High-affinity hemoglobin and blood oxygen saturation in diving emperor penguins

Extreme environments, particularly those with limited oxygen(O₂) availability, are especially challenging to most life forms,yet still host diverse and abundant life. Species that flourish in such environments are ideal models in which to examine the physiological, cellular and biochemical mechanisms underlying tolerance to hypoxia and O₂depletion. The emperor penguin (Aptenodytes forsteri), the consummate avian diver, thrives in the extreme Antarctic underwater environment, diving to depths greater than 500 m (Kooyman and Kooyman, 1995; Wienecke et al., 2007) and for durations longer than 23 min(Ponganis et al., 2007). Recent measurements of air sac and blood partial pressure of O₂(P_O2) in diving emperor penguins have revealed exceptional tolerance to low O₂ in this species(Ponganis et al., 2007; Stockard et al., 2005). It has been hypothesized that this underlying tolerance, well below the limits of many birds and mammals, may necessitate biochemical and molecular adaptations including a shift in the emperor penguin O₂–hemoglobin (Hb)dissociation curve, relative to other birds.

The O₂–Hb dissociation curve of whole blood of emperor penguins has not been determined. In general, the hemoglobin of birds has lower O₂ affinity than that of mammals. This may reflect a shift toward favoring O₂ unloading to the tissues, as the avian respiratory system is inherently more efficient at O₂ uptake(Piiper and Scheid, 1975; Powell, 2000; Powell et al., 1989). Consequently, the P₅₀ (P_O2at which Hb is 50% saturated) values of most birds (44–52 mmHg) are much higher than those of mammals (25–31 mmHg)(Christensen and Dill, 1935; Hirsowitz et al., 1977; Lenfant et al., 1970; Lutz, 1980; Wastl and Leiner, 1931). However, Adélie (Pygoscelis adeliae), chinstrap (P. antarctica), and gentoo penguins (P. papua), and the bar-headed goose (Anser indicus), a bird adapted to life at high altitudes, have P₅₀ values in the mammalian range(Black and Tenney, 1980; Lenfant et al., 1969b; Milsom et al., 1973; Petschow et al., 1977),favoring O₂ uptake from the lungs when P_O2 is low. The P₅₀ for isolated emperor penguin Hb (36 mmHg) has been measured(Tamburrini et al., 1994). Determination of the P₅₀ and dissociation curve in whole blood for this species remains necessary, however, as the P₅₀ of isolated Hb of this species was highly sensitive to the presence of various cofactors(Tamburrini et al., 1994).

We characterized the O₂–Hb dissociation curve of the emperor penguin in whole blood in order to (1) investigate the biochemical adaptation of Hb in emperor penguins, and (2) address blood O₂depletion during diving, by applying the dissociation curve to previously collected P_O2 profiles(Ponganis et al., 2009; Ponganis et al., 2007) to estimate in vivo Hb saturation (S_O2)changes during dives. It was hypothesized that the P₅₀ of the emperor penguin would be similar to that of the bar-headed goose and other penguin species, an adaptation which could contribute to tolerance to low O₂ in the emperor penguin (Black and Tenney, 1980; Lenfant et al., 1969b; Milsom et al.,1973; Ponganis et al.,2007). Hemoglobin with O₂ affinity in this range would also be consistent with blood P_O2 and O₂ contents of emperor penguins at rest(Ponganis et al., 2007). The magnitude of the Bohr effect was expected to be similar to those in other birds and penguins (–0.4 to –0.6)(Lenfant et al., 1969b). A similar range of Bohr effects, which should favor O₂ unloading to the tissues, has also been observed in a variety of marine mammals(Lenfant et al., 1970; Qvist et al., 1981; Willford et al., 1990). An accurate determination of the Bohr effect in emperor penguin whole blood also allows refinement of the estimation of S_O2 from P_O2 profiles since P_CO2 and pH data are available at various points in dives of this species(Ponganis et al., 2009; Ponganis et al., 2007).

The extent and rate of respiratory O₂ store depletion have been previously determined for the emperor penguin(Stockard et al., 2005). The addition of blood O₂ store depletion data provides the next component of the investigation of total O₂ consumption in this exceptional diver. It was hypothesized that S_O2profiles would reflect the wide range of P_O2values at the end of dives (Ponganis et al., 2009; Ponganis et al.,2007) and that the rate of decrease of blood O₂ would vary inversely with dive duration since heart rate in this species progressively decreases during longer dives(Meir et al., 2008). An inverse relationship to dive duration has been demonstrated in the respiratory O₂ store depletion rate of the diving emperor penguin(Stockard et al., 2005) and in blood O₂ store depletion rates of seals(Elsner et al., 1964; Stockard et al., 2007).

As in past studies (Kooyman et al.,1992; Ponganis et al.,2001), non-breeding emperor penguins (Aptenodytes forsteri Gray; ∼15 birds/season, 20–30 kg) were captured near the McMurdo Sound ice edge or at Terra Nova Bay in October 2003–2005,2007 and 2008 and were maintained for 6 weeks at an isolated dive hole enclosed within a corral at the Penguin Ranch on the McMurdo Sound sea ice(77°41′, 165°59′). All procedures were approved under a UCSD Animal Subjects Committee protocol and US Antarctic Treaty Permit. All birds were returned to the ice edge and released upon completion of the study.

P_O2 electrodes (Licox C1.1 Revoxode; Integra LifeSciences, Plainsboro, NJ, USA) and thermistors (model 554, Yellow Springs Instruments, Yellow Springs, OH, USA) were inserted percutaneously into the aorta or vena cava of emperor penguins under general isoflurane anesthesia as described previously (Ponganis et al.,2007; Ponganis et al.,2001; Ponganis et al.,2004; Ponganis et al.,2003). An additional arterial data set was obtained in 2008 from a bird equipped with only an arterial P_O2electrode as described in the previous study(Ponganis et al., 2009). The electrodes were connected to a custom-built P_O2/temperature recorder (UFI, Morro Bay, CA,USA) and an Mk9 time–depth recorder (TDR) was also attached (Wildlife Computers, Redmond, WA, USA) as previously described(Stockard et al., 2005). After overnight recovery from anesthesia, birds were allowed to dive at the isolated dive hole. The instrumented birds dived for 1–2 days, after which catheters, probes and recorders were removed under general anesthesia. All P_O2 values were corrected to 38°C for construction of the P_O2 profiles, as previously described (Ponganis et al.,2007).

O₂–Hb dissociation curves on fresh whole blood were determined with the mixing technique of tonometered blood(Scheid and Meyer, 1978). Blood samples were obtained from the Penguin Ranch emperor penguins during anesthesia, placed on ice, and processed immediately. All dissociation curve analyses were completed within 6 h of blood collection in order to prevent depletion of labile organic phosphates such as inositol pentaphosphate (IPP)which would influence the dissociation curve, and to avoid prolonged metabolism of these nucleated avian red blood cells. The mixing technique consisted of the volumetric mixing of 0% O₂-saturated blood and 100% O₂-saturated blood to achieve the desired S_O2 at various points (i.e. 90, 70, 50, 40, 20,10, 5% S_O2) along the curve with subsequent measurement of the P_O2 of the resulting mixture using an i-STAT blood gas analyzer (37°C; Abbott Point of Care, Princeton,NJ, USA) (Black and Tenney,1980; Johansen et al.,1987; Nörgaard-Pedersen et al., 1972; Qvist et al.,1981). Use of the i-STAT analyzer also allowed verification of pH and P_CO2, and Tucker chamber analyses(Tucker, 1967) provided verification of blood O₂ content. The CO₂ Bohr effect was determined by changing the CO₂ concentration of the gas in the tonometer in order to adjust pH. Dissociation curves were determined at pH values of 7.5, 7.4, 7.3 and 7.2. The log[S_O2/(100–S_O2)]vs log(P_O2) was plotted and linear regression analysis performed in order to generate the equation for the O₂–Hb dissociation curve at each pH(Nicol, 1991). As in similar studies (Nicol, 1991), all S_O2 points from all birds were combined in order to generate a general O₂–Hb dissociation curve for this species. This is justified as, with the exception of disease, the structure of hemoglobin varies among species, not individuals of a species. The Bohr coefficient was derived from linear regression of the log P₅₀ on pH (each point averaged from all data of all birds for pH 7.5, 7.4, 7.3 and 7.2) (Nicol,1991; Willford et al.,1990). In addition, the fixed acid Bohr effect was determined by titrating with HCl to pH 7.4, 7.3 and 7.2 while P_CO2 was maintained at the level required for the standard pH 7.5 curve in that penguin(Willford et al., 1990). The effect of lactic acid on the dissociation curve was also evaluated in whole blood by determining additional dissociation curve points after adding lactic acid (L6661, Sigma, St Louis, MO, USA) to the blood to a final concentration of 5 and 10 mmol l^–1(Nicol, 1991).

In order to validate the specific equipment and methods used in this study,dissociation curves were also determined with blood from an avian (chicken, Gallus gallus domesticus) and pinniped (Leptonychotes weddelli,Phoca vitulina, Zalophus californianus) species with previously published O₂–Hb binding data(Bartels et al., 1966; Hirsowitz et al., 1977; Lenfant, 1969; Lenfant et al., 1969a). Mixing technique tests were also conducted to verify that no hemolysis occurred during mixing.

Percent hemoglobin O₂ saturation(S_O2) values were obtained by applying data from P_O2 profiles [from previous studies(Ponganis et al., 2009; Ponganis et al., 2007) and current studies] to the linear regression equation generated by the log[S_O2/(100–S_O2)]vs log(P_O2) plot and solving for S_O2 (at the appropriate pH, see below).

Blood samples taken for the O₂–Hb dissociation curve were also used for hemoglobin analyses [cyanomethemoglobin technique(Ponganis et al., 1993)] in order to obtain Hb concentration. Oxygen content for initial and final(end-of-dive) time points for each dive were calculated from the corresponding S_O2 values, using a Hb concentration of 18.3 g dl^–1 ([this study, agrees well with previous findings(Kooyman and Ponganis, 1998; Ponganis et al., 1997a)] with the formula: O₂ content (ml O₂ dl^–1blood)= (1.34 ml O₂ g^–1 Hb)×[Hb](g dl^–1)×S_O2+(0.003×P_O2). Initial S_O2 (S_O2at the start of the dive) was estimated with the pH 7.5 dissociation curve,and the final S_O2 value (from the final P_O2 value, measured within the last 5 or 15 s of the dive depending on the specific recorder) was estimated with the pH 7.4 dissociation curve. The effect of pH is critical in estimation of final S_O2. For example, from data of Weddell seals(Qvist et al., 1981), a P_O2 of 10 mmHg corresponds to 8%S_O2 at pH 7.4, but only 2% at pH 7.0. Blood lactate concentration does not increase significantly during dives of emperor penguins, even during dives longer in duration than the previously measured aerobic dive limit [ADL; duration beyond which blood lactate concentration increases above resting levels (Kooyman et al., 1983); 5.6 min for the emperor penguin(Ponganis et al., 1997b)] for this species (Ponganis et al.,2009). The rise in blood lactate does not occur until the post-dive period in this species (Ponganis et al., 1997b; Ponganis et al., 2009). Based on lactate, P_CO2 and pH measurements obtained from emperor penguin blood samples drawn at various time points during diving (Ponganis et al.,2009), a pH of 7.4 was deemed most representative of conditions at the end of the dive.

The percentage of net O₂ content depletion in either the arterial or venous system (dependent on the site of insertion) for each dive of each penguin was calculated [% O₂ content depletion=(initial O₂ content–final O₂ content)/initial O₂content×100]. The rate of O₂ content depletion [(initial O₂ content–final O₂ content)/dive duration] was also calculated for the arterial and venous systems. This provides an overall depletion rate, not an instantaneous measure of O₂ consumption. Data from penguins with arterial P_O2 electrodes provided an estimate of the net depletion of O₂ in the arterial system; those with venous probes provided estimates of net venous O₂ depletion.

As outlined above, the dissociation curve, calculated using linear regression, was applied to P_O2 values to yield S_O2 values, and calculations of O₂depletion rate and percentage of O₂ depletion were made. Differences between arterial and venous results were determined with one-way ANOVA. Correlation between dive duration and the variables of final S_O2, pre-dive S_O2, percentage O₂ content depleted and depletion rate was addressed with Spearman rank order correlation tests. Statistical significance was assumed at P<0.05 and the significance level is quoted in the text. Values are expressed as means± s.d. P_O2 values are expressed in mmHg(as measured). Figures include the corresponding values in kPa, assuming 1 mmHg=0.133 kPa.

Dive behavior of emperor penguins at the Penguin Ranch for the dives associated with the P_O2 profiles in this study has been described previously (Ponganis et al., 2009; Ponganis et al.,2007). Dive duration and depth data are given in Table 1.

Complete O₂–Hb dissociation curves were determined with blood from 12 penguins, with additional P₅₀ values at specific pH levels and fixed acid Bohr effects obtained in an additional two penguins. The P₅₀ was 28±1 mmHg (N=12 penguins) at pH 7.5 (Fig. 1A). The CO₂ Bohr effect (slope of log P₅₀vs pH) was –0.45 (y=–0.45x+4.81, r²=0.98, P=0.008). The fixed acid Bohr effect(addition of HCl or lactic acid) was not significantly different from that of CO₂.

The resulting regression equations from the plots of log[S_O2/(100–S_O2)]vs log(P_O2) (all saturation points,all penguins combined) were:

• pH 7.5:log[S_O2/(100–S_O2)]= 2.92589 × log(P_O2) – 4.24338(N=43, r²=0.98, P<0.0001),

• pH 7.4:log[S_O2/(100–S_O2)]= 2.94767 × log(P_O2) – 4.39858(N=70, r²=0.98, P<0.0001),

• pH 7.3:log[S_O2/(100–S_O2)]= 3.04945 × log(P_O2) – 4.72019(N=38, r²=0.99, P<0.0001),

• pH 7.2:log[S_O2/(100–S_O2)]= 3.15958 × log(P_O2) – 4.97618(N=9, r²=0.99, P<0.0001).

Samples from other species (chicken and pinnipeds) run for dissociation curve method validation agreed with previously published values(Bartels et al., 1966; Hirsowitz et al., 1977; Lenfant, 1969; Lenfant et al., 1969a). Mixing technique verification tests confirmed that no hemolysis occurred while mixing blood samples in the syringes, based on clear plasma color in hematocrit tubes after centrifugation. The mean hemoglobin concentration in this study(18.3±1.1 g dl^–1, N=17) was equivalent to that of previous studies (Kooyman and Ponganis, 1998; Ponganis et al., 1997a). Oxygen content analyses performed with 100%O₂-saturated blood (Tucker chamber) agreed within 5% of the maximal O₂ content calculated using the equation in this study[O₂ content (ml O₂ dl^–1 blood)=(1.34 ml O₂ g^–1 Hb)×[Hb](g dl^–1)×S_O2+(0.003×P_O2)].

In addition to the arterial P_O2 data collected in previous studies (Ponganis et al., 2009; Ponganis et al.,2007), an arterial P_O2 and dive profile was obtained for one more emperor penguin in the 2008 Antarctic field season (Fig. 2A). This increased the arterial data sample size to six penguins (71 dives in total). The venous data were from nine penguins (130 total dives). These profiles have been adequately described in previous studies(Ponganis et al., 2009; Ponganis et al., 2007).

In vivo arterial Hb saturation (S_a,O2) often remained near 100% for much of the dive, reflecting the large increase in arterial P_O2 at the beginning of the dive and the fact that despite a subsequent decrease, P_O2 remained relatively high until the later dive phase (Fig. 2A)(Ponganis et al., 2007). A rapid decline in S_a,O2 generally began during the final ascent phase of the dive (Fig. 2A). S_v,O2 (venous Hb saturation)profiles reflected venous P_O2 profiles(Ponganis et al., 2007) in that they were quite variable among dives(Fig. 2B, Fig. 3), with marked fluctuations, transient increases during the dive, and a large range of final S_v,O2 values(Fig. 4). Re-oxygenation occurred rapidly upon return to the surface, as demonstrated by the return in both S_a,O2 and S_v,O2 to values consistent with those at rest within approximately 2 min post-dive(Fig. 2). Pre-dive and initial S_v,O2 values(Table 1, Fig. 2B, Fig. 3) were often higher than those corresponding to the mean venous P_O2 of emperor penguins at rest (Ponganis et al.,2007).

Final S_v,O2 was ≤20% in 28% of dives,≤5% in 15% of dives and ≤2% in 6% of dives. With the exception of only one dive, S_v,O2 decreased below 20% only in dives greater than the previously measured ADL(Ponganis et al., 1997b)(Table 1, Fig. 4).

Max S_O2 (the peak S_O2 during the dive), pre-dive S_O2, initial S_O2, final S_O2,Δ S_O2 (initial S_O2 minus final S_O2), percentage O₂ content depletion, and blood O₂ store depletion rate data are given in Table 1. Max S_O2, initial S_O2, final S_O2,Δ S_O2 and percentage O₂ content depletion were all significantly different between the arterial and venous compartments [one-way ANOVA: P≤0.001 for all exceptΔ S_O2 (P=0.035); F=144.40, 180.42, 59.14, 4.53, 10.93, respectively]. The blood O₂ store depletion rates between the two compartments, however,were not significantly different (one-way ANOVA: P=0.94; F=0.00627).

Both final S_a,O2 and S_v,O2 demonstrated a strong and significant negative correlation to dive duration (Spearman rank order correlation, Table 2, Fig. 4). The percentage of O₂ content depleted during the dive showed a strong and significant positive correlation to dive duration for both compartments (Spearman rank order correlation, Table 2). The pre-dive S_v,O2 had a significant, but weak positive correlation to dive duration, whereas pre-dive S_a,O2 was not significantly related to dive duration(Spearman rank order correlation, Table 2). Blood O₂ store depletion rate had a significant,but weak, positive relationship to dive duration in both the arterial and venous compartments (Spearman rank order correlation, Table 2).

As hypothesized because of its potential to contribute to tolerance to low O₂ in this species (Ponganis et al., 2007), the O₂–Hb dissociation curve of the emperor penguin is indeed left-shifted relative to most birds, similar to that of other penguin species and the high-flying bar-headed goose(Black and Tenney, 1980; Lenfant et al., 1969b; Milsom et al., 1973; Petschow et al., 1977). Left-shifted hemoglobin (higher Hb–O₂ affinity) is advantageous for the diving emperor penguin, as it implies that more O₂ is available at any given P_O2. This becomes particularly relevant for a breath hold diver that often experiences low P_O2 values during diving. For example, at a P_O2 of 20 mmHg, a pekin duck would be stripped of all its O₂, whereas the hemoglobin of the emperor penguin, with its left-shifted dissociation curve, is still 27%saturated (pH 7.5; 21.5% at pH 7.4) at this same P_O2 (Fig. 1B). This advantage could also assist in the prevention of deleterious effects such as shallow water blackout if P_a,O2 drops to low levels during diving.

An increased Hb–O₂ affinity also allows for more complete depletion of the respiratory O₂ store. As opposed to marine mammals whose lungs do not constitute a significant portion of total O₂stores, the respiratory store of the diving emperor penguin remains a significant contributor to O₂ stores while diving [19% of total O₂ stores (Kooyman and Ponganis, 1998)], and this species dives upon inspiration(Kooyman et al., 1971). The increased O₂ affinity of emperor penguin Hb allows for continued extraction of this O₂ from the respiratory O₂ store for use during diving. By contrast, a pekin duck forcibly submerged to the point of `imminent cardiovascular collapse' was left with 25% of its respiratory O₂ store unused, as the blood contained no O₂ at an air sac P_O2 near 30 mmHg(Hudson and Jones, 1986).

Presumably, the biochemical adaptation underlying the increased Hb–O₂ affinity is achieved with specific amino acid substitution(s), similar to that of other species. In two species of high-altitude birds with increased Hb–O₂ affinity (bar-headed and Andean geese), the same α₁β₁ intradimer contact site is disrupted, accomplished by two different amino acid substitutions (on the α-chain in one of the species, and on the corresponding β-chain site in the other)(Perutz, 1983; Weber and Fago, 2004). This same intradimer contact site is altered in high O₂ affinity fish Hb, and even reconstituted human Hb demonstrates significantly higher affinity for O₂ when this single site is altered(Jessen et al., 1991; Weber and Fago, 2004). Upon examination of the published amino acid sequence of emperor penguin hemoglobin(Tamburrini et al., 1994), it is revealed that this specific site is not altered in emperor or rockhopper(Eudyptes crestatus) penguins. However, as compared to human Hb,there are changes in six α-chain and five β-chainα ₁β₁ contact sites, and one α-chainα ₁β₂ contact site. It is possible that one of these modifications accounts for the emperor penguin's increased Hb–O₂ affinity, although other structural features might also be responsible.

We hypothesize that the small Bohr effect (–0.25) found in the previous study of isolated emperor penguin Hb may be secondary to experimental conditions and the sensitivity of the isolated Hb to co-factors(Tamburrini et al., 1994). It was hypothesized in that study that the low Bohr effect prevents `stripping'of O₂ during the long dives of emperor penguins(Tamburrini et al., 1994). However, such a low Bohr effect for Hb in whole blood was not found in this study, nor is it present in other diving penguins and marine mammals(Lenfant et al., 1970; Lenfant et al., 1969b). A small Bohr effect is a potential disadvantage in terms of unloading O₂ to the tissues in a diving animal, especially in those with relatively high Hb–O₂ affinities. Blood pH data during dives have also demonstrated that emperor penguin blood does not become very acidotic during the dive, indicating that the effects of the Bohr shift may not be particularly relevant during diving in this species(Ponganis et al., 2007).

Final S_v,O2 reached very low levels, ≤5%in 15% of dives (corresponding to O₂ contents <1.2 ml O₂ dl^–1 blood) and approached 0% in some dives,particularly in those with durations longer than the previously measured ADL(Ponganis et al., 1997b)(Table 1, Fig. 2B, Figs 3 and 4). This resulted in a large percentage venous O₂ content depletion during the dive(Table 1). The wide range of final S_v,O2 values for dives of similar duration (Fig. 4) and the fluctuations in the venous P_O2 and S_O2 profiles during dives(Fig. 3) suggest significant variability in tissue O₂ extraction and blood flow patterns during dives. As previously suggested (Ponganis et al., 2009), this may reflect differences or changes in the peripheral vascular response including regulation of blood flow to muscle and other organs as well as through arterio-venous (a-v) shunts.

Final S_a,O2 reached only as low as 47%(Table 1, Fig. 4; corresponding to an O₂ content of 11.5 ml O₂ dl^–1 blood) in dives as long as 12 min. Such high values contrast with the lower values found in elephant seals (Mirounga angustirostris)(Meir et al., 2009) and,again, are considered secondary to the size of the respiratory O₂store in emperor penguins and to the maintenance and efficiency of respiratory gas exchange during these dives. Such high S_a,O2 may well minimize the risk of shallow water blackout in these birds(Fig. 2A). It should be noted that many of these final S_a,O2 values are greater than would have been predicted from final air sac P_O2 values for dives of similar duration(Ponganis et al., 2009; Ponganis et al., 2007). It is unclear why final P_a,O2 values are higher than those recorded in the air sacs, especially since air sac P_O2 prior to dives and during dives is greater than P_a,O2 (Ponganis et al., 2009). Sample size, type of dive, circulatory lag time,and the location of the sampled air sac may contribute to these differences. Regardless of the explanation, it is apparent that emperor penguins can perform dives as long as 12 min in which final arterial P_O2 and S_O2 are relatively well preserved.

It should also be noted that there is considerable overlap between final arterial and many final venous S_O2 values of the dives in this study (Fig. 4). Such overlap may be secondary to the previously suggested intra-dive a-v shunting (Ponganis et al.,2009) and the close equilibration of air sac, arterial and venous P_O2 in the latter parts of such dives. Equilibration of arterial and venous P_O2 and S_O2 values toward the end of long breath holds has been reported in seals (Elsner et al.,1964; Stockard et al.,2007). As the sample size for arterial records was smaller than that of venous records, and because there were no dives longer than 12 min for arterial data (compared with >23 min for venous data), we believe that final S_a,O2 values in longer dives may well be lower than those documented in this study.

Both S_a,O2 and S_v,O2returned to values equivalent to those at rest within about 2 min post-dive(Fig. 2), consistent with P_O2 profiles of this species(Ponganis et al., 2009; Ponganis et al., 2007). In the previous study, no relationship was shown between surface interval duration and the time to return to the P_O2 value at rest(Ponganis et al., 2009). Thus,re-oxygenation occurred very rapidly upon return to the surface, and does not represent a limitation to the onset of the subsequent dive.

As previously discussed (Ponganis et al., 2009), the initial transient increase in P_a,O2 during dives reflected the compression hyperoxia observed during diving in the air sacs of emperor penguins(Stockard et al., 2005) and demonstrated the maintenance of pulmonary gas exchange and transfer of O₂ from the respiratory to the blood O₂ store. Even as P_a,O2 declined from the early peak value, its value was relatively high for much of the dive (Fig. 2A). As a consequence, S_a,O2 remained near 100% for much of the dive (Fig. 2A), preserving a high O₂ content in the arterial system for critical organs such as the brain. S_a,O2generally did not decrease significantly until the final ascent phase of the dive, consistent with the decline in ambient pressure and decrease in both air sac and arterial P_O2 during ascent(Ponganis et al., 2007; Stockard et al., 2005)(Fig. 2A). The net transfer of O₂ from the lungs to the blood also resulted in some final P_a,O2 values greater than the P_a,O2 of birds at rest and some final P_v,O2 values not only greater than that at rest, but even greater than P_v,O2 at the start of the dive (Ponganis et al.,2007). The corresponding S_O2 values resulted in negative values for ΔS_O2,percentage O₂ content depleted and blood O₂ depletion rates (i.e. an overall increase in S_O2 during the dive) in such dives (Table 1, Fig. 5). These findings are consistent with the significant role of the left-shifted dissociation curve and with the contribution of respiratory O₂ to total O₂ stores in this species.

By contrast, although there was also a transient increase in P_a,O2 and consequently S_a,O2 in diving elephant seals at the beginning of dives, max P_a,O2 in the emperor penguin was nearly twice that in the elephant seal [mean max P_a,O2=88±20 mmHg for the elephant seal(Meir et al., 2009) and 157±52 mmHg in the emperor penguin]. As in the emperor penguin, S_a,O2 peaked near 100% in the elephant seal, but this high level of S_a,O2 was not maintained and instead decreased rapidly after the peak (Meir et al.,2009). These findings reflect the difference in the magnitude of the respiratory O₂ store between these two species. Unlike emperor penguins, elephant seals dive upon expiration and the lungs do not constitute a significant portion of total O₂ stores(Falke et al., 1985; Kooyman and Ponganis,1998).

It should be noted that maximum S_O2 values in this study were probably slightly underestimated, because it was not possible to determine a true, critical P_O2value at 100% S_O2. In order to ensure complete saturation in the tonometry of the 100% O₂-saturated blood for O₂–Hb dissociation curve determination, a high P_O2 (>200 mmHg) was used for that sample. Thus, the actual inflection point in P_O2 at which S_O2 first reaches 100% was not determined(Fig. 1A). The lack of this data point in the regression equations results in a slightly lower calculated S_O2 for P_O2values that approach the inflection point of 100%S_O2. In turn, this implies that net depletion values and O₂ depletion rates are slightly underestimated.

Pre-dive and initial S_v,O2 values were often higher than those corresponding to P_v,O2 of emperor penguins at rest (Ponganis et al.,2007), reaching as high as 95%(Table 1, Fig. 2B, Fig. 3). Although maximum S_v,O2 during a dive was much more variable than S_a,O2 (Table 1), the mean of maximum S_v,O2 was also quite high at 81%, and, in some dives, max S_v,O2 values were almost 100%(Table 1, Fig. 2B, Fig. 3). These `arterialized'venous values imply some degree of arterio-venous (a-v) shunting or at least a lack of tissue O₂ uptake from the blood. In addition to previous findings in these species, including increases in P_v,O2during the descent period, the lack of blood lactate accumulation during the dive, muscle temperature profiles, dramatic bradycardias and a lack of association between heart rate and stroke frequency during diving, these results further support the concept that muscle is isolated from the circulation during dives of emperor penguins(Meir et al., 2008; Ponganis et al., 2009; Ponganis et al., 2007; Ponganis et al., 2003).

Such high values during surface intervals may be secondary to an increased gas exchange rate, enhanced reloading of O₂ stores and improved ventilation perfusion matching due to a more pronounced hyperventilation and tachycardia before such dives. In addition to the maximization of the blood O₂ store by the arterialization of its venous blood, these birds often made full use of this component of their O₂ store by decreasing S_v,O2 to near 0% in longer dives, as discussed above (Table 1, Fig. 2B, Fig. 3). For example, in the longest recorded dive to date for an emperor penguin (23.1 min), pre-dive S_O2 was 91–95%, initial S_O2 was 91%, and final S_O2 decreased to as low as 1%, demonstrating a near complete utilization of the venous blood O₂ store in this remarkable dive (Fig. 2B).

Based on pre-dive S_O2 values, the magnitude of a peripheral a–v shunt prior to the dive can be estimated with a shunt equation, as follows, assuming a 5 ml O₂dl^–1 a-v O₂ difference at rest(Ponganis et al., 2007):{venous O₂ content=[arterial O₂ content×(% a-v shunt)]–[(arterial O₂ content–5 ml O₂dl^–1)×(1–(% a-v shunt))]}. For example, the maximum pre-dive S_v,O2 in this study was 95.8%(Table 1). Using O₂contents calculated for this level of S_v,O2(23.5 ml O₂ dl^–1) and the maximal pre-dive S_a,O2 (24.4 ml O₂ dl^–1) an arterio-venous shunt of 82% would be necessary to reach this venous SO₂. For the pre-dive S_v,O2of 91.3% (22.4 ml O₂ dl^–1) prior to the 23.1 min dive (Fig. 2B), the magnitude of the a-v shunt would be 60%. These calculations should be considered slight overestimates since the S_v,O2 values are vena caval and not true mixed venous samples (myocardial O₂ extraction would decrease the mixed venous S_v,O2further).

As discussed in previous studies, the mean P_a,O2 of emperor penguins at rest (68±7 mmHg) is less than two-thirds of the mean value in the air sac (Ponganis et al., 2009; Ponganis et al.,2007), probably mainly because of ventilation–perfusion mismatch (Powell, 2000; Powell et al., 1989). Using the O₂–Hb dissociation curve and the classic pulmonary shunt equation [% shunt=(capillary O₂ content–arterial O₂ content)/(capillary O₂ content–venous O₂ content)], calculations can be made to estimate the size of the pulmonary shunt in the emperor penguin both at rest and in the pre-dive period. Calculating S_O2 from the previously measured mean P_a,O2 and P_v,O2 values at rest (Ponganis et al.,2007) and assuming capillary S_O2=100%, based on air sac P_O2 measurements of 120 mmHg from previous studies (Stockard et al.,2005), and a [Hb]=18.3 g dl^–1 (this study) to calculate O₂ content, the extent of the intrapulmonary shunt in a resting emperor penguin is 28%. Again, this value may be slightly overestimated since mixed venous values were not used in the calculation. Nonetheless, shunting of pulmonary blood flow is usually very small in birds,less than 1–2.7% of cardiac output in anesthetized artificially ventilated geese and ducks [representing the true intrapulmonary shunt(Burger et al., 1979; Powell and Wagner, 1982)], or 6.3–8% in ducks [representing intrapulmonary plus extrapulmonary shunts(Bickler et al., 1986)]. The high percentage shunt value in emperor penguins may also be due in part to an overestimation of the capillary O₂ content because capillary P_O2 and saturation may be decreased secondarily to the thickened parabronchial capillary blood-to-air barrier reported in emperor penguins (Welsch and Aschauer,1986). Using pre-dive S_a,O2 and S_v,O2 values (mean values for pre-dive S_v,O2 and S_a,O2 in this study; Table 1) instead of values at rest, however, the intrapulmonary shunt is reduced to 14.3%. This suggests that the hyperventilation and tachycardia characteristic of the pre-dive period in this species (Kooyman et al., 1971; Kooyman et al.,1992; Meir et al.,2008) improves ventilation–perfusion matching prior to the dive.

Since the blood volume of an emperor penguin of a given mass can be calculated (based on 100 mlkg^–1)(Kooyman and Ponganis, 1998; Ponganis et al., 1997a), a calculation of the contribution of the blood O₂ store to overall metabolic rate was made: blood O₂ store contribution=[(final O₂ content–initial O₂ content)/dive duration×blood volume], with the assumption that one-third of blood volume is arterial, and two-thirds venous in distribution(Kooyman, 1989). Using the mean values for all arterial and all venous dives combined, the blood O₂ store contribution alone to metabolic rate from the arterial compartment is 0.3±0.2 ml O₂ kg^–1min^–1 (range=–0.1–+0.7) and from the venous compartment is 0.6±0.9 ml O₂ kg^–1min^–1 (range=–4.1–+3.1). These values are only approximately 5% (arterial) and 10% (venous) of the resting metabolic rate measured for this species (Kooyman and Ponganis, 1994; LeMaho et al.,1976; Pinshow et al.,1977). If dives in which S_O2increased during the dive are excluded from the analysis, the blood O₂ store contribution alone to metabolic rate from the arterial compartment is 0.4±0.2 ml O₂ kg^–1min^–1 (range=0.0–+0.7) and from the venous compartment is 0.9±0.6 ml O₂ kg^–1min^–1 (range=0.1–+3.1). In comparison, respiratory O₂ store depletion rates were 2.3 and 5.3 times that in the venous and arterial blood O₂ store compartments, respectively[mean=2.1±0.8 ml O₂ kg^–1min^–1 (Stockard et al.,2005)].

As previously discussed, since S_O2 was sometimes higher at the end of the dive than at the initial time point,negative values resulted for ΔS_O2,percentage O₂ content depletion, and depletion rates(Table 1, Fig. 5). For example, the venous O₂ store contribution to metabolic rate was –1.8 ml O₂ kg^–1 min^–1 for a 1.6 min dive. Assuming a gain in the venous blood O₂ store of 1.8 ml O₂ kg^–1 min^–1 in this 1.6 min dive, this indicates a minimum loss of 2.9 ml O₂kg^–1 from the respiratory system during the dive. If the respiratory system initially has 20% O₂ at 69 ml kg^–1 (Ponganis et al.,1999), it contains approximately 14 ml O₂kg^–1 at the beginning of the dive. For the 1.6 min dive, a minimum of 21% (2.9/14) of the respiratory O₂ was transferred into the blood. This represents a minimum value, as it does not include respiratory O₂ that entered the blood and replaced any O₂ taken up by tissue from the blood. However, such calculations are useful to illustrate the maintenance of gas exchange during the dive and the potential magnitude of the transfer of respiratory O₂ into the blood.

These calculations illustrate the complexity of estimating the contribution of the net blood O₂ store to metabolic rate as the blood O₂ depletion rate in a penguin (or any diver with a large respiratory O₂ store) is a function of both the transfer of O₂ into the blood and the uptake of O₂ from blood by the tissues. Simultaneous air sac and blood P_O2data would allow calculation of the net contribution of these O₂stores to diving metabolic rate, but such measurements are not currently feasible. However, because the separately determined O₂contributions from the lungs and blood are both low, these data are consistent with (1) a significant contribution from the exceptionally large muscle O₂ store to diving metabolic rate in emperor penguins(Kooyman and Ponganis, 1998; Ponganis et al., 1997a), and(2) the low field metabolic rate and the true diving bradycardia (heart rate while diving significantly lower than that of heart rate at rest) exhibited by emperor penguins at the isolated dive hole(Meir et al., 2008; Nagy et al., 2001).

This investigation has revealed enhanced O₂ affinity of emperor penguin hemoglobin, similar to that of high-altitude geese and other penguin species. These species are distantly related (Sphenisciformes and Anseriformes) (Hackett et al.,2008) and have evolved in different environments under different constraints. The conditions of hypoxia that high-altitude birds and penguin species encounter are also different, ranging from that of extended hypoxia during sustained flight for the bar-headed goose, to transient, though perhaps frequent hypoxia experienced by diving penguins. Despite these differences,both bar-headed geese and emperor penguins have evolved hemoglobins with remarkably similar O₂ affinities. As neither of the two amino acid substitutions responsible for the increased Hb affinity in high-altitude birds is present in emperor penguin hemoglobin, the mechanisms underlying this shared trait are also different. Perhaps this level of O₂ affinity is closer to the physiological limit of maximizing O₂ uptake during hypoxia. As this is also the range of O₂ affinity optimized by mammalian hemoglobins, including those of diving mammals, it demonstrates the compromise between O₂ uptake from the respiratory system and O₂ unloading to the tissues intrinsic to the O₂–Hb dissociation curve.

S_O2 profiles during diving, based on the O₂–Hb dissociation curve, demonstrated (1) the maintenance of S_a,O2 levels near 100% throughout much of the dive, (2) a wide range of final S_v,O2 values and optimization of the venous blood O₂ store resulting from arterialization and near complete depletion of venous blood O₂during longer dives, and (3) that estimated contribution of the blood O₂ store to diving metabolic rate was low and highly variable. This pattern is due, in part, to the influx of O₂ from the lungs into the blood during diving, and variable rates of tissue O₂uptake.

LIST OF ABBREVIATIONS

 
• ADL

  aerobic dive limit

 
• Hb

  hemoglobin

 
• P_CO2

  partial pressure of carbon dioxide

 
• P_O2

  partial pressure of oxygen (P_a,O2 or P_v,O2: arterial or venous)

 
• P₅₀

  partial pressure of O₂ at which Hb is 50% saturated

 
• S_O2

  % hemoglobin saturation (S_a,O2 or S_v,O2: arterial or venous)