ABSTRACT
Techniques for the measurement of unidirectional flux rates in fish which require no anaesthesia or surgery are described.
Resting values for Cl− uptake at 10 and 17 °C were 8·03 ± 1·11 and 13·52±0·95μ-equiv. 200 g−1 h−1 (±S.E.), respectively; and for Na+ the rates were 15·49±0·40 and 26·30±0·36, respectively.
Hypercapnic acidosis caused an increase in Na+ uptake, presumably through Na+/H+ (or ) exchange. It is suggested that this is a com-pensation mechanism leading to the increase in blood buffering observed in response to hypercapnia.
Alkalosis was observed following acute temperature increase and was accompanied by an increase in the rate of exchange and also by an increase in Na+/H+ exchange.
The role of these branchial ion exchange mechanisms in overall acid-base regulation is discussed.
INTRODUCTION
In a theoretical discussion of CO2 exchange and pH regulation in aquatic poikilo-therms, Rahn (1967) has proposed that OH−/H+ ratio, rather than [H+], is maintained at a constant value as temperature changes. An inverse relationship between blood pH and temperature is now well established for a number of groups, including Crustacea (Howell et al. 1973; Truchot, 1973), amphibia (Howell, 1970), and fish (Howell, 1970; Randall & Cameron, 1973). Rahn (1967) suggested, and a recent reviewer (Albers, 1970) implies, that the decrease in pH at higher temperatures is accomplished by decreasing ventilation in order to increase the of the blood, as occurs in air-breathers. A constant bicarbonate concentration is implied. Randall & Cameron (1973) have recently demonstrated that the of the blood of rainbow trout does not increase significantly over a wide temperature range, but that instead the concentration varies inversely with temperature. These changes in with increased temperature were not accompanied by any changes in the ratio of ventilation to oxygen uptake (and presumably CO2 excretion), nor does increased ambient CO2 produce a lasting change in ventilation (Janssen, 1973; D. J. Randall & J. N. Cameron, unpublished data).
*Present Address : University of Texas at Austin, Marine Science Laboratory, Port Aransas, Texas 78373 U.S.A.
Hence ventilation and its effects on CO2 diffusion loss do not appear to have anything to do with regulating pH as the temperature changes. Neither does ventilation appear to participate in pH regulation in the face of environmental hypercapnia. Cameron & Randall (1972) detected no change in the CO2 gradient between blood and water at elevated ambient CO2 levels but found pH compensation occurring through an increase in plasma concentration after 24 h.
An alternative mechanism for pH and bicarbonate regulation must then be sought, and recent work on the exchange of for Cl− and on exchange of Na+ for H+ or in the gill suggests that these ion exchanges may possibly play a key role (Maetz & Garcia-Romeu, 1964; Kerstetter & Kirschner, 1972; DeRenzis & Maetz, 1973). An excess excretion of over H+ (in exchange for Cl− and Na+, respectively) would lead to acidification of the blood, and to decreased buffering in response to alkalosis. Conversely, an excess of H+ excretion over excretion would lead to compensation of acidosis. The dependence of the rates of these branchial exchanges on external ion concentrations has been demonstrated (Kerstetter & Kirschner, 1972; Maetz & Garcia-Romeu, 1964), and depletion of ions internally has been shown to lead to pH imbalance (DeRenzis & Maetz, 1973), but so far there has been no demonstration that acid-base disturbances lead to compensatory changes in the exchange rates. A primary purpose of the present study was to investigate the effects of acid-base disturbance on the rates of exchange and on Na+/H+exchange.
The second objective of this study was to determine the quantitative rates of excretion and ion fluxes under conditions that were as physiological as possible. It was of particular interest to determine whether the direct excretion of bicarbonate could account for a significantly large percentage of total CO2 excretion; if so, the assumptions of an earlier simulation model (Cameron & Polhemus, 1974) would not hold, and the dynamics of the CO2 excretion system in the gill would have to be re-examined. It was also of interest to measure the resting rates of ion fluxes under carefully defined conditions of handling and treatment, since such details are seldom generously described in the literature (cf. DeRenzis & Maetz, 1973 ; Kerstetter & Kirschner, 1972; Maetz & Garcia-Romeu, 1964) and are known profoundly to affect the blood pH, electrolytes, and water loss (Hickman & Trump, 1969; Houston et al. 1971; Wood, 1971). A considerable effort was therefore devoted to developing techniques for minimizing handling effects and defining the physiological state of the animals being measured.
MATERIALS AND METHODS
Symbols used : Ji, Jo, Jn Influx, outflux, and net flux − superscripts refer to ion species.
, Oxygen consumption, CO2 excretion rates. Others defined in text.
Experiments were carried out between October 1973 and February 1975 on Arctic grayling (Thymallus arcticus Pallas) obtained from the Chena River, interior Alaska, by electroshocking or by hook-and-line. Mortality from both methods was less than 3 %, and any fish appearing unhealthy after a few days was discarded. In most cases the fish were held in filtered recirculating river water at 10 °C, but some were held at 16 °C (see below). Natural levels of sodium and chloride were very low (⩽0·1 mm), so for analytical convenience, NaCl was added to the holding tanks to bring concentrations of both ions to 0·9 − 1·2 m-equiv./l, still well within the natural range for fresh waters inhabited by grayling. All tests were carried out with water from the holding tanks. The fish were fed brine shrimp and chopped liver, but not in the 1–3 days before tests.
Recirculating chamber apparatus
The main components of the chamber developed for the experiments are shown in Fig. 1. The holding chamber is constructed of blackened 12 mm Perspex and has a volume of about 1 1 (without fish). Water is pumped with a tubing pump from the chamber through the temperature jackets of a pH electrode and an oxygen electrode, then into an aeration reservoir, and finally back into the chamber. Oxygen tension is maintained by bubbling air through an air stone in the reservoir, the air first having passed through three washing flasks for humidification. The entire apparatus is immersed in a temperature bath, including the air-washing flasks. Total volume of the system with a 200 g fish in the chamber was usually between 1200 and 1400 ml.
This apparatus allows measurement of not only the salt (Na+ and Cl −) fluxes but also of the oxygen consumption rates by monitoring flow through the chamber and the difference between inflow and outflow. Since resting values for grayling were known from previous work (J. N. Cameron, unpublished data), it could be assessed whether or not the animal was in a resting state. Typical results are shown in Fig. 2, an experiment measuring the effects of temperature change (see Results).
Another important feature was that the animals were never anaesthetized, nor were any surgical procedures necessary. Handling of the animals was very limited : animals were gently transferred from the holding tank to a large bucket of water and from there to the chamber by cradling the fish gently in the hands. With practice, the fish could be lifted and quickly transferred without struggling and without having to actually grasp the fish, as there seemed to be a moment’s lag in their response. Those fish that did struggle and were handled less gently were returned to the holding tanks for at least 2 weeks.
In some of the early experiments, fish that received rougher handling sustained high net loss rates of Cl −, and after a time, the influx rates also rose, probably due in part to internal Cl − depletion. Later, we actually used the net flux rate as a measure of our handling skill ; poor handling could cause an increase of two-to five-fold in the rate of salt loss, and so these fish were rejected.
Protocol for the temperature experiments
Protocol for the first set of experiments on the effects of temperature increase was as follows: 500 ml of water from the holding tanks was added to the chamber, followed by introduction of the fish, as explained above. The lid was fastened, then the chamber and the remainder of the system was also filled, and the total water volume was noted h was allowed for acclimation to the chamber, and 1 − 3 μ Ci of Na36Cl was then added to the system. After a suitable mixing interval, periodic sampling was begun. For each interval, the chamber flow rate, inflow and outflow oxygen tensions and temperature were measured; samples were taken for analysis of the total chloride and36Cl radioactivity. Total chloride analyses were performed immediately so that we were constantly aware of the status of the animal.
After about 24 h at 1 ° C, provided the usual initial loss of chloride had been stopped for at least 12 h, the temperature was raised to 17 ° C during a period h. The sampling routine was then continued for about another 24 h.
As controls, one series of fish was followed at 10 ° C for about 48 h and a second series at 17 ° C for 48 h. This latter group had been acclimated to 16 ± 0·5 ° C for 2 − 4 weeks.
To document the alkalosis occurring after a temperature increase, catheters were implanted in the dorsal aortae of six fish, following the method of Smith & Bell (1964). After 1 − 3 days recovery, these animals were put through the procedure outlined above, except that approximately 0·2 ml of blood was taken for pH measurement at each sampling interval. Samples were replaced with an equivalent amount of saline having Cl − identical to that of plasma.
Protocol for hypercapnia experiments
The methods for the series of experiments testing the effects of hypercapnia were like those for the temperature experiments, except that instead of changing the temperature after the first 24 h, the air was replaced by a mixture of 1 % CO2 in air This mixture was provided by a gas mixing pump (Wösthof), and was checked for accuracy against standard gases.
Sodium fluxes were also measured in this series and in a second temperature series using 22Na. The only modifications necessary were the taking of extra samples for total sodium and 22Na radioactivity analyses, and changes in the isotope counting methods for dual-labelled samples.
At the end of each experiment, the following data were collected : fish weight, final chamber volume, haematocrit, plasma chloride, plasma sodium, plasma specific radioactivity of 36Cl and 22Na.
Chemical and isotope analyses
Oxygen tensions and pH were measured using temperature-jacketed electrodes (Radiometer-Copenhagen) incorporated into the chamber system. Chloride and sodium analyses were done by electrometric titration (Buchler − Cotlove) and atomic absorption spectrophotometry (Perkin-Elmer), respectively.36Cl activity was assayed by liquid scintillation counting, 22Na by a crystal scintillation detector. In dual-labelled samples, a curve was established for subtraction of the sodium radioactivity from the liquid scintillation counts to give chloride radioactivity.
Calculations
Chloride space was calculated as: where and are the initial and final radioactivities of Cl−in the bath (cpm ml−1), V = bath volume (ml), and is the radioactivity per ml of plasma (cpm ml−1) at the end of the experiment.
All computations and statistical analyses were performed with programs written for a 32K mini-computer (Nova 800, Data General).
RESULTS
Resting flux rates
Table 1 contains a summary of data obtained in the recirculating chamber apparatus from the fish acclimated and tested for 24−48 h at 10 °C. The first 1−4 h of each experiment were not included in the data analysis, since oxygen consumption and flux rates were often somewhat elevated during the initial period. Using the mean data for each fish, the proportion of CO2 excretion occurring via exchange was calculated, based on the assumption that (1) RQ = 0·9, and (2) there is a 1:1 stoichiometry between Cl− uptake and excretion (DeRenzis & Maetz, 1973).
Variability of the flux rates over time was fairly large, as has been reported previously (Mackay, 1974). As a consequence, trends or small changes in flux rates could often be detected only by statistical analysis of many observation periods.
A similar summary for 4 fish tested and acclimated to 17 °C is also given in Table 1. Chloride influx was somewhat higher, and the net loss was a little less than at 10 °C.
The proportion of CO2 excretion occurring as - exchange increased from 2·8−3·9%) but this difference was not statistically significant.
Effects of temperature increase on blood pH
The pH of samples of grayling blood equilibrated in vitro at a constant , of 2·26 torr changed only 0·040 pH units between 5 and 20 °C, a dpH/dT of 0·0027 °C. The time course of pH and OH−/H+ ratio following temperature change in 6 grayling is shown in Fig. 3. Data points immediately following the temperature increase resemble the in vitro data, resulting in a relative alkalosis at 17 °C. Compensation occurred over the next 24 h, however, so that the final pH was about 0·12 pH units lower than at 10 °C, yielding an in vivo slope of 0·017, which is the same as has been reported for rainbow trout (Randall & Cameron, 1973) and most other poikilotherms (Reeves, 1972). At this point the original OH−/H+ ratio was restored to the normal range. The severe immediate alkalosis (pH 8°2) following temperature change in one fish is unexplained and was not typical.
The passive pH change expected from the Henderson-Hasselbach equation as temperature is changed from 10 to 17 °C and total CO2 is held constant is small, probably from about 8·02 to 7·99 based on the pK’-temperature relationship for other fish (Albers, 1970). This is not known precisely because data relating pK’ for grayling blood and temperature are not available. In summary, though, relative alkalosis is a result of increased temperature in grayling ; in order to conform to a constant OH−/H+ ratio, must be reduced by about 15 % in response to a 7 °C temperature rise.
Effects of temperature increase in chloride fluxes
Branchial influx of chloride and the ratio of influx to outflux increased after temperature increase in all nine fish studied (Table 2). Following the temperature increase net flux of chloride averaged 7·56 μ-equiv. 200 g−1 h−1, which was also significantly greater than in either the 10°or 17 °C controls. There seemed to be a trend for higher influx rates in the first 12 h following temperature increase than in the second 12 h, which would match the time course of alkalosis shown in Fig. 3, but the large inherent variability in fluxes prevented statistical demonstration of such a trend.
Efflux rates (branchial + renal) and temperature were not related, but efflux was weakly correlated with oxygen consumption (R = 0·36, P < 0·05).
Effects of temperature increase on Na+ and CF−fluxes
When the changes in Na+ and Cl− uptake were measured simultaneously in a separate series of experiments, the results were more complicated (Table 3). Although the rate of exchange was also elevated in these experiments, and the ratio of decreased, the rate of Na+ uptake increased by a larger absolute amount. The four-fold increase in influx-to-outflux ratio for chloride is consistent with the idea that increased excretion is occurring to compensate for the relative alkalosis, but the Na+ uptake data indicate that an even greater amount of H+ (or ) exchange is occurring. Influx-to-efflux ratio for sodium rose from 0·95 to 1·42. At the same time the oxygen consumption rose from 31·4 ± 6·3 at 10°C to 48·4 ± 3·1 ml kg−1 h−1 at 17 °C.
Effect of hypercapnia on chloride fluxes
Results of seven experiments testing the effects of hypercapnic acidosis (1 % CO2, or 7·5 torr) on chloride fluxes alone are summarized in Table 4. In two of these experiments fluxes were studied for an initial control period (approx. 24 h) and a subsequent period of hypercapnia (approx. 24 h). In the other 5, the initial control period was shortened to 6−12 h, depending on how rapidly fluxes stabilized. Following the hypercapnic period (12−18 h) a second control period was employed. The reduction of influx between the first control period and the hypercapnic period was significant overall and in five of the seven individual experiments. Influx increased following return to the control conditions, but was significantly greater in only 3 of the 5 individual experiments.
Effect of hypercapnia on Na+ and Cl− fluxes
A series of four experiments was performed measuring both Na+ and Cl− fluxes over a 24-h control period and a subsequent 24-h hypercapnic period at 10 °C (Table 5). As in the previous series where Cl− flux alone was measured, changes in the chloride fluxes were small in response to hypercapnia; they were not quite significant statistically in this series. The changes in sodium fluxes were quite pronounced, however, with a large increase in the branchial influx and a significant decrease in the total (branchial + renal) efflux. The ratio of sodium-to-chloride efflux decreased, sodium-to-chloride influx ratio increased more than two-fold, and there was an increased net gain of sodium relative to chloride.
DISCUSSION
(A) Methodology
Ion fluxes in teleosts have been studied in three fundamentally different types of apparatus: aquarium preparations using intact fish; perfusion apparatuses using intact, anaesthetized fish; and various forms of isolated gill preparations. Choice of method depends on experimental aims, but from the standpoint of evaluating the quantitative role of various processes in the physiological functioning of a whole fish, the isolated gill method seems clearly inappropriate, and will not be discussed further here.
The second method, as described by Kerstetter, Kirschner & Rafuse (1970) was the first one I tried, since the small bath volume has distinct advantages. I found that the fish sustained high loss rates of salts, that evaporation was a constant problem, and in general that conditions were so unphysiological that the method was unsuitable. The recirculating chamber used subsequently allows natural ventilation, allows monitoring of oxygen consumption, minimizes the need for handling and abrasion of the fish, and does not require anaesthesia or surgery. The fish could also be held in it for long periods, so no experiment had to begin until it was determined that salt loss rates were not excessive and that the animal had reached the resting state (determined by ). Disadvantages of the system are the larger volume, which reduces sensitivity to small changes, and the lack of discrimination between renal and branchial ion efflux. Since ion uptake was of primary interest, I judged it more important to avoid the stresses associated with bladder catheterization or blockage.
(B) Role of exchange in CO2 excretion
The conversion of plasma bicarbonate to CO2 via carbonic anhydrase in the ery-throcyte (or gill tissue) and the subsequent diffusive movement of CO2 is the principal pathway for excretion of CO2 (Reaction 9–8–6–5, Fig. 4). Data presented in this study indicate that only 2·8 to 3·9% of the total , occurs via the branchial exchange pathway (Table 1), given that all the Cl− uptake is in exchange for as indicated by the recent work of DeRenzis & Maetz (1973). It has not been established whether the source of for exchange is the plasma pool or generated from CO2 within the epithelium.
Comparison with data from other fishes is extremely difficult due to differences in method, lack of temperature data, and the lack of simultaneous measurement of or, in other published work. The chloride fluxes reported per unit weight (200 g) in the goldfish range from 32 to 100 μ-equiv. h−1 in some studies (DeRenzis & Maetz, 1973; Maetz & Garcia-Romeu, 1964), but as low as 2μ-equiv. h−1 in another (Mackay, 1974). Kerstetter & Kirschner (1972) report rates of 22−39 μequiv. 200 g−1 h−1 for rainbow trout using their perfusion method, compared to 8 and 13 for grayling (Table 1). A problem with such comparisons is that fish have frequently been acclimated to low Cl− levels (0·1 mm or less), which may depress plasma [Cl−], and have then been tested at higher concentrations, often about 1 mm (DeRenzis & Maetz, 1973; Kerstetter & Kirschner, 1972).
With these problems in mind, it does not seem safe at present to say whether branchial . exchange activity is greater in goldfish than in grayling or trout.
(C) Role of branchial ion exchange in acid-base regulation
The adjustment of buffer status in response to hypercapnic acidosis is clearly not passive; and from in vitro measurements, calculations using the Henderson−Hassel-bach equation, and studies of more complex buffer systems (Reeves, 1972), the adjustment of buffer status following temperature change is not passive either. Ventilatory control of as temperature changes was ruled out by the work of Randall & Cameron (1973) and seems also ruled out in the case of hypercapnic acidosis (Janssen, 1973; D. J. Randall & J. N. Cameron, unpublished results). A consideration of the pathways for CO2 and reactions (Fig. 4) would suggest that alteration of the relative rates of branchial exchange and Na+/H+ exchange offers a mechanism by which the buffer capacity of the blood can be adjusted at constant .
In the present study, respiratory acidosis or alkalosis was induced independently of ion concentrations. The significant response of Na+ uptake to hypercapnic acidosis (Table 5) is strongly indicative of a key rôle for branchial Na+/H+ exchange in acid-base regulation. A Davenport diagram of acid-base events occurring during hyper-capnia is shown in Fig. 5. Point A was established using data from the grayling (this study). The slopes of AB and AC were taken from Cameron & Randall’s (1973) data for rainbow trout. The initial response to hypercapnia is a shift from A to B, following the in vitro buffer line (Cameron & Randall, 1972: Janssen, 1973). Over the next 24 h, acid-base status shifts along the .isoline toward Point C, yielding an A-to-C in vivo buffer line. The mechanism proposed for increase of bicarbonate buffering is retention of formed from metabolic CO2, and excretion of H+ ions in exchange for Na+ at the gill. Stoichiometric equivalence of Na+ uptake for H+, or its functional equivalent, is assumed (Maetz, 1973).
Assuming that space is about the same as Cl− space, or 25 % of body volume, we see that a 200 g fish needs to increase concentration from about 8·5 to 25 mm in 25 % of 200 (= 50) ml of fluid for a total increase of 875 μM, and needs to excrete the same amount of excess H+ ions. The difference between Na+ and Cl− uptake increased from 6·47 μ-equiv. 200 g−1 h−1 to 19·59 μ-equiv. 200 g−1 h−1, a change of 13·12 (Table 5). Since there was no change in the rate of oxygen consump-tion, presumably there was no change in the rate of metabolic acid production, so that this entire increase can be attributed to buffer readjustment. At this rate, 66 h would be required for grayling to achieve the degree of compensation implied by the in vivo slope AC (Fig. 6). Janssen (1973) has shown that the compensation process in the rainbow trout continues up to 72 h, and it seems likely that the grayling have a similar time requirement. Compensation would occur more rapidly if there is significant renal H+ secretion. Reduction of total Na+ efflux (Table 5) may be due to increased renal resorption of Na+ in exchange for H+. We do not at present have sufficient data to show how the acid secretion is partitioned between the gill and the kidney, but the present study indicates that most, if not all, of it is carried out by the gill during hypercapnic acidosis.
The results from the thermal alkalosis experiments (Tables 2 and 3) are not so clear. On the one hand, the exchange was stimulated by thermal alkalosis, and the ratio of sodium uptake to chloride uptake fell, as predicted by the hypothesis. On the other hand, the absolute increase in sodium uptake was greater than the increase in Cl−, which argues against the hypothesis. A further complication is that the rate of metabolism rose 54% (as measured by ), and so presumably the rate of metabolic acid production would also rise. Temperature per se undoubtedly has other effects on salt balance, such as permeability changes. Whether the observed increase in exchange activity contributed to compensation of the alkalosis, then, depends on knowing not only what is shown in Table 3, but also upon knowing the rate of metabolic generation of excess H+ and whether there is any change in the excretion of H+ or by the kidney. Acid-base events following temperature increase are summarized as a Davenport diagram in Fig. 6, but whether branchial ion exchange mechanisms are responsible for the compensatory shift (B to C) cannot be stated at present.
Of course the control of the rates of these branchial ion exchange mechanisms is not simply a matter of internal acid-base status. Uptake of both sodium and chloride has been shown to depend upon concentrations of the ions, both internally (DeRenzis & Maetz, 1973) and externally (Maetz & Garcia-Romeu, 1964; Kerstetter et al. 1970; Maetz, 1971). The total excretion of CO2 has also been shown to depend on the availability of external chloride (Dejours, 1969). Conversely, loading has been shown to stimulate Cl− uptake, although my own analysis of one of these reports (Kerstetter & Kirschner, 1972) shows that the difference is not statistically significant; and in the other (Maetz & Garcia-Romeu, 1964), it is difficult to say what the effect of the injection was: does intraperitoneally injected diffuse as the charged ion into the blood, leading to alkalosis, or does it react with H+ locally and then diffuse into the blood as CO2, causing acidosis ?
In summary, the following points emerge: first, that the rates of salt flux in un-stressed, resting fish are low in fresh water, leading to long pool turnover times; second, the normal rate of exchange does not constitute a major pathway for CO2 excretion, amounting to only a few percent of the total ; and third, at least in the case of hypercapnic acidosis, the relative rates of exchange and Na+/ H+ exchange change in a manner leading to compensation of the acidosis. The possibility remains that alkalosis can also be compensated for in this manner, but a more complete picture of renal activity and effects of temperature on metabolic acid production is needed to settle this point. It is tempting to speculate that the basic mechanisms at work in the fish gill are the same as those now studied intensively in the kidney of higher vertebrates (Steinmetz, 1974), and that one of the major physio-logical changes during the course of the evolution of land animals was a transfer of this entire function from the gill to the kidney.
ACKNOWLEDGEMENTS
I wish to express my appreciation for the expert technical assistance of Mrs Karen Henderson, and to acknowledge support of research grants made to me: National Science Foundation Grant GB-39785, and National Heart and Lung Institute Grant 5-RO1-HL-14822.