The Roles of Sodium Transport and Anion Permeability in Generating Transepithelial Potential Differences in Larval Salamanders^*

Transepithelial potential differences (TEPs) have been studied for many years, and our understanding of the ionic mechanisms underlying them is derived largely from studies on isolated frog skins and toad urinary bladders bathed on both sides by Ringer’s solution. Little attention has been paid to the origin of TEPs in intact animals bathed by dilute solutions. Barker-Jørgensen, Levi & Zerahm (1954) showed that when frogs were immersed in 3 –5 mM NaCl the TEP was 50–100 mV. body fluids positive to bathing solution. Lower potentials, the polarities of which were variable, were found when the animals were immersed in KC1. Brown (1962) found that the TEP in intact frogs depended on environmental NaCl concentration and that its concentration dependence resembled that of isolated skins. These observations are consistent with a role for active salt transport in generating the TEP, and suggest that the model developed for the in vitro preparation (Koefoed-Johnsen & Ussing, 1958; Ussing, 1960) may be applicable to the intact animal in its normal environment. However, some flux data obtained on animals in dilute solutions are hard to reconcile with this model. For example, Krogh (1937) showed that salt-depleted frogs would take up chloride from solutions with a non-penetrating cation (K⁺), while the model provides no EMF for anion transfer in the absence of sodium transport. Krogh’s animals also absorbed sodium from solutions with non-penetrating anions. Under these conditions the model requires that the TEP approach the EMF of the transport mechanism but that no net ion uptake occur. The observations were confirmed by Barker-Jørgensen et al. (1954), who showed that even when NaCl was absorbed the movement of neither ion satisfied the flux ratio criterion for diffusion. Hence, chloride appears to be actively transported into the animal from dilute solutions. These data suggest that the ionic basis for electrogenesis in dilute solutions needs further investigation.

Many of these observations have also been made in larval salamanders. The present investigation explores the relationship between ion uptake and electrogenesis in salamanders in dilute solutions. Special attention has been paid to establishing whether active sodium transport is part of the electrogenic mechanism and to the role played by anion absorption.

The experimental animals were larval salamanders, Ambystoma tigrinum captured in local streams. A few experiments were conducted on A. gracilus. and no inter-specific differences were noted. Most of the animals weighed between 10 and 50 g. They were stored in porcelain tubs in a lighted room at 15 ± 1° C. The storage medium was local tap-water.

Experimental animals were removed from the holding tanks and equilibrated with an appropriate experimental solution for a week before measurements were made. Most of the animals were equilibrated with artificial ‘pond water ‘containing 1–2 IDM/1. NaCl (Alvarado & Kirschner, 1963). One group was salt-depleted by immersing them in deionized water which was changed daily. This treatment augments active salt transport, and on being returned to pond water these animals absorb NaCl 2–3 times as rapidly as normal (i.e. pond-water-adapted) animals (Alvarado & Kirschner, 1961). Another group of animals was adapted to 40 mM/1. NaCl before being used. This depresses the transport mechanism (unpublished experiments). The animals were used to study changes in TEP under condition of normal, augmented and depressed ion transport. The salt-depleted animals were also used to demonstrate independence of Na⁺ and Cl⁻ transport systems.

Each salamander was anaesthetized with tricaine methane sulphonate (0·1 %). The skin and body wall were punctured with a dissecting needle, and a fine polyethylene tube filled with agar-3 M-KC1 was inserted into the body cavity. The animal was then transferred to the experimental solution which contained 0·03 % tricaine methane sulphonate. A second agar-KCl-filled bridge dipped into the bathing solution. Both bridges were connected through calomel electrodes to a recording potentiometer with an input resistance of 100 kΩ at balance. Potentiometric precision was about ± 1 mV. At the end of each experiment electrode asymmetry was measured by connecting the body fluids of the animal through a saturated KC1 bridge to the bathing solution. All potentials reported below have been corrected for the small asymmetries noted.

Many experiments were run in which animals were initially placed in distilled water, and the potential differences across the skins were recorded when NaCl was added to the bath. Although these experiments were adequate to demonstrate dependence of the TEP on external salt concentration, they had the disadvantage that both anion and cation concentration varied simultaneously. Since one of the goals of this investigation was to assess the role of each ion in electrogenesis, an alternative procedure was used in all later work. This involved using 5 mM/1. K₂SO₄, rather than distilled water, as the initial solution. Evidence to be presented below shows that both K⁺ and SO₄²⁻ have a low mobility in the skin. Theassumption underlying its use is that a non-penetrating salt should have no electrical effect on the animal. This is borne out by the fact that the transepithelial potentials in distilled water and in 5 mM/1. K₂SO₄ fell into the same range, 5–20 mV., body fluids negative to the external solution. By using Na₂SO₄ it was possible to vary the sodium concentration through the range 0·1−10mM/1. without changing the anion composition or the ionic strength of the medium. The effects of (NH₄)₂SO₄, Li₂SO₄, Ce₂SO₄ and MgSO₄ were also tested. Chloride concentration was varied through the same range by using KC1. In all experiments the animal was placed in the initial bathing solution and allowed to remain until the TEP was stable. Studies on the dependence of potential on ion concentration began by replacing a measured volume of bathing solution with an equal volume of 5 mM/1. Na₂SO₄ or KC1. When the new TEP was stable it was recorded. This procedure was repeated to give a series of readings through the range 0·1−10 mM/1. with each step increasing the test ion concentration about threefold. Concentrations were measured rather than calculated, because intermittent urination provided an uncontrolled source of ions. Sodium, potassium, and calcium were determined with a direct-reading flame-photometer. Precision of these measurements was about ± 1 %. Chloride was determined by silver titration with a Cotlove ‘chloridimeter’.

Net fluxes of Na⁺, K⁺ and Cl⁻ were measured on animals previously salt-depleted for i week. Each animal was placed in a plastic container with a known volume of solution, initially 0·6 HIM/1. Na₂SO₄ for one group and 1·2 mM/1. KC1 for another. Samples were removed periodically for analysis. After about 24 hr. each animal was transferred to pond water (1·2 mM/1. NaCl) and aliquots were taken for another 24 hr. Changes in concentrations in the bath were estimated by flame-photometry (Na⁺, K⁺) or silver titration (Cl⁻). Net uptake or excretion was calculated from these changes. Chloride determinations were not made on Na₂SO₄ solutions because of sulphate interference.

Sulphate permeability was measured by placing animals in media made from Na₂³⁵SO₄ and measuring disappearance of isotope from the bathing solution. Aliquots were plated on aluminum planchets with a drop of detergent. After drying they were counted with an ultrathin-window GM tube.

The ammonia-excretion experiments were carried out in cylindrical Lucite chambers stirred and aerated by a stream of compressed air. The pH of the bathing solution was adjusted to 6·8 to minimize distillation of ammonia in the air stream. Such losses were shown to be negligible by leading the air through a trap containing boric acid. No ammonia appeared in trap even after 6 hr. of aeration. Ammonia was determined by direct Nesslerization of aliquots removed from the bathing solution. Optical density was measured at 430 mμ.

All experiments were carried out at a temperature of 15 ± 1° C.

When salamanders are in a steady state in pond water about two-thirds of their salt loss is renal, hence the flux ratio across the body surface must be about three for both ions. Since blood [Na⁺] is about 96 mM/1. a TEP of 137 mV., body fluids negative, would be necessary to sustain a sodium steady state by diffusion. For chloride a TEP of 137 mV., body fluids positive, would be required. In several hundred measurements on animals in pond water the body fluids have always been positive, but values never exceeded 35 mV., and most were in the range 10–20 mV. Therefore, both ions must be actively transported into the animal under these conditions.

However, these data provide no information on whether the ion movements are coupled through a single mechanism or are independent. To answer this question the following experiment was performed. Four animals were salt-depleted for a week. Two were transferred to 1·2MM/L. KC1 and net movements of Na⁺, K⁺ and Cl⁻ were measured for 23 hr. The other pair were placed in 0·6 mM/I. Na₂SO₄ and net transfers of Na⁺ and K⁺ were measured. At the end of this period all four animals were transferred to pond water and ion movements were followed for another 20·5 hr.

Table 1 shows the total quantity of each ion transferred in each experimental period. It is clear that CL uptake from KC1 occurred. A net loss of sodium occurred during this period. Although unidirectional ion fluxes were not measured here,^*the data show that chloride uptake can occur without simultaneous sodium transport. Uptake of K⁺ occurs in some animals but is absent in many. Even when it can be demonstrated (as in Table 1) it is never equal to Cl⁻ uptake. When these animals were transferred to normal pond water, influx of Na⁺ markedly exceeded that of CL. There was a net loss of K⁺.

Animals placed in Na₂SO₄ showed a net uptake of Na⁺ while losing some K⁺. Chloride was not measured, but none was present initially and a net loss must have occurred. When this pair was transferred to pond water, Cl⁻ uptake was much greater than Na⁺ uptake.

Since Na⁺ and Cl⁻ appear to move independently it is necessary to account for the preservation of electrostatic neutrality. In the case of Na⁺ uptake an alternative to anion absorption is exchange with another cation. Fig. 1 shows that animals adapted to pond water excrete ammonia continually in that medium. During the first 2 hr. the untreated animals excreted more ammonia than those with cloacas sutured to prevent urination, but in the period 2·5–7·25 hr. average ammonia efflux was comparable in both groups: 0·52 μmole (10 g.⁻¹ hr.⁻¹ for untreated animals; 0·60 μmole (10 g.)⁻¹ hr.⁻¹ for animals with stitched cloacas. Since the high initial output was seen only in the control animals it probably is due to urinary ammonia. But the sustained efflux in both groups must represent movement across the body surface, and hence may be part of a cation exchange system.

No flux experiments relating to electrostatic neutrality during chloride uptake were conducted.

Since several functions would describe the data at least as well (a rectangular hyperbola would provide a better fit at lower concentrations), no theoretical significance is attributed to equation (1). However, it generates two easily assessed parameters useful in characterizing the effects of experimental variables. One of these is the TEP at [Na⁺] = 1 mM/1. (a₀), a concentration approximating that encountered by the animal in nature; the other is the slope of the line (b) which expresses the potential change per tenfold concentration change.

Abolition of the TEP in sodium-free solution, and its concentration dependence, both could be explained if the potential were generated by sodium transport. Such a working hypothesis suggests that animals with augmented ion transport should generate high TEPs while those with inactive transport systems should show low TEPs. Three groups of animals were prepared as described under Methods: (1) one group was salt-depleted to stimulate transport; (2) a ‘normal’ group was adapted to pond water; (3) transport was depressed in the third group by pre-adaptation to 40 mM/1. NaCl. The relationship expressed in equation (1) was examined, beginning with the animals in K₂SO₄ (5 mM/I.) as described under Methods. Thus, the anion concentration and total ionic strength of the medium remained constant; only [Na⁺] varied. Measurements on all animals in each group were pooled, and a₀ and b were estimated by the method of least squares. Table 2 summarizes the results, and establishes two facts. First, since b is positive in each case there is a positive correlation between TEP and [Na⁺] in all three groups. Second, animals with increased sodium uptake generated larger potentials than animals adapted to pond water. The difference between a₀ values is highly significant (P < 0·001), although no significance can be ascribed to the apparent differences in slopes (P > 0·1). Conversely, animals pre adapted to high salinities show much smaller TEPs over the same concentration range. In this case both and b are different (P < 0·001) from their values in animals adapted to pond water.

Chemical specificity of the cation requirements was examined by using a series of alkali and alkaline earth metals. Li+ generated a TEP whose concentration dependence was similar to that described by equation (1). A small TEP also appeared in the presence of low concentrations (0·1−0·3 mM/l.) of Ca²⁺ even in Na⁺-free solutions, but this potential was little changed by further increasing [Ca²+], Fig. 3 shows the TEPs generated by Ca²⁺ and Na⁺ in the same animal, and the differences in magnitude and concentration dependence are apparent. The small increase in TEP at higher [Ca²⁺] was probably due to a gradual rise in [Na⁺] through the experiment (shown by the figures in parentheses near each experimental point on the calcium curve). K⁺, Rb⁺, Cs⁺, Mg²⁺ and NH₄+ were also tested; none of them generated a TEP over this concentration range.

The metabolic inhibitors azide and dinitrophenol abolished the TEP. Fig. 4 shows the time-course of the dinitrophenol effect in one animal, as well as the fact that it acts reversibly. Azide inhibition was irreversible. Strophanthin had no effect on the TEP when added to the bath in concentrations to 10 μg./ml. or when injected into the animal.

NaCl had the same effect on the TEP as Na₂SO₄. This was shown in an experiment similar to that described in Table 2, except that aliquots of the initial K₂SO₄ solution were replaced with NaCl. Table 3 shows the mean a₀ and b values for salt-depleted, normal, and salt-loaded animals. In this experiment, too, TEPs were largest for the salt-depleted animals and smallest for the salt-loaded group. Statistical comparison of the parameters for each group in Table 3 with those for the corresponding group in Table 2 shows that there is no significant difference between them. Thus, the magnitude of the potential difference generated at any sodium concentration over this range is independent of the anion used.

These data suggest that there may be little difference between the diffusion mobilities of chloride and sulphate ions in the salamander skin. Sulphate permeability was measured by placing animals in artificial pond water made from Na₂^S6SO₄ and measuring the disappearance of isotope from the bath. Table 4 shows that no measurable uptake occurred; in another experiment a very small uptake (0·01–0·02 μ-equiv. (10 g.)⁻¹ hr.⁻¹) was noted. Net sodium uptake by animals under these conditions is in the range 0·5–1·0 μ-equiv. (10 g.)⁻¹ hr.⁻¹. Thus, skin permeability to sulphate is low, and this implies that leak permeability to chloride is also small.

Ussing (1949) has shown that dilute Cu²⁺ added to Ringer’s solution bathing the outside of the isolated frog skin augments the TEP by decreasing chloride permeability. Fig. 5 shows that low concentrations of cupric ions have little effect on the TEP generated by salamanders in pond water. A transient period of depression occurred as described by Ussing, but even after 15 hr. of exposure the potential difference was within a few millivolts of its initial value. Higher concentrations of the metal only depressed the potential. Lack of copper effect is also consistent with low chloride permeability in the untreated animal under these conditions.

Fig. 6 shows the results of an experiment in which the K₂SO₄ solution was replaced by KC1. In the absence of Na⁺ wide variations in [Cl⁻] had little effect on the TEP. The small change noted was almost certainly a concomitant of an increase in bath [Na⁺], probably from urination ([Na⁺] is shown by the figures in parentheses above the curve). In spite of this the internal electrode never became positive.

It is clear that when salamanders are in an ionic steady state in pond water both Na⁺ and Cl⁻ are actively transported inward. In addition, the data in Table 1 show that the transport systems can operate independently of each other. The fact that uptake of each ion in this experiment was not balanced by absorption of a counter ion can only mean that electrostatic neutrality must be maintained by a pair of ion-exchange systems. Fig. 1 shows that these animals excrete ammonia across the body surface. In 3 experiments involving 13 salamanders ammonia excretion was measured for periods of 5–7 hr. The mean efflux was 0·71 ± 0·27 (S.D.) μmole (10 g.)⁻¹ hr.⁻¹. Sodium uptake was not measured in these animals but a mean value of 1·3 μmole (10 g.)⁻¹ hr.⁻¹ was reported by Alvarado & Kirschner (1963). Thus, ammonia excretion is somewhat lower than sodium uptake, but would provide electrostatic balance for much of it. Na⁺/NH₄⁺ exchanges have been reported in the frog (Krogh, 1937), in the goldfish (Maetz & Garcia-Rameau, 1964), and in the crayfish (Shaw, 1960b), and ammonia is also excreted by Necturus maculosus (Fanelli & Goldstein, 1964). All of these animals have the same ionic regulatory problems as the larval salamanders used in our work, and it appears possible that such an exchange system operates to maintain electrical neutrality during Na⁺ uptake in all fresh-water animals. An alternative possibility in Na⁺/H+ exchange, with NH₃ diffusing out of the epithelial cells to maintain a constant cytoplasmic pH.

Our data leave little doubt that active sodium transport plays a central role in electrogenesis under the conditions studied. Of the alkali and alkaline earth metals tested only Na⁺ and Li+ generated a concentration-dependent TEP. The effective concentration range, the size of the TEPs and nature of the relationship, including its maximum at 20–30 mM/1., have also been described in the intact frog (Brown, 1962) and in isolated frog skins (Steinbach, 1933; Linderholm, 1952). In isolated frog skins the same chemical specificity and concentration dependence are characteristic of the sodium-transport mechanism, and provide part of the evidence that it is involved in generating the TEP (Kirschner, 1955; Ussing & Zerahn, 1951). The correlation between transport rate and TEP, suggested by the concentration dependence, was confirmed by comparing TEPs in salt-depleted (augmented transport), normal, and salt-loaded (depressed transport) animals.

Depression of the potential by metabolic inhibitors may also be a concomitant of transport inhibition. However, these compounds also effect the diffusion permeability of some preparations, for example the isolated frog skin (Nakajima & Hashimoto, 1960 ; confirmed by unpublished experiments in our laboratory). Such an increase in passive ionic conductance would also tend to depress the potential, and this makes interpretation of inhibitor studies equivocal. However, the results reported here are compatible with participation of sodium transport in electrogenesis.

Lack of effect of strophanthin in the TEP was the only observation apparently inconsistent with a direct role for the sodium pump. Addition of the inhibitor to the external bathing solution may have prevented it from reaching the transport locus, but injection of the compound in quantities sufficient to give blood concentrations of 1–10 μg./ml. were also ineffective. These observations are in contrast with similar studies showing that strophanthin inhibits sodium transport across many epithelia (Nakajima, 1960).

Our experiments also show that salamander skins have a low leak conductance for chloride in dilute solutions. Cl⁻ uptake is thermodynamically active, and can occur independently of simultaneous cation absorption. Fig. 6 shows that Cl⁻ uptake is not electrogenic. These observations suggest that a tightly coupled exchange against another anion must be involved. We have no evidence concerning the identity of the exchangeable anion. A C1⁻/HCO₃⁻ exchange across the body surface has been described in frogs (Krogh, 1937), crayfish (Shaw, 1960a) and fresh-water teleosts (Maetz & Garcia-Rameau, 1964). Such exchanges are also well known in other epithelia, and the ready availability of CO₂ makes bicarbonate a probable member of such an exchange system in the salamander.

Thus, the TEP in larval salamanders in dilute saline appears to be generated by active Na⁺ transport with electrostatic neutrality maintained by cation exchange. External anions play no role in electrogenesis. The experiments of Barker-Jørgensen et al. (1954) and Krogh (1937) show that this may also be the case in frogs. Yet there is little doubt that isolated frog skins bathed by Ringer’s solution have an appreciable diffusion permeability to Cl⁻ and that influx of this ion maintains electrostatic neutrality during Na⁺ transport. Either isolation of the skin or the use of Ringer’s solution outside seems to cause a change in leak permeability to chloride. This change is being investigated.

1. Larval salamanders in pond water actively transport both Na⁺ and Cl⁻ inwards. The two fluxes can occur independently, indicating that they are not linked obligatorily through a single mechanism. Ammonia is excreted extrarenally at rates comparable with active Na⁺ influx.

2. A potential difference exists across the body surface of larval salamanders under these conditions. Over the range 0·1−10mM the transepithelial potential (TEP) varies approximately logarithmically with [NaCl]. Only Na⁺ and Li⁺ salts generate such a TEP, while K⁺, Rb+, Ca²⁺, Mg²+, and NH₄+ salts have little effect. Azide and dinitrophenol abolish the TEP, while strophanthin has no effect. When active Na⁺ uptake is enhanced by salt-depletion, higher TEPs are generated. Conversely, animals with Na⁺ transport depressed by salt-loading show significantly smaller TEPs.

3. The magnitude of the TEP is uneffected by replacing NaCl with Na₂SO₄. Permeability of the body surface to SO₄²⁻ is very low, and the similarity in electrical behaviour of the two anions suggests that Cl⁻ penetration by diffusion is also small. This surmise is supported by two observations: (1) KC1 generates no TEP over a wide range of concentrations ; and (2) cupric ion has no effect on the TEP developed in dilute NaCl solution.

4. It is suggested that the TEP in salamanders is generated by active inward transport of Na⁺, but that in dilute solutions electrostatic neutrality is maintained by a cation-exchange system involving ammonium ion. Uptake of Cl⁻ under these conditions appears to occur by an independent mechanism and is non-electrogenic.