Returning on empty: extreme blood O₂ depletion underlies dive capacity of emperor penguins

Emperor penguins (Aptenodytes forsteri) are premier avian divers with routine dive durations of 5–12 min and a reported maximum dive duration of 22 min (Kooyman and Kooyman,1995; Robertson,1995). This breath-hold capacity is dependent, in part, on utilization and depletion of the blood O₂ store(Scholander, 1940). Since dissociation of all O₂ from hemoglobin (Hb) requires a low O₂ partial pressure (P_O2), effective utilization of the blood O₂ store implies exceptional hypoxemic(low P_O2) tolerance in these animals. Prior investigations of the diving physiology and behavior of emperor penguins at the corralled, isolated dive hole of our Penguin Ranch research camp on the sea ice of McMurdo Sound, Antarctica, have made such diving activity an ideal model in which to investigate blood O₂ depletion and hypoxemic tolerance.

While under the sea ice of McMurdo Sound, emperor penguins primarily feed on the sub-ice fish, Pagothenia borchgrevinki(Ponganis et al., 2000). Dive durations of 5–12 min are common, and dive depths are usually less than 100 m. During these dives, the birds undergo variable bradycardias(Kooyman et al., 1992),maintain aortic and vena caval temperatures near 37–39°C(Ponganis et al., 2001; Ponganis et al., 2004; Ponganis et al., 2003) and have an aerobic dive limit (ADL; dive duration associated with post-dive blood lactate accumulation) of 5.6 min (Ponganis et al., 1997b).

Particularly relevant to the current investigation was the recent successful application of an air sac P_O2electrode and backpack recorder to diving birds at the Penguin Ranch(Stockard et al., 2005). P_O2 profiles obtained via the O₂ electrode revealed that 42% of these voluntary dives of emperor penguins had end-of-dive air sac P_O2 values less than 20 mmHg (1 mmHg=133.3 Pa). Such low P_O2 values are significant in comparison with other birds for several reasons. First, the lowest of these air sac values in emperor penguins is less than inspired air values (23 mmHg) of birds at altitudes as high as 11 580 m (Black and Tenney, 1980). Second, these values in free-diving emperor penguins are also lower than the air sac P_O2values (∼30 mmHg) of pekin ducks (Anas platyrhynchos) forcibly submerged to the point of `imminent cardiovascular collapse'(Hudson and Jones, 1986). Third, because air sac P_O2 represents the maximum arterial P_O2 and, in fact, is usually greater than the simultaneous arterial value(Powell, 2000; Weinstein et al., 1985), these low air sac values in emperor penguins imply that blood P_O2 values are commonly less than 20 mmHg. This is remarkable in that blood O₂ content at a P_O2 of 22 mmHg is very low (less than 5 ml O₂ dl^–1 blood) in the high-altitude bird, the bar-headed goose (Anser indicus), and is even less in pekin ducks and pigeons (Columbia livia) (Black and Tenney, 1980; Hudson and Jones, 1986; Weinstein et al.,1985).

Therefore, in order to further investigate hypoxemic tolerance and blood O₂ depletion during dives, we equipped emperor penguins that were diving at the Penguin Ranch research camp with intravascular P_O2 electrodes and backpack recorders. In addition, we measured P_O2, O₂content, pH, P_CO2 and lactate concentrations in arterial and venous blood samples from birds at rest. Our goals were to determine (1) the baseline blood respiratory variables in birds at rest, (2)the relationship of the final blood P_O2 of a dive to dive duration, (3) the relationship of final blood P_O2 of a dive to the ADL, and (4) the hypoxemic tolerance of emperor penguins.

In the austral springs of 2001 and 2003–2005, non-breeding emperor penguins (Aptenodytes forsteri Gray; 20–30 kg in body mass)were captured on the sea ice of McMurdo Sound or at Terra Nova Bay near Cape Washington (74°36′, 165°24′). They were maintained at the corralled, isolated dive hole of the Penguin Ranch in McMurdo Sound(77°41′, 165°59′) for six weeks, as in past studies(Kooyman et al., 1992; Ponganis et al., 1997b), and were then released at the McMurdo Sound ice edge. Blood-sampling catheters or P_O2 electrodes/thermistors were inserted percutaneously into the aorta or vena cava of emperor penguins under general isoflurane anesthesia as described previously(Ponganis et al., 1997b; Ponganis et al., 2001; Ponganis et al., 2004) and below. Catheterizations were confirmed as arterial or venous by observations of obvious pulsatile flows from arteries, by transduction of intravascular pressure and/or by P_O2 values of >250 mmHg in arteries and <100 mmHg in veins during anesthesia with 100%O₂ ventilation (Ponganis et al., 2004). After overnight recovery from anesthesia, birds were allowed to dive at the isolated dive hole. After 1–2 days of diving and data collection, probes and recorders were removed under general anesthesia. Data from the recorders were downloaded to a personal computer and analyzed with Microsoft Excel (OriginLab Corp., Northampton, MA, USA) and SPSS (SPSS,Inc., Chicago, IL, USA) software. Means are expressed ± standard deviation (s.d.). All procedures were approved under a UCSD Animal Subjects Committee protocol and US Antarctic Treaty Permit.

For blood analyses in birds at rest, arterial samples were obtained via the femoral artery in one case and via the brachial artery in the wing in two cases. Venous samples were obtained from the femoral or axillary vein. The femoral artery was cannulated percutaneously with a 20 g epidural catheter (Perifix catheter; B. Braun Medical Inc., Bethlehem, PA,USA) via a 19 g thin-wall needle. Brachial arteries were catheterized percutaneously with 4.5 cm, 20 g catheters (RA-04020; Arrow International,Reading, PA, USA). Only short arterial catheters could be inserted in the wing secondary to the axillary arterial plexus in emperor penguins(Trawa, 1970). The femoral vein was catheterized percutaneously(Ponganis et al., 1999) with 5-Fr nylon catheters (Cook, Bloomington, IN, USA) or the 20 g epidural catheters via the same technique used for the femoral artery. In two cases, the axillary vein was catheterized with an epidural catheter via a 16 g catheter introducer (Becton Dickinson, Sandy, UT, USA) in the brachial vein of the wing. The femoral artery catheter and all venous catheters were inserted 15–30 cm into the aorta and vena cava,respectively. Catheters were flushed with heparinized (4 U heparin ml^–1) 0.9% saline. The portion of the catheter and stopcock external to the body was filled with a pre-measured volume of 40% ethanol/60%heparinized saline to prevent freezing of the external catheter solution(Ponganis et al., 1997b).

Blood samples from birds at rest were obtained between two and four hours after recovery from anesthesia. During this time, the birds were kept inside the penguin transport box (45×45×120 cm) near their corral at ambient Antarctic temperatures (–10 to –20°C). Distraction of the calm bird by one observer via the open lid of the box allowed the collection of the sample (after withdrawal of >3× the tubing deadspace) by a second researcher.

Blood gas (P_O2, P_CO2 and pH) and lactate concentration analyses were conducted with a Series 200 i-STAT Portable Clinical Analyzer (CG4+cartridge; Abbott Point of Care Inc., East Windsor, NJ, USA) at 37°C. O₂ content was determined with a Tucker chamber technique (Models SI 782 O₂ meter and 1302 O₂ electrode; Strathkelvin,Motherwell, Scotland, UK) (Tucker,1967). Samples were analyzed within 10 min after collection. Blood gas, O₂ content and [lactate] were stable in the blood gas syringes(Model 4041, Sims Portex, Keene, NH, USA) for as long as 4 h at room temperature.

For P_O2 studies, a P_O2 electrode (Licox C1.1 Revoxode; Integra LifeSciences, Plainsboro, NJ, USA) (manufacturer's specifications: 90%response time <1 min, temperature correction factor <5%°C^–1, sensitivity error of <1%, and probe drift <2%day^–1) and thermistor (model 554; Yellow Springs Instruments,Yellow Springs, OH, USA) (60% response time 0.2 s, sensitivity 0.05°C),evaluated and calibrated as previously described(Ponganis et al., 2001; Ponganis et al., 2004; Stockard et al., 2005), were inserted 11–20 cm percutaneously into the vena cava or aorta via the femoral vein or artery with a peel-away introducer (PLIP 4.5 or 5.0-18-9-DENNY introducer; Cook)(Ponganis et al., 2001; Ponganis et al., 2004; Stockard et al., 2005). Prior to insertion, the P_O2 electrode was heparin-coated with an aseptic one-min immersion in 7% TDMAC heparin solution(Polysciences, Warrington, PA, USA).

In order to evaluate the effect of O₂ consumption by the electrode on O₂ depletion and a decline in P_O2 in stagnant blood, the output of two O₂ electrodes, each immersed in 2 ml of saline equilibrated with air at 38°C, was monitored when the equilibration air supply was shut off.

The custom P_O2/temperature recorder (UFI,Morro Bay, CA, USA), protected in an underwater housing (250 g,15×6×3.5 cm) and connected to the electrode and thermistor with waterproof cables (Sea Con, El Cajon, CA, USA; Impulse Enterprise, San Diego,CA, USA), was mounted with a Velcro™ patch and cable ties as previously described (Stockard et al.,2005) while the bird was under anesthesia. It recorded both parameters at 15 s intervals. The bird was also equipped with a Mk9 time depth recorder (TDR, Wildlife Computers, Redmond, WA, USA) (sensitive to 0.5 m, 30 g, 6.5×1.7×1.7 cm; sample rate, 1 Hz)(Stockard et al., 2005).

In prior studies, mean aortic and vena caval temperatures during dives ranged from 38.3 to 39°C and from 37.2 to 38.3°C, respectively(Ponganis et al., 2004; Ponganis et al., 2003). Therefore, the output of the P_O2 electrode was temperature corrected to 38°C and all data in the P_O2 profiles are reported for a temperature of 38°C. For a P_O2 of 60 mmHg, a±1°C temperature difference between an in vivo temperature and 38°C would result in a very minor ±4 mmHg difference between the in vivo P_O2 and the P_O2 profile reported at 38°C(Ashwood et al., 1983). For a P_O2 of 4 mmHg, the difference would be less than 0.3 mmHg.

Results of blood gas analyses are reported at 37°C because previously measured mean aortic and vena caval temperatures of emperor penguins at rest were 37.3–38.0°C and 36.3–38.7°C, respectively(Ponganis et al., 2004; Ponganis et al., 2003). For a P_O2 of 60 mmHg, P_CO2 of 40 mmHg and a pH of 7.40 reported at 37°C, an in vivo temperature of 38°C would again result in minimal changes in the in vivo values(Ashwood et al., 1983; Kiley et al., 1979), i.e. 64 mmHg, 42 mmHg and 7.39 pH units.

Histograms of dive duration and maximum depth in emperor penguins equipped with the intravascular P_O2 recorder are shown in Fig. 1. Arterial P_O2 profiles were obtained from 12 dives in three birds. Venous profiles were obtained from 130 dives in nine birds.

Arterial P_O2, O₂ content, pH, P_CO2 and [lactate] in three emperor penguins at rest were 68±7 mmHg, 22.5±1.3 ml O₂dl^–1, 7.50±0.02 pH units, 42±6 mmHg and 1.0±0.2 mmol l^–1, respectively. Venous P_O2, O₂ content, pH, P_CO2 and [lactate] in nine birds at rest were 41±10 mmHg, 17.4±2.9 ml O₂ dl^–1,7.50±0.03 pH units, 49±4 mmHg and 0.7±0.3 mmol l^–1, respectively.

Arterial P_O2 profiles during dives were similar to those previously observed in the air sacs(Stockard et al., 2005). The rise and fall in air sac and arterial P_O2profiles are illustrated in shallow dives of similar duration from two birds in Fig. 2. Final arterial P_O2 values, recorded during the last 15 s of dives as long as 6.8 min, ranged from 44 to 92 mmHg. The sample size of arterial P_O2 profiles (12 dives in three birds)was limited due to thrombus formation on the electrode, saltwater leaks into underwater connections and the technical difficulty of arterial catheterization.

Venous P_O2 profiles during a short dive and a long dive are illustrated in Figs 2 and 3. The final venous P_O2, again recorded during the last 15 s of a dive, was obtained over a wide range of dive durations, including many dives beyond the ADL (Fig. 4A). Final venous P_O2 values reached values of 1–6 mmHg in some dives and were less than 20 mmHg in 29% of dives. An exponential regression was constructed from the highest final P_O2 during each 1-min interval of dive duration(Fig. 4A); it represents the minimum rate at which final venous P_O2 declined in relation to dive duration of emperor penguins at the isolated dive hole. The declines in final venous, arterial and air sac P_O2(Stockard et al., 2005) in relation to dive duration are compared in Fig. 4B.

Blood temperatures during dives were similar to those recorded in previous studies (Ponganis et al.,2004; Ponganis et al.,2003). Final temperatures recorded during the last 15 s of dives ranged from 36.3 to 39.4°C.

In the test-tube evaluation of the depletion of O₂ in saline due to the O₂ consumption of the electrode, the P_O2 declined from 5 to 8 mmHg over the first 4 min after the equilibration air supply was shut off to the test tubes. It did not decline further over 20 min for either of the two P_O2 electrodes tested.

In emperor penguins equipped with the intravascular P_O2 recorder, dive durations and maximum depths were similar to those in past studies(Kooyman et al., 1992; Ponganis et al., 2001) at the isolated dive hole (Fig. 1). Dives of >100 m in depth were rare, and 47% of dives were >5.6 min, the previously measured ADL (Ponganis et al.,1997b). In 2004, birds often performed long-duration dives of>16 min; this included a 23.1 min dive(Fig. 3), which is now the longest reported dive in an emperor penguin. Dive profiles and lack of hunting ascents to the undersurface of the sea ice during these exceptional dives suggested that the birds were foraging at 50–80 m depth.

In emperor penguins at rest, mean arterial P_O2 was 68±7 mmHg, less than two-thirds the mean value in the air sac of birds at rest(Stockard et al., 2005). Large air-sac-to-arterial differences in P_O2 in birds at rest are considered primarily due to ventilation–perfusion mismatch(Powell, 2000). It is unknown if this air-sac-to-arterial difference in P_O2in emperor penguins at rest is also partially secondary to the thickened parabronchial capillary blood-to-air barrier that has been reported in these birds (Welsch and Aschauer,1986).

Although these arterial P_O2 values in emperor penguins are in the lower range of values reported for birds at rest(Powell, 2000), the arterial O₂ content of 22.5±1.3 ml O₂dl^–1 still represented greater than 91% saturation of the previously measured average [Hb] of 18 g dl^–1(Kooyman and Ponganis, 1998). Arterial pH, P_CO2 and [lactate] in emperor penguins at rest were characteristic of other avian species(Powell, 2000) and are consistent with a lack of stress during the sampling procedure. Venous P_O2, O₂ content, pH, P_CO2 and [lactate] were consistent with the observed arterial values and resulted in an estimated 70% venous Hb saturation.

Arterial P_O2 profiles reflected the effects of compression hyperoxia and O₂-store depletion previously documented in the air sacs during dives(Stockard et al., 2005). For example, during a 5.3-min, 60-m deep dive, arterial P_O2 increased from an initial value of 98 mmHg to a peak value of 257 mmHg and then gradually decreased to a final value of 76 mmHg near the end of the dive. Fig. 2 demonstrates the similarity of arterial and air sac P_O2 profiles in two different birds during shallow dives of short duration. Near the ends of some dives, the compression hyperoxia resulted in arterial P_O2 values that were greater than the mean value (68±7 mmHg) of birds at rest.

Venous P_O2 did not always simply decline during dives but sometimes increased (Fig. 2). This accounts for final dive values(Fig. 4A) that are greater than those of birds at rest. In addition, there was a wide range of pre-dive venous P_O2 levels (Figs 2 and 3). Prior to a 23.1 min dive(Fig. 3), the blood O₂ store was optimized with an elevated pre-dive venous P_O2 of 63 mmHg. This value was not only greater than that of birds at rest but was nearly equivalent to arterial values of birds at rest. These findings support the concept that the blood O₂store of emperor penguins can be enhanced by `arterialization' of venous blood.

Although final venous P_O2 declined in relation to dive duration, the relationship was variable; e.g. at a dive duration of 5.6 min (the ADL), final venous P_O2values spanned a range of 40 mmHg. These final venous P_O2 data and the previously published final air sac P_O2 data provide evidence that the total body O₂ store is not depleted at the ADL. In fact, the body O₂ store is still not depleted even after many minutes beyond the ADL (Fig. 4). The wide range of final values for a given dive duration was consistent with variations in the rates of decline of P_O2 in the individual venous profiles and, presumably, was related to differences in rates of O₂ consumption during dives. The minimum rate at which final venous P_O2 declined in relation to dive duration of emperor penguins at the isolated dive hole is described by the exponential regression in Fig. 4A.

We also propose that venous P_O2 profiles and end-of-dive values, especially during the latter portions of long dives,approximate arterial values. Comparison of final air sac(Stockard et al., 2005) and venous P_O2 values from dives of emperor penguins reveals that final air sac values become indistinguishable from final venous values during longer dives (Fig. 4B). This is particularly apparent after dives beyond the ADL(Fig. 4B). Arterial P_O2 data, available over a limited range of relatively short dive durations, occupy the same range as air sac values and,in some cases, overlap venous values. Given the similar distributions of arterial, air sac and even some venous final P_O2 values for short dive durations, and the assumption that air sac P_O2 represents maximal arterial P_O2, we think that venous final P_O2 values approximate arterial final P_O2 values for long dives. Similar equilibrations of venous and arterial P_O2values have also been reported in seals(Elsner et al., 1964; Stockard et al., 2007). Thus,as indices of the entire blood O₂ store, the venous P_O2 profiles from the longer dives in this study demonstrate that emperor penguins clearly push the limits of hypoxemia and, indeed, are capable of `returning on empty' to the dive hole. In 29% of dives, final venous P_O2 values were less than 20 mmHg; in some dives, P_O2 reached values as low as 1–6 mmHg. These final P_O2 values are well below the arterial and venous thresholds (20–25 mmHg) for cardiovascular collapse in pekin ducks(Hudson and Jones, 1986) and are also less than arterial P_O2 (22 mmHg) in bar-headed geese at 11 580 m altitude(Black and Tenney, 1980). In the 23 min dive of an emperor penguin, P_O2 was less than 20 mmHg for 8 min and eventually reached 6 mmHg(Fig. 3).

One might question whether the extremely low final venous P_O2 data could be secondary to O₂consumption by the electrode in blood made stagnant by the bradycardia and low cardiac output of diving. We think this is unlikely for several reasons. First, low P_O2 values also occurred in the air sacs, which should not be affected by stagnant blood. Second, given the stroke volume and blood volume of emperor penguins(Kooyman et al., 1992; Ponganis et al., 1997a), even if heart rate were 5 beats min^–1 during the last 10 min of the 23 min dive, the entire blood volume would circulate during that time period. Presumably, there would still be some flow past the electrode in that situation. Third, low final venous P_O2 values also occurred during short dives, which have higher heart rates(Kooyman et al., 1992) that should be associated with higher blood flows. Fourth, during the initial post-dive portion of the surface interval, venous P_O2 often stayed the same(Fig. 3) or even decreased further (P.J.P., unpublished data). This lack of an immediate increase in venous P_O2 during the tachycardia(Kooyman et al., 1992) and presumed high blood flows of the initial surface period again support our argument that the low values during dives are not secondary to the localized depletion of O₂ in a stagnant layer of blood around the electrode. Fifth, in the test-tube evaluation of the potential effect of the P_O2 electrode itself on O₂ depletion in saline, there was only a minimal initial decline in P_O2 over 4 min, and then no change thereafter. This change in P_O2 in saline should be the maximum potential effect of the electrode since localized O₂depletion in blood would be buffered by release of O₂ from Hb. Therefore, we expect that the localized depletion of O₂ by the electrode in blood would be even less and that the low venous P_O2 values in this study are not secondary to O₂ consumption by the electrode.

These low P_O2 values in the blood and respiratory systems of diving emperor penguins are also remarkable in comparison to mammalian indices of hypoxemia, including (a) the typical arterial P_O2 criterion of 60 mmHg for treatment of human patients (Nunn,1977), (b) end-tidal P_O2 values of 35 mmHg from climbers on ambient air at the top of Mount Everest(West et al., 1983), (c) human thresholds for shallow-water blackout near 25 mmHg(Ferrigno and Lundgren, 1999),(d) mixed venous P_O2 values of 27–34 mmHg in terrestrial mammals exercising at maximal O₂ consumption(Taylor et al., 1987), (e)femoral venous P_O2 values of 20 mmHg in humans exercising at maximal O₂ consumption(Roca et al., 1992) and (f)arterial and end-tidal P_O2 values of 15–20 mmHg in free-diving Weddell seals (Leptonychotes weddellii) and bottlenose dolphins (Tursiops truncatus)(Ponganis et al., 1993; Qvist et al., 1986; Ridgway et al., 1969). The only P_O2 values equivalent to the extremes of hypoxemia found in these free-diving emperor penguins are the arterial and venous P_O2 levels (10 and 3 mmHg, respectively)found in harbor seals (Phoca vitulina) forcibly submerged to an electroencephalographic threshold for hypoxemic brain damage(Kerem and Elsner, 1973).

These findings of extreme hypoxemia in emperor penguins also suggest that,in contrast to the Hb of the pekin duck or pigeon(Hudson and Jones, 1986; Weinstein et al., 1985), the Hb of the emperor penguin is not stripped of all its O₂ at a P_O2 of 20 mmHg. In other words, the P₅₀ (P_O2 at 50%O₂ saturation of Hb) of whole blood in emperor penguins is probably much lower than the P₅₀ of pekin ducks (42–52 mmHg)(Black and Tenney, 1980; Lutz, 1980; Powell, 2000) and perhaps even lower than the P₅₀ of isolated, reconstituted emperor penguin Hb (36 mmHg) (Tamburrini et al.,1994). Rather, it is probably closer to the lowest whole-blood P₅₀ values (30–34 mmHg) found in Adelie, gentoo and chinstrap penguins (Pygoscelis adeliae, P. papua, P. antarctica) and in high-altitude-adapted birds such as the bar-headed goose(Black and Tenney, 1980; Milsom et al., 1973; Petschow et al., 1977). This suggestion of a relatively low P₅₀ in the emperor penguin is supported by the blood gas and O₂ content analyses of these birds at rest. At a mean venous P_O2 of 41 mmHg,mean O₂ content was 17.4 ml O₂ dl^–1,which represents approximately 70% saturation of an average Hb concentration of 18 g dl^–1 (Kooyman and Ponganis, 1998).

In comparison to the P₅₀ of the pekin duck, a lower P₅₀ in the emperor penguin would not only increase blood O₂ content during hypoxemia but it would also enhance dive capacity by allowing more complete depletion of the respiratory O₂ store. In the pekin duck forcibly submerged to the point of `imminent cardiovascular collapse', 25% of the respiratory O₂ store was still unused because the blood contained almost no O₂ at an air sac P_O2 near 30 mmHg(Hudson and Jones, 1986).

In conclusion, intravascular/air sac P_O2profiles in diving emperor penguins have revealed that their dive capacity is at least partially achieved through optimum management of the blood/respiratory O₂ stores and extreme hypoxemic tolerance. The blood P_O2 profiles provide insight into the nature and magnitude of physiological responses during the dive as well as into biochemical/molecular mechanisms underlying hypoxemic tolerance. In particular, a Hb with high O₂ affinity (low P₅₀) in penguins is essential not only to enhance blood O₂ content during hypoxemia but also to allow depletion of the respiratory O₂ store, which in emperor penguins constitutes 19% of the total body O₂ store (Kooyman and Ponganis, 1998). Other mechanisms of such extreme hypoxemic tolerance may include increased capillary densities, modifications in reactive O₂ species production and/or scavenging, and changes in the concentration and function of neuroglobin and cytoglobin. In addition to their relevance to the diving capacity and biology of emperor penguins, these potential cellular adaptations may also serve as models for improved understanding and treatment of human hypoxemic/ischemic pathologies.

Supported by NSF grants OPP 9814794 and 0229638 to P.J.P. J.U.M. was supported by a NDSEG fellowship. We thank J. Heil, K. Sato, E. Stockard, Y. Habara, J. St. Leger, T. Zenteno-Savin, and D. H. and C. Levenson for assistance in the field, H. Hanish and M. Loughry of UFI for design and construction of the P_O2 recorders, the SIO Machine Shop for underwater housings, and G. L. Kooyman for his support and manuscript review.