Bulk Flow of the Medium and Cutaneous Sodium Uptake in Frogs: Potential Significance of Sodium and Oxygen Boundary Layers

Although the unstirred layer or boundary layer is a ubiquitous component of transmembrane processes, many membrane physiologists have deemed such layers of dubious physiological significance other than as a nuisance to the unambiguous determination of membrane properties (Barry and Diamond, 1984). During the past decade, however, boundary layers have been recognized as having a physiologically significant impact on transmembrane processes in diverse circumstances. For example, in the gut the unstirred layer around the villi is the principal rate-limiting step to the absorption of nutrients (Levitt et al. 1990); in blood vessels the concentration of various agonists in the unstirred layer may be an important means by which endothelial cells sense and respond to shear stress (Dull and Davies, 1991); and unstirred layers are directly responsible for severe local hypoxia at the skin surface in whole animals undergoing cutaneous gas exchange (Pinder and Feder, 1990; Booth and Feder, 1991; Feder and Booth, 1992). The present study investigates whether unstirred layers have a similar impact on transmembrane ion flux in a major model system for membrane studies, the frog skin.

In dilute external media, frogs balance a considerable urinary Na⁺ loss with direct cutaneous uptake of Na⁺ from the medium. Epithelial Na⁺ uptake occurs through amiloride-blockable Na⁺ channels (Lindemann and Van Driessche, 1977). In dilute external media, extrusion of H⁺ by an apical H⁺-ATPase is partially responsible for creating the electrochemical gradient that drives Na⁺ inward; a basolateral Na⁺/K⁺. ATPase also contributes to this gradient and removes Na⁺ from the epithelial cell (Ehrenfeld et al. 1985; Harvey and Ehrenfeld, 1988). As a result of these processes, a transepithelial potential (TEP) develops, which can be highly correlated with Na⁺ influx under physiological conditions (Aceves and Erlij, 1971; Christensen, 1974; Salako and Smith, 1970).

Unstirred layers may be physiologically significant to ion uptake in frog skin in at least two non-exclusive ways. First, Na⁺ uptake is strongly dependent upon [Na⁺] in the external medium at [Na⁺] less than 1mmol l^{− 1} (Greenwald, 1971; Kirschner, 1988). If rates of Na⁺ removal from the unstirred layer (e.g. by cutaneous uptake) exceed rates of Na⁺ replenishment in the unstirred layer from the bulk medium or the skin itself, then [Na⁺] will fall in the boundary layer. If the steady-state [Na⁺] in the unstirred layer falls to less than approximately 1mmol l^{− 1}, Na⁺ uptake may be depressed acutely. Second, if rates of cutaneous O₂ uptake from the unstirred layer exceed rates of O₂ replenishment from the bulk medium, the will fall in the unstirred layer. This decline in can reduce the oxygen consumption of the entire organism and the skin itself (Vitalis, 1990; Burggren and Feder, 1986; Booth and Feder, 1991). Insofar as the uptake of Na⁺ is driven by epithelial ATPases whose activity is limited by ATP supply, hypoxic depression of ATP synthesis could be expected to depress Na⁺ uptake. In theory, many other diffusable species could also affect Na⁺ flux through accumulation or depletion in the unstirred layer. In any event, stirring of the medium or ventilation of the skin should ameliorate or abolish these effects by dissipating the unstirred layer. The present study reports diverse data that are consistent with these expectations for frogs of two species (Rana catesbeiana and Rana pipiens) exposed to different convective regimes and external [Na⁺]. A novel aspect of these findings is the suggestion that oxygen boundary layers have a significant effect on cutaneous Na⁺ transport.

Adult Rana catesbeiana and Rana pipiens were obtained from a commercial supplier (Amphibians of North America, Nashville, TN, USA). While awaiting experimentation, frogs were maintained in tap water derived from Lake Michigan, at room temperature.

Except where otherwise specified, artificial pondwater (Alvarado and Dietz, 1970) was used as a diluent. Because of substantial inter-individual variation in variables under study, whenever possible we examined differences within individual animals exposed to stirred and unstirred conditions and tested for significance with the t-test for paired samples. In other cases we employed the two-sample t-test. We present mean ± S.E. and sample size for measurements.

Rana catesbeiana and Rana pipiens were anaesthetized and prepared for TEP determinations as outlined above. The bath was aerated continuously, but the experimental animal was shielded from convection due to the aeration. When TEP had reached steady-state, the bath was alternately stirred or not stirred with the magnetic stirring apparatus. Unstirred periods were ended when the TEP attained an apparent steady state, typically 10–20min after cessation of stirring. Likewise, stirred periods were ended when the TEP reached an apparent steady state, typically 10–15min after initiation of stirring. Bath [Na⁺] was 0.75–1.0mmol l^{− 1}.

Rana pipiens (N=10; mass 35.6±1.1g) were prepared for Na⁺ influx determinations as outlined above. The initial bath concentration was 1mmol l^{− 1} Na⁺ in artificial pondwater. Beginning 5min after addition of ²²Na⁺ to each stirred bath container, animals were exposed to one of two treatments: 12h without stirring or 12h with stirring. In the unstirred treatment, the bath was stirred for the final 5min of the experimental period. Thereafter, a bath sample was taken and each animal was exposed to the other experimental treatment for an additional 12h period. The order of treatments was reversed in half of the experimental animals.

This experiment resembled experiment B except that animals were anaesthetized before (MS-222, 1:1000) and during (MS-222; 1:10000) experimentation. Flux periods were 6h instead of 12h. The initial bath concentration was 0.4mmol l^{− 1} Na⁺ in artificial pondwater. This concentration was chosen to increase specific activity of ²²Na⁺ during the relatively brief flux periods. Subjects were Rana pipiens (N=9; mass 33.4±1.7g)

Rana pipiens were housed individually in screened containers through which water could circulate. These containers prevented frogs from emerging from the bath but did not hinder pulmonary ventilation or restrict movement within the container. Containers with frogs were assigned to three experimental treatments: (1) deionized water; (2) 0.75mmol l^{− 1} Na⁺ in unstirred artificial pondwater; (3) 0.75mmol l^{− 1} Na⁺ in artificial pondwater continuously stirred through vigorous aeration. Bath volume was 1.25l per frog. The [Na⁺] of the bath was determined daily. After 14 days, frogs were removed from their containers and Na⁺ influx was determined as described above. At that time, their mass was 37.7±1.0g and did not vary significantly among experimental groups.

Initial bath [Na⁺] for flux determinations was 1.8–2.0mmol l^{− 1} and the bath was stirred continuously; these conditions are likely to yield near-maximum rates of Na⁺ influx. Otherwise, conditions for flux determinations resembled those in experiment B.

This experiment resembled experiment A except that the bath was stirred continuously and the treatments alternated among gassing the bath with air, 100% O₂ or 100% N₂. Within 10min of gassing the bath with O₂, the of the bath was greater than 97kPa. Within 10min of gassing the bath with N₂, the of the bath was less than 1.5kPa. Bath [Na⁺] was 1.0mmol l^{− 1}.

This experiment resembled experiment B except that the bath was stirred continuously and animals were exposed to various bath values. In the first group, frogs (N=6, mass 29.9±1.9g) were exposed to 12h of either hyperoxia (bath ⁠) or normoxia while was determined; during the subsequent 12h, the frogs were exposed to the alternative treatment. In a second group, frogs (N=6, mass 32.7±1.6g) were exposed to 12h of either hypoxia (bath or normoxia while was determined; during the subsequent 12h, the frogs were exposed to the alternative treatment. Bath was established by gassing the bath with air, 100% O₂ or 100% N₂. The initial bath concentration was 1.0mmol l^{− 1} Na⁺ in artificial pondwater.

This experiment was performed to assess whether flow regime affected TEP. In each trial, cessation of stirring decreased TEP and resumption of stirring increased TEP (e.g. Fig. 1). The decrease in TEP by 10min after cessation of stirring averaged 6.9mV (±1.1mV S.E.; N=9 trials) in Rana catesbeiana; at the TEP of frogs in stirred water (23.9±4.2mV), this decrease was 34.8±7.0%. The decrease in TEP by 10min after cessation of stirring averaged 5.8mV (±0.5mV; N=10 trials) in Rana pipiens; at the TEP of frogs in stirred water (47.9±3.9mV), this decrease was 13.5±2.1%. Upon resumption of stirring, TEP typically increased rapidly at first and slowly thereafter.

These experiments were performed to assess whether changes in actual paralleled the changes in TEP. Experiment B examined awake animals, which could breathe air and move (i.e. ventilate the skin) ad libitum. Experiment C examined anaesthetized apnoeic animals, which were incapable of ventilating their skins voluntarily and were treated similarly to the animals for which TEP was determined. In both experiments, animals in the stirred bath had significantly greater values than animals in the unstirred bath. In experiment B (initial [Na⁺]=1mmol l⁻¹), decreased from 200.4±33.9 μmolkg^{− 1} h^{− 1} to 144.3±23.1 μmolkg⁻¹ h⁻¹when stirring was halted. For the nine animals for which both stirred and unstirred values of were available, the unstirred rate averaged 85.9% of stirred values, a significant (P<0.05, t-test for paired comparisons) decrease. In experiment C (initial [Na⁺]=0.4mmol l⁻¹), decreased from 85.0±13.9 μμmolkg⁻¹ h⁻¹ to 40.5±11.2 μmolkg⁻¹ h⁻¹ when stirring was ceased. For the eight animals for which both stirred and unstirred values of were available, the unstirred rate averaged 44.5% of stirred values, a significant (P<0.01, t-test for paired comparisons) decrease.

Previous studies have shown that frogs upregulate after prolonged exposure to deionized water, which depletes them of salt (e.g. Greenwald, 1971). This experiment was performed to assess whether prolonged exposure to an unstirred 0.75mmol l⁻¹ Na⁺ bath, in which the skin might contact a salt-depleted boundary layer, has the same effect. During the 14-day exposure to the experimental treatments, actual [Na⁺] of the bath was less than 0.1mmol l⁻¹ in the deionized water group and fluctuated between 0.69 and 0.99mmol l^{− 1} in the other two groups. in the three experimental groups was as follows (Fig. 2): (1) deionized water, 407.1±35.5 μmolkg⁻¹ h⁻¹ (N=10); (2) unstirred 0.75mmol l⁻¹ Na⁺, 417.0±73.3 μmolkg⁻¹ h⁻¹ (N=5); (3) stirred 0.75mmol l⁻¹ Na⁺, 244.9±38.1 μmolkg⁻¹ h⁻¹ (N=8). The mean of the unstirred 0.75mmol l⁻¹ Na⁺ group was 70% greater than that of the stirred 0.75mmol l⁻¹ Na⁺ group (P<0.05, t-test) and did not differ significantly from that of the deionized water group (P=0.89, t-test).

This experiment was performed to assess whether changes in the of a well-stirred bath could directly affect TEP and thus to assess whether a hypoxic boundary layer might affect Na⁺ influx by depressing ATP-requiring transporters. Bath hyperoxia resulted in a modest (7.4%) increase in TEP in one frog (Fig. 3A), but no change in others (N=4). Bath hypoxia, by contrast, resulted in a considerable (9.0±2.3mV; N=6) decrease in TEP by 10min after the initiation of gassing with N₂ (Fig. 3B); at the TEP of frogs in normoxic water (41.8±3.2mV), this decrease was 21.7±5.0%. This decline continued as long as the bath remained hypoxic. Upon gassing the bath with air, TEP rapidly returned to normoxic levels.

This experiment was performed to assess whether -induced changes in actual Na⁺ uptake paralleled the -induced changes in TEP in experiment E. In the first experimental group, was greater in the hyperoxic bath (292.8±22.5 μmolkg⁻¹ h⁻¹; N=6) than in the normoxic bath (198.7±29.3 μmolkg⁻¹ h⁻¹; N=4). For the four animals for which both hyperoxic and normoxic values of were available, the hyperoxic rate averaged 128.1% of normoxic values, a near-significant (P=0.052, t-test for paired comparisons) change. In the second experimental group, was lower in the hypoxic bath (123.9±16.8μmolkg⁻¹ h⁻¹; N=6) than in the normoxic bath (250.5±29.5 μmolkg⁻¹ h⁻¹; N=6). The hypoxic rate averaged 52.3% of normoxic values, a significant (P<0.01, t-test for paired comparisons) decrease.

Numerous studies have examined the significance of environmental ion concentrations for organismal ion homeostasis in aquatic organisms, as well as the physiological mechanisms for maintaining ion homeostasis in the face of often unfavourable environmental ion concentrations in fresh and saline water [Boutilier et al. (1992) and Shoemaker et al. (1992) review the relevant information for amphibians]. The present study suggests that the convective environment of an aquatic organism (i.e. the range of flow velocities that an organism naturally encounters or induces through its own movements and whether these flows are laminar or turbulent) may be just as significant for ion homeostasis as ion concentrations in the bulk medium. Immersion of frogs in an unstirred bath resulted in a 14–65% decline in TEP and/or ⁠, and stirring of the bath alleviated this diminution (experiments A–C). Such changes have the potential to compromise Na⁺ homeostasis. Although the significance of the convective environment for cutaneous gas exchange is well known (Vogel, 1983; Feder and Pinder, 1988; Pinder and Feder, 1990; Booth and Feder, 1991; Feder and Booth, 1992), the present study is the first documentation of parallel phenomena in cutaneous ion exchange.

Exposure to natural or artificial environments of extreme ion concentrations can induce acclimatory or evolutionary adjustments of ion fluxes that compensate for environmental conditions. For example, several weeks’ exposure to deionized water, which ought to reduce ion concentrations in body fluids of amphibians, results in a substantial upregulation of in ranid frogs (Greenwald, 1971). Experiment D of the present study showed that exposure to unstirred water can have the same effect, in this case resulting in a 70% increase above the for frogs in stirred water of similar [Na⁺]. The mechanism of this increase is presently unknown but might involve changes in either the numbers or properties of the components of the Na+ uptake system or in the ability of epithelial cells to maintain ATP supply to the various ATPases in the face of a hypoxic boundary layer.

A layer of unstirred or poorly stirred medium surrounds all physical objects immersed in a fluid (Vogel, 1983). If Na⁺ uptake from this boundary layer exceeds Na⁺ replenishment from the bulk medium outside the boundary layer or via cutaneous efflux of Na⁺, then [Na⁺] in the boundary layer will decrease. Given the apparent saturation kinetics of Na⁺ uptake in frog skin, a large decrease in [Na⁺] in the boundary layer could account for the observed decline in and TEP in unstirred water, but is not the only possible explanation (see next section). In physiologically relevant conditions (i.e. open-circuit conditions and a dilute external medium), of frog skin has an apparent K_m of 0.2mmol l⁻¹ Na⁺ and approaches an upper limiting value at approximately 1mmol l⁻¹ Na⁺ (Greenwald, 1971; Kirschner, 1988). These kinetics appear to be due to coupling with outward H⁺ transport in open-circuit conditions (Harvey and Ehrenfeld, 1988; Kirschner, 1988) and are demonstrable in both whole frogs and frog skin (Greenwald, 1971; Kirschner, 1988).

At issue in support of the Na⁺ depletion hypothesis are the normal range of [Na⁺] of the bulk medium, how much [Na⁺] declines in the boundary layer in unstirred medium and whether the apparent saturation kinetics of are such that the steady-state at the steady-state [Na⁺] of the boundary layer will be less than that in well-stirred medium. Average values of [Na⁺] are 0.17mmol l⁻¹ for soft lake water and 0.39mmoll^{− 1} for river water (Schmidt-Nielsen, 1990). Although natural [Na⁺] can clearly be greater, these average values suggest that aquatic amphibians often (if not typically) experience bulk medium [Na⁺] sufficiently low for changes in [Na⁺] to affect markedly. How much [Na⁺] declines in the boundary layer is presently unclear. clearly declines in the boundary layer, sometimes as much as 19kPa (142mmHg) below its level in the bulk medium (Feder and Booth, 1992). Na⁺, however, has both a greater solubility and a greater diffusion coefficient in water than does oxygen, suggesting that its diminution in the boundary layer should be less than for oxygen. Calculation of the [Na⁺] at the mucosal surface from first principles is problematic because flow regime around intact frogs is poorly characterized. We have made a single measurement of [Na⁺] just outside the skin of an intact, MS-222-anaesthetized frog with a Na⁺-selective microelectrode [ionophore: ETH157 (Fluka); bulk medium: 3mmol l⁻¹ NaCl at normoxic ⁠; species: Rana catesbeiana]. [Na⁺] is 3mmol l⁻¹at 125 μm from the mucosal surface and declines to 1.33mmol l⁻¹ at some distance between that point and the mucosal surface. If proportional declines in [Na⁺] occurred in experiments A–C, they might well be sufficient to account for the observed declines in and TEP. Obviously, a definitive evaluation of the Na⁺ depletion hypothesis will need to await future studies; none of the present results, however, is inconsistent with this hypothesis.

As with Na^+, a decline in in the boundary layer sufficient to compromise the supply of ATP to the ATPases to which Na⁺ uptake is linked might account for or contribute to the observed decreases in and TEP in unstirred medium. The evidence for such an oxygen depletion effect is more substantial at present than for Na⁺ depletion, but is not yet definitive.

Previous studies convincingly document that declines considerably in the boundary layer surrounding the skin of amphibians in normoxic unstirred water (Pinder and Feder, 1990; Booth and Feder, 1991; Feder and Booth, 1992). In such circumstances, the steady-state is well below the critical (Booth and Feder, 1991) and, as a result, oxidative metabolism declines unless the affected cells have an alternative source of oxygen. Exposure to unstirred water directly decreases oxygen consumption of intact amphibians (Burggren and Feder, 1986; Pinder and Feder, 1990; Booth and Feder, 1991) and exposure to values of expected at the skin’s surface in unstirred water similarly affects the oxygen consumption of isolated frog skin (Vitalis, 1990). Zerahn (1956) firmly established the link between oxygen consumption and Na⁺ uptake in frog skin. Insofar as cutaneous oxygen uptake is affected by the convective environment, so should oxygen-dependent Na⁺ uptake.

A variety of experimental data is consistent with this hypothesis. Both mitochondrial inhibitors of oxidative phosphorylation and anoxia block Na⁺ uptake in frog skin (Harvey and Ehrenfeld, 1988). Frog skin apparently does not deploy anaerobic metabolism to mitigate affects of limited O₂ supply (Vitalis, 1990). A surprisingly large classical literature, in advance of an adequate mechanistic understanding of Na⁺ uptake, showed that hypoxia or anoxia reduce TEP (e.g. Taylor, 1936; reviewed by Vitalis, 1990), as does experiment E of the present study. Although we were unable to demonstrate a consistent or large effect of hyperoxia on TEP, Taylor (1936) reported just such an effect. increases modestly in a well-stirred hyperoxic bath and decreases considerably in a well-stirred hypoxic bath (experiment F).

A challenge to the oxygen depletion hypothesis is in demonstrating actual inhibition in mitochondrial ATP supply to the relevant ATPases in frog skin. The required to inhibit mitochondrial oxidative phosphorylation is low and the actually experienced by the mitochondria of mitochondria-rich cells in frog skin is unknown. On the one hand, the relevant ATPases reside in cells in the most superficial layers of the skin and the mitochondria of these cells may receive sufficient O₂ despite the reduction in in the boundary layer. Furthermore, the skin is richly vascularized and its cells may be sufficiently oxygenated regardless of the of the boundary layer. On the other hand, the outside the skin in the boundary layer is already low and the skin itself is a major resistance to the diffusion of oxygen (Piiper, 1982; Pinder and Feder, 1991). Moreover, the measurements of Vitalis (1990) suggest that the skin itself is primarily responsible for its own oxygen uptake, with little O ₂ delivered via the circulatory system. This situation could, however, reverse when the skin undergoes severe hypoxia. Direct measurements of intracellular and ATP concentrations would contribute to the resolution of these issues.

In conclusion, although the coupling of gas exchange and ionoregulation is scarcely a novel concept, the primary focus of interest has been on the interrelationship of CO₂ excretion, acid–base regulation and ion homeostasis, largely within the integument itself and inside the organism (Boutilier et al. 1992; but see Gonzalez and McDonald, 1992; Randall et al. 1991). The present study, however, suggests that a novel and physiologically relevant form of such coupling may involve oxygen and the convective environment and come about through the common effect of fluid dynamic boundary layers or unstirred layers on solute species outside the skin. Data on both natural convective environments and on several key components of the proposed coupling are still few. If borne out, however, the coupling of cutaneous ion and oxygen uptake in the boundary layer may have significant implications for both adaptation to natural convective environments and general models of transmembrane exchange.

We thank Renee Richer for technical assistance, Dr Richard Kraig for his assistance with ion-selective microelectrodes, Dr Richard Miller for facilitating access to instrumentation and Dr George Iwama for warning us of a particularly serious flaw in a preliminary experimental design. Tzvi Robbins was supported by a summer research fellowship from a Howard Hughes Medical Institute Undergraduate Education Initiative grant and Colleen Talbot was supported by NIH/NHLBI training grant 2-T32-HL7237. Research was supported by NSF grant DCB87–18264.