Intracellular and extracellular acid-base status as a function of temperature in the freshwater channel catfish, Ictalurus punctatus¹

Although the inverse relationship between temperature and the pH of fish blood has been known for about 15 years (Rahn, 1967) and subsequently verified for many species, complete acid-base parameters have only been measured in a few. In order for pH to decline at increasing temperature, either the total CO₂ must decline, or the rise, or a combination of both must occur. Those studies done to date or fish have shown a variable pattern : in the rainbow trout, the pH decrease is acconW panied by a decline in total CO₂, with virtually no change in (Randall & Cameron, 1973), while in the marine dogfish there is a substantial increase in the with a slight rise in the total CO₂ as well (Heisler, 1978; Heisler, Neumann & Holeton, 1980). The pattern observed in other water-breathers is intermediate, with some small rise in and a decline in total CO₂ (Cameron & Batterton, 1978 ; Cameron, 1978; Heisler, 1980). The air-breathing ectotherms appear to have a more predictable response, with the rising at higher temperatures at a rate that results in a constant total CO₂ content (Jackson, Palmer & Meadow, 1974). In the air-breathing ectotherms the ventilatory control of is apparently a primary regulator, but the high ventilation requirement in water in order to supply oxygen needs precludes ventilation as a means of regulating in water-breathers.

The changes in total CO₂ content are dependent upon changes in strong ion differences (SID), and are generally thought to involve linked ion exchanges in the gills (Maetz & Garcia-Romeu, 1964; De Renzis & Maetz, 1973; Cameron, 1976). Adjustment of pH by change in SID takes longer and requires more energy than if the change were effected by change in

Even fewer data have been published on the intracellular pH in fishes : the dogfish Scyliorhinus stellaris has been investigated in some detail (Heisler, Weitz & Weitz, 1976; Heisler, 1978; Heisler & Neumann, 1980; Heisler et al. 1980), and preliminary data have been published for Synbranchus marmoratus (an air-breather) and Cyprinus carpio (carp) (reviewed by Heisler, 1980). These data generally show that changes in intracellular pH parallel those in the blood, and indicate some role for intra-/ extracellular transfers in the adjustment to temperature.

The objectives of this study, then, were to investigate the changes in intra-and extracellular acid-base status in a further species of teleost fish as a function of temperature, to assess the relative importance of changes in and SID in bringing about pH adjustment, and to measure the branchial and renal acid-base-related excretion during temperature adjustments.

All the experiments were performed with channel catfish, Ictalurus punctatus Rafinesque, obtained from the Adams Fish Farm, Angleton, Texas. The fish were maintained in the laboratory without feeding, in running dechlorinated Port Aransas tap water, and were usually used within two weeks after purchase. The composition of the tap water was similar to local surface waters, with Na⁺ and Cl^- averaging about 3 m-equiv. I^-1, and Ca²⁺ about 1·6m-equiv. I^-1. The fish ranged in weight from 490 to 1500 g, and were held at temperatures ranging from 22 to 25 °C, at a partial pressure of CO₂ (⁠⁠) less than 0·5 torr.

The protocol for most of the experiments was as follows: after at least two days’ acclimation to the laboratory, a catheter was surgically placed in the dorsal aorta using the method of Soivio, Nyholm & Westman (1975), under 1:10000 MS 222 anaesthesia. At the same time, a urinary catheter was also implanted using a method lightly modified from Wood & Randall (1973). The urinary catheter was constructed om polyethylene tubing (PE 60), flared slightly at one end, and enlarged twice about 5 and 10 mm from the tip by heating to form spherical enlargements or ‘blebs’, that served to make the catheter fit snugly in the ureters. A 1·5 × 2 cm piece of surgical rubber sheet was laid over the papilla externally, its corners fastened with sutures, and the rubber glued to the fish with a cyanocrylate cement. This procedure avoided the use of a ligature around the urinary papilla, which in our experience caused rapid tissue erosion and early failure of the catheters by leakage or loss. Using our procedure, the catheters remained patent for several days.

Surgery usually took about 15– 30 min, after which the fish were placed in darkened lucite chambers provided with a copious flow of well-aerated water. At least 24 h recovery was allowed in these chambers before starting any experimentation. As a measure of full recovery from surgery, the blood pH was measured, and if it had not reached stable values after 24 h a second full day of recovery was allowed. At the end of the recovery period, measurements of urine volume flow were made continuously at 2 h intervals (longer at night), and the urine samples were saved for analysis of pH, total CO₂ (C_T), titratable acidity (TA), and ammonia concentration. At various intervals, usually at the mid-point of the urine collection periods, blood samples were taken from the dorsal aortic catheter for analysis of plasma pH, C_T and Following 24 h of these control measurements, the temperature of the chamber system in which the fish were held was changed over about a 30 min period to a higher or lower value, at which the fish were held for another 24– 48 h. The same blood and urine measurements were continued throughout this second temperature treatment.

At least once during the day following surgery, the fresh flow to the chamber was cut off, and the closed volume re-circulated with constant aeration. The volume of the closed recirculation system was about 3– 6 times that of the fish. During these periods, the total ammonia concentration and change in titratable acidity were measured in the water at half-hourly intervals for calculation of net branchial acidbase flux. Titratable acidity was measured by titrating water samples to a pH of 4·000 with 0·010 N-HCI, SO that differences from one interval to the next represent changes in the SID of the bath. This procedure was repeated on the second day after surgery, and again at several intervals following the temperature change. Some fish were put through the same protocol, except that no temperature change was made.

For intracellular pH measurements, approximately 10 μCi of [³H]mannitol and 4 μCi of ¹⁴C-labelled 5,5-dimethyl– 2,4-dioxazolidinedione (DMO) were infused over a 2 min period into the dorsal aorta in a total volume of 0·25 ml saline. The blood (plasma) concentrations and acid-base parameters were measured at half-hourly intervals beginning 1 h later, and at 4 h a final blood sample was taken and the fish quickly killed by a sharp blow on the skull. The blood samples were taken from undisturbed fish via the catheters, so that no acidosis produced by struggling (as in ‘grab and stab’ sampling) interfered with the results. Samples of white muscle, red muscle, heart and brain were quickly dissected out, weighed fresh, and placed in an oven at 65 °C for drying to constant weight. Six replicates of white muscle, 5 of red muscle were processed, but due to their small size the heart and brain each| yielded only a single value.

The acid-base measurements for plasma and urine were made with routine electrode techniques (pH, ⁠) except that the C_T was determined using a conductometric apparatus constructed by one of us, based on methods outlined by Maffly (1968). The estimation of titratable acidity (TA) in the urine was made by titrating an aliquot of the 2 h collection to the pH of the blood at the mid-point of the collection interval. In a few samples of urine which had pH values higher than blood, titration with base was performed, but since there was no significant titratable base present other than bicarbonate, data from these titrations were not used. Total ammonia concentration of the urine and water samples was measured using the phenolhypo-chlorite method (Solorzano, 1969).

The mean whole-body intracellular pH was calculated on the basis of the total DMO dose, the concentration in plasma at 4 h (previously determined to be adequate time for equilibration, Cameron, 1980), and the extracellular space determined by extrapolation of the mannitol clearance data to zero time, according to formulas given by Cameron (1980). The dried individual tissue samples were combusted (Packard Sample Oxidizer, Model 306) so that the ³H and ¹⁴C radioactivity could be assayed separately, whereas the plasma samples could be counted directly. All radioactivity assays were performed with a liquid scintillation counter (Packard 3255) using the external standardization method, and correcting all samples for counting efficiency.

In order to determine the non-bicarbonate buffer value of blood at varying haematocrit, samples were taken from 12 fish and either analysed at their normal haematocrit (which varied from 10 to 32%) or at a lower haematocrit produced by removing some of the cells by centrifugation, followed by re-suspension of the remaining cells. The samples were equilibrated at 0·2, 0·4, 1·0 and 2·0% CO₂ in air, and at each gas tension their pH and C_T were measured as described above. The slope of the plot of pH vs. C_T is the buffer value, β, in m-mole/pH, or slykes. Tissue samples were treated somewhat differently, according to the method of Bate-Smith (1938): fish were rapidly sacrificed, samples of tissue quickly dissected out and weighed immediately, and the tissue then ground in 10 ml of iced 0·9% NaCl with 20 strokes of a glass tissue homogenizer. These homogenates were then equilibrated in a stirred titration vessel with CO₂-free nitrogen, enough NaOH added to bring their pH to about 8·0, then titrated with HC1 to pH 6·5. The slope of the titration curve between 7·1 and 7·5 was taken as the buffer value, since this was the approximate range of intracellular pH.

The values for arterial pH, total CO₂ and obtained from 61 fish are shown in Fig. 1. Numbers of determinations shown in Fig. 1 vary slightly, since not all data sets were complete; each point represents the mean of from 1 to as many as 12 measurements from one fish. The regression equations describing the fish mean values were :

The data for the calculated mean whole-body intracellular pH (pH₁) are given in Fig. 2, along with data obtained from individual tissue samples of white muscle, red muscle, heart and brain. The statistics for the regression lines on temperature are given in Table 1. Only the slope for heart muscle is significantly different from the rest (5% level, analysis of covariance). While the difference in temperature slope between the mean whole body pH, and the individual white muscle pH] values was not significant, the line for the mean whole body pH! was shifted about 0·1 units upward, thus overestimating the pH of both red and white muscle. Heart and brain had the highest pHj values, so that at 25 °C the means were: blood 7·803, mean whole body 7·315, heart 7·486, brain 7·396, red muscle 7·221 and white muscle 7·216. That the mean whole-body estimate falls roughly in the middle of the various tissue values is probably coincidental, since the majority of the intracellular compartment consists of white muscle, the rest contributing only a minor fraction (heart and brain less than 1 % each).

The relationship between the haematocrit and the buffer value of the whole blood is shown in Fig. 3, the data in which are described by the linear regression: β = 5·76 + 0·34 Het, r = 0·94. The mean of the white muscle determinations (N = 5 fish, 12 determinations) was 35 + 2 slykes (± s.d.), or 35 m-mole.kg H₂O^-1.pH^-1 between 7·1 and 7·5. (At a range of 6·6 to 7·1, the buffer slopes of the tissue titrations averaged 55% higher.)

In order to assess the CO₂ pool size at various temperatures, some model calculations were carried out, utilizing the data given above, and the procedure described below. Mean values for extra-and intracellular pH are given in Table 2 at 15, 22 and 31 °C, representative temperatures in this study. The PH₁ data are taken from Fig. 1 and the pH₁ data used were the white muscle values given in Fig. 2. Also shown are the total CO₂ concentrations measured in the ECF (Fig. 1) and calculated for the ICF, as described in Cameron (1980). The values are also taken from Fig. 1.

In order to use these data to calculate the total CO₂ pool size, first the extracellular pool size was calculated by multiplying its concentration by the volume of the pool (186 ml.kg^-1, Cameron, 1980). A similar procedure was used for the intracellular pool (493 ml.kg^-1), and the sum of these two pools was the total (∑). Then this was done for each temperature, the differences (Δ) between 15 and 22 °C, and between 22 and 31 °C were calculated, and are given in Table 2. Thus there was a 0·264 m-mole.kg^-1 increase in the total CO₂ pool going from 22 to 15 °C, and a decrease of 0·244 m-mole.kg^-1 going from 22 to 31 °C. Nearly all of the total CO₂ present is in the form of HCO₃^-, a ‘volatile’ base, but these numbers may or may not correspond to the measured net H⁺ or OH^- excretion following temperature change, depending on the composition of the intracellular buffer pool (see Discussion).

As a measure of the contribution of changes in the various parameters, two further sets of calculations were carried out. The first employed the supposition that had not changed, and the resulting pool sizes and the magnitude of temperature-induced changes under these hypothetical conditions are also given in Table 2. Finally, the calculations were repeated using the same data except that the product of and solubility (i.e. ⁠) was held constant. The calculated pool changes would have been considerably larger had either of these conditions been met, which highlights the effect of variations of the in reducing the required acid-base-related fluxes.

The effects of temperature elevation were studied in 10 fish that had been acclimated to holding temperatures of 22 ± 1 °C. These fish were catheterized on Day 1, control measurements made at 22 °C on Day 2, then further control measure ments on the morning of Day 3, following which the temperature was increased in about to 31 °C. Further measurements were made throughout Day 3, and finally 24 h after the temperature change on Day 4. The net H⁺ or OH excretion from kidneys was measured by monitoring urine flow rate, pH, titratable acidity, total CO₂ and ammonia concentration. Branchial measurements were made as described above, and included ammonia and titratable acid excretion. The results of the net acid excretion measurements are shown in Fig. 4.

The increase in net OH^- excretion following the temperature increase was statistically significant for the gills at 0–2 h and 2–4 h, and for the kidneys from 0 to 6 h. The total excreted by the gills was obtained by summing the difference between the observed rates from o to 4 h and the control rate, taken as the average of the control periods just prior to the temperature increase, and 24 h afterwards. A total of 0– 18 in mole. kg^-1 net OH^- was excreted above the control rate, and after similar calculations the renal pathway contributed an additional 0·22 m-mole.kg^-1, for a total of 0·40 in mole, kg^-1 net OH^- excreted. This is about 60% more than predicted by the CO₂, pool calculations in Table 2.

Similar experiments were carried out with 12 fish acclimated to 24 ± 1 °C, with a control period at 25 °C, and a temperature decrease on Day 3 to 15 °C. The results of these experiments are shown in Fig. 5, and whereas there was again a significant change in the branchial excretion rate from 0 to 6 h following the temperature change (P < 0·05, one-way ANOVA), the renal data show a small but significant switch from net OH^- to net H+ excretion that persisted, rather than any pulse of net H+ output. The summed contribution of the gills was 0·166 m-mole.kg^-1 net H⁺ excreted, and if the 24 h value for the kidneys is averaged with the control rate for purposes of computing a baseline value, the renal pathway then contributed another 0·14 m-mole.kg^-1, for a total of 0·306 m-mole.kg^-1 net H+ excreted, or about 19% less than predicted for a 10 °C change using values from Table 2.

Unlike the response to infused acid and base loads (Cameron & Kormanik, 1982), the response to temperature did not involve any change in ammonia excretion by either gills or kidneys. There were no significant changes in the rate of ammonia excretion by either route during temperature increase or temperature decrease (one way ANOVA). The principal change in the urine following temperature increase was a rise in the urine bicarbonate concentration, and after a temperature decrease, a rise in titratable acid. With our methods the branchial response manifests itself as a change in titratable acidity (or alkalinity) of the water, but is not chemically identified.

Before and after the temperature change treatments, blood acid-base parameters were measured at frequent intervals (pH, and C_T), and conformed to the patterns shown in Fig. 1. Following the step increase or decrease in temperature, the new values were generally reached in less than 4 h, and remained stable for the remainder of the experiments. In a few cases the experiments were carried on for another day, and no difference was found between 24 h and 48 h values. Thus it appears that the adjustments of the acid-base pool to temperature change were complete within 4– 6 h.

The general pattern of blood acid-base regulation as temperature changed was similar to that described for several other species of fish (Rahn & Baumgardner, 1972; Cameron, 1978; Heisler, 1980), with a slope of pH_e vs. temperature very close to that often observed, i.e. – 0·013 pH/°C. The inverse relationship between pH and temperature was accomplished in the channel catfish by a greater increase in ⁠, and a smaller decrease in C_T than is usually described for teleosts, and so the channel catfish does not really maintain a ‘constant content’ system’ like an air-breathing ectotherm (Jackson et al. 1974), nor do they have a ‘constant partial pressure’ system. The data given here resemble somewhat the data given for ‘adult’ dogfish (Scyliorhinus stellaris) (Heisler et al. 1976), in which there was an increase in C_T at higher temperatures and an even larger increase in ⁠.

The of the blood in a water-breather is not thought to be under active control by gill ventilation, as it is by lung ventilation in the air-breathers. The evidence for this is partly that the ratio of ventilation to CO₂ production declines in the air breather at high temperature (Jackson et al. 1974) so that the gradient in the lung (and arterial ⁠) rises, whereas the ratio of ventilation to CO₂ production (or oxygen consumption) does not change at high temperatures in water-breathers (Randall & Cameron, 1973; Cameron & Batterton, 1978). The critical difference here is that oxygen availability in water dictates a very high rate of ventilation in the water-breather relative to the CO₂ excretion requirements, and so the ventilation rate is partially uncoupled from CO₂.

What then is the probable mechanism for the rise seen in the as the temperature changes? Since the total ventilation apparently does not change, there must be changes in either the blood flow in the gill, or the effectiveness of the blood and water flow distribution in the gills. It is, however, a bit puzzling that such effects have not been seen as a reduction in the efficiency of oxygen exchange. In the rainbow trout, the oxygen gradient across the gills apparently does not change between 6·8 and 20 °C (Randall & Cameron, 1973), but there are some changes in O₂ extraction in a water-breathing crab at high temperature (Cameron & Batterton, 1978). The change in in the rainbow trout is not as great as in the catfish, however, so the possibility of a reduction in oxygen extraction in this study, parallel to the increase in the CO₂ gradient, should not be ruled out.

The use of the ‘mean whole body’ method for estimating intracellular pH has been criticized on several grounds, principally that the number derived is not representative of any particular tissue compartment, and that a tissue compartment deviating from the mean whole-body value will bias the estimate in a non-linear way (Manfredi, 1963; Waddell & Bates, 1969). The error resulting from non-homogeneity can be shown to be rather small by a simple model calculation, and in any case the same criticism may be applied to the use of individual tissue determinations by the DMO method, since the intracellular compartment is itself heterogeneous. The comparison conducted in the present study, then, was of some interest, since to our knowledge values obtained by both the mean whole-body method and individual tissue analysis have not been compared in any poikilotherm. The mean whole-body values (Table 1, Fig. 2) were biased upwards by about 0°1 pH unit, but yielded a reasonably faithful estimate of the changes in pHj of the individual tissues as temperature changed.

Aside from heterogeneity of the various tissue pH₁ values, other errors can affect the mean whole-body estimate. Excretion or metabolic breakdown of the DMO seems the most probable, since the other variables that enter into the calculation (total dose administered, final plasma radioactivity, and final plasma pH) may be measured fairly precisely. In an earlier study it was determined that the mean whole-body estimates reach a plateau in 2– 4 h, giving an estimate of mixing time (Cameron, 1980), but the rate of excretion or metabolism of DMO has not been measured in any fish.

The individual tissue values for pH₁ (Table 1, Fig. 2) all conform in general to the pH-temperature relationship described for blood, except that pH₁ is about 0·5 units below the blood. Although most of the individual temperature slopes were not significantly different from one another (except heart), the pattern of red muscle > white muscle > heart, as reported for the dogfish (Heisler et al. 1976) and carp (Heisler, 1980), was observed. It is tempting to speculate that the difference for heart muscle may be related to the much higher concentration of myoglobin (acting as a protein buffer), but that argument would seem to be contradicted by the results from red muscle, which also has higher myoglobin content, and a greater temperature slope. The biochemical consequences of these different pH-temperature relationships are completely unknown, but the differences are sufficiently great that they should be noted by those conducting investigations into pH effects on enzymes of various tissues.

The calculations of the total CO₂ pool at various temperatures as shown in Table 2 agree satisfactorily with pool calculations conducted in an earlier study using the mean whole-body method (Cameron, 1980), and give an idea of the change in total CO₂ content attendant upon temperature changes. Since nearly all of the total CO₂ at these physiological pH values exists as bicarbonate ion, the adjustment to temperature would appear to require considerable transfer of bicarbonate between the intra-and extracellular compartments, and between the extracellular compartment and the environment. These transfers, of course, would be much greater were there no change in the as shown by the alternate calculations in Table 2. This suggests that one ‘function’ of the variations in is to reduce the total energy requirement for an adjustment to temperature change, since the transfer processes are almost certainly energy-requiring.

In all of the foregoing discussion, it is understood that when speaking of bicarbonate transfer one could just as well mean OH-transfer, or H+ transfer in the opposite direction. Experimentally it is impossible to distinguish between these alternative processes, and any of the alternatives must also be understood in terms of strong ion transfer between intra-and extracellular compartments. For example, the transfer of a Na⁺ ion from ECF to ICF without an accompanying strong anion or counter movement of a strong cation will necessarily result in an increase in the SID, and consequently in the intracellular HC0₃⁻ concentration, regardless of whether there is an actual HCO₃⁻ transport. Thus it may be technically incorrect to speak of HCO_S-transfer, as Stewart (1978) has emphasized, but it is a harmless convenience so long as the convention is understood. The terms ‘net H+ excretion’ and ‘net 0H^- excretion’ have been used here to indicate the end result without specifying the actual species transported.

Taken as a whole, then, the pattern of extra-and intracellular pH regulation in response to a temperature increase in the catfish may be summarized as follows: the pH of the blood decreases in a fashion similar to the change in the pK of water, involving a decrease in the total CO₂ (mostly bicarbonate), and varying degrees ofincrease of The decrease in plasma pH is similar to the decrease in intracellular pH, which also involves a decrease in intracellular HCO₃^- (or SID), and presumably a proportionate rise in the intracellular ⁠. The changes in intracellular HCO₃^- appear to be brought about by passive changes in the pKs of the dominant non-bicarbonate buffers, phosphates and imidazole compounds, and by transfer of strong ions (bicarbonate, in effect) between the intra-and extracellular compartments. Several features of the system appear to combine to operate in such a way as to minimize the degree of active inter-compartmental transport.

The authors wish to acknowledge the valuable technical assistance in all phases of the work of Mrs Laura Henry.