Vol. 49 No. 2/2002 459–470 QUARTERLY

Thermal stabilities of chicken, grey lag goose (Anser anser), turkey as avian hemoglobins (Hbs); and human, bovine, sheep and horse as mammalian Hbs in hemolysate form were investigated and compared with oxygen affinities taken from literature. The thermal stability was obtained from thermal profiles using temperature scanning spectrophotometry. The buffer conditions were 50 mM Tris, pH 7.2, and 1 mM EDTA. The average of the inverse temperature transitions, average hydrophobicity, total van der Waals volume, partial molal volume and hydration potential were calculated by computational methods. The hemolysed avian Hbs have a lower oxygen affinity, higher thermal stability and higher self association than the mammalian Hbs. These observations are based on amino-acid composition, influence of ionic effectors, and the presence of Hb D in several avian Hbs. The results indicate that the avian Hbs have a more tense (T) conformation than the mammalian Hbs.

The respiratory system of birds differs from that of mammals; whereas their hemoglobins are functionally similar.Oxygen affinities of mammalian and avian hemoglobins have been studied by a number of investigators who have found the presence of inositol pentaphosphate (IP 5 ) in the birds blood in-stead of 2,3-diphosphoglycerate, DPG, found in mammalian blood (Lenfant et al., 1969;Christensen & Dill, 1965;Bartels et al., 1966;Tuker, 1972;Knapp et al., 1999).The agent that is generally used to study the hemoglobin-phosphate interactions in vitro is inositol hexaphosphate (IHP) (Rollema & Bauer, 1979;Brygier et al., 1975;Coletta et al., 1993;Messana et al., 1998).Cho & Choy (1980) studied the thermal stabilities of various methemoglobin and metmyoglobin derivatives in the presence of different ligands and at different pH values by differential scanning calorimetry (DSC).As a result, they concluded that the thermal denaturation of Hb derivatives could be described by an activated two-state process and determined corresponding parameters of activation.The stability of normal and cross-linked hemoglobins was studied on the assumption of a two-state model for denaturation of Hb derivatives (Yang & Olsen, 1988).The relationship between thermal stability and hydrophobicity has been studied very extensively (Jiang et al., 2001;Irback & Sanderlin, 2000;Wagschal et al., 1999;Kumar et al., 2000).A number of researchers believe that this relationship should be the method of choice for studying the hemoglobins of birds and mammals (Irback & Sanderlin, 2000;Petruzzelli et al., 1996;Bull & Breese, 1973).For example, Bull & Breese (1973) have suggested a relationship between stability and hydrophobicity of hemoglobins from animal sources.It has been suggested that the stability of proteins at high temperature (60-70°C) is due to enhanced hydrophobic bonding interactions (Bull & Breese, 1973).Also Bigelow & Channon (1976) believed that the hydrophobic character of proteins could be the determining factor of their stability and proposed a method for the calculation of average degree of hydrophobicity of thermophilic proteins.Once the method was applied, they found the hydrophobicity of the thermophilic proteins was by 10% higher than that of normal proteins.It should be noted that the correlations between quaternary structures and stabilities and functions of proteins were also studied by the accessible and the buried surface area calculations (Chothia et al., 1976).Hydrophobic interactions affect also polymerization of biomolecules (Urry et al., 1992;Adachi et al., 1987;Adachi et al., 1993;Yohe et al., 2000).
The role of these interactions in increasing the association or polymerization of Hb S (Harrington et al., 1997) and lamprey Hb (Andersen & Gibson, 1971) has been already investigated.Increments in phosphate concentration and temperature were found to increase the rate of polymerization (Yohe et al., 2000).
Total free energy of unfolding is obtained by calculating the sum of two energy contributions, i.e conformational and hydration energy.Hydration affects the empirical conformational energy values.The free energy of hydration is composed of additive contributions of various functional groups of proteins.The hydration of each group is assumed to be proportional to the accessible surface area of the group (Ooi et al., 1987).The energy of unfolding is considered as a measure of protein stability (Jiang et al., 2001;Ooi et al., 1987;Koehl & Levitt, 1999).Koehl & Levitt (1999) have developed a protein design procedure that optimizes whole sequences for a given target conformation based on the knowledge of template backbone and on semiempirical potential energy function.This energy function is purely physical including steric interaction based on Lennard-Jones potential, electrostatic interaction based on a Coulomb potential, and hydrophobicity in the form of an environment free energy based on accessible surface area and interatomic contacts.The constants of proportionality, representing the free energy of hydration per unit area of accessible surface, have been evaluated for seven classes of groups (Koehl & Levitt, 1999).
In this study we have compared the stabilities and functions of several mammalian and avian hemoglobin samples by using various experimental and computational methods.
Hemoglobin samples from fresh blood of humans, bovine, sheep, horse and greylag goose were obtained from the Blood Transfusion Organization, Slaughter House, and The School of Veterinary Medicine, University of Tehran, respectively.

Preparation of Hb
The heparinized blood samples were washed three times with 0.9% NaCl.After each wash, the sample was centrifuged at 3 000 r.p.m.Afterwards, the packed cells were lysed with 5 portions of double distilled water, and the sample was centrifuged at 10 000 r.p.m.Subsequently, 5% (w/v) NaCl was added to the sample and the mixture was centrifuged at 18 000 r.p.m.All the steps were carried out at 4°C, and each centrifugation step was carried out for a period of 15 min.The method of Williams & Tsay (1973) was applied.All samples were dialyzed against a buffer: 50 mM Tris, pH 7.2, and 1 mM EDTA.
The dialyzed samples from chicken were purified by chromatography in 4.5 ´10 cm DEAE Sephadex A-50 (Cobb et al., 1992).The major component of Hb A was eluted immediately and then Hb D was eluted with the same buffer solution containing 0.03 M NaCl.After separation of Hb A and Hb D, each was concentrated by ultrafiltration through the Amicon filter.The chicken hemoglobin was separated from the organic phosphate by gel filtration on Sephadex G-25 in 0.1 M Tris/HCl, pH 8.5 and 0.5 M NaCl (Rollema & Bauer, 1979).

Spectrophotometry
The Gilford spectrophotometer (model 2400-2) was used for obtaining the thermal profiles of hemoglobin solutions, using wave-lengths of 280 and 700 nm for the conformational and the turbidity studies, respectively.The scan was run at a rate of 1 K/min in a degassed buffer solution.

Theoretical calculations
The single point energy was determined by the AMBER force field (scale factor 4) implemented in Hyperchem 5.02 version.Data for the human (PDB file is 1hho.pdb)and the bar-headed goose (PDB file is 1a4f.pdb)Hbs were used (Zhang et al., 1996).The conformational energies (E) in folded and unfolded or extended forms ((f, y, w) = (-155 o , 160°, 180°)) were obtained for the dimers and monomers.The hydration free energy, DG h, , was calculated by using the group hydration (g i ) values (Table 3) taken directly from (Ooi et al., 1987) for the corresponding Hb conformation.The difference between the dimers (in folded conformation) and the monomers (in unfolded conformation) as well as the differences between the hydration energy, DG h t , and the conformational energy, DE t, , were calculated as follows: Where, f, u, a, b, ab represent the folded and unfolded conformation, monomers and dimer subunits, respectively.The accessible surface area (ASA) was computed based on the methods of Shrake & Rupley (1973).This method was implemented in the Fantom (http://www.scsb.utmb.edu)and Naderi-manesh and coworkers (Naderi-manesh et al., 2001) programs.The total energy of unfolding was obtained as follows: (2) The buried surface area, DASA, for the oligomers is defined as the sum of the accessible area of the monomers minus that of the dimer, and correlates with the stability.The gap volume is a void space in the interface between the subunits.It was calculated for the same PDB files of human oxy Hb and the bar-headed goose Hb using the web site (www.biochem.ucl.ac.uk).
All of the physical parameters for proteins such as average hydrophobicity, Hf, average inverse transition temperature, <T t >, total van der Waals volume, V t , total hydration potential, HP t , and average partial molal volume, <v>, were calculated by using amino acid parameters, i.e. hydrophobicity index, HI (Argos et al., 1982) inverse of transition temperature, T t , (Urry et al., 1992), (T t º T (g) is the temperature at which the turbidity measured at 700 nm reaches the half-maximal value), van der Waals volume, V v (Chalikian et al., 1996), hydration potential, HP (Argos et al., 1982), and partial specific volume, u (Prakash, & Timasheff, 1985), respectively.The cited average parameters were calculated as the sum of the parameters for amino acids (Table 2) divided by the number of amino acids.

RESULTS
Figure 1 shows the denaturation profiles at 280 nm for the oxyhemoglobin samples.It points to the higher melting temperature, T m , considered as a stability criterion, for the avian as related to the mammalian Hb samples.Figure 2 shows the direct relationship between T m of hemoglobins in buffer condition and p 50 of mammalian and avian blood, taken directly from literature and collected in Table 1.The data indicate that lower oxygen affinity is consistent with the higher stability of hemoglobins.The hydrophobicity, Hf, of 33 hemoglobins (Hb A), downloaded from (Swiss-Prot) data base, was calculated as the sum of the amino acids hydrophobicity index, HI, which is tabulated in Table 2. Figure 3 shows the average hydrophobicity, Hf, and the average inverse transition temperature, <T t >, for the avian and the mammalian hemoglobins.This Figure also points to a higher hydrophobicity and a lower <T t > for the association criterion of the avian as compared with the mammalian Hbs.The variation of absorbance at 700 nm as a turbidity parameter versus temperature and (DTAB) (Fig. 4) points to a higher turbidity of the avian as compared to the mammalian Hbs.The experimental and calculated parameters for avian and mammalian hemoglobins are listed in Table 1.
The conformational energy, DE t , for the bar-headed goose and human Hbs were obtained by theoretical calculations.The free energy of unfolding, DG u , is the sum of the hydration and conformational energies (Ooi et al., 1987).The difference between the dimer and the monomer hydration energies were obtained from the group hydration data taken from Ooi et al. (1987) (see Table 3).The conformational, DE t , unfolding, DG u , electro-static and the van der Waals energies of the samples are tabulated in Table 4.
The results show that the stability parameter, DG u , has a higher value for birds (e.g.goose) than for mammals (e.g.humans).In general, a higher conformational stability for dimer goose Hbs, from our experiment, could be deemed as a characteristic feature for the avian compared with the mammalian Hbs.

DISCUSSION
Our results indicate a reversed relationship between the oxygen affinity and the melting temperature, T m , i.e., an increase in the stability is directly related to a decrease in the oxygen affinity (Fig. 2).In our experiments, we have used seven hemoglobin samples (avian and mammalian) in hemolysate form.Our data show that higher temperature Hbs stability in birds is related to lower oxygen affinity.The higher stability of avian as compared to mammalian Hb is believed to be due to the differences in their amino-acid composition (Kumar et al., 2000), the presence of ionic effectors (Rollema & Bauer, 1979;Brygier et al., 1975) and/or the influence of Hb D (Knapp et al., 1999).
Avian Hbs contains more hydrophobic amino-acid residues than mammalian Hbs (Fig. 3).This observation is consistent with the presence of a higher content of hydrophobic residues in avian hemoglobin, its higher thermal stability and attainment of a tense (T) state.The conservation of hydrophobic domains in proteins such as the avian Hbs might have, in fact, required the stabilization of tertiary structure, in order to maintain the function of the protein through a long period of evolution (Perutz, 1983).
It has been reported that an extra positive charge in the cavity site of avian Hb (Moss & Hamilton, 1974) led to a stronger binding of ionic effectors in birds than it has been observed in mammals (Brygier et al., 1975).It led also to a higher R®T transition for the avian Hb.In fact, the ionic effectors, which for mammals and birds are DPG and IP 5 , respectively, exhibit a stronger binding in deoxy forms leading to decreased Hbs oxygen affinity.The number of negative charges in IP 5 seemed to be higher than in DPG.As a result, the binding constant for IP 5 with Hb is higher than that for DPG (Rollema & Bauer, 1979;Brygier et al., 1975).It has been noted that the ionic effectors increase the Hill coefficient (Knapp et al., 1999), tetramer association  (Argos et al., 1982); T t , inverse of temperature transition (Urry et al., 1992); V v , van der Waals volume (Chalikian et al., 1996); HP, hydration potential; v (ml/g), partial molal volume (Prakash & Timasheff, 1985); a Values estimated from a direct correlation between van der Waals volume and partial molal volume.(Ooi et al., 1987).(Yohe et al., 2000) and stability of Hb (cf.data for IHP in Table 5).Knapp and coworkers (Knapp et al., 1999) have reported that the presence of Hb D component in Hb from several birds induces a lower oxygen affinity and that formation of a large tetramer-tetramer interface in chicken Hb D mediates a shift in the helix of ab subunits in such a way that the distal histidine is pushed further into the heme pocket, as it was also observed in lamprey deoxy Hb.Since the resulting position of the distal histidine hinders the oxygen binding, the oxygen affinity is reduced.Avian Hbs have a lower oxygen affinity than the mammalian Hbs.One reason for the reduction of oxygen affinity in the birds is the presence of Hb D. Our results also indicate that Hb D has a higher stability than Hb A (see  ab, a, b, refer to dimer and monomers, respectively.Human (H), bar-headed goose (G) oxyhemoglobin E v , E e , f and u, are van der Waals energy, electrostatic energy, folded and unfolded states, respectively.The energy of unfolding of the bar-headed goose hemoglobin was calculated because of the existence of the three dimensional data in Protein Data Bank and the high homology (%) relative to graylag goose hemoglobin.
Table 5).We may conclude that a lower oxygen affinity of hemoglobins corresponds to their higher stability (both in birds and mammals).Chothia and coworkers (Chothia et al., 1976), reported that the T state conformation was more stable (5 kcal/mol) than the R state long before our findings regarding the presence of forms of the T state in the avian Hbs (Chothia et al., 1976).This leads us to believe that the structural conformation of avian Hbs has a higher content of T state (deoxy form) in relation to R state (oxy form).
Based on the foregoing facts, we can conclude that the higher hydrophobicity of avian Hbs is a factor that contributes to higher stability of the T structure, nearly 20 kcal/mol (Chothia et al., 1976).Accordingly, the avian Hb may have a higher T/R state ratio than the mammalian Hb.
The gap volume between ab subunits was found to be higher in human Hb than in the bar-headed goose Hb.Table 1 also shows a higher V t for the avian Hb.Partial specific volume was also determined from the amino-acids constituents according to Zamyatnin The protein concentration for thermal profiles and DTAB curves was 6 mg/ml and 2 mg/ml, respectively.Chicken (o), turkey (m), goose (D), human (s), bovine (l), horse (n), sheep (6) hemoglobins.1972).Since the partial specific volume is associated with average hydrophobicity, Hf, and the compressibility factor is dependent on the partial molal volume v, accordingly the empirical formula can be obtained (Gekko & Hasegawa, 1986) as follows: Thus we can conclude not only that a higher Hf value corresponds to a higher partial specific volume, but we can also conclude that the partial molal volume of deoxy (T-state) is higher than that of the R state (Bureau & Banerjee, 1976).Thus, the higher stability of protein is probably related to a higher partial specific volume.It has been found that the partial specific volume has both a direct and a reverse relationship to the areas of the nonpolar and polar surfaces, respectively (Chalikian et al., 1996).Also our study indicates that the total hydration potential (HP t ) has a direct relationship with the polar surfaces, and it has been determined to be higher for the avian Hb.This fact leads to the conclusion that the avian Hbs have a lower tendency to be hydrated (Kharakoze & Sarvazyan, 1993).The buried surface area of the oligomers is associated with their stabilities.Accordingly, the calculated DASA values of alpha-beta dimers for the bar-headed goose Hb are higher than for the human Hb (Table 5).
The buried surface areas are apparently higher in T state than they are in R state (Chothia et al., 1976).Janin & Chothia (Janin et al., 1988;Janin & Chothia, 1990) reported twenty three oligomeric proteins with high stabilities which were assumed to be associated with the higher buried surface areas.It has been suggested that the stability of proteins at high temperatures (60-70°C) is due probably to increasing hydrophobic bondings in the protein (Goldsack, 1970).It is important to note that, in addition to hydrophobic interactions, the electrostatic and van der Waals forces also play substantial roles in the stabilization of protein structures (Chothia et al., 1976).Table 5 shows the calculated values for each energy contributing source for two Hb samples (bar-headed goose, human).Another reason for a higher T/R state form for the avian Hb in relation to mammalian Hb is a higher turbidity factor and self association of proteins which is believed to be due to a higher hydrophobicity.Figure 4 shows the variation of the absorbance at 700 nm with temperature and the effect of DTAB as a surfactant factor.We have assumed that these changes are associated with the degree of turbidity.In this case, the protein association seems to be higher for the avian Hb.It is believed that the surfactant either increases the surface hydrophobicity or, may be, neutralizes the surface charges of the hemoglobin of birds, so that this surface hydrophobicity is intrinsically higher in the avian Hb than it is in mammalian Hb.The literature data indicate that the increment of the surface hydrophobicity raises the association and polymerization of proteins such as Hb S (Harrington et al., 1997), and Hb of lamprey (Andersen & Gibson, 1971).These Hbs become more associated in the deoxy form (T state) because the surface hydrophobicity in this state is probably higher than that in the oxy forms.Accordingly, Hb D induces a turbidity and polymerization in the avian Hb.The results of differential scanning calorimetry (not shown) at three scan rates (2, 1, 0.5 K/min) based on the method of Cho & Choy (1980) gave identical results for three Hb samples.This finding leads us to believe that the kinetics of association of the avian Hb (including Hb D) is biphasic.This phenomenon has been also observed for Hb S (Reddy et al., 1996;Mu et al., 1998).We have also used T t to estimate the magnitude of self-association of Hbs.T t depends on the hydrophobicity of the constituent amino-acid residues (Urry et al., 1992) and seems to be lower for the hydrophobic amino acids.Once the value of T t is above that for the physiological temperature, the polypeptide or protein will be unfolded (disassembled); and once T t is below that for the physiological temperature, the polypeptide or protein will be folded (assembled).The calculated average T t , <T t >, given in Fig. 3 indicates that avian Hbs proteins characterized by higher hydrophobicity have lower <T t > values.Accordingly, the higher hydrophobicities of avian Hb have a lower <T t >.

Figure 3 .
Figure 3. Hf (s) and <T t > (l) values for 33 hemoglobins (from birds and mammals) calculated from amino-acid parameters, as described under Methods.
Figure 4. Variation of absorbance at 700 nm versus temperature for oxyhemoglobin as a measure of turbidity, and versus DTAB concentrations (inset).

Table 1 . Experimental and calculated parameters for avian and mammalian hemoglobins
a p 50 , pressure at half saturation of hemoglobin directly taken from literature

Table 2 . Parameters for amino acids used in this work
HI, Hydrophobicity index

Table 3 . The computed g coefficient for free en- ergy of hydration a
a Data are taken from