Blood Gases, and Extracellular/Intracellular Acid-Base Status as a Function of Temperature in the Anuran Amphibians Xenopus Laevis and Bufo Marinus

The early studies on the inverse relationship between extracellular pH and body temperature in ectothermic vertebrates (e.g. Robin, 1962; Rahn, 1967; Howell, Baumgardner, Bondi & Rahn, 1970; for further references see Heisler, 1986b) led to the formulation of two models: that of a constant relative alkalinity (Rahn, 1967) and the alphastat hypothesis (Reeves, 1972). These models were based on experimental findings which suggested that arterial plasma pH changed with temperature in a manner similar to that in neutral water (ΔpN/ Δt =−0·019 units °C⁻¹ from 5 to 20°C; ΔpN/Δt = −0·017 units °C⁻¹ from 20 to 35°C; Weast, Handbook of Chemistry and Physics; see Heisler, 1986b). However, as more experimental data on individual animals became available, it was increasingly evident that the magnitude of changes in pH with temperature (ΔpH/Δt) was quite variable, ranging from −0·008 to −0·021 in turtles (Robin, 1962; Malan, Wilson & Reeves, 1976), from −0·001 to −0·017 in lizards (Wood, Glass & Johansen, 1977; Wood & Moberly, 1970), from −0·011 to −0·021 in amphibians (Mackenzie & Jackson, 1978; Malan et al. 1976) and from −0·008 to −0·017 in fish (Walsh & Moon, 1982; Heisler, 1984b; for a complete tabulation of data for all orders of animals see Heisler, 1986b).

The alphastat hypothesis of Reeves (1972) predicts that the observed decrease in pH with increasing body temperature (Reeves, 1972; Malan et al. 1976) is based on adjustment towards a constant ionization of histidine-imidazole. As a central criterion of the model, Reeves proposed that pH adjustment was accomplished by regulation of ventilation and therefore⁠. However, it was postulated that the bicarbonate concentration would remain constant because of the non-titration of non-bicarbonate buffers and the lack of any transmembrane or transepithelial, acid— base-relevant ion transfer. This type of pH regulation was claimed for both the extracellular and intracellular body compartments with pH/temperature coefficients (ΔpH/Δt) close to the ΔpK/Δt of biological histidine-imidazole moieties [ΔpK_Im/Δt =−0·018 to −0·024 units °C⁻¹, depending on ligands and steric arrangement (Edsall & Wyman, 1958; see Heisler, 1986b)]. However, with constant bicarbonate concentration and semi-closed buffer system regulation (open only for CO₂, see Heisler, 1986a), constant imidazole dissociation can only be achieved with histidine-imidazole as the predominant non-bicarbonate buffer. The histidine-imidazole of haemoglobin is the predominant non-bicarbonate buffer of the extracellular space, but intracellular tissue buffering usually deviates considerably from this (Heisler & Neumann, 1980; Heisler, 1984a; see also Heisler, 1986b). Nevertheless, the magnitude of ΔpH_c/Δt is highly variable within and amongst different groups of ectotherms, with most data falling outside the predictions of the alphastat model (see Heisler, 1986b).

Considerably less is known about changes of intracellular pH with temperature, even though histidine-imidazole protein residues have been suggested as possible intracellular sensors for the regulation of the ratio of CO₂ production to ventilation and, thus, the regulated variable of the proposed system, (Rahn & Reeves, 1980). In fact, with amphibians and reptiles (the animals for which the alphastat hypothesis was originally conceived), intracellular pH as a function of temperature has been determined in only three species (Malan et al. 1976; Bickler, 1982; P. Neumann, G. M. O. Maloiy & N. Heisler, in Heisler, 1986b). In experiments on the turtle, Pseudemys scripta, ΔpH_i/Δt values in white muscle (− 0·0186 units °C⁻¹) and liver (−0·023 units °C⁻¹) were not significantly different from the range of the ΔpK/Δt of imidazole compounds and of simultaneously measured blood (−0·0207 units °C⁻¹), whereas ΔpH,/ Δt of heart (−0·0122 units °C⁻¹) was significantly lower (Malan et al. 1976). In parallel studies on the bullfrog, the ΔpH/ Δt of mixed types of striated muscle (−0·015 units °C⁻¹) was not significantly different from that of arterial blood (−0·0206 units °C⁻¹). More recently, Bickler (1982) found that brain intracellular pH of the lizard, Dipsosaurus dorsalis, was maintained constant over an 18–35°C temperature range, whereas other tissue compartments conformed to the alphastat model at least in the preferred temperature range. As with the findings of Bickler (1982) for Dipsosaurus brain, the intracellular pH of Varanus exanthematicus white muscle (ΔpH/Δt =−0·005 units °C⁻¹), heart (ΔpH/Δt = −0·003 units°C⁻¹) and oesophagus (ΔpH/Δt =−0·002 units °C⁻¹) changed little with changes in temperature (Neumann et al. in Heisler, 1986b). From these studies, data on intracellular pH in amphibians are available for only one muscle type in one species of Anura (Malan et al. 1976). This scarcity would be of little importance were it not for the primary role that the amphibian model has played in formulating the alphastat hypothesis (Reeves, 1972). Against this background we have examined the pH_i/temperature relationships of several tissue types in the aquatic anuran, Xenopus laevis, and in the semi-terrestrial toad, Bufo marinus.

Specimens of Bufo marinus (mass 210–423 g, N = 27) and Xenopus laevis (mass 47–65 g, N = 28) were purchased from a commercial supplier (Charles D. Sullivan Co. Inc., Nashville, TN) and transported to Germany by airfreight. Animals were held in captivity for at least 1 month prior to experimentation and fed on crickets and chopped liver with vitamin supplements. Ten days before implantation of femoral artery cannulae (see Boutilier, Randall, Shelton & Toews, 1979; Boutilier, 1984) animals were acclimated to either 10, 20 or 30°C in thermostatically controlled aquaria or terraria (±1°C). For cannulation, animals were anaesthetized by immersion in a 0·7 % solution of tricaine methane sulphonate (MS-222, Sigma) titrated to pH 7 with sodium bicarbonate. The entire operation including anaesthesia took approximately 1 h, whereupon the animals were returned to their respective acclimation temperature for a 48-h recovery period. The air and water phases of the experimental chambers were held to within ±0·5 °C of the acclimation temperatures.

Twelve hours before each experiment, [¹⁴C]DMO (5,5-dimethyl-2,4-oxazolidinedione) and [³H]inulin were injected into the femoral artery for subsequent estimation of intracellular pH by use of the DM0 distribution technique (Waddell & Butler, 1959). Blood samples (about 300 μl) were taken from the animals (four from Bufo and two from Xenopus) at their respective acclimation temperatures and analysed for plasma pH, total CO₂ concentration, and ⁠. In the case of Xenopus, the samples were taken shortly after a breathing period at the surface (see Boutilier, 1984). Immediately after the last blood sample (which included an additional 300 μl for plasma isotope analysis) had been taken, the animals were swiftly killed by anaesthetic overdose and tissue samples removed for subsequent analysis of tissue water compartments and radioactivity. The gastrocnemius and sartorius muscles from the non-cannulated limbs were dissected away, as were individual muscle bands associated with the pectoral area. Five samples (100–400 mg) from each skeletal muscle group of each animal were taken for analysis. Samples of heart tissue (50–150 mg) were taken from Bufo (five samples) and Xenopus (1–2 samples). Immediately after removal, tissues (and corresponding plasma samples) were weighed and then dried to a constant weight (at 110°C) before being prepared, by oxidation, for liquid scintillation counting.

Whole blood oxygen dissociation curves were constructed in vitro (at 10, 20 and 30 °C) on blood samples obtained from cannulated animals at each of the respective temperatures. Samples of blood were pooled to achieve the required volumes and distributed equally to tonometers supplied with either air/CO₂ or N2/CO₂ gas mixtures (the CO2 concentrations being adjusted at each temperature to approximate the corresponding levels of arterial seen in vivo). Blood samples of known O₂ saturation were prepared for and pH measurement as described by Hicks, Ishimatsu & Heisler (1987), using the mixing method (Haab, Piiper & Rahn, 1960) detailed by Scheid & Meyer (1978).

Statistical significance of differences between variables was determined by application of Student’s t-test at a level of 2P<0·05.

The blood pH/temperature relationships (ΔpH_e/Δt) in both Xenopus laevis and Bufo marinus were not significantly different from that of neutral water over the entire temperature range studied (10–30°C) (Fig. 1; Table 1). In both species the decreases in blood pH with increasing temperature were achieved primarily by increases in arterial ⁠; the changes in plasma [HCO₃⁻], although statistically significant, contributed less to the change in pH (Fig. 2).

The pH/temperature relationships of skeletal and cardiac muscle tissue compartments of Xenopus laevis and Bufo marinus are plotted in Fig. 3. Unlike arterial blood, ΔpH/Δt in the intracellular compartments varied considerably with temperature (Table 1) and was often markedly different from that of arterial blood. In Bufo marinus, for instance, ΔpH/Δt for skeletal muscle was lower in the temperature range 10–20°C than at higher temperatures. However, the reverse was true for Xenopus laevis (Table 1). At any given temperature there were no significant differences in pH among the three groups of skeletal muscle in Xenopus laevis,although pH in the pectoral muscle of Bufo marinus was significantly lower than in either the sartorius or gastrocnemius muscle at 10 and 20°C (Fig. 3; Tables 2, 3). In both species, the pH of cardiac muscle was higher than that of skeletal muscle at any given temperature (Fig. 3). There were marked differences in the pH/temperature response of heart tissue between the two species. Cardiac muscle in Bufo marinus (Fig. 3) showed the largest change in pH with temperature of all the compartments studied (Table 1). In Xenopus, however, heart muscle pH was constant in the 10–20°C temperature range, although in the 20–30°C range the ΔpH_i/ Δt of heart muscle tissue was not significantly different from that of arterial plasma (Table 1).

The values for the fractional water content of the tissues and extracellular space (Q_cm), obtained using the inulin distribution technique, are presented in Tables 2 and 3. There was no significant effect of temperature on either measurement in either species.

Arterial blood increased with temperature in both species. These values were obtained during periods when the animals were ventilating their lungs (Bufo marinus) or shortly after a breathing episode (Xenopus laevis). Whole blood oxygen equilibrium curves, determined in vitro for both species, are shown in Fig. 4 at each of the acclimation temperatures examined in the present study. levels of the equilibrating gas mixtures were set close to the values obtained in vivo at the respective temperatures (Fig. 2), resulting in a pH-induced shift to the right with increasing temperature. When the measured values of at each acclimation temperature in vivo are plotted on their corresponding O₂ equilibrium curve (Fig. 4) it is evident that arterial blood O₂ saturation remains relatively constant over the entire temperature range.

The changes in arterial blood with temperature in Xenopus and Bufo occur at a relatively constant arterial blood O₂ saturation (Fig. 4). This can be explained on the basis of existing models for O₂ transport in animals with cardiovascular shunts (Wood, 1982, 1984; Wood & Hicks, 1985). Arterial desaturation during air breathing results from systemic venous admixture with oxygenated pulmonary venous blood (right to left intraventricular shunt; Shelton, 1976; Shelton & Boutilier, 1982). Under these conditions, the arterial of a mixture of pulmonary and mixed venous blood becomes a function of the resulting O₂ saturation (Wood, 1982). The rightward shift of the oxygen dissociation curves with rising temperature (Fig. 4) results, therefore, in an increase in arterial with temperature. Within the temperature range examined in the present study, the arterial O₂ saturation of Xenopus (during or shortly after a breathing episode) was maintained virtually constant, whereas in Bufo a small but progressive desaturation occurred as temperature increased (Fig. 4). Such effects have been previously demonstrated in a number of reptiles (Wood, 1982; Glass, Boutilier & Heisler, 1983; Wood & Hicks, 1985) and in two amphibians (Wood, 1982; Kruhøffer, Glass, Abe & Johansen, 1987). Although the underlying mechanisms of such phenomena are yet to be determined, the collected data imply that the common denominator of oxygen homeostasis in ectotherms may be O₂ saturation and/or content, rather than partial pressure (Glass et al. 1983; Wood & Hicks, 1985). Intraventricular shunting and venous mixture has less of an effect on CO₂ partial pressures owing to different characteristics of the CO₂ dissociation curves.

The arterial blood pH values of Bufo marinus and Xenopus laevis vary inversely with temperature (ΔpH/ Δt = −0·015 and −0·017, respectively; Fig. 1; Table 1, range 10 –30°C). These values are similar to those reported for the arterial blood of other amphibian species (Howell et al. 1970; Reeves, 1972; Burggren & Wood, 1981; Moalli, Meyers, Ultsch & Jackson, 1981; Hicks & Stiffler, 1984; Kruhøffer et al. 1987). The changes in arterial pH with temperature in the present study are brought about primarily through adjustments in ⁠, though small but significant changes in plasma bicarbonate concentrations with temperature were also observed (Fig. 2). In all instances, it appears that the arterial blood pH of amphibians varies inversely with temperature by a specific constant from the pH of neutral water, or that the animals maintain a constant relative alkalinity (Rahn, 1967). The compiled data on arterial blood of amphibians also fit the alphastat hypothesis of Reeves (1972). The biological importance of this hypothesis rests, however, on the assumption that the pH values of intracellular compartments also change with temperature, more or less in parallel with that of arterial blood (constant relative acidity).

The pH_i/temperature relationships for skeletal and cardiac muscle in the present study do not follow the patterns predicted by the alphastat model (Fig. 3; Table 1). Previous measurements of intracellular pH in amphibians are restricted to one study on Rana catesbeiana (Malan et al. 1976) in which the pH, of striated muscle (apparently combinations of gracilis, sartorius and gastrocnemius) was determined as a function of temperature by the DM0 method. Malan et al. (1976) found ΔpH,/ Δt in striated muscle (−0 ·0152 units °C⁻¹) to be not significantly different from that of arterial blood (−0·0206 units °C⁻¹). In our study, average ΔpH_i/ Δt values for the range 10 –30°C varied in different groups of striated muscle from −0 −014 to −0 −023 ior Bufo marinus and from −0 ·014 to −0 ·017 ior Xenopus laevis (Table 1), the differences being even larger for smaller temperature intervals. ΔpH/ Δt tended to be lower at lower temperatures in Bufo, whereas the reverse was true for Xenopus laevis. These data suggest that intracellular pH may vary with temperature in a non-linear fashion, a conclusion also reached by Walsh & Moon (1982) on the basis of data obtained by acid-base/temperature studies in the eel. Similar non-linearities can also be found for the extracellular fluid of reptiles (for references see Heisler, 1986b). They are, however, not universal. For example, the ΔpH/ Δt of arterial plasma of both species is more or less constant, as is ΔpH,/ Δt of Bufo heart muscle and Xenopus sartorius muscle. In all of the other intracellular body compartments, however, there is a remarkable difference in ΔpH/ Δt between the lower and higher temperatures (Fig. 3; Table 1). This is particularly pronounced in heart muscle of Xenopus, where pH_i is held constant over the 10–20°C range, but falls by 0 ·14 pH units between 20 and 30°C (Fig. 3; Tables 1, 2). Similar observations were made in eel heart muscle (Walsh & Moon, 1982). These data indicate that mechanisms other than simple physicochemical dissociation processes are responsible for the fine adjustment of pH with changes of temperature.

Accordingly, there is no good a priori reason to assume that the pH/temperature relationship of any body compartment should follow a linear pattern.

Over the past few years, discussions about pH/temperature relationships in ectotherms and the validity of the alphastat hypothesis have become reduced to arguments about the relative magnitude of linearly interpreted ApH/At slopes. The present data on intracellular pH changes with temperature support data from the literature on non-amphibian ectotherms, showing a wide range of physiological ΔpH,/ Δt values in various body compartments (see Heisler, 1986b). These data also show that temperature-related intracellular acid-base regulation is relatively independent of extracellular acid-base regulation in amphibians, and clearly indicate that individual compartments may maintain unique pH levels and unique ΔpH/Δt relationships (see Fig. 3). The alphastat model, with adjustment of only organismic ⁠, cannot account for regulatory patterns such as these. Although less restricted than in fishes (see Heisler, 1984b), adjustment of in intermittently breathing animals is limited by the discontinuity of gas exchange (Boutilier & Shelton, 1986; Toews & Boutilier, 1986), and the resulting isothermal variability in arterial blood is likely also to be transmitted to intracellular body compartments. Intracellular pH regulation is, accordingly, passively affected by changes in and the temperature-dependent buffering characteristics, which are quite variable among species and among different tissues in individual species (see Heisler, 1986a,b). The active regulatory process, however, has to be sought in transmembrane and transepithelial ion transfer mechanisms. It should be emphasized that even concordance of ΔpH/ Δt with ΔpK/ Δt values of imidazole moieties (a relatively rare coincidence) (for references see Heisler, 1986b) provides little support for the alphastat model, if production of acid-base-relevant ions (by metabolic and buffering processes) and transfer of such ions across the borders of the respective compartments are neglected. Future studies must not, therefore, be limited to descriptions of pH/temperature coefficients, but should focus on detailed evaluation of more indicative features of acid-base regulation.

The authors gratefully acknowledge the skilful technical assistance of Mrs S. Glage and Mr G. Forcht. Supported by Alexander von Humboldt Stiftung and Deutsche Forschungsgemeinschaft.