ABSTRACT
Freshwater turtles (Chrysemys picta bellii Gray) were acclimated to temperatures of 5, 10, 20 and 30°C for at least 12 days, and pulmonary ventilation, oxygen uptake and arterial pH, and were determined in completely unrestrained specimens. Oxygen uptake increased overproportionately (6·7-fold) as compared to pulmonary ventilation (, 4·4-fold) when the temperature increased from 10 to 30°C. The observed rise in arterial from 13 (5 °C) to 32 mmHg (30°C) was the result of a decrease in , whereas an increase of arterial from 12Torr at 5°C to about 60Torr at 20 and 30°C mainly resulted from the effects of intracardiac blood shunting combined with temperature-dependent shifts of the oxygen dissociation curve. Arterial pH fell with rising temperature significantly less (ΔpH/Δt = −0·010U/°C) than required for constant relative alkalinity and for constant dissociation of imidazole. The changes of cerebrospinal fluid pH with temperature, calculated from the mean arterial values, were even smaller [ΔpH/Δtcsf = −0·008). It is concluded that the observed temperature dependence of the acid-base status is not in agreement with the alphastat hypothesis.
INTRODUCTION
Turtles encounter a wide range of environmental temperatures throughout the year and have accordingly often been used as model animals for studies concerning the effects of changes in body temperature on the acid-base regulation of ectothermic vertebrates (cf. Jackson, 1978). It was generally found that arterial plasma pH decreased with increasing temperature, the extent of decrease (ΔpH/Δt), however, was rather inconsistent, varying by a factor of two even when reported for the same species (cf. Jackson, Palmer & Meadow, 1974; Malan, Wilson & Reeves, 1976). Temperature-dependent regulation of pulmonary ventilation, and thus adjustment of arterial , has been suggested as the mechanism responsible for the regulation of pH with changes of temperature. Pulmonary ventilation as a function of temperature has been measured in several studies, but the results vary largely between the two extremes of no changes of pulmonary ventilation, and increases in ventilation by a factor of more than two per 10 °C temperature rise (Q10>2) (cf. Glass, Hicks & Riedesel, 1979; Jackson, 1971).
The variability of the data obtained to date may partially result from various methodological problems and the non-resting conditions of the experimental animals. Therefore, the aim of the present study was to determine simultaneously the effect of various temperatures on arterial pH, and , on pulmonary ventilation, and on oxygen uptake in the freshwater turtle Chrysemys picta bellii, paying particular attention to the constancy of the experimental conditions and to the achievement of a resting state.
MATERIALS AND METHODS
Specimens of Chrysemys picta bellii Gray were kept in captivity in large aquaria (20–30°C) equipped with heat lamps and dry basking areas for several months prior to experimentation. The animals were fed on chopped beef liver, chicken meat and commercially available turtle food pellets. At least seven days before implantation of an arterial catheter into the brachial artery under halothane anaesthesia (for details of the procedure see Glass, Boutilier & Heisler, 1983) the animals were acclimated to temperatures of 5°C (number of animalsN= 8, average weight = 572g), 10°C (N=6, = 600g), 20°C (N=8, = 708g) and 30°C (N = 7, = 579g).
After recovery from surgery, the turtle was introduced into the experimental apparatus, which consisted of an aquarium filled with water thermostatted to the respective temperature and shielded against visual disturbances. The animals had no access to the air except in a breathing funnel (for details see Glass et al. 1983). This cone-shaped chamber above the water surface was flushed with a constant flow of air (50–250 ml min−1, depending on temperature and size of the turtle). Deviations from the basal flow rate through the breathing funnel represented pulmonary ventilation and were monitored at the outflow by pneumotachography (Model 17212, Godart Statham, Bilthoven, Netherlands). Oxygen consumption was determined from the integrated flow rate through the breathing funnel and the difference in fractional oxygen concentration between inflowing and outflowing gas measured by means of a differential oxygen analyser (Model 5-3A, Applied Electrochemistry Inc., Sunnyvale, CA, U.S.A.).
Pulmonary ventilation and oxygen uptake were monitored for at least 5 days, and blood sampling was not started before the normal periodic breathing pattern and resting O2 uptake rate were re-established. Then pulmonary ventilation, length of ventilatory and non-ventilatory periods, frequency of ventilation and oxygen consumption were determined (at 10, 20 and 30°C). Arterial pH, and were measured (at 5, 10, 20 and 30°C) several times (up to 10 times) in each animal during ventilatory and non-ventilatory periods during a time period of 72 h. Care was taken to withdraw the arterial blood samples without disturbance of the animals, which would have elicited changes in the ventilatory pattern.
Plasma bicarbonate concentration was calculated/On the basis of the Henderson-Hasselbalch equation using values for pKí″ determined on plasma of Chrysemys picta bellii (Nicol, Glass & Heisler, 1983) and for CO2 solubility, adapted from Reeves (1976) for the respective temperature.
Calculation of ventilation was based on inspired volumes.
RESULTS
Pulmonary ventilation increased by a factor of 4·4 with a rise in body temperature from 10 to 30°C, but less than the oxygen uptake which increased 6·7-fold (Fig. 1).
The rises in pulmonary ventilation, , from 10 to 20°C, and from 20 to 30°C were the result of different mechanisms: from 10 to 20°C, tidal volume remained essentially constant (10°C: 12·7 ± 1·7; 20°C: 10·7 ± l·2ml BTPS kg−1), and the average breathing frequency increased 4-fold (10°C: 0·47 ±0·05; 20°C: 1·90 ± 0·27min−1), whereas from 20 to 30°C tidal volume increased almost 2-fold (30°C: 18·3 ± 1·9) and the average breathing frequency fell at 30°C to about 80% of the value at 20°C (30°C: 1·47 ± 0·11). The observed large changes in average breathing frequency were mainly the result of changes in the ratio of the ventilatory period to the length of the total breathing cycle ( 10°C: 0·031 ± 0·005; 20°C: 0·084 ± 0·011; 30°C: 0·084 ± 0·013); in contrast, the breathing frequency during the ventilatory periods increased with temperature only to a minor extent (10°C: 161 ± 2·7; 20°C: 19·6 ±2·2; 30°C: 21·3 ±2·4).
Arterial pH and changed little during breathing or diving. The slope of the linear regression obtained from, all pH measurements during breathing periods at 10, 20 and 30°C [pH = −0·011tB+7·958; tB, body temperature (°C), N = 41, r = 0·82] was identical with that of the regression for diving periods (pH = −0·01 ltB+7·939, N = 43, r=0·82). Based on all measurements during breathing as well as diving periods, the change in pH was less between 5 and 10 °C (ΔpH/Δt =−0·008 U/°C) than between 10 and 20°C (ΔpH/Δt =−0·011) and between 20 and 30°C (ΔpH/Δt = −0·010).
The change in pH with temperature (Fig. 2, upper panel) resulted from changes of both and plasma bicarbonate concentration. Arterial increased over the whole temperature range from about 13 mmHg at 5 °C to about 32 mmHg at 30°C (Fig. 2, middle panel), whereas plasma bicarbonate increased with rising temperature from 5 to 10°C, but then decreased again at 20 and 30°C. Arterial increased linearly from 5 to 20°C (12 to 64mmHg), but remained essentially constant between 20 and 30°C (Fig. 3).
DISCUSSION
Regulation of ventilation with changes in temperature in turtles is still a matter of controversy. Jackson (1971) and Hitzig (1982) have reported that pulmonary ventilation is maintained virtually unchanged in the freshwater turtle Pseudemys scripta over a wide range of temperatures (Fig. 4). In contrast, studies on the closely related species Pseudemys floridana (Kinney, Matsuura & White, 1977) and Chrysemyspicta bellii (present study), as well as on Terrapene omata (Glass et al. 1979) and Chelonia mydas (Kraus & Jackson, 1980), closely agree in reporting considerable increases of pulmonary ventilation with rising temperature (Fig. 4). However, even when ventilation increases with temperature, this rise is not large enough to maintain a constant ratio of pulmonary ventilation to pulmonary oxygen uptake, (often called ‘air convection requirement’). This is illustrated for Chrysemys picta bellii inFig. 5, based on data obtained in the course of the present study. The ratio of has been extensively studied in the turtle Pseudemys scripta and the effects of changes in body temperature on arterial and pH have been related to the temperature-dependent changes of (Jackson, 1971, 1978).
Alveolar and are inversely related (cf. Jackson, 1978) and temperature-dependent changes in account for the rise in arterial with increasing temperature in Chrysemys.
The drastic reduction of arterial at low temperature (Fig. 3) can only be understood on the basis of intracardiac right-left shunts with the result of large proportions of cardiac blood flow by-passing the pulmonary circulation. A partial arterial desaturation always occurs in turtles, since the intermediate of the three interconnected chambers in the hearts of turtles is common to both pulmonary and systemic circulation, allowing considerable venous admixture to the arterialized blood, similar to the conditions in Varanus (Heisler, Neumann & Maloiy, 1983). Arterial then becomes a function of the mixed systemic venous and the arterialized pulmonary venous blood oxygen content, the shunted blood fraction, and the position of the O2 dissociation curve of the blood. The right shift of the O2 dissociation curve with increased temperature (Glass et al. 1983) then causes an increase of as has been recently pointed out by Wood (1983). Above 20°C, the oxygen loading in the lungs is reduced and the oxygen, content of the systemic return falls, resulting in the observed plateau of . AS a result of these mechanisms, the changes of arterial with temperature in turtles have to be considered to be primarily a function of central vascular shunting events rather than being determined by the ratio of , whereas arterial is little affected by these factors because of the different characteristics of the CO2 dissociation curve.
Values for relative alkalinity of plasma in Chrysemys presented in Fig. 6A clearly show that a constant relative alkalinity is not maintained in this species. The deviation is highly significant, and the data compiled in Fig. 6B from the other turtle species studied to date indicate that constant relative alkalinity is rarely achieved, except for some species in the range of higher temperatures. This may be partially due to the above mentioned fact that the neutral pH of water (pN) changes less with temperature at higher than at lower temperatures.
An alternative hypothesis for temperature-dependent acid-base regulation has been provided by Reeves (1972). His imidazole alphastat hypothesis claims that ventilation, and thus , is regulated in such a manner that the fractional dissociation of peptide-linked histidine imidazole is kept constant. Alphastat regulation thus implies that ΔpH/Δt in the arterial blood is regulated to equal the value for ΔpK/Δt of the imidazole buffer system. It also implies that, since the histidine imidazole of haemoglobin and plasma proteins represents by far the most important non-bicarbonate buffer system of the extracellular space, the bicarbonate concentration of the compartment is kept constant, especially when transmembrane and transepithelial acid-base relevant ion transfer is excluded (Reeves & Malan, 1976).
The pK value of imidazole changes with temperature (ΔpK/Δt) by −0·018 to −0·024U/°C (25°C), depending on ligands and steric arrangement (Edsall & Wyman, 1958). The value for ΔpH/Δt determined for Chrysemys in the course of this study in the temperature range of 10–30°C (−0·011) is only about 60% of the lower limit of the above range and significantly different at a high level (P < 0·001). This difference is even more pronounced between 5 and 10°C (ΔpH/ΔT = −0·008).
Comparison of the literature data available on ΔpH/Δt in turtle blood reveals that nine out of ten studies report values lower than required for constant imidazole dissociation (Table 1). The average of all studies yields a ΔpH/Δt value of −0·012, a value very close to the ΔpH/Δt typically found in water-breathing fish (see Heisler, 1980, 1984a,b). These values do not support the concept of imidazole alphastat regulation. Is this disagreement caused by directional errors introduced by the experimental approach? It turns out that most conceivable sources of error tend to increase ΔpH/Δt rather than reduce it.
(1) If body temperature is measured with a thermometer in the body cavity after killing the animals and taking blood and tissue samples (e.g. Malan et al. 1976), then the temperature can only have changed towards room temperature, i.e. animals acclimated to lower temperatures are warmed and vice versa. This would result in an underestimate of the real temperature difference between experimental groups and accordingly to an overestimate in ΔpH/Δt.
(2) Underestimates of the real temperature range could also result from the temperature difference of 1–3 °C between body cavity and environment, which is found in many reptiles (N. Heisler, P. Neumann, H. Weitz & A. Weitz, unpublished data on Tupinambis nigropunctatus, Varanus exanthematicus, Testudo horsfieldi). This difference is expected to increase at higher temperatures due to higher metabolism. The resulting underestimate of the temperature difference consequently also increases theΔpH/At value.
(3) Deviations from the normal ventilatory pattern as a result of disturbances prior to blood sampling or due to non-resting conditions may also interfere. From our experience, disturbances of turtles result in a considerable hyperventilation at low temperatures (5–15°C). This would cause a rise in pH and thus also an increase of the ΔpH/Δt value.
(4) The blood sampling procedure may considerably influence the pH pattern. If lactic acid formation occurs as a result of struggling before sampling, then H+ ions would be extruded from the intracellular space at an extremely high rate (Benadé & Heisler, 1978; Holeton & Heisler, 1983; Holeton, Neumann & Heisler, 1983), lowering blood pH. This effect is least important at lower temperatures because of metabolic pathway limitations and slow H+ extrusion kinetics, but may result in considerable acidification at higher temperatures, increasing the ΔpH/Δt value. This effect on arterial plasma pH values in fish is well documented (cf. Ali et al. 1980; Holeton et al. 1983; Holeton & Heisler, 1983). It is probably not fortuitous that the highest ΔpH/Δt values in turtles (−0·021 and −0·016: Table 1) have been obtained from measurements in blood sampled by heart puncture.
The evaluation of directional error sources suggests that the averageΔpH/ Δ t of about −0·012 represents an upper rather than a lower limit for this parameter. Evidently alphastat regulation is not typical of the extracellular compartment of turtles. This does not exclude the possibility that alphastat regulation may be achieved in certain intracellular compartments. In fact, the only study available on the relationships of intracellular pH and temperature in turtles (Malan et al. 1976) reportsΔpH/ Δt values in the intracellular space of white muscle (−0·0186 U/°C) and liver (−0·0233 U/°C) in the range of the ΔpK/ Δt values of imidazole compounds. These data, however, have been obtained in the study with the highest extracellularΔpH/At value ((−0·021 U/°C), which is about 170% of the average value determined in turtles. If the extracellularΔpH/Δt value was in fact affected by the sources of error discussed above, then the intracellularΔpH/At values would have been overestimated by an even larger factor because of the properties of the DMO method for measurement of intracellular pH (see Appendix).
Regardless of these considerations, the change in observed in the present study for Chrysemys is far too small to cause a change of plasma pH with temperature in parallel with the changes in pKim (for model calculations see Heisler, 1978, 1984a,b; Heisler & Neumann, 1980).
In intracellular compartments, theΔpH/Δt value is less influenced by the change in than by the ratio of imidazole-like (ΔpKim/Δt∼−0·020 U/°C) to phosphate-like (ΔpKph/Δt∼−0·002 U/°C) non-bicarbonate buffer values (βim/βph) (Heisler & Neumann, 1980; Heisler, 1984a,b). In the extracellular space this ratio is in the range of 10–30, whereas the intracellular ratio has been estimated (Reeves & Malan, 1976) and determined (Heisler & Neumann, 1980; Heisler, 1984a,b) to be in the range of 1 to 4 in various tissues. Intracellular bicarbonate concentrations are lower by factors of 2 to 5, according to the lower pH values than those in extracellular compartments. In addition, intracellular non-bicarbonate buffer values are expected to be much higher on the basis of data from other vertebrate species (Heisler & Piiper, 1971, 1972; Heisler & Neumann, 1980; Heisler, 1984a). As a consequence of these conditions, intracellular ΔpH/Δt values are rather independent of (see Heisler, 1984a,b). IfΔpH/Δt in any intracellular compartment is regulated to a value different from that predetermined by the buffer values ratio (βim/βph),this has to be performed mainly by transmembrane transfer of acid-base relevant ions (i.e. HCO3−, OH−, or H+ in the opposite direction). Consequently, an overall imidazole alphastat acid-base regulation in all body compartments by adjustment of only pulmonary ventilation, as propounded by Reeves (1972) and Reeves & Malan (1976), cannot be expected.
Histidine-imidazole protein residues as sensors for the regulation of ventilation per se would not be positioned optimally in an intracellular compartment. Minor non-respiratory disturbances of pH would result in extreme changes in ventilation, which would nevertheless not be sufficient to maintain the imidazole dissociation due to the steep slope of the intracellular CO2-buffer curve. In contrast, the cerebrospinal fluid (CSF), which is claimed to be the environment for respiratory receptors in higher vertebrates, contains virtually no non-bicarbonate buffers, except for a small concentration of phosphates, and is known to have extremely close diffusive contact with the arterial blood. Accordingly, it would be an optimal site for an alphastat sensor and in fact Hitzig (1982) has reported thatΔpH/Δt in the CSF of Pseudemys was −0·015 U/°C, not significantly different from the lower limit of -0-018 U/°C for ΔpK/Δt of physiological imidazole compounds.
In addition, the model condition of excluded ionic transfer between CSF and other compartments is a most unlikely condition in vivo. The ion transfer mechanisms which have been shown to exist for the structures surrounding the CSF in higher vertebrates (e.g. Ahmad & Loeschke, 1982a,b) probably also exist in lower vertebrates and interfere with the changes in pH induced by the temperaturedependent changes in .If constant imidazole dissociation is achieved by means of ionic transfer, this must clearly diminish the role of this parameter for the regulation of pulmonary ventilation. Accordingly, the adjustment of ventilation very probably follows other criteria.
It can be concluded that the rise in arterial with increasing temperature in Chrysemys is a result of readjustment of the ratio , whereas the temperaturedependent alterations of arterial have to be mainly attributed to the interrelationship between the right shift of the oxygen dissociation curve with increased temperature and the partial desaturation of arterial blood as a result of central vascular R-L shunting. The observed temperature-dependent fall in plasma pH, which is considerably and significantly smaller than required for constant pKim, is, however, only partially due to the rise in with increasing temperature, but is also effected by changes in plasma bicarbonate. Analysis of the changes in with variation of temperature in various species of turtles suggests that, if constant dissociation of histidine-imidazole is ever achieved in any body compartment, closed-system buffering is less important for this type of acid-base regulation than transmembrane and transepithelial transfer of acid-base relevant ions.
APPENDIX
ACKNOWLEDGEMENTS
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 and the Danish National Science Research Council.