Introduction

Virtually all ectothermic animals reduce arterial pH (pHa) when temperature increases (Austin et al. 1927; Rahn et al. 1971; Randall and Cameron 1973; Wood et al. 1978; Smatresk and Cameron 1982; Cameron and Kormanik 1982; Boutilier et al. 1987; Amin-Naves et al. 2004; Truchot 2012; Fobian et al. 2014). The regulated variable underlying this ubiquitous physiological pattern, be it ionization of the α-imidazole-groups on histidines or the relative alkalinity of water, remains as disputed as it is unresolved (Austin et al. 1927; Reeves 1972; Hazel et al. 1978; Heisler 1980; Wang and Jackson 2016). Nevertheless, given the omnipresence of the pHa fall with temperature, there seems to be general consensus that the reduction in pHa with increased temperature (ΔpHa/ΔT) is a tightly regulated response, although direct evidence is limited.

The temperature-mediated pHa reduction can be achieved by either transepithelial exchange of acid–base relevant ions or respiratory modulation of arterial PCO2 (PaCO2) by altering ventilation relative to metabolic CO2 production (i.e., air convection requirement) (Wang and Jackson 2016). The importance of these two distinct mechanisms differs between water and air breathers. Thus, due to the high ventilation rates imposed by oxygen’s low water solubility, water-breathing vertebrates are typically obliged to alter plasma [HCO3 −] ([HCO3 −]pl) by transepithelial ion exchange. In contrast, air breathers normally regulate pHa by altering ventilation rates to change PaCO2.

Bimodal breathers, such as air-breathing fishes and many amphibians, also reduce pHa as temperature increases (Johansen 1970; Smatresk and Cameron 1982; Graham 1997; Amin-Naves et al. 2004; Wright and Turko 2016), but the underlying mechanisms are likely to differ according to the plethora of respiratory structures and species-specific dependence on aerial respiration. In this context, the air-breathing catfish Pangasianodon hypophthalmus represents an appropriate model to examine how pHa regulation is affected by water versus air-breathing. This species tolerates temperatures over a wide range from 20 to 39 °C and is in possession of both large gills and well-vascularized air-breathing organ (Phuong et al. 2017). At temperatures between 27 and 39 °C, P. hypophthalmus can accommodate all gas exchange by its well-developed gills in normoxic water (Andersen et al. 2015), but resorts to air-breathing by virtue of its modified swim-bladder when exposed to aquatic hypoxia (Lefevre et al. 2011). This response is different compared to other bimodal breathing fishes that typically increase air-breathing frequency with temperature in normoxia.

We hypothesized that P. hypophthalmus regulates pHa via active ion-exchange only (i.e., PaCO2 is constant and [HCO3 −]pl is reduced) when exposed to increases in temperature in normoxic water where P. hypophthalmus can accommodate gas exchange entirely by water-breathing. In contrast, we hypothesized that the transition to air-breathing in hypoxic water would lead to pHa-regulation being of respiratory origin (i.e., PaCO2 increases and [HCO3 −]pl remains constant), as PaCO2 would increase passively due to an increased resistance in CO2 unloading associated with air-breathing. Thus, by disentangling the effects of air-breathing, we also address the more general question as to whether the reduction in pHa with increased temperature is indeed maintained during both air and water-breathing; If pHa and ΔpHa/ΔT are indeed similar during both conditions, it would provide strong evidence for pHa actually being the sensed variable, and that new set-point is regulated when body temperature changes.

Materials and methods

Animals

Juvenile Pangasianodon hypophthalmus were obtained from a local supplier (Credofish, Denmark) and raised at Aarhus University at 27 °C for at least 6 months in 1 m3 tanks supplied with oxygenated water from a recirculating system. The fishes were fed to satiety with commercial dry pellets every day. All procedures described below were in accordance with the Danish rules for animal experimentation (2016-15-0201-00865).

Animal handling and experimental protocol

Pangasianodon hypophthalmus (n = 20, mass = 1590 ± 93 g, length = 48 ± 1 cm; means ± SEM) were anaesthetized in 0.15 g l−1 benzocaine. When unresponsive to handling, they were transferred to a surgical table where the gills were irrigated with well-oxygenated water containing 0.08 g l−1 benzocaine (Phuong et al. 2016). A polyethylene catheter (PE50) was inserted into the dorsal aorta (Soivio et al. 1975) and another catheter (PE50) was placed in the opercular cavity through the operculum. The fish were allowed to recover for 24 h, which is sufficient for complete acid/base recovery in P. hypophthalmus (Phuong et al. 2016), in individual 100 l normoxic holding tanks at 35 °C supplied with water from a central 500 l tank in which water PO2 (PwO2) and temperature were regulated. This setup allowed fast equilibration of both temperature and PwO2 in the individual holding tanks, while the air-phase above water in those tanks remained normoxic. On the following day, ventilatory parameters were monitored for 30 min in each fish, whereupon a blood sample was withdrawn. Temperature was then decreased by 5 °C on each of the following days, and ventilation and blood parameters were measured at each temperature. Temperature was changed by ~ 2.5 °C h−1, allowing ~ 22 h for compensation to the temperature reduction, which is sufficient for pHa regulation in other air-breathing fishes (Smatresk and Cameron 1982). This protocol was completed in both normoxia (PwO2 > 130 mmHg, n = 13) and hypoxia (PwO2 = 25 mmHg, n = 7). The critical oxygen tension of P. hypophthalmus is 57 mmHg at 27 °C (Lefevre et al. 2011), and the hypoxic PwO2 level was chosen to represent 50% of the Pcrit value, where the species is expected to use air-breathing to provide for almost all of its metabolic demand.

To remove a possible influence of the direction of temperature change (i.e., heating from 20 to 35 °C versus cooling from 35 to 20 °C) on our data interpretation, six of the normoxic fish were recovered at 20 °C, and temperature was then increased rather than decreased on the following days. We did not observe any effect of direction of temperature change on pHa (a proxy for post-anesthesia acid/base recovery) all hypoxic fishes were cooled from 35 to 20 °C. This allowed us to calculate a Ventilation Index (VI) relative to 35 °C normoxia on day 1 (see calculation for VI below), where the hypoxic fish recovered from anesthesia in normoxia at 35 °C and ventilatory parameters were measured on the first day, after which the water was bubbled with pure N2 in the central tank to obtain aquatic hypoxia.

Water oxygen level and temperature were monitored and regulated through a controller box (Respirometer, v. 1.5.0.c Aarhus University, Denmark) coupled with an optical O2 probe (VisiFerm DO, Hamilton Process Analytics, NV, USA).

Blood analysis

Immediately after blood sampling, pHa and PaO2 were measured on a GEM Premier 3500 automated blood gas analyzer (Instrumentation Laboratory, Bedford MA) using previously validated temperature compensation algorithms (Malte et al. 2014). Blood hemoglobin concentration was measured spectrophotometrically after conversion to cyano-methemoglobin using Drabkin’s reagent. Plasma concentration of CO2 ([CO2]pl) and blood concentration of O2 ([O2]bl) were measured in duplicate using the methods described by Cameron (1971) and Tucker (1967), respectively. The remaining blood was centrifuged at 4000g for 3 min and the separated plasma stored at −80 °C for subsequent measurements of plasma Cl− concentration ([Cl−]pl) using a chloride titrator (Sherwood model 926S MK II Chloride analyzer, Sherwood Scientific Ltd., Cambridge, UK). PaCO2 was calculated by rearranging the Henderson–Hasselbalch equation (Damsgaard et al. 2014) using pK´ from (Boutilier et al. 1984) and αCO2 from (Dejours 1981). [HCO3 −]pl was calculated by subtracting dissolved CO2 from [CO2]pl ([HCO3 −]pl=[CO2]pl − αCO2·PaCO2). The concentration of hemoglobin-bound O2 ([HbO2-sat]) was calculated by subtracting plasma-dissolved O2 from [O2]bl ([O2]bl − αO2·PaO2) using PaO2 from GEM and αO2 from (Dejours 1981). Fractional arterial saturation of hemoglobin with O2 (HbO2-sat) was calculated as the concentration of hemoglobin-bound O2 relative to blood hemoglobin concentration. The measured PaO2 using the GEM 3500 is overestimated by up to 25%, but due to the low αO2 compared to [O2]bl, such errors will overestimate HbO2-sat by less than 0.33% (Malte et al. 2014). For the same reason, PaO2 was calculated at all four temperatures from the HbO2-sat and the Hill equation (HbO2-sat = PaO2 n50/(PaO2 n50 + P50 n50); P50, partial pressure of oxygen at half saturation; n50, Hill’s cooperativity constant) using data from whole blood O2 equilibrium curves (Damsgaard et al. 2015b), and these calculated PaO2 values are similar to the values determined with a Radiometer PO2-electrode in a previous study (Damsgaard et al. 2015b). The values used for this model were n50 = 1.5, and P50 = 1.3, 3.2, 8.3, and 20.1 mmHg at 20, 25, 30, and 35 °C, respectively, and the same P50 was used for both the normoxic and hypoxic groups as the typical hypoxia-induced reduction in P50 due to lowered intra-erythroid [ATP] is not expected in P. hypophthalmus with ATP irresponsive Hb (Damsgaard et al. 2015b).

Measurements of cardio-respiratory parameters

The opercular and dorsal aortic catheters were connected to pressure transducers (PX600, Irvine, CA, USA) and pressure data were collected with a MP100 BIOPAC system (Biopac Systems Inc., CA, USA) at 200 Hz. The pressure transducers were two-point calibrated against a static water column on a daily basis with pressures similar to in vivo values. The fish tanks were covered with dark plastic to avoid visual disturbances during the experiment. Air-breathing frequency (f AB) was determined from the characteristic changes in the pressure profile from the opercular catheter. This AB-signal was determined after observation of several ABs in different individuals prior to the experimental protocol and were easily distinguishable from normal ventilation. Movement of the fish did produce some noise on the signal, but only rarely did this interfere with the analysis. Total gill ventilation was expressed as a 30 min average for each fish at the given temperature, and expressed as the product of frequency (f V) and amplitude (V AMP). VI was calculated for each temperature (n), relative to 35 °C in normoxia:

$${\text{V}}{{\text{I}}_{{\text{temp}}\left( n \right)}}=\frac{{{{\left( {{\text{amplitude}}\; \times \;{\text{frequency}}} \right)}_{{\text{temp}}\left( n \right)}}}}{{{{\left( {{\text{amplitude}}\; \times \;{\text{frequency}}} \right)}_{{\text{temp}}\left( {35^{\circ}{\text{C}}} \right)}}}}\; \times \;100{\text{\% }}$$
(1)

Calculations and statistics

The temperature coefficient, Q 10, for a biological process (R) between two temperatures T 1 and T 2 was calculated as:

$${Q_{10}}={\left( {\frac{{{R_2}}}{{{R_1}}}} \right)^{\frac{{10^{\circ}{\text{C}}}}{{{T_2} - {T_1}}}}}$$
(2)

The effects of temperature and hypoxia were assessed by a mixed model analysis of variance using individual fish as random effect, temperature as repeated factor and hypoxia as treatment factor in R software (R Core Team, 2015, v. 3.2.2). Output from this analysis is tabulated in Table 1. A pairwise comparison test between normoxia and hypoxia at each temperature was performed, and P-values were adjusted using a Benjamini-Hochberg correction in InVivoStat software (InVivoStat v. 3.6.0). Data were inspected for homoscedasticity using a residual plot, and when heteroscedastic, data was rank transformed. Normality of all data was assessed from a normal probability plot. Significance level was set to 0.05.

Table 1 Output of the mixed model ANOVA showing the effect of temperature and hypoxia on respiratory parameters
Full size table

Results

Acid/base regulation and oxygen saturation

In normoxic water, pHa decreased with an average gradient of − 0.0086 ± 0.0018 U oC−1 (P < 0.001; Fig. 2a) between 20 and 35 °C. However, this pHa change appears non-linear and the change between 20 and 25 °C was not significant (P = 0.45). Between 25 and 35 °C, the pHa decrease was linear with a slope of − 0.013 ± 0.002 U oC−1 (Fig. 1). This was associated with a reduction in [HCO3 −]pl (− 0.17 ± 0.080 mM oC−1; P < 0.05; Figs. 1, 2c) at constant PaCO2 (P = 0.64; Figs. 1, 2b). The fall in [HCO3 −]pl was attended by an increase in [Cl−]pl (0.50 ± 0.11 mM oC−1; P < 0.001, Table 2).

Fig. 1
figure 1

Davenport diagram showing extracellular acid–base regulation in cannulated Pangasianodon hypophthalmus in response to 5 oC-temperature changes between 20 and 35 °C during aquatic normoxia (PwO2 > 130 mmHg; blue-lined symbols) and aquatic hypoxia (PwO2 = 25 mmHg; red-lined symbols). Isopleths are color coded for temperature (blue to red for increasing temperature). Isopleths were generated using pK´ from (Boutilier et al. 1984) and αCO2 from (Dejours 1981). Data are means ± standard error of mean (n = 13 for normoxia, 7 for hypoxia). (Color figure online)

Full size image
Fig. 2
figure 2

Extracellular acid–base regulation in cannulated Pangasianodon hypophthalmus in response to temperature changes between 20 and 35 °C during aquatic normoxia (PwO2 > 130 mmHg; blue symbols) and aquatic hypoxia (PwO2 = 25 mmHg; red symbols). a arterial pH (pHa), b arterial PCO2 (PaCO2), and c plasma bicarbonate concentration ([HCO3 −]pl). Asterisks indicate significant difference in a parameter between the normoxic and hypoxic groups at a specific temperature. Data are means ± standard error of mean (n = 13 for normoxia, 7 for hypoxia). (Color figure online)

Full size image
Table 2 Plasma concentrations of Na+, K+, and Cl−, and osmolarity in Pangasianodon hypophthalmus during acclimation to 20–35 °C and normoxia or hypoxia
Full size table

Arterial pH was not affected by water oxygen levels (P = 0.20; Fig. 2a). During hypoxia, PaCO2 rose approximately three-fold from 20 to 35 °C (0.13 ± 0.011 mmHg oC−1; P < 0.001; Figs. 1, 2b) attended by a rise in [HCO3 −]pl of 3 mM (− 13.3 ± 1.9 slykes (P < 0.001); 0.20 ± 0.019 mM oC−1 (P < 0.001); Figs. 1, 2c). Arterial HbO2-sat was higher during normoxia (P < 0.001) and decreased with temperature during hypoxia (P < 0.001), but not in normoxia (P = 0.87; Fig. 3).

Fig. 3
figure 3

a Arterial oxygen saturation (HbO2-sat) and b calculated arterial PO2 (PaO2) in cannulated Pangasianodon hypophthalmus in response to temperature changes between 20 and 35 °C during aquatic normoxia (PwO2 > 130 mmHg; blue-lined symbols) and aquatic hypoxia (PwO2 = 25 mmHg; red lined symbols). Asterisks indicate significant difference in a parameter between the normoxic and hypoxic groups at a specific temperature. Data are means ± standard error of mean (n = 13 for normoxia, 7 for hypoxia). (Color figure online)

Full size image

Cardioventilatory responses to temperature in normoxic and hypoxic water

In normoxic water, the ventilation index increased [however, not significantly (P = 0.20)] with a Q10 of 1.4 through a significant elevation in fV as temperature rose (1.2 ± 0.38 min−1 °C−1 (P < 0.001)), while there were no changes in rel. V AMP (P = 0.23) or fAB (P = 0.49) (Fig. 4). Gill ventilation frequency was unaffected by hypoxia (P = 0.23), but rel. V AMP and f AB were higher in hypoxia (both P < 0.001) (Fig. 4), resulting in an overall higher VI in hypoxia (P < 0.01), that, however, was only significantly different at 25 °C (P < 0.01). As in hypoxia, increases in temperature resulted in higher fV (1.1 ± 0.24 min−1 °C−1 (P < 0.001)), but not in Rel. V AMP (P = 0.25) and VI (P = 0.53, Q 10 = 0.88). Air-breathing was absent at 20 °C during normoxia and hypoxia, but fAB increased with temperature during hypoxia (1.4 ± 0.15 h−1 °C−1 (P < 0.001)), resulting in higher f AB at 25, 30, and 35 °C (P = 0.025, < 0.001, and < 0.001, respectively) (Fig. 4a).

Fig. 4
figure 4

Ventilatory regulation in cannulated Pangasianodon hypophthalmus in response to temperature between from 20 to 35 °C during aquatic normoxia (PwO2 > 130 mmHg; blue symbols) and aquatic hypoxia (PwO2 = 25 mmHg; red symbols). a Air-breathing frequency (f AB). b Branchial ventilation frequency (f V), c relative ventilation amplitude (rel. V AMP), and d Branchial Ventilation Index (VI). Relative ventilation amplitude and Ventilation Index for individual fish were calculated relative to the value in normoxia at 35 °C. Asterisks indicate significant difference in a parameter between the normoxic and hypoxic groups at a specific temperature. Data are means ± standard error of mean (n = 13 for normoxia, 7 for hypoxia). (Color figure online)

Full size image

Both fH and MAP increased with temperature (2.7 ± 0.4 min−1 °C−1 (P < 0.001), and 0.33 ± 0.05 cmH2O °C−1 (P < 0.001), respectively), and were unaffected by aquatic hypoxia (P = 0.65, and 0.31, respectively) (Fig. 5).

Fig. 5
figure 5

Cardiovascular regulation in cannulated Pangasianodon hypophthalmus in response to temperature changes between 20 and 35 °C during aquatic normoxia (PwO2 > 130 mmHg; blue symbols) and aquatic hypoxia (PwO2 = 25 mmHg; red symbols). a heart rate (f H) and b mean arterial blood pressure (MAP). Asterisks indicate significant difference in a parameter between the normoxic and hypoxic groups at a specific temperature. Data are means ± standard error of mean (n = 13 for normoxia, 7 for hypoxia). (Color figure online)

Full size image

Discussion

The reduction in pHa as temperature rose (ΔpHa/ΔT) in P. hypophthalmus, resembles the archetypical fall of ectothermic vertebrates (Rahn et al. 1971; Randall and Cameron 1973; Wood et al. 1978; Smatresk and Cameron 1982; Cameron and Kormanik 1982; Boutilier et al. 1987; Amin-Naves et al. 2004; Fobian et al. 2014), and pHa and ΔpHa/ΔT did not depend on whether P. hypophthalmus was exclusively water-breathing in normoxic water or bimodal-breathing in hypoxic water, where air-breathing frequencies approached the high range observed in other air-breathing fishes (Graham 1997). However, the acid–base status was strikingly different. Thus, pHa fell exclusively through a reduction in [HCO3 −]pl in normoxia, whereas elevated PaCO2 accounted for almost all of the temperature associated pHa change in hypoxic water. P. hypophthalmus, therefore, exhibits an intraspecific shift of the classic distinction between water- and air-breathers in the acid–base response to altered temperature. This shift bears resemblance to the transition in regulation between air- and water breathers within the sarcopterygian lineage and is normally ascribed to the emergence of central chemoreception for CO2/pH in amniotes allowing for ventilatory regulation of PaCO2. There is, however, no indication of central chemoreception affecting ventilation in P. hypophthalmus (Thomsen et al. 2017), and we suggest that the rise in PaCO2 with elevated temperature in hypoxia is a passive response arising from the increased aerial respiration and temperature-independent branchial ventilation in hypoxia rather than a regulated process per se.

P. hypophthalmus relies almost exclusively on branchial gas exchange in normoxic water (Lefevre et al. 2011), and it is not surprising that it follows the typical piscine pattern of achieving the ΔpHa/ΔT by means of decreasing [HCO3 −]pl (Randall and Cameron 1973; Heisler 1980). Given the reduction in [Cl−]pl, it seems that branchial HCO3 −/Cl− exchange mediates the reduction in [HCO3 −]pl with increased temperature, even though it does not follow the classical equimolar exchange rate (Grosell et al. 2009), which may have been due to non-equimolar exchange rates, H+/Na+-exchange, and/or experimental error (Evans et al. 2005; Perry and Gilmour 2006; Damsgaard et al. 2015a; Hvas et al. 2016). The temperature independence of PaCO2 during normoxia was consistent with the matching of gill ventilation to the rise in standard metabolic rate (SMR), where Q 10 for gill VI of 1.4 (25–35 °C) is similar to the Q 10 of 1.6 for SMR [27–36 °C; (Andersen et al. 2015)]. In addition, there was no increase in air-breathing events with increased temperature in normoxic water. This pattern contrasts to the air-breathing gar and South American lungfish, where air-breathing frequency increases as temperature stimulates metabolism (Smatresk and Cameron 1982; Amin-Naves et al. 2004), but resembles the air-breathing Alaska blackfish (Lefevre et al. 2014).

In hypoxia, P. hypophthalmus exhibited a marked branchial hyperventilation (significant at 25 °C), which is atypical for a bimodal breather with high capacity for aerial oxygen uptake. The increased gill ventilation in hypoxia, where CO2 is excreted at a higher rate compared to in normoxia, explains the lower PCO2 in hypoxia at the lower temperatures. Due to the higher capacitance coefficient for oxygen in air compared to water, a shift from water to air-breathing provides for decreased ventilation rates, leading to elevated PCO2 in blood and tissues (Rahn 1966), and explains why PaCO2 increases in many facultative air-breathing fishes as they switch to aerial respiration in hypoxic water (Shartau and Brauner 2014; Wright and Turko 2016). In these cases, there is a typical reduction in branchial CO2 excretion caused by reduced perfusion of the secondary branchial lamellae. We did not measure the partitioning of gas exchange in P. hypophthalmus in this study, but the rise in air-breathing with temperature indicates that the increase in PaCO2 with elevated temperatures in hypoxic water is a passive consequence of the progressive transition to aerial respiration that was enforced by the increased metabolism and associated CO2 production with increased temperatures. As such, P. hypophthalmus behaves akin to skin-breathing salamanders where increased PaCO2 causes a pHa reduction with increased temperature through a fortuitous matching of the temperature effects on CO2 production and the cutaneous conductance for CO2 (Moalli et al. 1981). In fact, in P. hypophthalmus, it appears that the rise in PaCO2 with temperature elevation would have caused a larger fall in pHa than predicted by normoxic pHa regulation. Pangasianodon hypophthalmus, therefore, regulates pHa by transepithehial ion exchange to elevate [HCO3 −]pl, which is a response also observed in other bimodal breathing fishes (Smatresk and Cameron 1982; Amin-Naves et al. 2004). This shows that despite of having markedly different PaCO2 in normoxia and hypoxia, P. hypophthalmus still regulate pHa to achieve the same pHa set-point, and hence the pHa set-point is PaCO2 independent. This suggests that it is pHa that is sensed by a hitherto unidentified pHa sensor and that this sensor regulates pHa by altering trans-epithelial ion-exchange rates, where positive deviations in set-point pHa result in enhanced HCO3 − excretion rates (and vice versa). The changes in pHa with temperature follows the changes in pK of α-imidazole, and since pHa might be the sensed parameter, this points to the α-imidazole on a surface-exposed histidine being the sensor of pHa and the mediator of temperature-induced pHa regulation.

Pangasianodon hypophthalmus had lower HbO2-sat and PaO2 across all temperatures during aquatic hypoxia. Interestingly, PaO2 was lower than PwO2 at all temperatures showing that blood did not fully equilibrate to PwO2 during the branchial transit. Branchial oxygen loss can, therefore, not account for the arterial desaturation, which has been suggested to be a problem for air-breathing fishes during aquatic hypoxia (Randall 1981; Graham 1997; Scott et al. 2017). In contrast, it is plausible that perfusion of the secondary lamellae is highly reduced during hypoxia curtailing such branchial loss, and that the low oxygen levels in the dorsal aorta represents the mixture of oxygenated blood from the swim-bladder and systemic venous return that merge at the inflow to the heart. This would resemble the bowfin (Amia calva), which displays a similar 25% desaturation, and where flow measurements indicated this was caused by the mixing of fully saturated blood leaving the ABO and venous blood (Randall et al. 1981). Regardless of the cause of the low HbO2-sat, branchial oxygen uptake could have been augmented through increases in pHa to left-shift the hemoglobin O2 binding curve. This illustrates the tradeoff between pHa- and oxygen homeostasis at higher temperatures in bimodal- and water-breathing vertebrates, where pHa regulation is prioritized over oxygen uptake at higher temperatures in P. hypophthalmus compromising blood–oxygen transport. A similar compromise between oxygen delivery and acid–base balance by virtue of lowering ventilation relative to metabolic CO2 production at high temperature has also been proposed for amphibians and reptiles (e.g., Stinner 1987; Wang et al. 1998).

In conclusion, we demonstrate P. hypophthalmus exhibits intraspecific shift in the pattern for temperature-induced pHa modulation, but that ΔpHa/ΔT and the actual pHa are independent whether the temperature-induced fall in pHa is achieved by changes in [HCO3 −]pl during water-breathing or via changes in PaCO2 during hypoxia. This was achieved without central chemoreception and limited the requirements for transepithelial ion-exchange. This finding provides strong evidence that pHa is indeed regulated around a new set-point when temperature changes, and since the changes in pHa follow the change in pK of α-imidazole, our data lend support to the idea of surface-exposed histidines sense pHa and mediate the acid/base regulation as originally proposed in Reeves’ alphastat hypothesis (Reeves 1972). Second, our study provides insight to the physiological tradeoffs at higher temperatures between interconnected physiological systems, such as O2 transport and acid/base regulation. We emphasize the need to understand how such systems are limited and prioritized at higher temperatures to obtain a more holistic understanding of how such elevations are likely to affect organismal performance in a future warmer world.