Abstract
The extinct Steller’s sea cow (Hydrodamalis gigas; †1768) was a whale-sized marine mammal that manifested profound morphological adjustments to exploit the harsh coastal climate of the North Pacific. Yet despite first-hand accounts of their biology, little is known regarding underlying physiological specializations to this extreme environment. Here, the adult-expressed hemoglobin (Hb) of this species is shown to harbor a fixed amino acid replacement at an otherwise invariant position (β/δ82Lys→Asn) that is predicted to profoundly alter multiple aspects of Hb function. To unravel the functional and evolutionary consequences of this substitution, we recombinantly synthesized Hb proteins of Steller’s sea cow, a H. gigas β/δ82Asn→Lys Hb mutant, the dugong (Dugong dugon), the last common dugongid ancestor, and the Florida manatee (Trichechus manatus). Our detailed functional analyses demonstrate that the Hb–O2 affinity of Steller’s sea cow evolved to become less affected by temperature compared to other sea cows. This phenotype presumably safeguarded O2 delivery to cool peripheral tissues and largely arises from a reduced intrinsic temperature sensitivity of the H. gigas protein owing to the β/δ82Lys→Asn substitution. We further confirm that this same exchange also underlies the secondary evolution of a reduced blood–O2 affinity phenotype that is moreover unresponsive to the intraerythrocytic allosteric effector 2,3-diphosphoglycerate. This radical modification, which augments O2 offloading by the protein, and presumably impacted maternal/fetal O2 exchange, is the first documented example of this phenotype among mammals. Notably, this replacement also increases protein solubility and is consistent with increased Hb concentrations within both the adult and pre-natal circulations that may have contributed to the elevated metabolic (thermoregulatory) requirements and fetal growth rates associated with their cold adaptation.
Introduction
During periods of submergence apnea when atmospheric gas exchange has ceased, the underwater foraging time by mammals is dictated by onboard oxygen stores and the efficiency of their use. Thus, evolutionary increases in oxygen stores, in the form of increased hemoglobin (Hb) and myoglobin—located within erythrocytes and skeletal/cardiac muscle, respectively—are nearly ubiquitous among aquatic species1. Notable exceptions to this rule are extant sirenians (sea cows), a group of strictly aquatic, (sub)tropical herbivores encompassing only four members, three species of manatee (family Trichechidae) and the dugong, Dugong dugon (family Dugongidae). While sirenians are proficient divers, they do not exhibit the strongly elevated body O2 stores or an enhanced dive reflex common to other lineages of marine mammals2,3. Rather, previous work revealed that the sirenian’s secondary transition to aquatic life coincided with a rapid evolution of their Hb encoding genes due, in part, to gene conversion events with a neighboring globin pseudogene4. The resulting high blood–O2 affinity phenotype presumably allows extant sea cows to maximize O2 extraction from the lungs during submergence while curbing the rate of O2 offloading, thus lowering overall metabolic intensity and fostering a prolonged breath-hold capacity4.
While the relatively weak thermoregulatory capacity of extant sea cows confine them to (sub)tropical waters5,6, fossil evidence and first-hand accounts of the sub-Arctic Steller’s sea cow (Hydrodamalis gigas) provide tantalizing insights into both the biological and morphological adaptations of this titanic sirenian to the harsh coastal conditions of the North Pacific, where they persisted from the Miocene until their demise in 17687–9. The retrieval of ancient genetic material from museum specimens has since been instrumental in clarifying the phylogenetic affinities and population history of this species, while providing additional details regarding the evolution of key morphological and physiological attributes4,10–15 (Fig. 1A). For example, characterization of “resurrected” Steller’s sea cow recombinant Hb demonstrated that the blood–O2 affinity of this lineage secondarily decreased following their divergence from dugongs in the mid Oligocene4. Although sirenians do not possess the capacity for non-shivering thermogenesis due to pseudogenization of the UCP1 gene11, this affinity shift in Steller’s sea cow Hb would have promoted increased O2 offloading that was speculated to fuel an increased metabolism to help cope with exposure to cold sub-Arctic waters. This decrease in Hb–O2 affinity was hypothesized to arise from a highly unusual 82Lys→Asn exchange in the chimeric β-type (β/δ) chain encoding the Hb of this species4. Data mined from more recent ancient DNA studies10,14 further confirms that this substitution was fixed in the last remaining Steller’s sea cow population (Fig. 1B). This replacement is intriguing not only because β82Lys is invariant among known mammalian Hbs, but because human variants with substitutions at this position display profound alterations in both structural and functional properties16–19. The human Hb Providence (β82Lys→Asn) variant in particular has been well characterized as asparagine at this position progressively undergoes post-translational deamidation to form aspartic acid (Asp), which over the life-time of the erythrocyte, constitutes ca. two-thirds of the Hb Providence proteins20–22. Both isoforms exhibit a decreased inherent Hb–O2 affinity, a markedly reduced sensitivity to allosteric effectors (e.g. 2,3-diphosphoglycerate [DPG], Cl−, and H+, which preferentially bind and stabilize the deoxy-state protein, reducing Hb–O2 affinity), and an increased Hb oxidative stability and solubility16,17,20,23,24. Notably, the 82Lys→Asn replacement is expected to benefit an aquatic lifestyle in cold marine climates if this exchange induced parallel changes in Hb–O2 affinity and Hb solubility of Steller’s sea cow Hb. However, a reduction of anionic effector binding may be non-adaptive as it is predicted to detrimentally increase the effect of temperature on Hb–O2 binding and release.
The weak covalent bond between O2 and the heme iron, which absorbs free energy upon breakage, dictates an inverse relationship between Hb–O2 affinity and temperature25. In temperate and Arctic endotherms this inherent attribute of Hb potentially impedes O2 delivery to the limbs and flukes, which are maintained at substantively lower temperatures to minimize heat loss and hence daily energy requirements26. Accordingly, these heterothermic mammals generally possess Hbs whose O2 binding properties are less affected by temperature than the Hbs of non-cold-adapted species, thereby maintaining sufficient O2 offloading in the face of decreasing tissue temperatures. This reduction in thermal sensitivity (the overall enthalpy of oxygenation, ΔH’), appears to predominantly arise from an increased interaction between allosteric effectors and the Hb moiety (an exothermic process), which releases additional heat to assist with deoxygenation25. Hence, the β/δ82Lys→Asn replacement, which deletes integral binding sites for both the heterotropic ligands Cl− and DPG17, is expected to maladaptively increase the effect of temperature on O2 uptake and release in the blood of the Arctic Steller’s sea cow.
To unravel the combined effects of evolved amino acid replacements on hemoglobin function in relation to the extreme thermal biology of the extinct Steller’s sea cow, we synthesized recombinant Hb proteins of this species together with those of the dugong (Dugong dugon) and Florida manatee (Trichechus manatus latirostris), and measured their O2 binding properties, relative solubilities, responses to allosteric effectors, and thermal sensitivities. We also synthesized a H. gigas β/δ82Asn→Lys Hb mutant to assess the specific effects of this exchange, together with the Hb of the last common ancestor shared between the dugong and Steller’s sea cow (from ~21.2 million years ago; Fig. 1A) in order to assess the directionality of evolved physicochemical changes in Hb function.
Results and Discussion
O2 Affinity of Sirenian Hbs
Measured O2-equilibrium curves of the five examined Hbs revealed marked differences in intrinsic O2 affinity (Figs. 2A, S1, S2 and Table S1). In the absence of allosteric effectors (pH 7.2, 37°C), the P50 (the O2 tension resulting in 50% Hb–O2 saturation) of Steller’s sea cow Hb (P50 = 8.8 mmHg) is ~2 times higher than that of dugong (3.5 mmHg), ancestral dugongid (4.3 mmHg), and manatee (5.4 mmHg) Hbs under the same conditions (Figs. 2A, S1 and Table S1). We used site directed mutagenesis to revert this amino acid to the ancestral state (β/δ82Asn→Lys) to test whether the low intrinsic O2 affinity of H. gigas Hb arises from this substitution. These experiments reveal that the increased intrinsic P50 of Steller’s sea cow Hb predominantly arises from β/δ82Asn, as the β/δ82Asn→Lys mutant exhibits an intrinsic P50 similar to that of the ancestral dugongid (Fig. 2A, S2). Of note, the O2 affinity of dugong, ancestral dugongid, and manatee Hbs was reduced in the presence of Cl− and DPG (P50 = 10.2, 9.9, and 10.9 mmHg, respectively) by a similar degree to that of Asian elephant Hb27. This finding extends previous studies conducted on sirenian Hbs4,28,29, and reveals that the high O2 affinity of dugong and manatee blood is not attributable to decreased allosteric effector sensitivity. Conversely, Steller’s sea cow Hb was shown to be markedly less responsive to allosteric effectors, as only a moderate reduction in O2–affinity was observed in the presence of Cl− and DPG (P50 = 14.3 mmHg). When the effects of these allosteric effectors were measured individually, Steller’s sea cow Hb exhibits markedly reduced DPG, Cl−, and H+ (Bohr) effects relative to those of the dugongid ancestor and extant sirenians (Fig. 2C-E; Table S1). These data confirm that a high intrinsic (i.e. in the absence of allosteric effectors) Hb–O2 affinity is an ancient sirenian trait that likely aided the transition of the group to the aquatic environment, and that Hb–O2 (and hence whole blood) affinity was secondarily reduced in the Steller’s sea cow lineage4.
The single most distinct feature of H. gigas Hb is the lack of a discernable effect of DPG on P50 (ΔlogP50(0.2 mM DPG – 0.0 mM DPG) = 0.02 at 37°C and pH 7.2; Fig. 2C and Table S1), relative to the Hbs of the ancestral dugongid (0.27) and the extant manatee and dugong (0.14 and 0.30, respectively). This intracellular effector generally occurs in equimolar concentrations to Hb30 and strongly influences the O2 affinity of most mammalian Hbs via its electrostatic interactions with the α-NH2 group of 1Val of the β2 chain and 2His, 82Lys, and 143His of β1 and β2 chains31 (Fig. S3). Unlike the other DPG-binding residues, which form single bonds with this organophosphate, β82Lys is strongly cationic within the physiological pH range, enabling DPG to form three bonds with β182Lys and two with β282Lys31. Presumably arising from its indispensable role in DPG binding, this residue is uniformly conserved in mammalian Hbs (Fig. S3), with the exception of several heterozygous adult human HbA carriers with substitutions at this position18,19,21. Given that none of the other six β/δ-chain replacements that evolved on the Steller’s sea cow branch (Fig. S4) are implicated in DPG binding, the replacement of a positive charge at β/δ82 in Steller’s sea cow Hb by a neutral (Asn) residue is consistent with its lack of DPG sensitivity. This conclusion is further supported by our measurements on the Steller’s sea cow β/δ82Asn→Lys mutant, which show that reversion to the ancestral state restores the DPG effect to the same level observed in ancestral dugongid Hb (Fig. 2C). In contrast, and likely owing to the high DPG sensitivity of human HbA (0.53)20, Hb Providence-Asn exhibits a notably higher DPG effect (0.08) than the H. gigas protein17,20. Accordingly, the whole blood O2 affinity of Hb Providence carriers is ~5 mmHg higher than that of the general population (P50 = ~21 mmHg vs. ~26 mmHg, respectively)20,23, whereas the results presented here provide strong evidence that the blood O2 affinity of Steller’s sea cows would have been lower than that of extant sirenians. These observations emphasize that the phenotypic effect of specific amino acid substitutions are conditional on the genetic background in which they occur32.
The insensitivity of H. gigas Hb to DPG it notable as it would have markedly reduced their capacity to modulate blood–O2 affinity in vivo (e.g., seasonally), and is the first demonstrated example of this phenotype among mammals. While eastern moles (Scalopus aquaticus) are a possible exception33, feliformid carnivores, ruminants, and two species of lemurs have also traditionally been placed in this category30 despite the fact their Hbs are moderately responsive to DPG in the absence of Cl−17,34–36. Nonetheless, red blood cell DPG concentrations of these ‘DPG insensitive’ species are markedly reduced relative to mammals whose Hb–O2 affinity is regulated by DPG (<0.1-1.0 mM vs. 4-10 mM, respectively), presumably owing to the energetic cost of DPG synthesis (that bypasses production of one of the two ATP molecules generated via glycolysis within the erythrocytes)30,33. Accordingly, Steller’s sea cow blood likely also contained low levels of this organophosphate.
Hb Solubility
Ectopic expression of human Hb Providence β82Asp mutants in E. coli has been shown to improve soluble protein production by 47-116% relative to the expression of human Hb variants not carrying this substitution24. Consistent with this observation, we found that Steller’s sea cow Hb is more soluble than those of other sirenians and the engineered H. gigas β/δ82Asn→Lys mutant (Fig. 2F and S5). While the precise mechanism underlying this phenomenon is unknown, both β82Asn and β82Asp have been shown to strongly reduce irreversible oxidative damage of nearby β93Cys that initiates Hb denaturation16,37,38. These β82 replacements thereby presumably decrease the rate of Hb turnover and increase the half-life of the protein38, and may be the underlying cause of mild polycythemia in humans carrying this substitution20,21. Blood with an elevated O2 carrying capacity is typical of most mammalian divers, where it increases onboard O2 stores and extends dive times1, but is not observed in extant sirenians29,39,40. However, solubility driven increases in blood Hb levels resulting from the β82Lys→Asn exchange would have allowed Steller’s sea cows to maintain an elevated rate of tissue O2 delivery to meet their metabolic demands during extended underwater foraging intervals. Although this species was presumably unable to completely submerge9,41, this conjecture is corroborated by Steller’s account that, “they keep their heads always under water [foraging], without regard to life and safety”9.
Thermal Sensitivity
The invariant energy change associated with forming the weak covalent bond between O2 and the heme iron (i.e. the enthalpy of heme oxygenation; ΔHO2) is exothermic (−59 kJ mol−1 O2)42, and only moderately opposed by the endothermic solubilization of O2 (; 12.55 kJ mol−1 O2), resulting in an inverse relationship between temperature and Hb–O2 affinity. However, the heat of the T→R conformational change (ΔHT→R), and the oxygenation-linked dissociation of H+ (ΔHH+), Cl− (ΔHCl-), and DPG (ΔHDPG) may offset this relationship, such that the overall enthalpy of Hb oxygenation (ΔH’) can become greatly minimized or even endothermic25,43. By facilitating adequate oxygenation of cool peripheral tissues, Hbs with numerically low ΔH’ values are interpreted to be adaptive for cold-tolerant, regionally heterothermic mammals. The evolution of this phenotype has predominantly been attributed to the formation of additional heterotropic ligand binding sites on the protein moiety, as has previously been demonstrated for the woolly mammoth, Mammuthus primigenius25,27. Conversely, Steller’s sea cow Hb that lacks heterotropic binding of DPG and displays lower Bohr (H+) and Cl− effects than all other sirenian Hbs measured (Fig. 2; Table S1), exhibits a ΔH’ value (−18.8 kJ mol−1 O2; Fig. 2B) that is close to that of mammoth Hb (−17.2 kJ mol−1 O2)27. This convergence largely arises from the inherently low ΔH of stripped Steller’s sea cow Hb (−34.2 kJ mol−1 O2) relative to dugong, ancestral dugongid, and manatee Hbs (range: −50.1 to −58.6 kJ mol−1 O2) at pH 7.8—where oxygenation-linked binding of protons is minimal—and indicates that structural differences modifying the T→R transition largely underlie the low thermal sensitivity of H. gigas Hb. Recent studies that have similarly shown that a large positive ΔHT→R may similarly contribute to the low ΔH’ of deer mouse, cow, shrew, and mole Hbs33,44–47 suggesting that this mechanism of temperature adaptation may be more widespread than previously appreciated. Our experiments with the Steller’s sea cow β82Asn→Lys mutant further implicate substitutions at this position as a key factor underlying the inherently low ΔH of the protein, as this modified protein displays a greatly increased ΔH in the absence of allosteric effectors relative to the wild-type Hydrodamalis protein (Fig. 2B). Interestingly, despite these inherent ΔH differences between the mutant and wild-type Steller’s sea cow Hbs, their ΔH’ values are indistinguishable in the presence of allosteric effectors (Fig. 2B). These data suggest that β82Asn uncouples thermal sensitivity from DPG concentration, permanently conferring the H. gigas protein with a numerically low ΔH’ by genetic assimilation while simultaneously eliminating the energetic cost of DPG production within the red blood cells.
Paleophysiology of Steller’s sea cows
The posthumously published behavioral and anatomical accounts of the last remaining H. gigas population by naturalist Georg Wilhelm Steller while stranded on Bering Island (55°N, 166°E) in 1741/1742 provide a rich tapestry to interpret the paleophysiology of this colossal marine herbivore. For example, their protective thick bark-like hide and extensive blubber layer give credence to the extreme nature of their shallow rocky and (during winter) ice strewn habitat10. Here, as Steller9 remarked, they used fingerless, bristle-covered forelimbs for support and to shear “algae and seagrasses from the rocks”, which they masticated “not with teeth, which they lack altogether, but with” large, ridged keratinized pads located on the upper palate and lower mandible. Although they became visibly thin during winter when “their spinous processes can be seen”, Steller9 noted that “(t)hese animals are very voracious, and eat incessantly” such that their stupendous stomach (“6 feet [1.8 m] long, 5 feet [1.5 m] wide”) and enormous intestines—which measured a remarkable 5,958 inches (~151.5 m) from esophagus to anus, equivalent to “20 times as long as the whole animal”—are constantly “stuffed with food and seaweed”. These observations of a proportionally larger gut41 are consistent with relatively high energetic requirements relative to extant manatees, which owing to their low metabolic intensity become cold stressed and die if chronically exposed to water temperatures below 15°C48. Reductions in insulative blubber thickness during the winter months would have compounded the rate of heat loss of these behemoths to sub-zero degree Celsius air and water, though may have been compensated for by arteriovenous anastomoses that regulated blood flow to the skin, and by countercurrent rete supplying the flippers and tail flukes, the latter of which are well developed in manatees and presumably other sirenians49,50. These structures conserve thermal energy by promoting profound cooling at the appendages and periphery26, and presumably underlie the low thermal dependence of Steller’s sea cow Hb relative to those of extant sea cows.
Reductions in blood–O2 affinity accompanying the H. gigas β/δ82Lys→Asn substitution is expected to have further augmented tissue O2 delivery with only negligible effects on lung O2 uptake, thereby helping to fuel increased thermogenesis to maintain a stable core temperature. The latter was presumably supplemented by a substantive heat increment arising from fermentation and other post-prandial processes50. Although the attendant increase in the rate of O2 consumption would have mandated a reduction in breath-hold endurance—likely reflecting the relatively short submergence times (4 to 5 minutes) observed by Steller9—our results suggests this may be been counteracted by an elevated Hb concentration (blood–O2 carrying capacity) that was potentially coupled to a greater lung volume41. Underwater foraging times were presumably further defended by key components of the dive reflex, namely bradycardia and peripheral vasoconstriction. Indeed, Steller inadvertently was the first to (indirectly) describe this phenomenon as he observed his crew hunting the animals with spears and knives, “the blood from the wounded back spurted up like a fountain. As long as he kept his head under water the blood did not flow out, but as soon as he raised his head to breathe the blood leaped forth anew”.
A final compelling aspect of Steller’s sea cow evolution was their immense increase in size—up to 11,000 kg in mass and 10 m in length—relative to extant sirenians7. While it is unfortunate Steller does not provide measurements of “their tender little offspring”, Gerhard Friedrich Müller, who edited Steller’s manuscript prior to publication, noted calves “weighed 1200 pounds [544 kg] and upwards”51. This value is ~10 to 50 times the mass of new born manatees and dugongs (~10-50 kg)52 and is suggestive of rapid prenatal growth during the ~1 year gestational period proposed by Steller9. Given that all mammals except ruminants and anthropoid primates express the same Hb isoform during both the fetal and post-natal stages of development, the novel DPG insensitive phenotype of H. gigas is particularly notable in this regard owing to its potential impact on maternal/fetal O2 delivery. In the vast majority of mammalian species, the O2 affinity of fetal blood is elevated relative to that of the maternal circulation30. This increase is typically achieved via maintenance of low intracellular DPG concentrations in fetal blood cells; exceptions are the above ruminants and anthropoid primates which instead express a higher O2 affinity fetal-specific hemoglobin isoform30. However, an elevated fetal blood O2 affinity is expected to lower blood-to-tissue PO2 gradients within the fetal circulation (i.e. lowered O2 offloading potential), which together with the immersion foraging strategy and high maternal blood O2 affinity of manatees and dugongs may promote intermittent fetal hypoxia, which has experimentally been shown to reduce the birth weight of rat pups53,54. By contrast, and owing to the inability of Steller’s sea cows Hb to respond to DPG in either the fetal or adult circulations, this species would represent a rare example (feloids and eastern moles are others) in which fetal and maternal blood have the same (lower) O2 affinity. However, by stipulating O2 offloading occurs at a relatively high PO2, the DPG insensitive Steller’s sea cow Hb phenotype may have increased both placental O2 transfer and fetal O2 delivery. Together with increases in Hb solubility/reduced susceptibility to oxidative damage arising from β/δ82Lys→Asn that conceivably also elevated the O2 carrying capacity of fetal blood, these attributes presumably fostered an enhanced fetal growth rate of these immense sirenians. The resulting increased thermal inertia and relatively low surface-area-to-volume ratio following birth, together with an adaptively reduced Hb thermal sensitivity and thick a ‘bark-like’ skin arising from inactivation of lipoxygenase genes10, were presumably central components of Steller’s sea cows successful exploitation of the harsh sub-Arctic marine environments of the North Pacific.
Methods and Materials
Construction of Recombinant Hb Expression Vectors
The adult-expressed Hb genes (HBA and HBB/HBD) of the Florida manatee, dugong, and Steller’s sea cow, and the most recent common ancestor shared by Steller’s sea cow and the dugong (‘ancestral dugongid’) have previously been determined4. As the H. gigas β/δ82Lys→Asn exchange is not known to occur in any living species, we mined recently deposited genomes for 13 additional Steller’s sea cows (PRJNA484555, PRJEB43951) to test for the prevalence of this replacement in the population. Briefly, we first searched SRA files of each specimen using the megablast function against a known H. gigas HBB/HBD gene sequence (GenBank accession #: MK562081). All hits were then downloaded, trimmed of adapters and low quality regions using BBDuk (Joint Genome Institute), and assembled to H. gigas HBB/HBD using Geneious Prime 2019 software (Biomatters Ltd, Auckland, New Zealand).
Globin sequences for the above four species were optimized for expression in E. coli and synthesized in vitro by GenScript (Piscataway, NJ). The resulting gene cassettes were digested with restriction enzymes and tandemly ligated into a custom Hb expression vector54 using a New England BioLabs Quick Ligation Kit as recommended by the manufacturer. Chemically competent JM109 (DE3) E. coli (Promega) were prepared using a Z-Competent E. coli Transformation Kit and Buffer Set (Zymo Research). We also prepared a H. gigas β/δ82Asn→Lys Hb mutant via site-directed mutagenesis on the Steller’s sea cow Hb expression vector by whole plasmid amplification using mutagenic primers and Phusion High-Fidelity DNA Polymerase (New England BioLabs), phosphorylation with T4 Polynucleotide Kinase (New England BioLabs), and circularization with an NEB Quick Ligation Kit (New England BioLabs). All site-directed mutagenesis steps were performed using the manufacture’s recommended protocol.
Hb expression vectors were co-transformed into JM109 (DE3) chemically competent E. coli alongside a plasmid expressing methionine aminopeptidase55, plated on LB agar containing ampicillin (100 µg/ml) and kanamycin (50 µg/ml), and incubated for 16 hours at 37°C. A single colony from each transformation was cultured in 50 ml of 2xYT broth for 16 hours at 37°C while shaking at 200 rpm. Post incubation, 5 ml of the culture was pelleted by centrifugation and plasmid DNA was isolated using a GeneJET Plasmid Miniprep Kit (Thermo Scientific). The plasmid sequence was verified using BigDye 3.1 sequencing chemistry and an ABI3130 Genetic Analyzer. The remainder of the culture was supplemented with glycerol to a final concentration of 10%, divided into 25 ml aliquots and stored at −80°C until needed for expression.
Expression and Purification of Recombinant Hb
25 ml of starter culture (above) was added to 1250 ml of TB media containing ampicillin (100 µg/ul) and kanamycin (50 µg/ul) and distributed evenly amongst five 1 L Erlenmeyer flasks. Cultures were grown at 37°C while shaking at 200 rpm until the absorbance at 600 nm reached 0.6-0.8. Hb expression was induced by supplementing the media with 0.2 mM isopropyl β-D-1-thiogalactopyranoside, 50 µg/ml of hemin and 20 g/L of glucose and the culture was incubated at 28°C for 16 hours while shaking at 200 rpm. Once expression had completed, dissolved O2 was removed by adding sodium dithionite (1 mg/ml) to the culture, which was promptly saturated with CO for 15 minutes. Bacterial cells were then pelleted by centrifugation and Hb purified by ion exchange chromatography according to Natarajan et al.55
It should be noted that the β82Asn residue of human Hb Providence is relatively uncommon in that it slowly undergoes post-translational deamination in vivo to form aspartic acid, with the latter residue comprising ~67-75% in mature mixed blood20,22. While it is unknown to what degree H. gigas β/δ82Asn was catalyzed into Asp in nature, O2 binding data (see below) of this species was collected from freshly purified recombinant samples for which only one peak— presumably β/δ82Asn—was resolved during chromatography (data not shown). Additionally, this reaction is dependent on the local protein environment56, specifically two nearby residues β143His and β83Gly22. Importantly, the latter residue was replaced by β/δ83Ser on the Steller’s sea cow branch (Fig. 1B), which is expected to slow (but not stop) the rate of deamidation55. Regardless, since the two Hb Providence isoforms have similar O2 affinities and functional properties17,20,23 it is unlikely that presence of β/δ82Asp in Steller’s sea cow blood would meaningfully alter the results and interpretations presented herein.
Functional Analyses of Hbs
O2-equillibrium curves for Hb solutions (0.25 mM heme in 0.1 M HEPES buffers) were measured at 25 and 37°C using the thin film technique described by Weber57. Hb solutions varied in their pH (6.9, 7.4, and 7.9), chloride concentration (0 or 0.1 M KCl), and organic phosphate concentration (0 or 2-fold molar excess of DPG relative to tetrameric Hb concentrations) in order to test the influence of these cofactors on Hb function. Each Hb solution was sequentially equilibrated with three to five different oxygen tensions (PO2) that result in Hb– O2 saturations between 30 to 70%. Hill plots (log[fractional saturation/[1-fractional saturation]] vs logPO2) constructed from these measurements were used to determine the PO2 (P50) and the cooperativity coefficient (n50) at half saturation, from the χ-intercept and slope of these plots, respectively. By this method, the r2 determination coefficients for the fitted curves exceed 0.995 and the standard errors (SEM) are less than 3% of the P50 and n50 values47. P50 values at 25 and 37°C were used to assess the thermal sensitivity of sirenian Hbs by calculating the apparent enthalpy of oxygenation using the van’t Hoff isochore: where R is the universal gas constant and T1 and T2 are the absolute temperatures (°K) at which the P50 values were measured. All ΔH values were corrected for the heat of O2 solubilization (12.55 kJ mol−1 O2).
Solubility assay
Ammonium sulfate was added to Hb solutions (0.074±0.004 mM Hb4) to generate final concentrations that ranged from 0 to 3.5 M. These solutions were incubated for 60 minutes at 37°C and the remaining soluble Hb was measured via Drabkin’s reagent, according to manufacturer’s instructions (Sigma-Aldrich).
Acknowledgements
We thank Mike Gaudry and Elin Ellebæk Petersen for technical assistance, Chandrasekhar Natarajan for providing us with a hemoglobin expression plasmid, and J. Storz for constructive feedback on an earlier version of this manuscript. Authorization to use paintings by C. Buell was kindly provided by J. Gatesy. This study was supported by NSERC (Canada) Discovery and Accelerator Supplement Grants (K.L.C.; RGPIN/238838-2011, RGPIN/412336-2011, and RGPIN/06562-2016), an NSERC Postgraduate Scholarship (A.V.S), the Faculty of Science and Technology, Aarhus University (REW), and the Danish Council for Independent Research (AF; DFF - 4181-00094).