|Home | About | Journals | Submit | Contact Us | Français|
This study is aimed at investigating the molecular basis of environmental adaptation of woolly mammoth hemoglobin (Hb) to the harsh thermal conditions of the Pleistocene Ice-ages. To this end, we have carried out a comparative biochemical-biophysical characterization of the structural and functional properties of recombinant hemoglobins (rHb) from woolly mammoth (rHb WM) and Asian elephant (rHb AE) in relation to human hemoglobins Hb A and Hb A2 (a minor component of human Hb). We have obtained oxygen equilibrium curves and calculated O2 affinities, Bohr effects, and the apparent heat of oxygenation (ΔH) in the presence and absence of allosteric effectors [inorganic phosphate and inositol hexaphosphate (IHP)]. Here, we show that the four Hbs exhibit distinct structural properties and respond differently to allosteric effectors. In addition, the apparent heat of oxygenation (ΔH) for rHb WM is less negative than that of rHb AE, especially in phosphate buffer and the presence of IHP, suggesting that the oxygen affinity of mammoth blood was also less sensitive to temperature change. Finally, 1H-NMR spectroscopy data indicates that both α1(β/δ)1 and α1(β/δ)2 interfaces in rHb WM and rHb AE are perturbed, whereas only the α1δ1 interface in Hb A2 is perturbed compared to that in Hb A. The distinct structural and functional features of rHb WM presumably facilitated woolly mammoth survival in the Arctic environment.
Human normal adult hemoglobin (Hb A) is a hetero-tetramer consisting of two α-subunits and two β-subunits (1). Within the bloodstream, this respiratory protein switches cyclically between high and low O2-affinity states to control the delicate balance between the uptake of O2 at the alveoli and its optimal unloading at the tissues. However, because Hb oxygenation is exothermic, this balance is disrupted by changes in temperature such that O2 affinity decreases sharply with increasing temperature and increases with decreasing temperature (1). Although this trait is often considered “adaptive” for enhancing O2 offloading at warm (exercising) muscles, it may pose considerable challenges for O2 delivery to cool extremities and peripheral tissues of heterothermic mammals and hence is often countered (2). However, the molecular mechanisms by which the thermal sensitivity of the O2 affinity of Hb, which is dictated by the overall enthalpy of oxygenation (ΔH), has been lowered through evolution within the various cold-adapted mammalian lineages are poorly understood (3-5). Elephantids are a particularly good model system to investigate the effects of temperature on Hb structure-function relationships as they possess both warm and cold-adapted members. Specifically, the group evolved in warm sub-temperate climates, where extant Asian and African lineages are still found, with the ancestors of the extinct woolly mammoth (Mammuthus primigenius) abruptly invading high-latitude environments of Eurasia near the start of the Pleistocene ice-ages some 1.2–2.0 million years ago (6). Consistent with this environmental diaspora, the functional properties of woolly mammoth Hb were found to differ substantially from those of Asian elephant Hb, with the most notable change being a significant reduction in the overall ΔH of the mammoth protein in the presence of naturally occurring allosteric effector molecules (6). Interestingly, this functional shift comes about despite the fact that the primary sequences of Asian elephant Hb differ from those of the mammoth at only one position in the α-globin chain (K5N) and at three positions (T12A, A86S, and E101Q) in the β-type globin chain, respectively [(Table 1); (6)]. Of these latter residues, β/δ12Ala of the mammoth protein is near the 2,3-bisphosphoglycerate (BPG) binding cleft, β/δ86Ser is in the heme pocket, and β/δ101Gln is located between 99Asp and 102Asn of the same chain in the inter-subunit α1(β/δ)2 interface, which are all critical to the function of Hb (1). Indeed, naturally occurring human mutations at β101 have illustrated that changes at this residue position alter both the intrinsic and allosteric properties of the protein, though the mechanism by which this residue exerts these effects is still unclear (7).
The Hb of elephantids is unusual in that the β-type subunit is encoded by a chimeric β/δ-fusion gene (HBB/D) that arose via an unequal crossing over event long before the radiation of this group (6, 8). Consequently, the N- and C- terminal halves of the hybrid elephantid polypeptide are orthologous to the ‘β’ and ‘δ’ globin chains of other mammals, respectively. Thus, it would be interesting to probe the structural significance of this evolutionary change, as the physiochemical properties of human Hb A and Hb A2 (a minor α2δ2 tetramer component of human blood) – which differ at only 10 residue positions [(Table 1); (9)] – show important differences in regards to anti-sickling properties and thermal stability (10).
The primary objective of this study is to evaluate and compare the biochemical factors that alter the temperature dependence of O2 binding to Asian elephant and mammoth Hb, with special reference to human Hbs A and A2. Additionally, we aim to provide new insights into the structural mechanisms that affect Hb function, thereby aiding the design of a new generation of medically relevant Hb-based oxygen carriers (blood substitutes) for use, e.g., during certain hypothermia-dependent cardiac and brain surgery applications (11). To this end, we obtained the 1H-NMR spectra of the four Hbs and measured their oxygen affinity, sensitivity to various allosteric effectors, and cooperativity at three different temperatures (37, 29, and 11 °C) and over a range of pH (pH 8.5 to 5.5).
Hb A and Hb A2 were isolated and purified from human normal blood samples obtained from the local blood bank using the procedures routine in our laboratory (12-13). Chemicals and restriction enzymes were purchased from major suppliers, such as Fisher, Sigma, Bio-Rad, Boehringer Mannheim, New England BioLabs, Pharmacia, Promega, and United States Biochemicals, Inc., and were used without further purification.
The plasmid (pHE27E) that expresses recombinant Hb AE was constructed by replacing the human α- and β-globin genes in Hb expression plasmid pHE2 (12) with Asian elephant α-like and β/δ-like cDNAs. The plasmid (pHE27M) for expressing rHb WM was constructed by introducing the mammoth-specific residue mutations (αK5N, β/δT12A, β/δA86S, and β/δG101Q) into pHE27E using site-directed mutagenesis. The desired mutations were confirmed by DNA sequencing. For additional details, see (6).
Plasmids for rHb AE (pHE27E) and rHb WM (pHE27M) were transformed into E. coli strain JM109 and the cells were grown in a minimal medium in a 20-L fermentor (B. Braun Biotech International, Model Biostat C) at 32 °C until the optical density reached ~10 at 600 nm. Isopropyl β-D-thiogalactopyranoside (IPTG) was then added at a concentration of 24 mg/L to induce the expression of the rHbs. After the addition of hemin (25 mg/L), growth was continued for at least another four hours. Harvesting was accomplished through centrifugation. Cell paste was stored at –80 °C until it was needed. The purification of rHbs followed the previous procedure (12, 14). The purity of the rHbs was then checked with an electrospray ionization mass spectrometer and Edman degradation as described previously (12, 14). All rHbs used in this study had the correct molecular weights and contained less than 5% methionine at the amino termini.
Oxygen-equilibrium curves for these four Hbs were measured in both 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) and 0.1 M sodium phosphate buffers at 11, 29, and 37 °C with a Hemox Analyzer. Experiments were conducted in presence and absence of three-time molar concentrations of inositol hexaphosphate (IHP), in the pH range 5.5 to 8.5. The sample concentration of Hbs used 100–120 μM (in terms of heme) was selected to avoid tetramer-dimer dissociation. Each sample was checked before and after each measurement for met-hemoglobin (met-Hb) in a spectrophotometer. Any sample with greater than 5% met-Hb was discarded. The partial pressure of O2 at 50% Hb saturation (P50), a measure of oxygen affinity, and the Hill coefficient (n50), a measure of the cooperativity of the oxygenation process, were calculated for each oxygen-equilibrium curve. The P50 values (mmHg) have an accuracy of ±5% and the n50 values have an accuracy of ±10%.
To detect changes in the tertiary or quaternary structure of the two elephantid Hbs, 1H-NMR spectra were obtained on Bruker Avance DRX-300 (results not shown) and DRX-600 NMR spectrometers. Samples consisted of aqueous solutions of Hb at a concentration of 5% (3.1 mM in terms of heme) in 0.1 M sodium phosphate buffer at pH 7.0 in 95% water and 5% deuterium oxide (D2O) and were assessed at 11, 29, and 37 °C. A jump-and-return pulse sequence was used to suppress the water signal. 1H chemical shifts were indirectly referenced to the methyl proton resonance of the sodium salt of 2,2-dimethyl-2-silapentane-5-sulfonate (DSS) through use of the internal reference of the water signal at 4.76 ppm downfield of DSS at 29 °C.
The functional properties of the four Hbs were studied by measuring their oxygen-binding affinities as a function of buffer, pH, temperature, in the presence or absence of IHP (Figure 1). Each of the Hbs studied exhibits distinct functional properties, with P50 values varying widely between Hbs. Notably, rHb WM has a lower O2 affinity (higher P50 value) than that of rHb AE under the various experimental conditions (Figure 1). The P50 values of Hb A and Hb A2 are very similar and always higher than those of rHb WM and rHb AE. The P50 values are also affected remarkably by buffer condition. For instance, in MES buffer and in the absence of IHP (“stripped” condition), the differences of the P50 values between rHb WM and rHb AE are very small (Figure 1A), but become significant in phosphate buffer (Figure 1B) due to their different response to phosphate ion. Similarly, the addition of IHP to Hbs significantly decreases the O2 affinities in both MES buffer and phosphate buffer, but the relative IHP effect is much stronger in MES buffer. Thus, in the presence of IHP, the O2 affinities of these Hbs are always lower in MES buffer than in phosphate buffer (Figure 2). The effects of the buffer and IHP on the O2 affinities are also illustrated by the dramatic shift of the oxygen-binding curves. For example, at pH 7 and 37 °C, in the absence of IHP, the O2-binding curves for these Hbs are very close in MES buffer (Figure 3A), but become scattered in phosphate buffer (Figure 3B). Upon the addition of IHP, all of the binding curves are shifted to the right in both buffers. Finally, it is noted that the addition of IHP affects the O2 affinities of Hb A and Hb A2 the most, followed by rHb WM, whereas rHb AE is affected the least by IHP (Figure 3).
The cooperativity of the oxygenation process for all Hbs is measured by the Hill coefficient, n50. The n50 values for Hbs under various experimental conditions are taken as the slope of the Hill plots at 50% saturation and are summarized in the Supporting Information Table 1S. The n50 values change significantly in the presence of IHP, but are only slightly affected by buffer, temperature, and pH. For example, as shown in Figure 4, the n50 values for both Hb A and Hb A2 are maintained around the expected value of 2.8–3.0 in the range of pH 5.8–8.4 in 0.1 M phosphate buffer at 29 °C, while n50 values for rHb WM and rHb AE are found to have lower values (2.0–2.5) under the same experimental conditions. Upon the addition of IHP, a broader spectrum of differences in the n50 values is observed with n50 being generally reduced at lower pH values and close to n50 values measured in the absence of IHP at higher pHs. Of note, the n50 values of rHb WM are lower than those of the other Hbs under all experimental conditions (Figure 4).
The usual Hb concentration used for our O2-binding measurements is 100–120 μM (in terms of heme). In order to investigate if rHb WM and rHb AE exhibit unusual tetramer-dimer dissociation, we have also carried out a concentration dependent study of the O2-binding measurements of these two Hbs together with Hb A over a Hb concentration range of 25 to 100 μM in 0.1 M sodium phosphate buffer at pH 7.4 and 37 °C. No observable differences in both P50 and n50 values among these concentrations were found, indicating that there is no significant tetramer-dimer dissociation over the range of experimental conditions studied (see Supporting Information Table 2S).
The pH dependence of the oxygen affinity of Hbs is measured over a range of pH 8.5 to 5.5 and a significant difference is found in the amplitude of the alkaline Bohr effect (ΔlogP50/ΔpH between pH 6.8 and 8.0) for these Hbs (Table 2). In the “stripped” condition, the O2-binding affinity of Hb A shows a stronger Bohr effect (Δlog P50/ΔpH = –0.53), while rHb WM and rHb AE are characterized by lower Bohr coefficients (ΔlogP50/ΔpH = –0.38 and –0.28, respectively). However, upon the addition of IHP, Bohr coefficients of rHb WM and rHb AE reach their maximum (ΔlogP50/ΔpH = –0.81), which is comparable to that of Hb A (ΔlogP50/ΔpH = –0.79). The Bohr effect of rHb WM is slightly increased (–0.46) in phosphate buffer, and is further increased to about –0.70 upon the addition of IHP, very close to that of Hb A and A2 under the same experimental conditions (Table 2). However, the Bohr effect of rHb AE remains low (–0.32) in phosphate buffer and only increase slightly (–0.47) in the presence of IHP, a value still considerably lower than that of the other Hbs.
The oxygen-binding affinity for Hbs is affected not only by allosteric effectors and pH, but also by temperature (1, 15). The temperature dependence of the oxygen affinity of Hbs has been measured at 37, 29, and 11 °C, in the pH range of 5.5 to 8.5 in MES and phosphate buffers, respectively, and the results are summarized in Figure 5 and Supporting Information Figure 1S. As expected, P50 values of Hbs are lower at 11 °C than at 37 °C, indicating that O2 binds tighter to Hbs at lower temperatures (Figure 5). In the absence of IHP, the P50 values of these Hbs are in the same range, increasing from ~ 3 mmHg at 11 °C to ~ 16 mmHg at 37 °C. However, the P50 values of rHb WM exhibit a smaller increase for the same experimental temperature change. As illustrated in Figure 5A, rHb WM has a lower ΔlogP50/ΔT value (0.023 ) in the MES buffer, compared to Hb A (0.034) and rHb AE (0.028), where ΔT represents the difference between two temperatures (in °C) and ΔlogP50 is the difference in P50 (mmHg) at the two temperatures. In phosphate buffer, the ΔlogP50/ΔT values for these Hbs are all reduced, through that of rHb WM (0.019) remains consistently lower than that of Hb A (0.023) and rHb AE (0.024). Upon the addition of IHP, all of the P50 values at each experimental temperature increase, but the absolute values vary for each Hb because of their different response to IHP. In order to illustrate the temperature effect more clearly, the apparent heat of oxygenation, ΔH, is calculated from the van't Hoff equation (1) on the basis of ΔlogP50 at the two different temperatures (T1 and T2, in °K):
ΔH includes the heat of oxygen solvation (–3 kcal mol-1). The results of the exothermic oxygenation enthalpy (ΔH, kJ mol-1, where 1 cal=4.184 J) presented in Table 3 have been corrected for this value. It is noted that the ΔH values for the four Hbs are all pH dependent (Figure 6). For example, in phosphate buffer, the ΔH value for rHb WM changes from –41.0 kJ mol-1 at pH 5.8 to –18.7 kJ mol-1 at pH 7.4. At pH 7.0 in the “stripped” condition, the ΔH values of rHb WM and rHb AE are –25.9 and –30.6 kJ mol-1, respectively, much lower (in absolute terms) than that of human Hb A (–46.1 kJ mol-1), i.e., rHb WM and rHb AE show a smaller temperature effect than Hb A. In phosphate buffer, the ΔH values of rHb WM are less negative than those of rHb AE, while the ΔH values of Hb A and Hb A2 are very similar over the range of experimental conditions. When IHP is present, the ΔH values of the four Hbs are all significantly decreased (in absolute terms), with that of rHb WM remaining slightly lower than that of rHb AE. For all four Hbs, the lowest ΔH values (in absolute terms) are observed when both phosphate and IHP are present as functional modulators, suggesting an additive effect of phosphate and IHP to the ΔH value.
1H-NMR spectra were recorded for the Hbs in the deoxy and the CO forms in the absence and presence of IHP at 11, 29, and 37 °C, respectively. The exchangeable proton resonances of α103His and α122His of the four Hbs studied in the CO and deoxy forms are shown in Figures 7. In Hb A, the resonance at 12.2 ppm has been assigned to the side chain Nε2H group of α103His (16-19), and the resonance at 12.9 ppm has been assigned to the side chain Nε2H group of α122His (18). In the CO form of Hbs (Figure 7A), the peak at 12.2 ppm (α103His) is not present in the spectra of rHb AE or rHb WM, but appears in the spectra for Hb A and A2. Additionally, the peak at 12.9 ppm (α122His) in Hb A is shifted downfield in the spectra of rHb AE and rHb WM (which have α- and β/δ-chains) and Hb A2 (which has α- and δ-chains). In the spectrum of deoxy-Hb A (Figure 7C), the resonance at 14 ppm was assigned to the H-bond between α42Tyr and β99Asp, and has been used as a T-state marker at the α1β2 interface (20). In the presence of IHP, the T-marker starts to appear in the spectra of the CO form of Hb A and Hb A2 at a lower temperature (11 °C), but is not seen in the spectra of the CO form of rHb AE and rHb WM (Figure 7B). Significant differences are observed in the deoxy state of the Hbs (Figure 7C). As reported previously, the T-marker in deoxy-Hb A2 is present at the same spectral position as that of deoxy-Hb A (13). These T-state markers shift 0.1–0.3 ppm downfield in the spectra of rHb WM and rHb AE, and the peak at 11.1 ppm assigned to β37Trp also shifts 0.2 ppm downfield at all temperatures, indicating a perturbation of the α1(β/δ)2 interface in rHb WM and rHb AE (Figure 7C).
Ring-current-shifted resonances at –1.5 ppm to –2.0 ppm (Figure 7D) provide information about the tertiary structure around the heme pocket (16). The resonances at –1.75 and –1.82 ppm have been assigned to the γ2-CH3 group of E11Val of the α- and β-chains of HbCO A, respectively (21-22). For rHb WM and rHb AE, the E11Val methyl resonances of both the α- and β/δ-chains are shifted upfield to –2.01 ppm, while the resonances of Hb A2 are not affected.
Mammals can be divided into two groups based on how the O2-binding affinities of their Hbs respond to allosteric effectors. The O2-binding affinities of Hbs from humans, pigs, dogs, and most primates are decreased in the presence of organic phosphate, whereas those from cats and ruminants are not sensitive to organic phosphate, but exhibit significant effects in the presence of chloride ions (23-25). Our study has examined the properties of rHb WM and rHb AE with respect to their interactions with phosphate and IHP. Our results have confirmed that rHb WM and rHb AE belong to the group of mammalian Hbs that exhibit an intrinsically high oxygen affinity, which can be modulated by phosphate anions (i.e. phosphate and IHP). Comparisons between the P50 values obtained in MES and sodium phosphate buffers show that the O2-binding affinities of Hb A, rHb WM, and rHb AE significantly decrease in phosphate buffer and/or in the presence of IHP (Figure 2). However, the IHP effect is stronger in MES buffer than in phosphate buffer, suggesting that the allosteric effects of phosphate and IHP are not synergistic with respect to the O2 binding. Conversely, our results indicate that these two anions compete for the same binding sites. Thus, the changes of P50 values of these Hbs upon the addition of IHP reach their maximum in MES buffer, and cannot be increased further in phosphate buffer.
On the basis of our current studies, rHb WM exhibits a stronger response to allosteric effectors (H+, phosphate, and IHP), than rHb AE, despite possessing only three amino-acid residues in their β/δ-chain sequences (6). The functional properties of β/δ101Gln in rHb WM and β/δ101Glu in rHb AE are of particular interest because this residue is located in the α1(β/δ)2 contact region. In liganded Hb A, β101Glu is in close contact with α94 residue, while in the deoxy-Hb A, it has contacts with both β94Asp and α96Val residues (1, 26). Thus, this residue is involved in the stabilization of both unliganded and liganded tetramers of Hbs. Various mutations of Hb A at β101Glu have been studied, but the role of this residue is still not clear (7). All mutant Hbs at this position have an elevated O2 affinity relative to Hb A, except for the mutant Hb Rush (β101Glu→Gln), which is the only one β101 mutant having a higher sensitivity to chloride ion than Hb A. Consequently, despite possessing a high intrinsic O2 affinity (as in the other β101 mutants), the O2 affinity of Hb Rush is decreased to a slightly lower value than that of Hb A in the presence of 0.1 M chloride (7). As suggested by Campbell et al. (6), the β/δ101Glu→Gln substitution in rHb WM might play the same role as in Hb A, which could allow the positive charge of β/δ104Arg to form an additional Cl– binding site within the α1(β/δ)2 interface of the deoxy-state molecule. This proposed additional chloride-binding site in rHb WM might also be responsible for the increased allosteric effect caused by the addition of phosphate and IHP, as shown by the drastic decrease in O2 affinity of rHb WM in the presence of these allosteric effectors.
Oxygen-binding experiments show that both inorganic phosphate and IHP affect the functional behavior of rHb WM and rHb AE, not limited only to the absolute value of the oxygen affinity, but also affecting the amplitude of the Bohr effect (ΔlogP50/ΔpH). As shown in Table 2, in the “stripped” condition (i.e. MES buffer and in the absence of IHP), the ΔlogP50/ΔpH values of rHb WM and rHb AE are lower (in absolute terms) than that of Hb A between pH 6.8 and 8.0. Bohr effects of the two elephantid rHbs are slightly influenced by the presence of inorganic phosphate, but are markedly affected by the presence of IHP. For Hb A, it is known that there are a number of amino acid residues that contribute to the observed Bohr effects, including the N-terminal residues (27-28) and a large number of surface His residues (29-32). In 0.1 M HEPES buffer with 0.1 M chloride, β146His contributes the most to the alkaline Bohr effect (63% at pH 7.4) among those surface histidyl residues, while β143His contributes the most to the acid Bohr effect (71% at pH 5.1) (29-32). These two histidyl residues are also present in rHb WM and rHb AE, and presumably also contribute to the Bohr effect of these proteins. It is noted that the number of histidyl residues in rHb WM and rHb AE is the same as in Hb A, but the locations of these residues are different (Table 1). β2His and β116His in Hb A, which exert moderate negative and moderate positive contributions, repectively, to the Bohr effect (31-32), have been replaced by β/δ2Asn and β/δ116Arg in rHb WM and rHb AE. Thus, the contributions from these two histidyl residues to the Bohr effect are absent in rHb WM and rHb AE. Consequently, the low Bohr effects observed for these two Hbs (Table 2) may arise from substitutions at β/δ44(Ser→His) and β/δ56(Gly→His), which presumably contribute negatively to the Bohr effect of these two proteins. Although rHb AE has the identical histidyl residues as rHb WM, it exhibits a weaker Bohr effect in phosphate buffer, even in the presence of IHP, suggesting that the contributions from the histidyl residues might not account for all of the Bohr effect of these two rHbs. The lower Bohr effect of rHb AE is consistent with its weaker response to phosphate and IHP relative to rHb WM (Table 2).
The oxygenation of Hb is exothermic, with increasing the temperature lowering the O2 affinity directly by weakening the hydrogen bonds between Hb and O2 (1, 33-34). Arctic ruminants (e.g. musk-ox and reindeer) are routinely subjected to extremely low environmental temperatures. Hence, to reduce the heat loss under these conditions, these animals exploit countercurrent heat exchangers in the extremities that allow them to maintain markedly lower tissue temperatures at these sites (3). Thus, a decrease in the temperature sensitivity of oxygen binding could be a functional strategy which allows unloading O2 to the cool peripheral tissues. This phenomenon has been observed from the previous studies for the Hbs of a number of sub-arctic and arctic mammals, including Eskimo dog, musk-ox and reindeer (5, 35-36). The apparent heat of oxygenation, ΔH, is used to evaluate the temperature effect on the O2 affinities of Hbs. Based on the ΔH values, the Hbs of various species can be broadly divided into two groups. The Hbs of the first group, including reindeer Hb and musk-ox Hb, have intrinsically low ΔH values even in the absence of allosteric effectors (5). The second group of Hbs can be further divided based on whether the large negative ΔH of these Hbs becomes less negative in the presence of 2,3-bisphosphoglyceric acid (2,3-BPG), such as pig Hb, and fetal human Hb (Hb F); or whether the ΔH is not very sensitive to the addition of 2,3-BPG, such as Hb A (5). In the “stripped” condition, rHb WM and rHb AE have a less negative ΔH of oxygenation when compared with Hb A (Table 3 and Figure 6A). In phosphate buffer, the ΔH values of rHb WM and rHb AE decrease slightly and remain in the same range as in MES buffer, while that of Hb A dramatically decrease (in absolute terms) compared to that in MES buffer, (Table 3 and Figure 6B). Further comparisons between rHb WM and rHb AE show that although similar ΔH values are observed in MES buffer, rHb WM has much less negative ΔH values in phosphate buffer (Figure 6B), which can be attributed to the fact that the O2 affinity of rHb WM has a stronger response to phosphate, a weak allosteric effector. When IHP is present, the ΔH values of Hbs are greatly affected as the result of a change in the P50 values. Differences between the ΔH values of the various Hbs remain, but become smaller due to the strong allosteric effect of IHP (Figure 6). The less negative values of the ΔH values observed upon the presence of phosphate and/or IHP suggest that the apparent temperature effect is dependent on the ability of each of the Hbs to respond to the allosteric effectors. On the basis of the previous studies on the Hbs of various arctic animals (5), a potential structural explanation has been proposed, suggesting that the amino acid residues at β8, β76, and β77 positions form an “additional” chloride-binding site (relative to Hb A), which is responsible lowering ΔH of these proteins (37-38). Additionally, the mutation of β76Ala→Lys might be responsible for the synergistic effect of 2,3-BPG and chloride on the O2 affinity of these Hbs (38). In the Hbs of arctic animals, such as reindeer Hb and ox Hb, the amino acid residue at β76 is Lys, and the ΔH values of these Hbs are less negative than that of horse Hb, which has Ala at β76. Both rHb WM and rHb AE have β/δ76Lys, the same as in reindeer Hb. Our previous study of rHb WM suggests the E101Q substitution in rHb WM creates an additional binding site for allosteric effectors, which contributes to the less negative ΔH of oxygenation relative to rHb AE, thus forming part of the physiological adaptation of woolly mammoth (6). The less negative ΔH values of rHb WM observed in the present work confirm the lower temperature effect of rHb WM compared to that of rHb AE (6).
On the basis of our measurements, the ΔH values of rHb WM are in the same range as those of Hbs from other arctic animals obtained under similar experimental conditions (5). For example, in 0.1 M HEPES buffer with 0.1 M NaCl in the presence of 2,3-BPG, the ΔH of Hb reindeer is –14.0 kJ mol-1 (5), whereas the values calculated for rHb WM are –18.7 kJ mol-1 in 0.1 M sodium phosphate buffer (current result) and –19.3 kJ mol-1 in the presence of Cl- and BPG (6). These low ΔH values of rHb WM are consistent with life in the arctic environment since the O2 affinity of its Hb is less affected by low temperature and, therefore, it would be easier to unload the O2 to the cool peripheral tissues (6). The ΔH value of rHb AE is –30.6 kJ mol-1 in phosphate buffer, which is more negative than that of rHb WM. However, it is reduced by the addition of IHP (–15.9 kJ mol-1), a value only slightly higher than that of rHb WM (–10.4 kJ mol-1). This suggests that a lower ΔH value observed from rHb AE might be attributed to the stronger allosteric effect of IHP, and that the reduction in the ΔH of rHb AE is less pronounced in the presence of 2,3-BPG (6).
In the stripped condition, the ΔH value for Hb A (–46.1 kJ mol-1 at pH 7.0) is more negative than that of rHb WM (–25.9 kJ mol-1), and it is only slightly increased by 2, 3-BPG in the presence of chloride ions (5, 39). In our study, when Hbs are saturated with IHP, the ΔH values for Hb A are significantly increased to the same range as that of rHb WM and rHb AE, suggesting that the O2 affinity and the ΔH value of Hbs are dependent on the property and the concentration of the allosteric effectors. It could also be the case for Hbs in the blood cells. For example, Asian and African elephant blood have slightly different functional properties, with the O2 affinity of Asian elephant blood being lower than that of African elephant due to an increased 2,3-BPG concentration in the red cells (40-41).
1H-NMR spectra of Hbs in 0.1 M sodium phosphate buffer at pH 7.0 were assessed at 11, 29, and 37 °C. One significant change shown in the 1H-NMR spectra is that the peak at 12.2 ppm disappears in the CO and deoxy-forms of rHb WM and rHb AE (Figure 7). In Hb A, this resonance has been assigned to the side chain Nε2H group of α103His, which is hydrogen-bonded to β131Gln (16-19). The resonance at 12.9 ppm has been assigned to the side chain Nε2H group of α122His (18), which forms a water-mediated H-bond with the side chain of β35Tyr. Both H-bonds are located at the α1β1-subunit interface of Hb A. Our previous study of the mutant rHb (β131Gln→Glu) shows that the peak at 12.2 ppm in the 1H-NMR spectrum disappears, suggesting that the H-bond between α103His and β131Gln does not exist due to the replacement at the β131 position (42). In rHb WM and rHb AE, the amino acid residue of α103His is the same as that in Hb A, but β131Gln existing in Hb A is changed to β/δ131Glu (Table 1). Thus, the disappearance of the resonance at 12.2 ppm in the 1H-NMR spectra of rHb WM and rHb AE is likely to be caused by this mutation, as reported for the mutant rHb (β131Gln→Glu) (42). Our previous studies for rHb (β131Gln→Glu) also show a slight downfield shift of the α122His resonance to 13.1 ppm, indicating that the perturbations caused by the replacement at β131 are not just confined to the mutation site, but also affect the environment of α122His (42). This also could be true for the case of rHb WM and rHb AE. Thus, the changes in 1H-NMR spectra on the interface histidyl resonances of α103His and α122His could be attributed to the replacement of β131Gln→Glu in rHb WM and rHb AE. In the deoxy state of rHb WM and rHb AE (Figure 7C), the T-state markers located at 14.2 ppm and the peak of β37Trp at 11.1 ppm are shifted downfield, indicating a perturbation of the α1(β/δ)2 interface as well (Figure 7). Thus, in rHb AE and rHb WM, both α1(β/δ)1 and α1(β/δ)2 interfaces are perturbed in both the deoxy- and CO forms in the absence and presence of IHP as compared to those of Hb A and Hb A2. From the previous functional studies of the mutant rHb (β131Gln→Glu), it was shown that this mutation has only a small impact on the oxygen affinity of the Hb molecule (42). Thus, although a significant change is observed in the 1H-NMR spectra, the replacement of β131Gln→Glu might not have an important effect on the function of rHb WM and rHb AE.
On the basis of the previous studies of Hb A, the non-exchangeable ring current-shifted proton resonances at –1.75 and –1.82 ppm have been assigned to the γ2-CH3 group of the α62Val and β67Val of HbCO A, respectively (15, 20, 21). These resonances provide information about the geometry/environment of the heme pocket. The ring-current-shifted resonances for these two valyl residues are resolved in the CO form of Hb A and Hb A2, but both shift upfield to –2.0 ppm in the CO form of rHb WM and rHb AE (Figure 7D). These results suggest that the distal heme pockets of rHb WM and rHb AE are altered as compared to those of Hb A and Hb A2. Comparing the amino acid sequences of Hbs, several replacements in rHb WM and rHb AE can be found near α62Val and β67Val, including α63Ala→Gly, α64Asp→Glu, β65Lys→Glu, and β69Gly→Thr. Previous studies have indicated that the mutations in the heme pocket of Hb A can change the O2 affinity by affecting the tertiary structure of the protein (34, 43-44). Our NMR studies reported here provide direct evidence showing that the geometry/environment of the heme pockets of rHb WM and rHb AE is different with respect to those of human Hbs, and that these changes may be related to the higher O2-binding affinity of these Hbs.
Hb A2 exhibits a remarkable structural similarity to Hb A. The α-chains are the same in these two Hbs, and there are only 10 amino acid substitutions in the δ-chain of Hb A2 compared to the β-chain of Hb A. On the basis of the ring-current-shifted resonances of E11Val, no significant change in the distal heme pocket was detected due to the δ-chain replacement (Figure 7D). In the 1H-NMR spectra of Hb A2, the resonance at 13.1 ppm shows a slight downfield shift, similar to that of rHb WM and rHb AE, while the resonance at 12.2 ppm is unchanged, suggesting that the δ-chain replacement in Hb A2 conserves the αlδ2 interface, but only slightly perturbs the αlδ1 interface (13). The perturbation on the intra-dimer interface is shown more clearly in the 1H-NMR spectra of rHb WM and rHb AE due to the combination effect of the β/δ-chain replacement and the mutation of β/δ131Gln→Glu.
Perutz and Raidt reported that Hb A2 was more resistant to thermal denaturation than Hb A (45). They suggested that two amino acid residues in the helices G (δ116Arg) and H (δ126Met) of Hb A2 might be responsible for its higher thermal stability. It is speculated that δ116Arg could make an extra hydrogen bond at the α1δ1 interface and δ126Met (H4) could make intra-subunit non-polar contacts with δ11Val of helix A (45). However, a recent X-ray crystal studies of Hb A2 did not support these additional interactions, but non-polar contacts formed between helices A and H were found, which presumably contribute to the higher thermal stability of Hb A2 (10, 46). Notably, rHb WM and rHb AE, also possess Arg at position 116 of their β-type chains (Table 1), which may similarly elevate the thermal stability of these two Hbs. The X-ray structural studies also suggest that the higher O2-binding affinity of Hb A2 may result from the slighter larger α1δ2 interface of Hb A2 and an additional hydrogen bond at the α1δ2 (or α2δ1) interface between α94Asp and δ37Trp in Hb A2 (10). This could also be true for rHb WM and rHb AE.
It is important to understand how the oxygen transport of Hbs is controlled by the combined action of temperature and ligands. Our studies on rHb WM and rHb AE have confirmed that these two Hbs respond differently to changes in experimental conditions. In general, the O2 affinity of rHb WM has a larger response to allosteric effectors and is less sensitive to temperature change than rHb AE. These features are related to its structure and appear to arise from at least two of the three mutations found on the β/δ chain of this extinct species. It should be noted that our biochemical-biophysical study is limited to only the Hb molecule. In red cells, the factors that affect the oxygen affinity are more complex. Thus, the negative influences, such as a decrease in oxygen unloading due to a lower temperature, may be compensated by the action of one or more other factors. For example, it has been shown that blood of African and Asian elephants possess slightly different functional properties, with the O2 affinity of Asian elephant blood being lower than that of African elephant blood due to an increased 2,3-BPG concentration in the red cells (40-41). It is possible that this subtle difference could also exist in woolly mammoth blood. The distinct structural features of rHb WM provide a part of the basis for woolly mammoth survival in the arctic environment. Further investigations are needed for applying these structural features to the design of a new generation of medically relevant Hb-based oxygen carriers.
We thank Dr. E. Ann Pratt for helpful comments on our manuscript.
†This work is supported by a research grant from the National Institute of Health (R01GM084614).
SUPPORTING INFORMATION AVAILABLE
Table 1S shows the Hill coefficient (n50). Table 2S and Figure 1S show the concentration and temperature dependence of the oxygen-binding properties (P50) of Hbs under various experimental conditions, respectively. This material is available free of charge via the internet at http://pubs.acs.org.