Studies on Hemoglobin from the Hagfish Eptatretus

SUMMARY Ultracentrifugation studies on hemolysates from the erythrocytes of the hagfish, Eptatretus burgeri, showed that in dilute solution oxygenated hemoglobins were in a monomeric form with a molecular weight of approximately 18,000. In deoxygenated hemolysates or in concentrated solutions of oxygenated hemolysates, an aggregate product appeared which was considered to be a tetramer, as judged from its sedimentation coefficient. Contrary to a previous report, the hemolysate of the hagfish showed weak but significant heme-heme interaction (n = 1.2) depending on its hemoglobin concentration and pH. In the hemolysate, four hemoglobin components were found by electrophoresis, and other hemoglobins, if any, were very small in amount. The four components

where ij is the partial specific volume (0.743 was used), p the solution density, and w the angular speed of rotation. Oxygen Equilibria-The method of Asakura et al. (12) was used with minor modifications.
Spectra between 500 and 700 nm were obtained at 22" with a Hitachi EPR-2 recording spectrophotometer.
Fractional saturation was calculated from extinctions at two wave lengths, the P-band and the minimun between the o(-and P-bands, to cancel small base line shifts. When concentrated solutions of hemoglobin were used, spectra between The value of n, the slope of the line relating log y/(1 -y) to log p, where y is the fractional saturation and p is the oxygen pressure, was estimated at SOY6 oxygenation.

Molecular
Weight and Heterogeneity of E. burgeri Hemoglobin-The molecular weight of E. burgeri hemoglobin was estimated by sedimentation equilibrium.
Plots of log f versus r2 for the oxygenated hemolysate are shown in Fig. 1. The molecular weight of oxygenated hemoglobin was calculated by means of the slope of the straight line of Fig. 1, and the value of 18,000 was obtained.
Where log f is larger than 0.9, meaning that the protein concentration was higher than 0.18 g per dl, the slope became steeper, probably because of aggregation of the protein.
The ratio of heme to protein was determined.
The concentration of heme was measured as described under "Materials and Methods," and protein was estimated by the method of Lowry (13), with human hemoglobin as a standard. A ratio of 19 mg of protein per micromole of heme was found, indicating that a single heme prosthetic group was present on each molecule. In the hemolysate which was obtained from more than 30 hagfish, at least four types of hemoglobin were found by electrophoresis (Fig. 2~). We designated these four hemoglobins as Fl, F2, F3, and F4, according to the order of electrophoretic mobility from the origin. A very faint band was also found between Fl and F2. Hemolysates of four hagfish which were studied individually showed almost the same electrophoretic pattern as seen in Fig. 2a, but among them, two hemolysates were completely lacking in the band between Fl and F2. Thus, it seems reasonable to state that E. burgeri commonly possesses at least four differeut hemoglobin molecules and some individuals have another hemoglobin, which is present in very small amounts.
The relative amounts of these four hemoglobins were estimated by densitometric measurements, and the ratio of 17:13:35:35 for Fl:F2:F3:F4 was obtained. Fractionation of E. burgeri IZemoglobins-Hemoglobin solution of about 1 mM heme in 5 InM Tris-HCl, pH 8.0, was applied on a DEAE-cellulose column (2 x 30 cm) previously equilibrated with the same buffer and eluted with a linear gradient of KC1 from 0 to 0.1 M in 5 mM Tris-HCl, pH 8.0. As shown in Fig. 3, E. burgeri hemoglobin was separated into four peaks. Electrophoresis showed that the first and the second peaks contained only Fl and F2, respectively.
The third peak, however, was a misture of F2 and F3 and the fourth was a mixture of F3 and F4.
Components other than these four were not found in these four peak fractions.
To separate F3 from F2, and F4 from F3, the third and the fourth peak fractions were collected and con centrated.
They were dialyzed against 0.005 M phosphate buffer, PI-I 7.0, and applied to hydroxylapatite columns (1.2 X 5 cm) equilibrated with 0.005 M phosphate buffer, pH 7.0. Hemoglobins were separated by elution with a linear gradient from 0.005 to 0.1 M phosphate buffer, pH 7.0. As shown in Fig. 4, F3 was successfully separated from F2. The separation of F4 from F3 was not complete, but the fractions eluted later than No. 31 in Fig. 4b were essentially free from F3, and pure F4 was obtained by collecting these fractions. Electrophoretic patterns of separated components are shown in Fig. 2, b to e. Arnillo acid analyses (Table I)  In this way, it was easily removed from the hemolysate.
The remainder, consisting of F2, F3, and F4, was analyzed in both the oxy-and deoxygenated form by ultracentrifugation.
The amounts of F3 and F4 in this preparation were nearly equal, as can be seen in Fig. 2h, and their sum accounted for 85% of all hemoglobins, whereas F2 was estimated to be only 15%.
Although the schlieren pattern of the oxyhemoglobin in concentrated solution showed two peaks corresponding to that of monomer and tetramer, the deoxyhemoglobin gave only one asymmetric peak, whose S value was higher than 4.0 and seemed to be the same as that of a tetramer.
It was seen that the slowly moving peak of the deoxygenated hemolysate disappeared when Fl was removed.
These The sedimentation coefficients of isolated Fl, F2, F3, and F4 hemoglobin were also measured, and the results are summarized in Table III. The values for Fl, F2, and F4 were all approximately 2.0 S in both the oxy-and deoxygenated states, whereas that of F3 was significantly higher than 2.0 S, especially in the deoxygenated state.
These results indicate that Fl, F2, and F4 did not aggregate by themselves but existed in a monomeric form both in the oxy-and deoxygenated states, whereas F3 was in an 6 Two components were mixed in equal amount. equilibrium state of association-dissociation, probably monomerdimer, and this equilibrium was changed by the binding of oxygen. As shown in Table III, it is clear that the mixture of F3 and F4 formed a tetrameric molecule, but there seems to be no interaction between F2 and F4, and between F2 and F3. Because F3 aggregated by itself, the relatively high S value of 2.7 S for the mixture of F2 and F3 did not necessarily imply an interaction between them.
It is possible that F2 takes part in the formation of the tetramer in such a form as (F2F3F4F4) or (F2F3F3F4).
However, we did not study this possibility because of the experimental difficulties involved. b Each fraction was mixed in equal amount.

Functional
Aspects of E. burgeri Hemolgobins-Oxygen equilibria of E. burgeri hemoglobins were measured. The oxygen pressure at 50% oxygenation, pg, and n, the slope of Hill's plot, were obtained and are summarized in Table IV. It was found that E. burgeri hemolysate showed weak but significant hemeheme interaction at high hemoglobin concentrations and at pH values below 7.0. A slight Bohr effect was observed.
Oxygen affinity also changed, depending upon the concentration of hemoglobins.
These results strongly suggest molecular interactions of E. burgeri hemoglobins.
When we used the hemoglobin preparation lacking Fl, whose properties were described in the preceding section, the heme-heme interaction and Bohr effect became more remarkable (Table IV). DISCUSSION Hagfish, E. burgeri, were found to have at least four different hemoglobin molecules. All four individual hagfish tested showed similar electrophoretic patterns. Although another hemoglobin was found in two individuals, its amount was very small.
Ohno and Morrison (7) found that the hagfish, E. stoutii, captured at the west coast of the United States had four to sis hemoglobin phenotypes, and they postulated that these hemoglobins were controlled by genes at four loci. The relationship between the hemoglobins of these two species of hagfish is uncertain, but E. burgeri, captured at the Pacific coast of Japan, seems to be genetically more hemogeneous.
In Table IV, the functional properties of isolated Fl, F2, F3, and F4 are also shown.
All of them showed no significant hemeheme interaction.
Oxygen affinity of Fl and F2 were nearly equal and very low as judged from pg, whereas F3 and F4 exhibited relatively high affinity. It is interesting that the monomeric hemoglobins, Fl and F2, showed such a low affinity.
In previous reports, the heme-heme interaction and Bohr effect of hagfish hemoglobin were described as extremely slight or undetectable (5,6) and their sedimentation coefficient was approximately 2.3 S (1, 2). However, these must be revised because of the fact that hagfish hemoglobins are heterogeneous and their components are different from each other both in structure and function. Oxygen equilibria of the combination of two fractions from F2, Sedimentation behavior of the hemolysate from E. burgeri n-as F3, and F4 were investigated.
As shown in Heme-heme interaction was evident much more when Fl was removed from the hemolysate. Fl did not aggregate with any other hemoglobins and remained in a monomeric state under the experimental conditions. Hemoglobins of E. burgeri studied here can be divided into two classes; those of the first exist always as monomer, and those of the second aggregate with one another to form tetrameric molecules.
Fl belongs to the first class, and although the precise aggregation propreties of F2 are still obscure, F2 probably belongs to t'his class also.
There is close resemblance between Fl and F2 in their amino acid compositions, sedimentation behaviors, and functional properties. Chromatographic and electrophoretic properties distinguished them, probably because of their different surface charges due to a few amino acid replacements.
Their low oxygen affinity is interesting, for the oxygen affinity of all other vertebrate hemoglobins, including lampreys, is known to be relatively high when they are in a monomeric state.
Physiological meaning of these hemoglobins, however, is obscure at present.
The second class of E. burgeri hemoglobins consists of F3 and F4. The amounts of F3 and F4 in the hemolysate were nearly equal and the mixture of F3 and F4 formed tetrameric molecules and showed heme-heme interaction.
The tetramer of E. burgeri hemoglobin was a hybrid molecule, and all hemoglobin fractions in their isolated state neither formed tetramers nor had any significant heme-heme interaction. Sea lamprey, Petromyxon marinus, was shown to have six species of hemoglobin (14). Among them, Fractions III, IV, V, and VI aggregated in the deoxygennted state to form a self-tetramer or a hybrid tetramer (15). The self-tetramer showed heme-heme interaction (16). Snother lamprey, Entosphenus japonicus, has essentially one species of hemoglobin which is self-associated to form a tetramer.l Although in the deoxygenated form both E. burgeri and lamprey hemoglobin showed a marked tendency toward aggregation, in the oxygenated form lamprey hemoglobin hardly aggregated (4), whereas in E. burgeri the aggregate product of oxygenated hemoglobin is easily observed at high hemoglobin concentrations. This suggests that E. burgeri hemoglobin tends to aggregate more easily than lamprey hemoglobin. In this respect E. burgeri hemoglobin, F3 and F4, are more closely related to higher vertebrate hemoglobin than the lamprey's. It may be postulated that F3, which aggregated slightly by itself, corresponds to the /3 chain of human hemoglobin and F4 to the o( chain. Of course, there are differences between F3-F4 and higher vertebrate hemoglobins; heme-heme interaction of the latter are very extensive and their association-dissociation equilibria are more favorable to association in comparison with E. burgeri hemoglobins. These differences, however, may be in degree rather than in quality.
Although our experimental conditions were not physiological, it may be said that tetrameric hemoglobins both in oxygenated and deoxygenated form exist in E. burgeri erythrocytes where the hemoglobin concentration is more than 10 mar in heme.