Body temperature increases in ectothermic vertebrates characteristically lead to both increases in arterial PCO2 (PaCO2) and declines in resting arterial pH (pHa) of about 0.017 pH units per 1°C increase in temperature. This ‘alphastat’ pH pattern has previously been interpreted as being evolutionarily driven by the maintenance of a constant protonation state on the imidazole moiety of histidine protein residues, hence stabilizing protein structure–function. Analysis of the existing data for interclass responses of ectothermic vertebrates shows different degrees of PaCO2 increases and pH declines with temperature between the classes, with reptiles>amphibians>fish. The PaCO2 at the temperature where maximal aerobic metabolism (V̇O2,max) is achieved is significantly and positively correlated with temperature for all vertebrate classes. For ectotherms, the PaCO2 where V̇O2,max is greatest is also correlated with V̇O2,max, indicating there is an increased driving force for CO2 efflux that is lowest in fish, intermediate in amphibians and highest in reptiles. The pattern of increased PaCO2 and the resultant reduction of pHa in response to increased body temperature would serve to increase CO2 efflux, O2 delivery and blood buffering capacity and maintain ventilatory scope. This represents a new hypothesis for the selective advantage of arterial pH regulation from a systems physiology perspective in addition to the advantages of maintenance of protein structure–function.
Body temperature influences blood acid–base balance in a very predictable pattern in ectothermic vertebrates, with a decrease of about 0.017 pH units per 1°C increase in temperature (Howell et al., 1970; Reeves, 1972). The regulation of ventilation with temperature has been proposed as a mechanism to regulate arterial PCO2 (PaCO2) and thus arterial pH (pHa) with temperature changes in ectotherms (Glass et al., 1985). In most ectotherms studied, increased ventilation does not match the temperature-induced increase in metabolism, and this relative hypoventilation leads to an increase in PaCO2 that decreases pHa from the generation of carbonic acid. Interestingly, this pattern parallels the effect of temperature variation on the pH of water. The rate of change in pH for both water and pHa is about −0.017 pH units °C−1, and because the arterial blood of ectotherms is about 0.6 pH units greater than that of water at any temperature, the phenomenon was frequently referred to as maintaining ‘relative alkalinity’ (Rahn, 1967), and the regulatory process to achieve this as ‘alphastat pH regulation’ (Reeves, 1972). The prevailing hypothesis for the advantage of alphastat pH regulation is maintenance of a constant ratio of OH− to H+ despite variation in pH. This alphastat pH pattern maintains a constant fractional protonation state on the imidazole moieties of histidines in proteins (Reeves, 1972, 1977). This has been argued to better maintain protein structure and function and preserve cellular function with varying body temperatures.
Reeves’ (1972, 1977) hypothesis for alphastat regulation of blood pH suggests that ventilation, and thus PaCO2, is regulated to maintain a constant fractional dissociation of histidine imidazole residues on proteins. This hypothesis implies that the change in pH with temperature is regulated to equal the change in the pK with temperature of the imidazole buffer system, which is about −0.018 to −0.024 U °C−1 (Edsall and Wyman, 1958). Although there is some support for the alphastat hypothesis for regulation of blood pH in ectotherms, there are several studies showing that the change in blood pH with temperature is significantly lower than the change in pK with temperature required for alphastat pH regulation (see Glass et al., 1985). Thus, although alphastat regulation is an attractive hypothesis for explaining the pattern of blood pH regulation in ectotherms, Cameron (1989) pointed out that as a realistic predictor of protein behavior, alphastat needs to be revised to accommodate both advances in protein chemistry and the evident heterogeneity of physiological findings. The pattern of increased PaCO2 and decreased pHa with increasing temperature has also been interpreted as a means of depressing metabolism via ventilation during bouts of torpor or hibernation in both endotherms and ectotherms (Malan, 2014).
Given the heterogeneity of the physiological data and in an attempt to provide an integrative metric of organismal function, we present an argument for the consideration of an organ system-level advantage related to O2 and CO2 fluxes during periods of increased aerobic demands associated with both increased temperature and increased activity for an increase in the regulated PaCO2 and consequential decrease in pHa with increases in temperature. Standard and maximal rates of aerobic metabolism of all ectotherms are temperature sensitive, with a range of Q10 values of about 1.5–3 (Hedrick et al., 2015). Maximal rates of aerobic metabolism during activity at an organ-system level reflect the maximal rates of oxygen delivery to working muscle and the maximal rates of CO2 removal from working muscle to the environment. The cardiovascular system is the principal limitation to maximal oxygen delivery in vertebrates (Hillman et al., 2013), whereas the respiratory system appears to be the principal limitation to CO2 efflux in ectotherms (Hillman et al., 2013; Hedrick et al., 2015). Consequently, co-adaptations that enhance the capacity for both O2 delivery and CO2 efflux will enhance aerobic metabolic capacity.
There is a shift in PaCO2 and pH regulation in the evolutionary transition from fish to amphibians and reptiles associated with the differences in O2 and CO2 capacitances of water and air (Dejours, 1975). Fish primarily regulate pH across their gills via ion exchangers (Na+/H+, Cl−/HCO3−) but CO2 is exchanged by diffusion (Heisler, 1986). Amphibians and reptiles primarily achieve pH regulation via ventilatory regulation of PaCO2. From an organismal metabolic perspective, how might an alphastat pH pattern of reduced pH and increased PaCO2 increase O2 delivery and CO2 removal with increases in temperature? We suggest that: (1) the regulated hypoventilation-associated increase in temperature would preserve ventilatory capacity; (2) the resultant increase in PaCO2 would increase the driving force for CO2 efflux; (3) the increase in PaCO2 would increase HCO3− and buffering capacity of the blood; and (4) the decrease in pHa would increase the delivery of O2 (Bohr effect) and the efflux of CO2 (Haldane effect) both at rest and during activity with increased body temperature.
If increasing the regulated PaCO2 with increased body temperature is selectively advantageous for enhancing organ system gas exchange, there are a variety of predictions that might follow: (1) increased temperature should increase PaCO2 and decrease pHa within the different classes of ectothermic vertebrates, and (2) interclass variation of the PaCO2 responses to temperature should correlate with interclass variation of the aerobic metabolic capacity. If these predictions hold, it suggests that there may be an alternative or additional evolutionary explanation to protein structure–function driving the evolution of this alphastat pH pattern of changes in PaCO2 and pH with temperature.
MATERIALS AND METHODS
Venous PCO2 (PvCO2) directly reflects the actual driving force for CO2 diffusional efflux across the respiratory surface, assuming that alveolar PCO2 remains the same. The difference between PvCO2 and PaCO2 is small at rest and in many cases they are almost indistinguishable, but resting PaCO2 represents a minimal estimate of the potential driving force across the respiratory surface. There are more data available for resting PaCO2 than for PvCO2; thus, we have used resting PaCO2 values throughout in our analysis. Although using resting PaCO2 may underestimate the actual driving force for CO2 efflux, especially during activity, increases in PaCO2 clearly reflect physiologically regulated increases in the net driving force for PCO2 efflux.
To evaluate the consistency of both blood pH (pHa) and PaCO2 with temperature for each group of ectotherms, we have used the summary data of Ultsch and Jackson (1996), which primarily selected data based on cannulated sampling rather than heart punctures for resting animals. Data for PaCO2 of resting mammals and birds were taken from Lahiri (1975), Tenney and Boggs (1986), Gleeson and Brackenbury (1984), Cushing and McClean (2010), Murrish (1983), Ponganis et al. (2007), Peters et al. (2005) and Scott and Milsom (2007).
Metabolic data (resting and maximal) for each class were taken from the summaries within Hedrick et al. (2015). Aerobic generation of CO2 is the result of aerobic metabolism and its efflux can be quantified as the product of conductance and the driving force for CO2 (i.e. GCO2×ΔPCO2) Our hypothesis is that the increase in PaCO2 with temperature reflects an increase in the physiologically regulated driving force for CO2 efflux. Consequently, to test whether the Q10 for the rate of resting CO2 efflux parallels the Q10 for the PCO2 driving force, we calculated the ratio of resting PaCO2 at different temperatures. We used the resting PaCO2 regressions, summarized in Figs 1 and 2, to determine the ratio of PaCO2 differences between two temperatures, analogous to the calculation of Q10 for reaction rates [i.e. (Rate 2/Rate 1)10/(T2−T1)] or (PaCO2 at T2/PaCO2 at T1)10/(T2−T1) (see Jackson, 1978).
Least squares regression was used to determine slopes and significance using Prism v.5 (Graphpad software, Inc. La Jolla, CA, USA).
RESULTS AND DISCUSSION
There were significant increases in resting PaCO2 with increased temperature for fish [F1,58=7.1; P=0.0098; PaCO2 (kPa)=0.0067±0.002°C+0.246, r2=0.11], amphibians [F1,38=51.1; P<0.0001; PaCO2 (kPa)=0.0538±0.008°C+0.0305, r2=0.57], and reptiles [F1,68=29.8; P<0.0001; PaCO2 (kPa)=0.0691±0.013°C+1.18, r2=0.30] (Fig. 1A). The slope of this relationship for fish, although significant, was about 10-fold lower than the slope for amphibians or reptiles. This would be expected given the low PaCO2 in fish due to the high CO2 capacitance in water.
There was a significant effect of temperature (P<0.0001) on resting blood pH for fish (F1,90=39.6; pHa=8.04−0.010±0.002°C, r2=0.31), amphibians (F1,44=70.0; pHa=8.09−0.013±0.002°C, r2=0.61), and reptiles (F1,78=258; pHa=7.96−0.014±0.001°C, r2=0.77) (Fig. 1B). Taken together, these results are consistent with a resting CO2-mediated decrease in blood pH with increasing body temperature.
At any particular temperature, PaCO2 for reptiles was approximately double that of amphibians, and PaCO2 for amphibians was 3–4 times that of fish (Fig. 1A). The elevated PaCO2 of reptiles would therefore account for the lower pH for this group at any body temperature (Fig. 1B).
The temperature at which V̇O2,max occurs is lowest in fish (20°C), intermediate in amphibians (25°C) and reptiles (35°C), and highest in mammals and birds (see fig. 4 of Hedrick et al., 2015). The temperature at which resting PaCO2 corresponds with V̇O2,max for five vertebrate classes (Hedrick et al., 2015) is presented in Fig. 2A. There was a significant, linear relationship (P<0.0077; r2=0.93) between PaCO2 and the temperature at which V̇O2,max occurs, indicative of an increased driving force for CO2 efflux with increased temperature at V̇O2,max for these vertebrate groups.
The relationship between V̇O2,max and PaCO2 where V̇O2,max occurs for all vertebrate groups is presented in Fig. 2B. Resting PaCO2 increased with the greatest V̇O2,max for the ectothermic classes, but was independent of V̇O2,max in the endothermic classes. A plateau of approximately 5 kPa PaCO2 seems to occur for vertebrates in general; reptiles at 35°C are near this apparent plateau.
Enhancing ventilatory scope
Our analysis of the resting PaCO2 patterns with temperature in fish, amphibians and reptiles (Fig. 1) revealed that, at a given temperature, resting PaCO2 is greatest in reptiles, intermediate in amphibians and lowest in fish. There were significant increases in resting PaCO2 with temperature in all three groups. The alveolar ventilation (V̇A) equation predicts alveolar PCO2, and thus PaCO2, to be inversely related to the air convection requirement (ACR) ratio in air-breathing ectotherms (i.e. V̇I/V̇O2 or V̇E/V̇O2, where V̇I and V̇E are inspiratory and expiratory ventilation, respectively) and the increased PaCO2 (and decreased pH) with temperature can be explained by an unequal response of minute ventilation (V̇I or V̇E) relative to metabolism. This approach would also apply to fish, substituting water for air. The hypoventilation (decreasing V̇A) will increase the ventilatory scope available during activity. Assuming consistent interclass Q10 effects on metabolism, the magnitude of the hypoventilation can be estimated as V̇A=1/PaCO2. The mean decrease in V̇A for the temperature intervals from 10 to 20°C and 20 to 30°C for fish is 16%, for amphibians 34%, and for reptiles 24%. This estimate reflects the potential increase in ventilatory scope available to enhance gas exchange with activity compared with that obtained if these groups maintained a constant ACR and pHa. Although the alphastat hypothesis implies that the reduced ACR with increased temperature is necessary to maintain a constant fractional dissociation of imidazole residues, we suggest that the reduced ACR with temperature may also be important for preserving ventilatory capacity with increased metabolism associated with both temperature and activity.
There are additional arguments that support this hypothesis. First, the pattern of pH regulation we observed for fish, amphibians and reptiles in this study does not fit the traditional alphastat hypothesis proposed by Reeves (1972). The slopes for the change in pH with temperature for the air-breathing ectotherms, amphibians (−0.013 U °C−1) and reptiles (−0.014 U °C−1) were about 25–30% lower than the approximate −0.017 U °C−1 required for alphastat regulation, and similar to the values found previously for a number of reptile species (Glass et al., 1985). Second, previous work in reptiles has shown that V̇I or V̇E increases about 3- to 4-fold with a temperature increase from 10 to 30°C whereas V̇O2 increases 6- to 7-fold over the same temperature range (Funk and Milsom, 1987; Glass et al., 1985). This is the basis for the reduced ACR, but if minute ventilation were matched to metabolism, thus maintaining a constant PaCO2 and pH (i.e. pH stat regulation), the resulting increase in minute ventilation would leave less scope for further increases with increased temperature or during bouts of activity as described above. We showed previously (Hillman et al., 2013) that at maximal exercise, CO2 extraction at the respiratory surface increases significantly in all vertebrates, and the ratio of V̇I to blood flow at the respiratory surface increases about 3-fold to support the increase of CO2 extraction at V̇O2,max. This requires a ventilatory capacity from rest to activity to support the increased CO2 extraction to maintain maximal CO2 efflux. Even with this level of ventilatory increase, PaCO2 increases at V̇O2,max in fish and amphibians, indicating that ventilation does not keep pace with the needs for CO2 efflux (Hillman et al., 2013).
Enhancing Bohr and Haldane effects
The relative hypoventilation with increased PaCO2 and reduced pHa pattern also takes advantage of Haldane and Bohr effects for increasing CO2 and O2 transport, respectively, with increased temperature. The delivery of O2 from hemoglobin (Hb) is influenced by the decline in arterial pH as O2+Hb↔HbO2+H+; hence, by mass action, an increase in [H+] at the tissue level (from elevated PCO2 and lactic acid) favors unloading of Hb (Bohr effect) and enhanced O2 delivery at the muscle. The increase in [H+] also enhances the uptake of CO2 at the tissue as a consequence of the formation of carbamino CO2 on the Hb molecule (Haldane effect). The increase in [H+] also favors the release of CO2 at the respiratory surface by mass action from the following reaction: H++HCO3−↔H2O+CO2. The advantages of the Haldane and Bohr effects for gas transport would not be fully realized without the regulated increase of PaCO2 and reduced pHa in ectotherms. Although the increase in PaCO2 and CO2 efflux is due, in part, to adjustments in the ACR, the impact on O2 transport is primarily caused by the right shift of the O2 dissociation curve with increased temperature and reduced pH (Bohr effect), and its interaction with intracardiac shunts that increase PaO2 and systemic O2 transport. Taken together, we suggest that the regulated hypoventilation relative to metabolism provides several identifiable benefits to systems gas transport independent of any effects on alphastat pH regulation.
Enhancing the CO2 efflux driving force
As indicated above, CO2 efflux is the product of GCO2 and ΔPCO2. In order to increase CO2 efflux with increased metabolic demands, either or both of these variables can be increased. For resting animals, the temperature-mediated ratios for the relationship of PaCO2 with temperature are 1.2 for fish, 1.5–1.9 for amphibians and 1.2–1.4 for reptiles, all generally lower than the Q10 values of 2–3 for standard and maximal metabolism (see summary in Hedrick et al., 2015). This indicates that changing the driving force for CO2 efflux by raising PaCO2 does not explain an intraclass limitation on V̇CO2 with changes in temperature and, instead, suggests the potential for co-adaptations in respiratory conductance and/or ventilatory capacity. Based on the resting PaCO2 values in Fig. 1A, the driving force for CO2 efflux is increased 42% for fish, 128% for amphibians and 73% for reptiles with body temperature increasing from 10 to 30°C. This indicates that the PaCO2 response to temperature in each class would enhance the driving force for CO2 efflux during maximal activity by increasing the regulated resting PaCO2, but is not sufficient to account for the Q10 during maximal activity.
An interesting intraclass test of the driving force hypothesis can be found in fish, a truly bimodal group (water versus air) in terms of gas exchange. The four obligate air-breathing species of fishes in the summary of Ultsch and Jackson (1996) have a PaCO2 of about 3.3 kPa compared with 0.42 kPa for water-breathing fish at equivalent temperatures. We interpret this as the necessity to increase the driving force for CO2 efflux when the gas bladder conductance is probably lower than the gill conductance combined with the decrease in CO2 capacitance of air compared with water.
From a maximal aerobic metabolic perspective, what might be the effect of interclass variation in the magnitude of the PaCO2 response to increased temperature on the capacity to enhance O2 delivery and CO2 efflux during activity? Based on the data from Fig. 2A, the ratio of interclass PaCO2, at their respective temperatures for V̇O2,max, between fish (20°C) and amphibians (25°C) is 17.1 and that between amphibians and reptiles is 2.5. The large phylogenetic ratio for PaCO2 between fish and amphibians is consistent with the Q10 of 13.4 for V̇O2,max between fish and amphibians at 20 and 25°C, respectively, and a Q10 of 1.2 for V̇O2,max between amphibians and reptiles at 25 and 35°C, respectively (Hedrick et al., 2015). The correspondence of Q10 values between V̇CO2,max and the ratios for interclass PaCO2 is consistent with an increase in PaCO2 playing a significant role in explaining interclass variation in V̇CO2,max, unlike the resting condition where increased conductance (GCO2) appears to provide the increase in resting V̇CO2. As noted above, fishes and amphibians, increase PaCO2 at V̇O2,max (see Hillman et al., 2013), which would enhance CO2 efflux by increasing the driving force for PCO2 to a greater extent than our estimates here using resting PaCO2. For reptiles, PaCO2 at V̇O2,max does not appear to increase over resting values (Hillman et al., 2013); thus, our estimates of CO2 efflux based on resting PaCO2 values for this group are probably more accurate.
These data may also indicate that increasing the PaCO2 driving force to increase CO2 efflux in vertebrates has limits. For example, increases of PaCO2 greater than 5 kPa, which appears to be near the upper limit for reptiles and endotherms, may cause significant changes in pH that potentially compromise protein function, suggesting that endotherms use alternative adaptations such as increased respiratory conductance and ventilatory capacity to achieve the greater fluxes of O2 and CO2.
Enhancing the blood buffering capacity
The increase in PaCO2 also leads to increased concentrations of HCO3− (Ultsch and Jackson, 1996). An increase in [HCO3−] would increase the buffering capacity of the blood. Lactic acid begins to accumulate in the blood when aerobic power outputs during activity are 50–70% of maximal (Davis et al., 1979; Seeherman et al., 1983; Gleeson and Brackenbury, 1984; Taigen and Beuchat, 1984; Goolish, 1991). Consequently, an added selective advantage of the increase in PaCO2 and [HCO3−] with increased V̇O2,max is less disruption of pHa during high metabolic power outputs. Malan (2014) has also suggested increased buffering as a benefit of the hypercapnic acidosis associated with hibernation and torpor.
The hypothesis presented here requires a linkage between body temperature and the regulation of ventilation. The regulation of increased PaCO2 and reduced ACR with increased temperature implies a receptor linked to ventilation operates to maintain ventilation within narrow limits as temperature changes. It is well known that PaCO2 is tightly regulated by the complex interactions of central and peripheral chemoreceptors in vertebrates (Milsom, 2002). A ventilation-mediated mechanism that controls ventilation and, therefore, arterial PaCO2 and pHa with changes in temperature provides a convenient negative feedback mechanism. Recent work with bullfrogs (Lithobates catesbeianus) and monitor lizards (Varanus exanthematicus) has shown the presence of CO2/pH chemosensitive neurons of the locus coeruleus (LC), a putative ventilatory control region (Santin et al., 2013; Zena et al., 2016). Lithobates catesbeianus has been characterized as a typical alphastat regulator (Reeves, 1972; Santin et al., 2013), whereas V. exanthematicus is a pH-stat regulator with little change in pHa over a broad range of temperatures (Zena et al., 2016). In L. catesbeianus, cooling increased, and warming decreased, the firing rate of LC chemosensitive neurons (Santin et al., 2013). Moreover, cooling reduced CO2/pH chemosensitivity in a temperature-dependent fashion; thus, the magnitude of the chemosensitive response was temperature dependent (Santin et al., 2013). By contrast, chemosensitive LC neurons in V. exanthematicus increase firing rates with increasing temperature and have a large Q10 effect compared with bullfrog chemosensitive LC neurons (Zena et al., 2016). Varanus exanthematicus also have populations of LC neurons that are excited or inhibited by CO2 and the proportion of CO2-inhibited neurons increases with cooling (Zena et al., 2016). The finding that populations of CO2/pH chemosensitive neurons in the LC of bullfrogs and lizards that are modulated by temperature provides a parsimonious explanation for ventilatory regulation of PaCO2 and pHa with changes in body temperature.
We suggest that the pattern of arterial pH and PaCO2 initially described by Howell et al. (1970) and later interpreted from a solely biochemical structure–function perspective (Reeves, 1972, 1977; White and Somero, 1982) may additionally, or primarily, have its evolutionary basis in the enhancement of systems-level gas transport. Increased temperature increases aerobic demands for O2 influx and CO2 efflux both at rest and during activity. The alphastat pattern of hypoventilation relative to aerobic metabolic demand, leading to an increase in PaCO2 and [HCO3−] and a decline in pHa, preserves ventilatory capacity, increases blood buffering capacity and enhances both CO2 and O2 fluxes that would be associated with increases in body temperature and activity. We suggest this hypothesis deserves consideration along with potential (as-yet undocumented) imidazole-mediated protein structure–function considerations.
We thank P. Bickler, L. Hartzler, J. Podrabsky, L. Crawshaw and R. Putnam for useful discussions in the development of this manuscript and multiple reviewers for strengthening the focus of the manuscript.
Conceptualization: S.S.H., M.S.H.; Methodology: S.S.H., M.S.H.; Formal analysis: S.S.H., M.S.H.; Writing - original draft: S.S.H., M.S.H.; Writing - review & editing: S.S.H., M.S.H.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
The authors declare no competing or financial interests.