses during shrinkage may be an increase in [Cl−]i. To test this hypothesis, we determined the [Cl−]i dependence of Na-H exchange, both under normo- and hypertonic conditions. We found that, even under normotonic conditions, the Na-H exchanger is inactive in the absence of Cl−. Increasing [Cl−]i causes a monotonic rise in JNa-H but, at a given [Cl−]i, JNa-H is always greater under hypertonic conditions. Thus, an increase in [Cl−]i is not the primary shrinkage signal. In a second series of experiments, we asked whether the Cl−-dependent step in the activation of the Na-H exchange is before or after the G-protein step. We found that, even in BMFs depleted of Cl−, we could activate Na-H exchange with GTPγS, AlF3, or CTX. Thus, the Na-H exchanger does not require Cl− per se. Moreover, the Cl−-dependent step precedes or is concurrent with the G-protein step in the signal-transduction cascade.
Barnacles were obtained from Bio-marine Enterprises (Seattle, WA) and kept in an aerated aquarium at 4°C. After dissection, barnacle muscle fibers were kept at 4°C for a period of up to 36 h in our standard artificial seawater (ASW) (see Solutions, below). Before experiments, fibers were incubated for at least 1 h in a Ca++-free artificial seawater (see Solutions, below) to prevent contraction during the subsequent cannulation (see below). This Ca++-free solution also contained 0.5 mM SITS (4-acetamido-4′-isothiocyanostilbene-2,2′-disulfate) (United States Biochemical Corp., Cleveland, OH) to permanently block the activity of the Na+-driven Cl-HCO3 exchanger (Boron, 1977). The fibers were cannulated in a Ca++-free (or a Ca++- and Cl−-free) ASW also containing SITS.
All ASWs were nominally HCO−3 free. The standard ASW, in which fibers were incubated before the experiments, consisted of (mM): 440 Na+, 10 K+, 11 Ca++, 45.5 Mg++, 558 Cl−, 5 EPPS− (the anionic form of N -(2-hydroxyethyl)piperazine-N ′-3-propanesulfonic acid; Sigma Chemical Co., St. Louis, MO), and 5 of the neutral form of EPPS (pKEPPS ≅ 8.0). The usual Ca++-free ASW was made by replacing the Ca++ in the standard ASW with Mg++. ASWs containing 0 Na+ were made by replacing Na+, mole for mole, with N-methyl-d-glucammonium (NMDG+) that was produced by using HCl to titrate the free base N-methyl- d-glucamine (Sigma Chemical Co.). The ASW containing 50 mM Na+ was made by diluting the standard ASW with the 0-Na+ ASW.
Cl−-free ASWs were made by replacing Cl−, mole for mole, with gluconate (Sigma Chemical Co.). In Cl−-free ASWs that also lacked Na+, we generated NMDG+ salts of gluconate by titrating NMDG free base with gluconic acid lactone. We made the 50-Na+/0-Cl− ASW in the same way we made the 0-Na+/0-Cl− ASW, except that we replaced 50 mM NMDG+/gluconate with Na+/ gluconate.
We adjusted the pH values of all ASWs to pH 8.00 at 22°C. We decreased pH, in Cl−-containing ASWs, with either EPPS or HCl, and in Cl−-free ASWs, with EPPS. We increased pH, in Na+-containing ASWs, with NaOH, and in Na+-free ASWs, with NMDG -free base. The osmolalities were determined using a vapor pressure osmometer (Wescor Inc., Logan, UT), and adjusted to 975 ± 10 mosM/kg with mannitol or water. Solutions were made hypertonic (1,600 ± 10 mosM/kg) by adding mannitol. Solutions were delivered to the chamber by peristaltic or syringe pumps, as previously described (Boron, 1985). Fibers were superfused at a rate of 1 ml min−1.
All the dialysis fluids (DFs) were Na+ free. The standard pH-7.2 DF contained 34 mM Cl−, and consisted of (mM): 243 K+, 7 Mg++, 175 glutamate, 34 Cl−, 2 EGTA, 44 of the anionic form of HEPES (United States Biochemical Corp.), 56 of the neutral form of HEPES, 0.5 phenol red, and 4.0 Tris/ATP. The standard pH-6.5 DF also contained 34 mM Cl−, and consisted of (mM): 255.5 K+, 7 Mg++, 160 glutamate, 34 Cl−, 2 EGTA, 71.5 of the anionic form of MES (2-(N-morpholino)- ethanesulfonic acid; Sigma Chemical Co.), 28.5 of the neutral form of MES, 0.5 phenol red, and 4.0 Tris/ATP. Cl−-free DFs were made by replacing all of the Cl− with l-glutamate. DFs with [Cl−] values above 34 mM were made by replacing glutamate, mole for mole, with Cl−. Dialysis fluid pH was adjusted upward with KOH and downward with either HCl (for Cl−-containing DFs) or l-glutamic acid (for Cl−-free DFs). In the DFs containing GTPγS, 1 mM ATP was substituted with 1mM GTPγS, keeping the total nucleotide concentration at 4 mM. Aluminum fluoride was added to the DFs as 10 mM KF + 100 μM AlCl3.
Measurement of pHi and Membrane Potential
The technique for measuring intracellular pH (pHi) in internally dialyzed muscle fibers has been published elsewhere (Russell et al., 1983). Single, isolated BMFs were horizontally cannulated in a Ca++-free solution. A length of cellulose-acetate dialysis tubing with a molecular weight cutoff of ∼6 kD was threaded into one cannula, through the muscle fiber, and out the opposite cannula. The pH electrodes were fabricated according to the design of Hinke (1967). Glass microelectrodes used for measuring membrane voltage (Vm) were filled with 3 M KCl when the dialysis fluid contained Cl−, but with 1 M K-glutamate when the DF was Cl− free.
After cannulating the fiber and inserting the dialysis tubing, we began dialysis with a pH-7.2 DF at a rate of 5 μl min−1. All DFs were Na+ free. The pH and Vm electrodes were inserted through opposite cannulas so that their tips were within ∼500 μm of each other. The central region of the fiber was isolated from the cut ends with grease seals. We then began superfusing the fiber with a Na+-free ASW at a rate of 1 ml min−1.
Studies on the [Cl−]i dependence of Na-H exchange.
In these experiments, after an initial ∼30-min period of dialysis with a DF having a pH of 7.2 and a [Cl−] between 34 and 194 mM, dialysis continued for an additional ∼50–60 min with a DF that was otherwise identical, except for having a pH of 6.5. Thus, the total time for dialysis was ∼80–90 min. Previous work has shown that 60–90 min is sufficient for either 22Na (Russell et al., 1983) or 36Cl (Boron et al., 1978; Russell and Brodwick, 1988) in the DF to achieve isotopic equilibrium. After dialysis was halted, pHi was allowed to stabilize for ∼15 min in a Na+-free ASW.
Studies with GTPγS or AlF3 in Cl−-depleted cells.
The protocol was similar to that above, except that the initial period of dialysis with the pH-7.2/Cl−-free DF continued for ∼140 min to deplete the cell of Cl−. During this time, the fiber was superfused with a Na+- and Cl−-free ASW. This pH-7.2/Cl−-free DF was then switched to an identical solution that also contained either 1 mM GTPγS or 10 mM AlF3. After ∼40 min, we switched to an identical solution in which the pH of the DF was lowered to 6.5 to acid load the fiber. We continued dialyzing with this solution for ∼60 min. Thus, the total dialysis time was ∼240 min.
Studies on Cl−-depleted cells injected with CTX.
This protocol was similar to the one in the GTPγS and AlF3 experiments. The major difference was that, ∼160 min after initiating dialysis, we microinjected the BMFs with CTX, and only then inserted the microelectrodes. The injection fluid was the pH-7.2/0-Cl− DF containing the dithiothreitol-activated CTX to a final intracellular concentration of ∼3 × 10−6 M. During the microinjection and electrode insertion, the fiber was briefly exposed to an ASW that lacked Ca++ (to prevent contraction). Also, during the microinjection and electrode insertion, the fiber was dialyzed continuously with the pH-7.2/0-Cl− DF. After an additional ∼110 min dialysis with this DF, and an additional ∼60 min with a pH-6.5/ 0-Cl− DF, we halted dialysis and allowed the fiber to stabilize for 120 min before assaying as described above.
Values are reported as means ± SEM. Groups of data were compared using a two-sample t test assuming unequal variance.
Effect of Increasing Internal Cl− under Normotonic Conditions
Fibers dialyzed with 34 mM Cl− under normotonic conditions
Fig. 1,A illustrates an experiment in which a muscle fiber was acid loaded by internally dialyzing it with a Na+-free DF containing 34 mM Cl− at pH 6.5. As noted in methods, all fibers were pretreated with SITS to eliminate Na+-driven Cl-HCO3 exchange, and all were superfused with a Na+-free ASW for ∼90 min. The terminal portion of this dialysis period is shown in Fig. 1,A. After we halted dialysis (Fig. 1,A, a), pHi continued to drift downward (Fig. 1,A, ab), probably because the tip of the pH electrode in this experiment was rather distant from the dialysis tube. The mean pHi at b was 6.72 ± 0.01 (n = 20). Inasmuch as the fiber had been dialyzed with a Na+-free DF, and superfused with a Na+-free ASW, [Na+]i should have been extremely low. Exposing the fiber to an ASW containing 50 mM Na+ produced a slow increase in pHi (Fig. 1,A, bc), due mainly to the basal activity of the Na-H exchanger (Davis et al., 1994). We chose to use a modest level of Na+, 50 mM, because this concentration is high enough to support Na-H exchange, but not so high as to compete with amiloride for binding sites on the transporter. We computed the total acid-extrusion rate (JTotal) as the product of the pHi recovery rate of segment bc and the previously measured intrinsic buffering power (Davis et al., 1994). The pHi increase of Fig. 1,A, bc was inhibited by adding 1 mM amiloride to the 50-Na+ ASW (Fig. 1,A, cd). The broken lines in the figure emphasize the slopes in the absence and presence of amiloride. The delay between the application and action of amiloride presumably reflects the time for the drug to reach the interstices of the BMF. For the purposes of this study, we will present values for Na-H exchange rate (JNa-H) as the amiloride-sensitive component of Jtotal; that is, the difference in flux values between bc and cd in Fig. 1,A. For 20 fibers, the mean JNa-H was 23 ± 4 μM min−1 (Table I).
Fibers dialyzed with 194 mM Cl−under normotonic conditions.
To address the question of whether increasing [Cl]i stimulates Na-H exchange under normotonic conditions, we performed the experiment shown in Fig. 1,B, in which we dialyzed the fiber with a Na+-free DF containing 194 mM Cl−. After we halted dialysis, pHi drifted upward very slowly (Fig. 1,B, ab). The mean pHi at point b was 6.71 ± 0.02 (n = 12). Exposing the cell to 50 mM Na+ produced a rapid intracellular alkalinization (Fig. 1,B, bc) that was largely blocked by amiloride (Fig. 1,B, cd). The mean JNa-H for these 12 experiments was 138 ± 26 μM min−1 (see Table I). Therefore, under normotonic conditions (975 mosM/kg), increasing [Cl−] in the DF from 34 to 194 mM dramatically stimulates the Na-H exchanger, increasing JNa-H from 23 to 138 μM min−1.
Effect of Increasing Internal Cl− under Hypertonic Conditions
Effect of 34 mM Cl− under hypertonic conditions.
Fig. 2,A illustrates 1 of 13 experiments in which we examined the effect of hypertonicity on Na-H exchange in fibers dialyzed with 34 mM Cl−. The first part of the protocol was identical to that of Fig. 1,A: the fiber was acid loaded by dialyzing with a Na+-free DF containing 34 mM Cl−, dialysis was halted, and pHi was allowed to stabilize (Fig. 2,A, ab). The mean pHi at point b was 6.70 ± 0.02 (n = 13). As in Fig. 1,A, exposing the cell to an ASW containing 50 mM Na+ elicited, at most, a very slow alkalinization (Fig. 2,A, bc).2 At point c, we switched to a Na+-free ASW made hypertonic (1,600 mosM/kg) by the addition of mannitol. The self-limited increase in pHi (Fig. 2,A, cd) was presumably due to the concentration of intracellular buffers, as previously described. Indeed, in a previous study, we found that exposing a muscle fiber to the same hypertonic solution caused intracellular buffering power, measured at a pHi of ∼6.8, to double (Davis et al., 1994). After pHi stabilized, exposing the cell to a hypertonic ASW containing 50 mM Na+ produced a slow pHi increase (Fig. 2,A, de). Applying amiloride not only blocked this pHi increase, it unmasked a slow acidification (Fig. 2,A, ef). The difference in the slopes of the pHi recoveries in segments de and ef indicates that there was a modest rate of Na-H exchange when cells dialyzed with 34 mM Cl− were shrunken. In a total of 13 similar experiments, the mean JNa-H was 86 ± 17 μM min−1 (see Table I), a figure that takes into consideration the increased buffering power in hypertonic solutions. This mean JNa-H, obtained under hypertonic conditions, is ∼3.7-fold higher than the mean JNa-H (see Fig. 1 A) obtained under normotonic conditions in cells dialyzed with 34 mM Cl−.
Effect of 194 mM Cl− under hypertonic conditions.
To determine whether increasing [Cl]i also stimulates Na-H exchange under hypertonic conditions, we performed the experiment shown in Fig. 2,B. This experiment is identical to that shown in Fig. 2,A, except that the DF contained 194 rather than 34 mM Cl−. The mean pHi at point b was 6.73 ± 0.01 (n = 8). Exposing the cell to a normotonic ASW containing 50 mM Na+ produced a rapid alkalinization (Fig. 2,B, bc), as observed above for fibers dialyzed with 194 mM Cl− (Fig. 1,B). After we switched to a Na+-free hypertonic solution, and pHi stabilized (Fig. 2,B, cd), increasing [Na+]o to 50 Na+ produced an even more marked alkalinization (Fig. 2,B, de) that was largely blocked by amiloride (Fig. 2,B, ef). The mean JNa-H for these eight experiments was 345 ± 43 μM min−1 (see Table I). Therefore, under hypertonic conditions, increasing the [Cl−] in the DF from 34 to 194 mM increased JNa-H fourfold, from 86 to 345 μM min−1.
[Cl−]i Dependence of Na-H Exchange under Normo- and Hypertonic Conditions
We have already seen that, under normotonic conditions, increasing the [Cl−] of the DF from 34 to 194 mM caused an increase in JNa-H (Fig. 1, A vs. B). To determine the [Cl−]i dependence of Na-H exchange under normotonic conditions, we performed additional experiments identical to those in Fig. 1 except that the [Cl−] of the DFs was 74, 114, or 154 mM. The data are summarized by the open circles in Fig. 3. In plotting the data, we have assumed that [Cl−]i is the same as the [Cl−] of the DF.
We also performed additional experiments to determine the [Cl−]i dependence of Na-H exchange under hypertonic conditions. These experiments were identical to those in Fig. 2, except that the [Cl−] of the DFs was 74, 114, or 154 mM. We have assumed that the cells behaved as perfect osmometers, so that increasing the osmolality from 975 to 1,600 mosM/kg increased the [Cl−]i by a factor of 1,600/975, or 1.64. Thus, [Cl−]i values for the hypertonic data, plotted as closed circles in Fig. 3, have values 1.64-fold higher than the corresponding [Cl−]i values for the normotonic data, plotted as open circles. Fig. 3 shows that, under both normo- and hypertonic conditions, increasing nominal [Cl−]i causes an increase in Na-H exchange activity. For normotonic conditions (Fig. 3, ○), a nonlinear least-squares curve fit (Hill coefficient = 2) produced an apparent K m for internal Cl− of 127 mM, and an apparent Vmax of 201 μM min−1. For hypertonic conditions (Fig. 3, •), the apparent K m was ∼112 mM, and the Vmax was 375 μM min−1. Therefore, we conclude that increasing the nominal [Cl−]i stimulates Na-H exchange activity under both normo- and hypertonic conditions. Furthermore, although hypertonicity had little effect on the apparent Km for internal Cl−, it increased the Vmax for Na-H exchange by nearly 90%.
Effect of Using Cl− Substitutes other than Glutamate
From the above data, it would appear that increasing [Cl−]i stimulates Na-H exchange, both under normo- and hypertonic conditions. However, to create DFs of increasing [Cl−], we simultaneously lowered [glutamate]. Thus, it is possible that the increased JNa-H we observed in high Cl DFs was due to decreasing [glutamate], rather than increasing [Cl−]. Therefore, we examined the effect of using two other Cl− substitutes, gluconate and sulfamate. Our standard DF contained 34 mM Cl− and 160 mM glutamate. In the present series of experiments, done on matched fibers, we dialyzed BMFs with DFs containing 34 mM Cl− plus 160 mM of either glutamate, gluconate, or sulfamate. We assayed the fibers for Na-H exchange as in Fig. 1 A, under normotonic conditions. If glutamate were inhibiting Na-H exchange at a [Cl]i of 34 mM, then replacing the glutamate with either gluconate or sulfamate should substantially increase JNa-H, that is, produce approximately the same JNa-H that we saw above at 194 mM Cl− (i.e., ∼138 μM min−1), when we substituted 160 mM glutamate for 160 mM Cl−. However, we found that although JNa-H was 16 ± 5 μM min−1 (n = 4) in fibers dialyzed with 160 mM glutamate, it was no higher in fibers dialyzed with either 160 mM gluconate (JNa-H = 0 ± 7, n = 4) or 160 mM sulfamate (JNa-H = 14 ± 4, n = 5). Thus, because glutamate is not inhibitory, Cl− must be stimulatory.
Where Does Cl− Play a Role in Activation of Na-H Exchange?
Because we can use GTPγS, AlF3, or CTX to activate the heterotrimeric G protein that ultimately activates the Na-H exchanger, we are in a position to ask whether the Cl−-dependent step in the shrinkage signal-transduction cascade is after the G protein. Our approach was, first, to verify that complete Cl− removal (i.e., removing Cl− from both DF and ASW) does indeed block Na-H exchange, and then to determine whether the aforementioned G-protein activators are capable of stimulating Na-H exchange in the absence of Cl−.
Effect of complete Cl− removal on Na-H exchange under normotonic conditions.
We Cl− depleted fibers by exposing them to a Cl−-free ASW and dialyzing them with a Cl−-free DF for a minimum of 160 min, and an average of ∼180 min. In the experiment shown in Fig. 4, we pretreated the fiber with SITS, and then dialyzed for 120 min with a pH-7.2 DF that was free of both Na+ and Cl−. During this time, the ASW was also free of Na+ and Cl−. We then switched the pH of the DF to 6.5 for an additional 60 min to acidify the cell. Fig. 4 picks up the experiment during this latter period of dialysis. After we halted dialysis, pHi stabilized (Fig. 4, ab). Exposing the cell to an ASW containing 50 mM Na+ did not significantly alter the trajectory of pHi (Fig. 4, bc). Neither was the pHi trajectory affected by applying 1 mM amiloride (Fig. 4, cd). In a total of six similar experiments, the mean JNa-H was −3 ± 7 μM min−1 (see Table I), which is not significantly different from zero. Thus, Cl− depletion completely blocks Na-H exchange under normotonic conditions.
Effect of GTPγS on Na-H exchange in Cl−-depleted cells.
We had previously shown that, in the presence of Cl− and under normotonic conditions, GTPγS activates Na-H exchange in barnacle muscle fibers (Davis et al., 1992a). To determine whether GTPγS also activates the exchanger in Cl−-depleted fibers, we performed a series of experiments similar to the one shown in Fig. 5,A. Our protocol was the same as for Fig. 4, except that the DF contained 1 mM GTPγS for the final ∼95 min of dialysis. In these experiments, the duration of dialysis with the Cl−-free DF, before the introduction of GTPγS, was as long as 145 min, and averaged 135 min. Fig. 5,A picks up the experiment during the latter part of dialysis with the pH-6.5 DF containing GTPγS. When dialysis was halted, the pHi continued to drift downward (Fig. 5,A, ab) in this particular experiment. Exposing the cell to a Cl−-free ASW containing 50 mM Na+ produced a substantial alkalinization (Fig. 5,A, bc) that was blocked by amiloride (Fig. 5,A, cd). For the eight fibers in this study, the mean JNa-H was 164 ± 23 μM min−1 (see Table I), which is significantly greater than the above value for Cl−-depleted cells in the absence of GTPγS, −3 ± 7 μM min−1 (P < 0.0001). Thus, even in Cl−-depleted fibers, GTPγS markedly stimulates Na-H exchange.3
Effect of AlF3 on Na-H exchange in Cl−-depleted cells.
To obtain further evidence that G -protein activation will stimulate Na-H exchange even in the absence of Cl−, we examined the effect of introducing AlF3 into Cl−- depleted fibers. Our protocol was similar to that in Fig. 5,A, except that AlF3 (10 mM KF plus 100 μM AlCl3) replaced GTPγS for the final 100 min of dialysis. Before the introduction of AlF3, muscle fibers in this group of experiments were Cl− depleted for as long as 175 min, and an average of 140 min. Fig. 5,B shows the terminal portion of dialysis with the pH-6.5, AlF3-containing DF. When we halted dialysis, pHi quickly stabilized (Fig. 5,B, ab). However, when we exposed the fiber to a Cl−-free ASW containing 50 mM Na+, pHi rose very rapidly, as was the case for cells dialyzed with GTPγS (see Fig. 5,A). The rapid pHi increase in the AlF3-dialyzed cells was greatly inhibited by amiloride (Fig. 5,B, cd). In a total of six experiments, the mean JNa-H was 296 ± 28 μM min−1 (see Table I), significantly greater than the control flux of −3 ± 7 μM min−1 in Cl−-depleted cells not dialyzed with AlF3 (P < 0.0001). Thus, AlF3 markedly stimulates Na-H exchange, even in Cl−-depleted fibers.
Effect of cholera toxin on Na-H exchange in Cl−-depleted cells.
In previous experiments from this laboratory, we had shown that injecting CTX into BMFs the day before the experiment stimulates Na-H exchange. These previous studies were performed on cells dialyzed with 34 mM Cl− and superfused with an ASW containing Cl− (Davis et al., 1992a). To determine whether CTX can activate Na-H exchange in the absence of Cl−, we needed to modify the protocol from the previous study so that the Cl− depletion, the CTX injection, and the Na-H exchange assay could all be done on the same day. Because the total time between CTX injection and Na-H exchange assay in the present study would be substantially less than in the previous one, we increased the final intracellular CTX concentration to 3 × 10−6 M. The protocol for the first portion of the experiment was similar to those shown in Fig. 5. The differences are detailed in methods; for example, the microelectrodes were not inserted until after CTX injection. In the experiment shown in Fig. 6,A, we exposed a fiber to a pH-7.2 DF that was free of both Na+ and Cl−. The Cl− depletion time before injection with CTX was as long as 170 min, and averaged 160 min. After we injected the CTX and inserted the microelectrodes, we continued to dialyze with the pH-7.2, 0-Cl− DF for an additional ∼110 min. We then switched to a DF with a pH of 6.5 to acidify the cell. After halting dialysis, we allowed the fibers to incubate for an additional 120 min before assaying for Na-H exchange. Thus, we assayed the fibers ∼290 min postinjection. Fig. 6,A picks up the experiment ∼75 min after dialysis had been halted. Exposing the cell to a Cl−-free ASW containing 50 mM Na+ caused an increase in pHi (Fig. 6,A, bc) that was reversed by amiloride (Fig. 6,A, cd). The mean JNa-H for five similar experiments was 100 ± 35 μM min−1 (see Table I), significantly greater than the control flux of −3 ± 7 μM min−1 in Cl−-depleted cells not injected with CTX (P < 0.02).
In the above CTX experiments, the JNa-H of ∼100 μM min−1 was substantially less than in comparable experiments with GTPγS (164 μM min−1) or AlF3 (296 μM min−1). One reason JNa-H may have been relatively low in the CTX experiments is that, even though we assayed ∼290 min after injecting the CTX, we may not have allowed enough time for the CTX to have its maximal effect. Therefore, we asked whether, under our assay conditions (Fig. 6,A), CTX would activate Na-H exchange similarly in the presence and absence of Cl−. To answer this question, we performed a second series of experiments, identical to that shown in Fig. 6,A, except that the DF contained 34 mM Cl−, and the ASW contained 558 mM Cl−. Because increasing values of [Cl−]i increase Na-H exchange, we chose a DF with a relatively low [Cl−] to minimize “background” Na-H exchange activity; that is, Na-H exchange independent of CTX. As with Cl−-depleted cells (see Fig. 6,A), cells dialyzed with 34 mM Cl− and injected with CTX exhibited significant Na-H exchange (see Fig. 6,B). For five experiments, the mean JNa-H was 135 ± 38 μM min−1 (see Table I). The background Na-H exchange rate for 34 mM Cl− (see discussion of Fig. 1) was ∼23 μM min−1. Subtracting 23 from 135 μM min−1 produces a CTX- dependent JNa-H of 112 μM min−1 for Cl−-containing cells. This figure is indistinguishable from the JNa-H in CTX fibers that were Cl− depleted (i.e., 100 μM min−1). Thus, under the conditions of our experiments, CTX produces a similar activation of Na-H exchange in the presence and absence of Cl−.
Role of Intracellular Cl− in the Activation of Na-H Exchange
The first indication that Cl− may play a role in the activation of the Na-H exchanger was Parker's observation, made in dog erythrocytes, that the Na-H exchanger fails to respond to cell shrinkage when extra- and intracellular Cl− is replaced with either thiocyanate or nitrate (Parker, 1983). However, Cl− is not required for Na-H exchange per se. When Parker activated the Na-H exchanger by shrinking the cells in the presence of Cl−, and then fixed the cells by briefly exposing them to glutaraldehyde, the Na-H exchanger remained activated even in the absence of Cl− (Parker, 1984). Similarly, Motais et al. (1989) found that Cl− is required for the cAMP-dependent activation of Na-H exchange by isoproterenol in trout erythrocytes. In these cells, once Na-H exchange was activated by cAMP in the presence of Cl−, the exchanger remained active even after NO−3 replaced Cl− (Motais et al., 1989). Thus, two signal-transduction processes leading to activation of the Na-H exchanger require Cl−, even though the exchanger itself does not require Cl−. More recently, a Cl−-dependent acid–base transporter, presumably an amiloride-resistant Na-H exchanger, has been found in the apical membrane of colonic crypt cells (Rajendran et al., 1995). There is one example in which Cl− inhibits Na-H exchange. In salivary acinar cells, carbachol stimulates Cl− channels, leading to Cl− efflux and cell shrinkage. The simultaneous presence of carbachol, a decreased [Cl−]i and cell shrinkage results in activation of Na-H exchange (Foskett, 1990; Robertson and Foskett, 1994).
Cl− dependence of Na-H exchange under normotonic conditions.
Perhaps our most unexpected result was that, even under normotonic conditions, the Na-H exchanger is markedly stimulated by increasing [Cl−]i. As summarized by the open symbols in Fig. 3, increasing [Cl−]i from its “normal” level of 34 mM to 194 mM caused JNa-H to increase six-fold. This result suggests that the exchanger may be the target of a Cl−-dependent signal-transduction system even under normotonic or “basal” conditions. If this were the case, then one might predict that reducing [Cl−]i to zero would eliminate Na-H exchange under normotonic conditions, a prediction verified in Fig. 4. Thus, it appears that, at least in barnacle muscle, Cl− is required to elevate the Na-H exchanger to even its basal level of activity, and to maintain this basal activity.
Cl− dependence of Na-H exchange under hypertonic conditions.
Three aspects of the data obtained under hypertonic conditions are noteworthy. First, as was the case under normotonic conditions, increasing the [Cl−]i produces a graded increase in JNa-H (Fig. 3, •). Second, at any given [Cl−]i, the JNa-H is always greater under hyper- than under normotonic conditions. For example, at a [Cl−]i of 114 mM, JNa-H was 89 μM min−1 under normotonic conditions (Fig. 3, a). Increasing the osmolality, while holding [Cl−]i fixed at 114 mM, would be expected to increase JNa-H approximately twofold to ∼191 μM min−1 (interpolated point b). Thus, an increase in [Cl−]i cannot be the primary signal for triggering the shrinkage-induced increase in JNa-H. Third, shrinkage not only increases JNa-H by shifting the exchanger from the normo- to the hypertonic curve in Fig. 3, it also increases JNa-H because the loss of cell water increases [Cl−]i. Thus, in a cell dialyzed to a [Cl−]i of 114 mM, increasing the osmolality from 975 to 1,600 mosM/kg, will increase [Cl−]i to 187 mM. As shown in Fig. 3, this elevation in [Cl−]i would increase JNa-H by nearly 45%, from ∼191 μM min−1 (Fig. 3, b) to 275 μM min−1 (Fig. 3, c). Thus, the increase in [Cl−]i that accompanies shrinkage is an auxiliary shrinkage signal.
As summarized in Fig. 3, the K m values for intracellular Cl− are similar under normo- and hypertonic conditions (127 vs. 112 mM). This observation is consistent with the notion that Cl− plays similar roles in activating the Na-H exchanger at normal and low cell volume. We propose the following model for barnacle muscle fibers: the “shrinkage signal,” which leads to activation of the Na-H exchanger, is amplified in such a way that the gain of the hypothetical amplifier increases with increasing [Cl−]i. In euvolemic cells, the basal shrinkage signal is small, but greater than zero. At a “normal” [Cl−]i of 34 mM, the amplification of this small signal is weak. Thus, the combination of a euvolemic cell and a [Cl−]i of 34 mM produces a very low JNa-H (i.e., 23 μM min−1).4 However, raising [Cl−]i to 194 mM in a euvolemic cell increases the amplification of even this weak shrinkage signal, producing a modestly high JNa-H (i.e., 138 μM min−1). In shrunken cells, the shrinkage signal is large. However, at a normal [Cl−]i of 56 mM (produced by shrinking a cell with an initial [Cl−]i of 34 mM), the amplification is weak, producing a rather modest JNa-H (i.e., 86 μM min−1). However, the combination of a shrunken cell and a [Cl−]i increased to 318 mM produces a robust JNa-H (i.e., 345 μM min−1).
Activation of Na-H by GTPγS, AlF3, and CTX in the Absence of Cl−
In previous work, we showed that, at a pHi of 6.8, Cl− is required for the shrinkage-induced activation of Na-H exchange (Davis et al., 1994). Studying barnacle muscle fibers at a pHi of ∼7.2, we also showed that a G protein is involved in this process (Davis et al., 1992a). In particular, we showed that dialyzing with GDPβS blocks the shrinkage-induced activation of the exchanger. The present study extends the previous G -protein work by demonstrating that GTPγS and CTX both activate the exchanger at pHi 6.8. In addition, we extend the earlier work by demonstrating that AlF3 also activates the exchanger. Because AlF3 stimulates heterotrimeric G proteins, but not low-molecular-weight G proteins (Kaziro et al., 1991; Bigay et al., 1985), our new observations with AlF3 imply that activation of a heterotrimeric G protein can activate the exchanger.5 In earlier work, we showed that activating the PKA or PKC pathways fails to stimulate the exchanger (Davis et al., 1992a). Thus, the most straightforward explanation for our data is that the signal-transduction cascade triggered by cell shrinkage includes a heterotrimeric G protein.
One of the goals of the present study was to determine the order of the Cl−-dependent and G -protein steps in the shrinkage signal-transduction cascade. One possibility is that Cl− acts at a step somewhere in the G -protein cycle. Indeed, Cl− appears to increase the affinity of αo for GTPγS (Higashijima et al., 1987). An increase in affinity of α for GTP would stabilize α in the GTP-bound or active state. A Cl− requirement of the G protein in the signal-transduction cascade could explain the Cl−-dependent activation of Na-H exchange in dog and trout RBCs, and in BMFs. To determine the order of the Cl−-dependent and G-protein steps in our BMF experiments, we attempted to activate Na-H exchange with GTPγS, AlF3, and CTX in Cl−-depleted cells. As noted above, we found that Cl− depletion blocks Na-H exchange under normotonic conditions, at pHi ≅ 6.8 (Fig. 4). Even in such Cl−-depleted cells, introducing GTPγS (Fig. 5,A), AlF3 (Fig. 5,B), or CTX (Fig. 6) activates the Na-H exchanger. Because each of these three agents acts at the level of the G protein,6 and is nevertheless able to bypass the blockade introduced by Cl− removal, we can conclude that the Cl−- dependent step must precede or be concurrent with the G -protein step.
We thank Dr. Catherine Berlot for helpful discussions, Drs. Raphael Zahler and Gordon Cooper for assistance in performing the curve fitting, Mr. Duncan Wong for computer programming and assistance in preparing the figures, and Mr. Francisco Rodriguez for technical assistance. We also thank Dr. Mark Bevensee for help with the final revisions.
This work was supported by National Institutes of Health grant NS18400.
Portions of this work have been published in preliminary form (Hogan, E.M., B.A. Davis, and W.F. Boron. 1995. Biophys. J. 9:A356).
Abbreviations used in this paper: ASW, artificial seawater; BMF, barnacle muscle fiber; CTX, cholera toxin; DF, dialysis fluids; NMDG+, N-methyl-d-glucammonium.
The slope of the pHi recovery in bc reflects not only the Na-H exchange rate, which tends to increase pHi, but also the presence of background acid loading, which tends to decrease pHi. In experiments in which the cell was later exposed to a hypertonic solution (c–f), it was not possible to inhibit the bc pHi increase with amiloride because the amiloride is not fully reversible. Thus, we can draw no conclusions about the Na-H exchange rate in bc in this experiment.
Because we added the 1 mM GTPγS as the tetra-lithium salt, the DF containing GTPγS also contained 4 mM Li+. Although previous work from this laboratory (Davis et al., 1992b) showed that 100 mM Li+ stimulates Na-H exchange, here we found that dialyzing with 4 mM Li+ (in the absence of GTPγS) fails to stimulate Na-H exchange at pHi ∼7.2 (not shown).
In principle, introducing GTPγS could also result in the de novo generation of ATPγS within the muscle fiber. Such ATPγS could lead to the generation of long lived phosphoproteins. However, our demonstration that cholera toxin and AlF3 also activate the Na-H exchanger argues that, regardless of a possible phosphorylation, activation of a G protein can lead to activation of the exchanger.
We would also predict that this “basal” shrinkage signal is also G-protein dependent so that dialyzing with GDPβS should have the same effect as dialyzing with a Cl−-free solution: inhibition of basal Na-H exchange.
A recent report shows that AlF3 can bind to a ras-GAP-GDP complex, inducing a shift in the absorbance spectrum of a fluorescent GDP analog (Mittal et al., 1996). The crystal structure of this ras-GAP-GDP-AlF3 complex has been solved (Scheffzek et al., 1997). However, there is no evidence that AlF3 activates a ras-related protein.
Address correspondence to Walter F. Boron, Department of Cellular and Molecular Physiology, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06520. Fax: 203-785-7678; E-mail: firstname.lastname@example.org