Na⁺ and Cl⁻ Uptake Kinetics, Diffusive Effluxes and Acidic Equivalent Fluxes Across the Gills Of Rainbow Trout: I. Responses To Environmental Hyperoxia

The ability of freshwater salmonids to alter dynamically the unidirectional Na⁺ and Cl⁻ fluxes across the gills has now been demonstrated under several different types of acid-base disturbance (Cameron, 1976; McDonald et al. 1983, 1989; Wood et al. 1984; Perry et al. 1987; Wood, 1988; McDonald and Prior, 1988). These ionoregulatory changes aid in the compensation of the acidosis or alkalosis by direct or indirect exchange of Na⁺ or Cl⁻ with acidic or basic equivalents, respectively. At present, little is known about the mechanisms involved.

A fundamental technique for understanding the function of transport systems has been to analyze their transport kinetics, an approach pioneered in vivo in freshwater crayfish by Shaw (1959). This involves measuring the transport rate at a variety of substrate (i.e. external ion) concentrations; the resulting MichaelisMenten relationship in turn yields estimates of the kinetic parameters - K_m (inverse of affinity) and J_max (maximum transport rate). Traditionally, K_m is related to the binding affinity of the actual operational site (e.g. transport enzyme, carrier or channel) for the substrate, and J_max is considered to be an index of the absolute number of operational sites available. Changes in either, or both, can theoretically be responsible for observed changes in the transport rates. To date, in fish, this technique has been applied only to long-term changes in transport associated with exposure to water of altered ionic compositions (e.g. Kerstetter et al. 1970; Kerstetter and Kirschner, 1972; de Renzis and Maetz, 1973; Wood and Randall, 1973a; Laurén and McDonald, 1987; Avella et al. 1987) and not to the dynamic changes in Na⁺/acid and Cl⁻/base exchange which might occur during acid-base disturbance. In the present and the following study (Goss and Wood, 1990), a technique for the rapid determination of Na⁺ and Cl⁻ uptake kinetics in individual rainbow trout has been applied to the study of such dynamic changes.

Recently, it has been appreciated that, in addition to influx modulation, adjustment of the differential efflux of Na⁺vs CP across the gill may also play a significant role in acid-base regulation (Wood, 1988; McDonald et al. 1989). If total cation efflux (of which Na⁺ is normally the major component) exceeds total anion efflux (mainly Cl⁻), the extracellular strong ion difference (SID) is reduced, with a concomitant reduction in pHe (Stewart, 1978). The reverse will be true if anion efflux exceeds cation efflux. The results of several other studies (Cameron, 1976; Wood et al. 1984; Perry et al. 1987) can be interpreted as providing evidence for the involvement of differential diffusive efflux in acid-base regulation. However, in all these investigations, differential efflux measurements by radio tracer techniques may have been complicated by the presence of exchange diffusion (e.g. Maetz, 1972; de Renzis and Maetz, 1973; Wood et al. 1984). Therefore, a novel method, again patterned after the original technique of Shaw (1959), has been developed to measure Na⁺ and Cl⁻ diffusive efflux rates during acid-base disturbances, without the complicating effects of exchange diffusion or changes in membrane permeability.

Earlier, environmental hyperoxia was shown to be a very useful probe for determining the role of ionic exchanges in acid-base regulation (Wood et al. 1984). The first goal of the present study was to determine possible changes in the branchial transport kinetics of Na⁺ and Cl⁻ during the normoxia-hyperoxianormoxia regime characterized by Hõbe et al. (1984), Wheatly et al. (1984) and Wood et al. (1984). A second objective was to assess the role of differential diffusive efflux of Na⁺ and Cl⁻ in acid-base regulation during this regime.

Rainbow trout (Oncorhynchus my kiss (Walbaum, 1792), 250–400g) were obtained from Spring Valley Trout Farm, Petersburg, Ontario, Canada, and acclimated for at least 2 weeks to 15±0.5°C in flowing dechlorinated Hamilton tapwater (composition in Table 1). The fish were starved during this period to minimize any variations in acid output due to feeding (Wood and Caldwell, 1978). The trout were then anaesthetized (MS-222 1:10000; Sigma) and fitted with a urinary bladder catheter (Wood and Randall, 1973b) to separate the branchial from the renal component of ion and acid-base regulation. The fish were allowed to recover for 72 h in separate, darkened, well-aerated acrylic boxes continuously supplied (0.51min⁻¹) with temperature-controlled (15±0.5°C) tapwater. The small volume of the boxes (1.3–1.71 plus the volume of fish) was necessary to determine accurately changes in each of the measured variables over short periods (see Discussion for assessment of precision), and could easily be reset to a premeasured level. Vigorous aeration, which was separated from the fish by an acrylic wall, thoroughly mixed the water without direct exposure of the fish to the bubbling. Complete equifibration time of substances added to the box was less than 2 min.

Acclimation and experimental series I were performed in dechlorinated Hamilton city tapwater. Experimental series II and III were performed using a NaCl-free artificial medium designed to duplicate the Ca²⁺, Mg²⁺ and titratable alkalinity (TAlk) composition of Hamilton tapwater as closely as possible (Table 1). Ca²⁺ and Mg²⁺ are important because of their well-known role in regulating membrane permeability and transepithelial potential (TEP) (Potts, 1984; McDonald and Rogano, 1986) and TAlk in providing water buffer capacity, HCO₃⁻ availability and the correct pH. This NaCl-free water was made by passing dechlorinated Hamilton tapwater through a deionizing cartridge (Anderson Water Corp.) then adding back appropriate amounts of Ca(CO₃)₂ and (MgCO₃)4-Mg(OH)₂.5H₂O to bring Ca²⁺ and Mg²⁺ to levels normally seen in Hamilton tapwater. Because these salts are relatively insoluble in water at neutral pH, the water was bubbled with 100% CO₂ overnight to achieve complete solution. The water was then vigorously aerated for 24 h to drive off the CO₂. The resulting medium had appropriate levels of Ca²⁺, Mg²⁺ and TAlk (titratable alkalinity to fixed endpoint pH=4.00), but 0 mmol 1⁻¹ Na⁺ or Cl⁻ (Table 1). The temperature was 15±0.5°C.

After 72 h of recovery from catheterization, a series of control measurements was made under normoxia ⁠. The aeration source to the boxes was then replaced with pure O₂, which elevated the water within 0.33 h to 70.9±1.9kPa. Measurements were then made over several periods during the next 72 h of maintained hyperoxia. At 72 h, the O₂ was replaced with air, quickly returning to normoxic levels, and a further series of measurements was made.

To avoid confusion, the term outflux (unidirectional outflux) is used to refer to outflux determined as the measured difference between influx and net flux in series I and III. The term efflux (diffusive efflux) is used to refer to the novel measurement technique in series II which did not involve radioisotopic measurements.

Series I was designed to characterize the time course and extent of changes in net and unidirectional ion fluxes (Na⁺, Cl⁻) and net acid-base fluxes [titratable acidity (TA)], ammonia (Amm) and net acidic equivalent fluxes occurring during the hyperoxic regime. Wood et al. (1984) reported only average flux values over 4–12 h periods, whereas the present study was designed to resolve fluxes over successive 0.5 h intervals. A particular goal was to detect periods when Na⁺ and Cl⁻ influxes and net acidic equivalent fluxes might be relatively stable for 4h, to allow the planned kinetic measurements of series III. Problems of radioisotopic backflux and specific activity limitations precluded continuous 0.5 h measurements over the entire 5 day experiment. Based on the results of Hõbe et al. (1984) and Wood et al. (1984), the following periods were anticipated to be of greatest interest and therefore investigated in 0.5h detail: normoxic control; the first 8h of hyperoxia (0–8h), representing the time of maximal acidosis; the final 7h of hyperoxic exposure (65–72 h), representing a time when internal pH had been restored almost to control values; and the first 8h of return to normoxia, representing the time of maximal alkalosis.

At the start of each measurement segment, the boxes were closed and either ²²Na⁺ (1.0 μCi) or Cl⁻ (1.0μCi) was added. These amounts were doubled in the normoxic recovery period to prevent error that might be caused by isotopic loading, increased internal specific activity and the associated backflux of radioactivity. The ²²Na⁺ (A=8) and ³⁶Cl⁻ (N=8) experiments were performed on different sets of fish; in all cases, the same individual fish were followed through all four measurement segments. After an initial 10 min mixing period, water samples (40ml) were taken at 0.5 h intervals for analysis of external [Na⁺] ([Na⁺]_c), [Cl⁻]_e, total ammonia (Amm), TA and ²²Na⁺ or ³⁶Cl⁻ ctsmin⁻¹. The boxes were flushed after each 3 h of closure with water of the appropriate to ensure that ambient ammonia levels did not exceed 200μequivl⁻¹. Following the flush, the appropriate amount of isotope was re-added to the box. In the period between the measurement segments, the boxes were served with flowing dechlorinated tapwater of the correct

In series II, a novel technique inspired by the original approach of Shaw (1959) was developed to measure the simple diffusive effluxes of Na⁺ and Cl⁻ from trout (N=12) in the absence of any possible exchange diffusion component (cf. Shaw, 1959) (see Discussion for an assessment of the validity of this approach). Two control efflux measurements (−3h, −l h), five hyperoxia measurements (1, 4, 8, 68 and 71 h after start of hyperoxia) and three measurements during normoxic recovery (73, 76 and 80h; i.e. 1, 4 and 8h after return to normoxia) were performed. At each time period, the boxes were flushed with NaCl-free water, to lower the external NaCl concentration ([NaCl]_e) to below 10μequivl⁻¹, and closed at a precise volume. Note that the pH, Ca²⁺, Mg²⁺ and TAlk concentrations were maintained at normal levels (Table 1). Because this [NaCl]_e was far below the K_m of either the Na⁺ or Cl⁻ transporter, and exchange diffusion should not occur owing to the unavailability of the external ion. Therefore, the initial appearance of Na⁺ and Cl⁻ in the external water should reflect the simple diffusive efflux of these ions. After an initial 2min mixing period, samples were taken at 5 min intervals for 0.5 h and analyzed for [Na⁺]_e and [Cl⁻]_e. Diffusive efflux was calculated from these values at each interval, and in most cases remained constant up to 0.5 h. The value obtained from 0 to 10 min was used in the calculation of the mean.

A separate experiment was performed to determine if there were any changes in transepithelial potential [TEP: blood relative to water as zero (0)] during exposure to NaCl-free media (Table 2). Fish were fitted with dorsal aortic catheters (see Goss and Wood, 1990) to allow TEP measurement without disturbance, and then held in Hamilton tapwater (Table 1) under control, normoxic condition. TEP between blood and external water was monitored via the catheter using standard techniques (Perry and Wood, 1985). KCl (3moll⁻¹)-agar bridges were connected via Ag/AgCl electrodes to a high-impedance voltmeter (Radiometer pHM82). Control determinations were made in normal Hamilton tapwater, and then the box was flushed with either the NaCl-free medium (N=7) used in the diffusive efflux measurements or with distilled water (N=5) for comparison. Measurements were made at 2 min intervals for 10 min, representing the period used for diffusive efflux measurements.

Series III was designed to measure the influx kinetics of either Na⁺ (N=12) or Cl⁻ (N=11) over increasing [NaCl]_e in the water at various times in the exposure regime. It also investigated relationships between [NaCl]_e and unidirectional efflux rates of the ion in question, Amm fluxes, TA fluxes and net acidic equivalent fluxes. Four periods were regarded as reasonably stable from the results of series I: ‘control’ normoxia; 4–8h of hyperoxia ( ‘initial’); 68–72h of hyperoxia ( ‘final’); and 4–8 h after return to normoxia (i.e. 76–80h; ‘recovery’).

In each period, flow was stopped and the boxes were flushed with NaCl-free water so that [NaCl]_e was below 10μequivl⁻¹. Again, pH, Ca²⁺, Mg²⁺ and TAlk remained at normal levels. At this point, NaCl was added to the boxes from a common 1 moll⁻¹ NaCl stock solution containing approximately 1.0μCi ml⁻¹ of either ²²Na⁺ or ³⁶C1⁻. This allowed for the maintenance of a relatively constant specific activity over the entire range of [NaCl]_c. After the stock had been added, 10min was allowed for mixing and then a sample was withdrawn to begin the flux. After 0.5 h, another sample was taken to end the flux and more stock was added to the box to increase the water [NaCl]_e further. Six flux periods of increasing [NaCl]_e (nominally 50, 150, 300, 600, 1200 and 2400μequivl⁻¹) were measured. Each lasted 0.5 h, except for the final flux period lasting 0.67 h. Samples (20 ml for ³⁶Cl⁻ experiments, 40ml for ²²Na⁺ experiments) were analyzed for Amm, TA (Amm and TA in ²²Na⁺ experiments only), [Na⁺]_e, [Cl⁻]_e and ²²Na⁺ or ³⁶Cl⁻ ctsmin⁻¹

Water [Na⁺]_e was measured by atomic absorption (AAS, Varian AA1275), water [Cl⁻]_e by the mercuric thiocyanate method (Zall et al. 1956) and water ammonia by the salicylate-hypochlorite method (Verdouw et al. 1978). Titratable alkalinity (TAlk) was determined as described by McDonald and Wood (1981). Duplicate 5 ml ³⁶Cl⁻ or ²²Na⁺ water samples were measured directly by scintillation counting (LKB Rackbeta, model 1217).

Unidirectional influxes (J_in), outfluxes (J_out) and net fluxes (J_nct) were calculated as described by Wood (1988). Correction for radioisotope backflux (Maetz, 1956) was not necessary because internal specific activity (SA) never exceeded 5 % of the external SA.

Net acidic equivalent flux was calculated as the sum of the TA flux and the Amm flux, signs considered, as outlined by Maetz (1973). This method does not distinguish ammonia movement as either NH₃ or NH₄⁺ or net acidic equivalent flux as a result of either acidic or basic equivalent movement (i.e. basic equivalent loss=acidic equivalent uptake and vice versa) but this does not matter in terms of the net acid-base status of the fish.

All values are presented as means±l S.E.M. (N). The significance of difference for individual means was assessed by the paired Student’s two-tailed t-test, using each fish as its own control. Comparisons between multiple means were performed by a one-way analysis of variance (ANOVA), followed by Duncan’s new multiple range test to determine individual differences in cases where the F-value of the ANOVA indicated significance. Linear regression relationships were generated by the method of least squares, and the strength of the relationship was assessed by a Fisher (z) transformation of Pearson’s simple correlation coefficient. This trans-formation is necessary for a bivariate normal distribution. Tests of significance for the slope and intercept were performed as outlined in Remington and Schork (1985). A significance level of P<0.05 was employed throughout.

Exposure to hyperoxia caused a significant 50 % reduction in over 65–72 h of exposure (Fig. 1). This decrease in was not significant in the initial 8 h period of exposure. Concurrently, was significantly depressed by about 60 %, resulting in the maintenance of a small positive throughout the exposure period.

Like ⁠, chloride influx was also significantly reduced by 50% upon exposure to hyperoxia (Fig. 2). However, this reduction took only 3–4 h to develop fully; remained depressed for the duration of the exposure ⁠. was unchanged during the initial exposure to hyperoxia, resulting in a significantly negative at this time. However, by 65–72 h, was also significantly reduced with respect to control levels, resulting in a not significantly different from 0.

is the sum of two components J^Amm and J^TA, signs considered. There was a 50 % reduction in J^TA during hyperoxia which took about 3–4 h to develop fully (Fig. 3). J^TA remained depressed at 65–72 h of continuous hyperoxia. There was no significant change in J^Amm (approx. −200 μequiv kg⁻¹ h⁻¹). A reduced J^TA and unchanged J^Amm resulted in a negative during initial hyperoxic exposure, which continued throughout the exposure period (Fig. 3).

The return to normoxia (recovery) caused a further fall in (Fig. 1). The overall decrease was significant compared to both the control and final time periods. a remained depressed throughout the recovery period (Fig. 1). was immediately stimulated by 60 % relative to the original control value, resulting in a large negative As the recovery progressed, was reduced and thus was also reduced.

Return to normoxia resulted in a highly significant threefold stimulation of to a level 50% higher than the original control value (Fig. 2). The increase was not immediate but took 3 – 4 h to develop fully. was greatly stimulated in the first 4h followed by stabilization over the next 4h of recovery. The increases in were significant with respect to both the final period of hyperoxia and the original control levels. The large and immediate stimulation of in combination with a delayed increase in resulted in a negative in the initial 4 h following return to normoxia, followed by a switch to a positive as became fully stimulated. This changeover was statistically significant. Return to normoxia resulted in a fourfold increase in J^TA, which also took 3 – 4 h to develop fully (Fig. 3). This was significantly higher than both the original control value and the final hyperoxia value. Throughout the recovery period, J^Amm was again not significantly altered (Fig. 3). The resulting effect was a highly positive during this time of metabolic alkalosis.

In summary, exposure to hyperoxia resulted in a positive Na⁺ balance, a negative Cl⁻ balance and a net acidic equivalent excretion, while recovery resulted in a negative Na⁺ balance, a positive Cl⁻ balance and a large net acidic equivalent uptake. The alterations in net Na⁺ and Cl⁻ balance resulted from complex changes in both influx and outflux components. The changes in were accomplished by changes in the J^TA component only, j^Amm remaining invariant.

The relationship between and based on all individual simultaneous flux measurements during initial exposure (0–8 h) to and recovery from (72–80 h) hyperoxia, is shown in Fig. 4. Based on the work of Wood et al. (1984) and Hõbe et al. (1984), these periods were chosen for analysis because they were the periods of maximal internal acid-base disturbances. The overall relationship was highly significant (r=−0.535, N=208, P<0.001) as described by the regression equation:

Exposure to hyperoxia produced no significant changes in the diffusive effluxes of either Na⁺ (Fig. 5A) or Cl⁻ (Fig. 5B). The net difference between Cl⁻and Na⁺ efflux was not significantly different from zero. This pattern continued throughout hyperoxia without significant change (Fig. 5).

Return to normoxia resulted in a significant stimulation of both Na⁺ and Cl⁻ efflux. However, Na⁺ efflux was increased to a much greater extent than Cl⁻ efflux, resulting in a net differential flux of about −150μequivkg⁻¹h⁻¹ immediately following return to normoxia, stabilizing at − 90itequivkg⁻¹h⁻¹ for the remainder of the recovery period. This differential flux was significantly different from the values measured both during the original control period and during the final period of hyperoxia. Qualitatively, trends in diffusive effluxes throughout the regime were similar to those in and (outfluxes, see Figs 1, 2) determined at comparable times in series I, but absolute values were quantitatively smaller in the former, suggesting the presence of exchange diffusion components in the radiotracer outfluxes.

Measurements of TEP during 10 min of exposure to the NaCl-free media used in the diffusive efflux determinations demonstrated that TEP remained invariant (Table 2). Thus, changes in TEP were not a factor biasing the diffusive efflux results. In contrast, 10 min of exposure to pure distilled water resulted in an immediate shift to highly negative TEP values (Table 2).

A particular goal in series I had been to determine 4 h periods within the overall protocol when Na⁺ and Cl⁻ influxes and net acidic equivalent fluxes might be reasonably stable. These periods could then be employed for the planned uptake kinetic measurements. Based on the results shown in Figs 1, 2 and 3, the following periods were selected: control (−4 to 0 h); initial hyperoxia (4–8h); final hyperoxia (68–72 h); and recovery normoxia (76-80 h).

In all fish in all periods, increases in [NaCl]_e resulted in increases in and (cf. Figs 6, 7) which followed typical Michaelis-Menten curves. These data could be linearized by Eadie-Hofstee regression analysis to obtain estimates of K_m and J_max of the transporters for individual fish. No consistent departures from linearity were observed. The means (Figs 8, 9) of the K_m and J_max values obtained from individual fish were substituted into equation 1 to generate the curves fitted to the mean data in Figs 6 and 7.

Exposure to hyperoxia (initial, final time periods) caused reduced values of at lower levels of [Na⁺]_e, but was unaltered at higher [Na⁺]_e (Fig. 6). Return to normoxia (recovery time period) lowered by at least 50 % at almost every [Na⁺]_e (Fig. 6). In contrast, was not greatly altered at any [Cl⁻]_e during exposure to hyperoxia (Fig. 7). However, the Cl⁻ influx values were greatly elevated at each [Cl⁻]_e in the recovery period (Fig. 7). Regression analyses of individual fish yielded estimates of the average K_m and J_max for each ion (Figs 8, 9). The affinity of the Na⁺ transporter was significantly decreased (i.e. increased fourfold from 114 to 445μequivl⁻¹) upon initial exposure to hyperoxia and remained so throughout the exposure period (Fig. 8A). Return to normoxia resulted in a further loss of affinity as increased to 559μequivl⁻¹. Concurrently, was not significantly altered during hyperoxia but return to normoxia resulted in a decrease in by 30 % (Fig. 8B), which was significant compared to both the control and final periods.

Exposure to hyperoxia had no initial effect on the affinity of the Cl⁻ transporter, but later decreased it significantly as increased by 50% (Fig. 9A). Return to normoxia reversed the effect, significantly increasing the affinity compared to the final time period as dropped from 250 to 137gequivl⁻¹, a value not significantly different from the original control. remained unaltered at about 291μequivkg⁻¹ h⁻¹ during the hyperoxic exposure but increased by 65% upon return to normoxia, reaching a value (445μequivkg⁻¹h⁻¹) that was significantly higher than both the control and the final exposure period values (Fig. 9B).

The effect of increasing [NaCl]_e on and and its components was also examined in the kinetics experiments (series III). As [NaCl]_e was increased, there were no detectable changes in either or and and became increasingly positive (data not shown). This suggests that the exchange diffusion components saturated at very low values of [NaCl]_e or else were absent. However, it must be noted that the accuracy of determinations progressively decreases at higher substrate concentrations. Small changes in J_out would be undetectable with the short flux periods used here.

J^Amm, J^TA and in each of the four exposure periods were also not significantly altered as [NaCl]_e was increased (Fig. 10). However, in all treatments except recovery, there was a tendency for J^Amm to increase with [NaCl]_c at the three lowest concentrations and for a general increase in acidic equivalent excretion (negative ⁠) over the whole range of [NaCl]_e tested. Interestingly, during recovery, when net acidic equivalent uptake predominated, there was a similar tendency for positive to increase at higher [NaCl]_e.

Shaw (1959) developed a method for measuring the diffusive efflux of Na⁺ and Cl⁻ from an aquatic animal which involved replacing the external water with deionized water and monitoring the appearance of the ion in question. As both influx and exchange diffusion are theoretically eliminated, the measured net flux should represent the true diffusive efflux, provided that influx and exchange diffusion blockade are complete and that changes in TEP and gill permeability are not induced by the changes in external water NaCl. In series II of the present study, a modification of this approach has been employed for the first time to measure the differential diffusive efflux of Na⁺ and Cl⁻ and its possible contribution to acid-base homeostasis. Shaw’s original method has been specifically modified to avoid possible changes in membrane permeability and TEP associated with the use of distilled water (e.g. Potts, 1984; McDonald and Rogano, 1986). An artificial medium was developed providing normal levels of Ca²⁺, Mg²⁺, TAlk and pH, but lacking Na⁺ and Cl⁻ (Table 1). The maintenance of Ca²⁺ and Mg²⁺ in this NaCl-free medium allowed for measurements of efflux in the absence of any detectable changes in TEP (Table 2). In contrast, exposure to distilled water resulted in an immediate highly negative TEP (Table 2), table from the absence of hardness cations (Potts, 1984).

Recently, it has been shown that rainbow trout body mucus may have up to 10 times the concentration of ions compared to the bulk water, although the activity coefficients are much lower (Handy, 1989). It is not known whether similar relationships apply to gill mucus, or to what levels these concentrations would be lowered during exposure to NaCl-free water. Thus, it is possible that the present method may not completely eliminate the influx or the exchange diffusion component owing to incomplete removal of the ions from the gill ment. It is also possible that Na⁺ and Cl⁻ removal from the apical surface may have an effect on the tight junction permeability, although we are not aware of any evidence that this effect occurs for the very small absolute decreases in [NaCl]_e used in the present study. Furthermore, the lack of change in TEP when the fish were placed in NaCl-free media (Table 2) suggests that such alterations in permeability did not occur.

The experimental protocols for both steady-state and kinetic analyses used in the present study required determination of both ion and isotopic fluxes over very short periods (0.5 h) to minimize errors and resolve trends resulting from the changing dynamic state of the fish during acid-base disturbance. These short flux determinations therefore placed great demand on analytical precision. Exceptional care was taken in sampling and analysis to determine accurately the small changes in concentrations and radioactivity occurring over these very short periods. The final protocols arose out of preliminary experiments with various experimental periods and analytical techniques.

In determining changes in total Na⁺ and Cl⁻ concentrations, which entered directly into net flux calculations, the most important precautions were: (i) to read the samples undiluted, wherever practical; (ii) to use background correction and bracketing standards in the range of measurement; (iii) to replicate (duplicates for Na⁺ and triplicates for Cl⁻); and (iv) to read the start and end samples for a flux period as pairs immediately after one another. In practice, the latter was the most important precaution, because it is the difference between start and end values, not the absolute values, which enters into the net flux calculation. To illustrate, the AAS had a reproducibility of ±about 1 % when calibrated over the range 550–650μequiv 1⁻¹ [Na⁺]_e. A typical net flux (80jzequivkg⁻¹h⁻¹) over 0.5 h (Fig. 1) might change the external Na⁺ concentration by 9μequivl⁻¹. While this difference was only 1.5 % of the total background of 600 μequivl⁻¹, it was 9 % of the calibration range and, in practice using the pairing approach, could be reproduced to 8–10μequivl⁻¹. Therefore, the overall reproducibility of the net flux measurements was about ±10%.

For influx determinations, the influence of such errors in the total ‘cold’ concentration measurements was small; the critical measurement was again the difference, this time in radioactivity, between start and end samples. Here counting statistics and pipetting were the major sources of error. Statistically, the error associated with radioactive counting can be reduced through increasing the counting time (and thus the total counts per sample), replicating the samples, repeating the count procedure on the same samples, and generating weighted means from the results. A complete description of the error calculations is beyond the scope of this paper, but can be found in Wang et al. (1975) and Remington and Schork (1985). In practice, we counted at least duplicate water samples each twice for 10 min. Statistically, error in count separation accounted for variations in influx determination from 2 to 12 % in the kinetic series (series I) and 5 to 10 % in the steady-state flux determinations. Pipetting error was reduced by using individual volumetric pipettes for the beginning and end of each flux period. As a result, overall reproducibility of about ±10% was achieved for influx measurements using 0.5 h periods.

The results of series I confirm and extend those of Wood et al. (1984), who found that branchial Na⁺vs acidic equivalent and Cl⁻ vs basic equivalent exchanges were dynamically adjusted to achieve acid-base homeostasis during the hyperoxic regime. A positive Na⁺ balance and negative Cl⁻ balance were associated with net acidic equivalent excretion, while a negative Na⁺ balance and positive Cl⁻ balance were associated with net basic equivalent excretion (=acidic equivalent uptake). While the general agreement of the two studies was good, some points of difference with the result of Wood et al. (1984) are noted subsequently.

An important new contribution of the present study was the precise characterization of the time course of these responses, for fluxes were measured over 0.5 h Intervals. Changes in occurred slowly during hyperoxia but rapidly during recovery (Fig. 1), while overall changes in took 3–4h to develop fully in both situations (Fig. 2). Similarly, J^TA and also took 3–4 h to respond to either internal acidosis or alkalosis (Fig. 3). The importance of the time course of these changes with reference to the mechanisms involved will be discussed below.

Theoretically, changes in J^TA could result from changes in NH₃ (but not NH₄⁺), H⁺ and/or HCO₃⁻(OH⁻) fluxes. However, the constancy of the total ammonia flux (J^Arnm) in these experiments in the face of large changes in (Fig. 3) and the non-correspondence of J^TA changes with (Fig. 2) suggest that the alterations in J^TA mainly reflected changes in HCO₃⁻(OH⁻) fluxes. The time course and pattern of changes in J^TA and followed each other very closely, presenting strong evidence for the linkage between influx of Cl⁻ and the outflux of HCO₃⁻. Throughout the experiment, changes in the net flux of acidic equivalents appeared to be largely the result of changes in J^TA, with no significant changes in J^Amm. This lack of an increase in J^Amm during hyperoxia contrasts with several previous studies on freshwater teleosts (Wood et al. 1984; Claiborne and Heisler, 1986; Perry et al. 1987), which found that respiratory acidosis in the arterial blood was accompanied by increases in the flux of ammonia to the water. The reasons for this difference are unknown. In any event, it must be emphasized that the ammonia and titratable components of are not necessarily dependent or independent parameters, for the latter may or may not be affected by the former.

During exposure to hyperoxia, both and were reduced significantly. While the decrease in was in accord with the results of Wood et al. (1984), the reduction in was not. Another difference in the present study was the continued negative (acidic equivalent excretion) at 72 h, rather than complete compensation by this time. We speculate that lamellar vasoconstriction and gill hypoventilation was larger in the present study, resulting in greater respiratory acidosis and reduced mixing in the gill boundary layer. This could decrease Na⁺ and Cl⁻ availability in the interlamellar water, thereby reducing the overall ion turnover rates and speed of acid-base correction. This non-specific effect may have been the explanation for the slow decrease in observed during hyperoxia (Fig-1).

A gradual decrease in was not the appropriate response for correction of a respiratory acidosis by Na⁺/acidic equivalent exchange. Wood et al. (1984) found only a very small increase in during initial exposure to hyperoxia and a decrease later in the exposure period. Perry et al. (1987) found a similar decrease in during hypercapnic acidosis. Based on these studies, the relative contribution of Na⁺/acidic equivalent exchange during compensation from respiratory acidosis appears to be minimal. However, inhibition of this exchange may have played an important role following return to normoxia (Fig. 1), in agreement with the results of both Wood et al. (1984) and Perry et al. (1987). The observed reduction in at this time is the expected response, as it would result in the retention of acidic equivalents and aid in compensation of the metabolic alkalosis present during normoxic recovery. The pattern of changes in was in general agreement with those of previous workers (Cameron, 1976; Wood et al. 1984,Perry et al. 1987), who found that was similarly inhibited during acidosis and greatly stimulated during alkalosis. This is the expected response if the exchange of Cl⁻ was for basic equivalents (HCO₃⁻, OH⁻).

In summary, these results suggest that and its associated exchange with acidic equivalents play a minor role relative to and its exchange with basic equivalents during compensation from hyperoxia-induced respiratory acidosis. To compensate an acidosis, the fish increases the plasma SID, thereby retaining plasma [HCO₃⁻]. This change in SID is usually not accomplished by increasing plasma [Na⁺]. Therefore, manipulation of the C1⁻/HCO₃⁻ exchanger offers the most direct method for compensation. However, to compensate an alkalosis, such as during post-hyperoxic recovery, both and are adjusted. Thus, post-hyperoxic alkalosis is a more powerful stimulant than hyperoxic acidosis for eliciting changes in ion/acidic equivalent fluxes, in agreement with the findings of Wood et al. (1984).

The flux of acidic equivalents across the branchial surface was stoichiometrically (−1:1) proportional to the difference between the net Na⁺ and Cl⁻ fluxes over a wide range of the two variables (Fig. 4). This excellent correlation agrees with the findings of many (but not all, see Goss and Wood, 1990, for references) previous workers. Heisler (1982), McDonald (1983), McDonald et al. (1983, 1989) and Wood et al. (1984) all found a similar stoichiometry. According to SID theory popularized by Stewart (1978), the differential flux of strong ions from one compartment to another will necessarily constrain a change in acid-base status in that compartment. However, the −1:1 stoichiometry says nothing about the mechanisms involved in these fluxes, only that Na⁺ and Cl⁻ are the major ions involved.

Exposure to hyperoxia produced no significant changes in the diffusive efflux of either Na⁺ or Cl⁻, and the Na⁺-Cl⁻ efflux difference was not significantly different from 0 (Fig. 5). During recovery from hyperoxia, there was a large stimulation of both Na⁺ and Cl⁻ efflux. This stimulation of Cl⁻ efflux was paradoxical, given that retention of anions would aid in correcting the acid-base disturbance. It is probable that a general non-specific stimulation of both Na⁺ and Cl⁻ efflux was the result of increased gill perfusion and ventilation upon return to normoxia. There was, however, a greater increase in Na⁺ efflux compared to Cl⁻ efflux at this time (net cation loss) via the diffusive efflux mechanism (Fig. 5C). A net loss of cations would act to decrease SID, thereby decreasing pH. Therefore, there appears to be a definite differential diffusive efflux mechanism for acid-base regulation across the branchial surface during recovery from hyperoxia. Diffusive efflux measurements (series II) and acidic equivalent flux measurements (series I) were made on different batches of fish, so direct comparison is difficult. Nevertheless, it is apparent that the differential efflux during normoxic recovery was large enough to account for up to 50 % of the positive at this time (cf. Figs 3 and 5). The mechanisms controlling the rate of diffusive efflux are not known but evidence is mounting that paracellular passage of ions across many epithelia is an active, controlled process (Madera, 1988).

The present study is only the third to report measurements of K_m and J_max for Na⁺ transport in intact, unanaesthetized rainbow trout (Wood and Randall, 1973a,b; Laurén and McDonald, 1987) and the first to report comparable values for Cl⁻ transport. However, there exist several previous determinations of these parameters in perfused head preparations (Perry et al. 1985; Avella et al. 1987) and irrigated/anaesthetized whole-trout preparations (Kerstetter et al. 1970; Kerstetter and Kirschner, 1972), while the data of McDonald and Rogano (1986) can be re-analyzed to estimate these parameters in intact trout. Table 3 compares the present measurements with those in the literature.

Affinity values obtained in vivo are clearly higher (i.e. lower K_m) than those obtained using either the perfused head technique or irrigated gills in the anaesthetized trout. This may be the result of limitations in the ‘artificial’ preparations. Incomplete ventilation of the gill lamellae could cause a depletion of [NaCl]_c in the boundary layer, resulting in reduced apparent affinity. K_m values could also be elevated by the presence of the high concentrations of epinephrine (10⁻⁷mol⁻¹) necessary to maintain the perfused head preparation (Perry et al. 1984). However, estimates of J_max for both the Na⁺ and Cl⁻ transporters appear to be relatively uniform using all techniques.

Previous descriptions of branchial transport kinetics have been concerned only with long-term adjustments to altered environmental ion levels (see Introduction). The present study (series HI) is the first to employ specific internal acid-base disturbances as possible stimuli for dynamic changes in the transport kinetics of Na⁺ and Cl⁻, and to relate the latter to the measured changes in and during acid-base compensation. The results demonstrate that alterations in Na⁺/acidic equivalent and Cl⁻/basic equivalent exchanges (Figs 1, 2, 3) occur through complex adjustments in both K_m and J_max for the two transport systems. Specifically, decreased Cl⁻ uptake/base excretion during the compensation of hyperoxic respiratory acidosis was achieved by decreased Cl⁻ affinity (increased K_m) without alterations in (Fig. 9). However, increased Cl⁻ uptake/base excretion during the compensation of post-hyperoxic metabolic alkalosis was mediated by both decreases in and increases in The opposite changes in Na⁺ uptake/acid excretion at this time were achieved by exactly opposite alterations of and

Changes in the K_m of branchial Na⁺ (Maetz, 1972; Avella et al. 1987) and Cl⁻ transporters (de Renzis and Maetz, 1973) have been observed in several previous kinetic studies. The explanation generally offered has been that alterations in K_m resulted from changes in the Ca²⁺ levels of the external water, acting as a competitive or uncompetitive inhibitor. This cannot be the explanation for the observed K_m shifts in the present study, for normal levels of water Ca²⁺ were maintained throughout all experimental treatments. The results instead point to direct or indirect effects of acid-base status on carrier affinity, though non-specific effects of increased oxygen should not be discounted. For example, it has been suggested earlier that decreased ventilation and perfusion of the gill during hyperoxia may have resulted in decreased Na⁺ and Cl⁻ availability in the interlamellar water. This could well have contributed to the observed increases in K_m at this time (Figs 8, 9).

There are many possible mechanisms for acid-base effects on K_m, all of which are highly speculative. These include structural changes in the transport molecules themselves, or changes in the charge of the channels. There could be direct effects of changes in pHi, or in pHe via an intracellular messenger, or there might be a hormonal or neural level of control. Large numbers of innervated neuroepithelial cells have been described in both elasmobranch and teleost gills (Dunel-Erb et al. 1982; Donald, 1987). These cells occur over the entire region of the gill filament and respond to hypoxia by degranulation (Laurent, 1985), though their exact role remains to be elucidated. Frain (1987) has suggested that K_m changes may reflect the existence of fundamentally different populations of transporters which are activated under different conditions. This whole topic of carrier affinity regulation in relation to acid-base status is clearly worthy of further detailed study.

Large changes in J_max also occurred, but only during the metabolic alkalosis of the recovery period (Figs 8, 9). Furthermore, these changes were in the opposite direction for the Na⁺ and Cl⁻ transporters. Classically, changes in J_max have been interpreted as increases in the number of transport sites available (Shaw, 1959). Theoretically many possible mechanisms could be involved. Laurent et al. (1985) showed that there was an increase in the number of chloride cells present in fish adapted to more dilute water. However, it seems unlikely that the actual number of chloride cells might increase or decrease in the relatively short time (4-8 h) over which responses occurred in the present study. Avella et al. (1987) demonstrated a strong correlation between increased relative surface area of the chloride cells and an elevated Na⁺ uptake, though the time course of these changes was not evaluated. Further study is clearly warranted to see whether these changes can occur over such a short time course. The possibility of non-specific effects resulting from the perfusive and convective changes occurring during hyperoxia (Wood and Jackson, 1980), or from the action of the oxygen itself, cannot be eliminated. Such non-specific mechanisms could act to alter the number of chloride cells actually perfused or ventilated on their apical exposure, thereby changing J_max for the Na⁺ and Cl⁻ transporters. However, one important argument against all these explanations is that the fish must simultaneously reduce the J_max of one transporter while increasing that of the other. Clearly, this would be difficult if changes in chloride cell number, morphology or ventilation/perfusion were the only mechanisms available. An alternative possibility recently suggested by Frain (1987) is the production of new transporters or activation/deactivation of nascent transporters already present on the cell membrane.

Kirschner (1988), using isolated frog skins, has suggested that apparent saturation of Na⁺ influx is not due to saturation of the Na⁺ element in the transporter, but instead is the result of a limitation in the flux of the counterion. Indeed, stimulation of proton efflux was found to increase ⁠. If this idea is substantiated, it suggests that the one-substrate Michaelis-Menten model which we and all previous workers on freshwater fish have employed may be overly simplistic. A model taking into account both the external substrate (Na⁺, Cl⁻) and the internal substrates (acidic, basic equivalents) will be required. As a first step in this process, we are currently examining the possibility that the availability of the acidic and basic equivalent counterions determines the rates of and during acid-base disturbances in trout.

This study was supported by a NSERC operating grant to CMW. Many thanks to M. Kovacevic for excellent technical assistance. Dr D. G. McDonald is thanked for his helpful comments and access to original data.