Correlation between functional and structural changes of reduced and oxidized trout hemoglobins I and IV at different pHs A circular dichroism study Rosita Gabbianelli 1 , Giovanna Zolese 2 , Enrico Bertoli 2 and Giancarlo Falcioni 1 1 Dipartimento di Biologia M.C.A., Universita ` di Camerino, Camerino, Italy; 2 Istituto di Biochimica, Facolta ` di Medicina, Universita ` Politecnica delle Marche, Ancona, Italy Circular dichroism (CD) spectra of two major hemoglobin components (Hb), HbI and HbIV, from Oncorhyncus mykiss (formerly Salmo irideus) trout were evaluated in the range 250–600 nm. HbI is characterized by a complete insensitivity to pH changes, while HbIV presents the Root effect. Both reduced [iron(II) or oxy] and oxidized (met) forms of the two proteins were studied at different pHs, 7.8 and 6.0, to obtain information about the pH effects on the structural features of these hemoglobins. Data obtained show that oxy and met-HbI are almost insensitive to pH decrease, remaining in the R conformational state also at low pH. On the contrary, the pH decrease induces similar structural changes, characteristics of ligand dissociation and R fi T transition, both in the reduced and in the oxidized HbIV. The structural changes, monitored by CD, are compared with the peroxidative activity of iron(II)-Hb and met-Hb forms and with the superoxide anion scav- enger capacity of the proteins. Keywords: trout hemoglobin derivatives; hemoglobin per- oxidase activity; superoxide anion; circular dichroism; pH effect. The hemoglobin system of the Oncorhyncus mykiss (for- merly Salmo irideus) trout is made up of four electro- phoretically distinct components, two of which [trout hemoglobin (Hb)I (% 20%) and trout HbIV (60%)] repre- sent quantitatively a large fraction of the whole pigment. In the last years, the properties of these two major components (HbI and HbIV) have been investigated in considerable detail under various experimental conditions [1,2]. Their structural and functional characterization has indicated some striking differences between the two proteins that have been related to their different physiological role [2]. HbI is characterized by the presence of cooperative phenomena and complete absence of pH and organic phosphate effects, while in HbIV, oxygen affinity and cooperativity depend on pH and organic phosphates (Root effect). In air (pO 2 ¼ 155 mmHg) and at pH 7.8, iron(II) HbIV is entirely in the oxygenated form, while at pH 6.0 it is almost completely in the deoxy form, as a consequence of the Root effect [2]. On the contrary, iron(II) HbI is fully oxygenated at both pHs. Moreover, both hemoglobins are stable towards dissoci- ation even in the ligated form; the value of the tetramer– dimer dissociation constant is between 10 and 50 times less than that of human hemoglobin measured under similar conditions [2]. The main function of HbI is to assure the basic level of O 2 to active tissues, providing a normal oxygen supply in emergency, while a basic role of HbIV is to release O 2 against high hydrostatic pressure in the swim bladder. Our previous study by circular dichroism (CD) demon- strated large structural differences in HbI and HbIV ([3] and references cited therein), which are likely related to their different physiological roles as gas carriers. Trout HbIV and HbI are also characterized by different peroxidative activities [4,5], which are dependent from the pH of the medium and/or the iron oxidation state. In fact, it is known that the hemoglobin molecules present peroxida- tive properties [5–7], which may be important for the life span of red blood cells (RBCs), continuously exposed to extracellular and intracellular sources of reactive oxygen species (ROS), which are a potential cause of oxidative injury and could have a role in erythrocytes senescence [7]. The toxicity of H 2 O 2 , a reactive oxygen species involved in cellular injury under various pathophysiological condi- tions, is known to be enhanced in the presence of hemoglobin [8]. H 2 O 2 binds to and reacts with Hb, generating the highly reactive ferrylhemoglobin intermedi- ate, which in turn oxidizes the substrate [6]. The final products of the reaction are superoxide radical and met-Hb. However, our previous studies on human hemoglobin demonstrated that the presence of Hb reduces the level of superoxide in the medium [7]. Met-Hb was shown to be more efficient in reducing the level of this radical with respect to oxyHb, and this difference was more marked at low pH values [7]. It is known that, during reversible oxygen binding, Hb undergoes a slow autoxidation to met-Hb, producing superoxide, which is released into the heme pocket [9] The superoxide released in the heme pocket reacts with globin, producing a secondary radical [9]. Correspondence to R. Gabbianelli, Dipartimento di Biologia M.C.A., Universita ` di Camerino, Via Camerini 2–62032 Camerino (Mc), Italy. Fax: + 39 073 7636216, Tel.: + 39 073 7403208, E-mail: rosita.gabbianelli@unicam.it Abbreviations: HbI, hemoglobin I; HbIV, hemoglobin IV; ROS, reactive oxygen species; RBCs, red blood cells. (Received 23 December 2003, revised 15 March 2004, accepted 25 March 2004) Eur. J. Biochem. 271, 1971–1979 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04109.x As heme’s interaction with globin appears to be import- ant both for autoxidation [9] and H 2 O 2 binding [6], a systematic study of the heme–globin interaction, at different pH and Hb oxidative state [iron(II)- or met-Hb], could be important to increase the knowledge of the structural basis of these interactions. Our previous CD study [3] demonstrated that the heme– globin interaction in oxy-HbI and oxy-HbIV are quite different. The aim of the present paper will be to charac- terize the heme–globin interaction in iron(II)- and met- HbIV and HbI, at different pH values, by CD. These studies will be compared with the peroxidative and superoxide anion scavenger activities of these proteins. The CD spectroscopy will be particularly useful because this tech- nique permits the evaluation of the optical activity of heme proteins [10,11], which result from different kinds of heme interaction with the protein matrix. Three different wavelength regions (near-UV, far-UV and visible) can offer different degrees of information. These CD regions are largely used to study the tertiary and quaternary structure of heme proteins [10,11]. A direct comparison for differences in structure between trout HbI and HbIV in solution will increase the knowledge on the physiological roles of these Hbs and on the molecular adaptation mechanisms of these aquatic organisms living under particular environmental conditions. Materials and methods All reagents were of analytical grade. Preparation of trout hemoglobin components were performed as previously described [12]. Iron(III) hemoproteins were obtained by the addition of K 3 [FeCN) 6 ] (molar ratio 2 : 1) to the oxygen- ated derivative; excess oxidizing agent and ferrocyanide were removed by gel filtration through a Sephadex G-25 column eluted with 50 m M Tris/HCl pH 7.8 or 50 m M Bis/ Tris, pH 6.0. The hemoglobin concentration was deter- mined by the pyridine–hemochromogen method [13]. Circular dichroism CD spectra were recorded on a Jasco spectropolarimeter under nitrogen flux at 6 °C. In the near UV and Soret region the Hb concentration was 0.1 mgÆmL )1 , while in the range 490–670 nm the concentration was 0.5 mgÆmL )1 . Measure- ments were carried out in 1 cm path length quartz cuvettes. The molar ellipticity is always expressed, on a molar heme basis, as degreeÆcm 2 Ædmol )1 . Readings were carried out against a reference cuvette containing the same components without protein. Data were acquired at 6 °C, in order to increase Hb stability [14] and to mimic environmental living conditions of the trout. All spectra are the averages of four experiments, where three different recording were accumu- lated for each sample. Peroxidase activity assay The assay for peroxidase activity was performed as reported by Everse et al. [6] using guaiacol as substrate. Fifty millimolar sodium phosphate/citrate buffer (0.9 mL) at pH 5.4 and containing Hb and 10 m M guaiacol was used. The reaction was started by the addition of 157 m M H 2 O 2 solution (0.1 mL) and monitored by absorbance changes at 470 nm. The absorbance change was due to Hb-catalyzed oxidation of the substrate by hydrogen peroxide. Chemiluminescence measurements Chemiluminescence measurements were performed by lucigenin as chemiluminogenic probe, and superoxide radicals were produced by xanthine/xanthine oxidase sys- tem as previously described [15]. Briefly, the chemilumines- cence (CL) was measured in automatic LB 953 (Berthold Co., Wildbad, Germany) in a reaction mixture containing 0.9 UÆmL )1 xanthine oxidase, 40 lgÆmL )1 of hemoprotein and 150 lmolÆL )1 lucigenin in 1 mL of the chosen buffer. The reaction was started by the injection of xanthine at the final concentration of 50 l M andfollowedfor60sas previously described [16]. Values obtained are expressed as counts per second (c.p.s.). Results Circular dichroism L-Band (260 nm region) and 270–300 nm region. The 270–300 nm region of CD spectra was used to study changes in the Hb quaternary structure at the a 1 b 2 interface [3,17]. Within the near UV region (1250–300 nm) CD bands are due to aromatic amino acids, S-S bridges and heme groups [11], and are poorly characterized. Near this region, the L -band (centered around 260 nm) is considered to be sensitive to the interactions between the heme and the surrounding globin, being influenced by the attached ligand and thereby by the spin state of the iron atom [11]. According to Perutz et al. [18], the region around 285 nm is considered as indicative of the R fi T transition: in the T-form, the ellipticity is more negative than in the R-form [10,11,19]. The ellipticity change in this band is independent from the ligand state of heme, but it is indicative only of R fi T transition [10]. CD spectra acquired in the near-UV and Soret regions (250–320 nm and 320–470 nm, respectively) for iron(II)- HbI and iron(II)-HbIV at both pHs and in air are shown in Figs 1 and 2. In the range 250–320 nm (Fig. 1), CD spectra are baseline corrected to zero ellipticity at 320 nm, accord- ing to Henry et al. [20]. In the region 320–470 nm (Fig. 2), spectra are corrected to zero ellipticity at 470 nm. In line with our previous data [3] iron(II)-HbI (Figs 1B and 2B) and iron(II)-HbIV (Figs 1A and 2A) spectra in the range 250–470 nm show similar positive bands, resembling the dichroic characteristics of other oxy-hemoglobins [21]. In our experimental conditions, the ellipticity of the L band (centered around 260 nm) is directly related to pH (Fig. 1A,C) for both iron(II)-HbIV and met-HbIV. Only small changes between iron(II)- and met-HbIV, at both pH values, are evident (Fig. 1A,C). L-Band ellipticity is directly related to pH decrease also in iron(II)-HbI and met-HbI (Fig. 1B,D), although this effect is more evident for met- HbI. Following iron oxidation (Fig. 1D) almost no changes in this band were observed at pH 7.8 [comparing this value to the L-band of iron(II)-HbI at the same pH]. In both iron(II)- and met-HbIV, the change from pH 7.8 to 6.0 induces a shift towards a more negative ellipticity (correlatedtoRfi T transition) in the region of 285 nm 1972 R. Gabbianelli et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Fig. 1. Near-UV CD spectra of iron(II)-HbIV (A), iron(II)-HbI (B), met-HbIV (C) and met-HbI (D) in 50 m M Tris/HCl, pH 7.8, (solid line) and 50 n M Bis/Tris, pH 6.0 (broken line). Temperature ¼ 6 °C. Fig. 2. Soret CD spectra of iron(II)-HbIV (A), iron(II)-HbI (B), met- HbIV (C) and met-HbI (D) in 50 m M Tris/HCl, pH 7.8 (solid line) and 50 n M Bis/Tris, pH 6.0 (broken line). Temperature ¼ 6 °C. Ó FEBS 2004 Structural changes in trout hemoglobins (Eur. J. Biochem. 271) 1973 (Fig. 1A,C). For iron(II)-HbIV, this behavior is in agree- ment with the presence of the Root effect, where the pH decrease induces deoxygenation and transition to the T-form. However, met-HbIV, compared with iron(II)- HbIV at the same pH, always shows a slightly more negative ellipticity (Fig. 1A,C). In trout iron(II)-HbI, the pH decrease induces hardly any changes in the negative ellipticity value at 285 nm (Fig. 1B); while in met-HbI, at pH 7.8 (Fig. 1D) a slight increase in the negative ellipticity compared with the same protein at pH 6.0 is evident. On the other hand, met-HbI at pH 6.0 shows a similar ellipticity to iron(II)-HbI at both pHs. Visible region. The Soret region for both iron(II) and met- HbIV at pH 6.0 and 7.8 are shown in Fig. 2A,C. Changes in the CD Soret region (near 400 nm) were related to the interaction of the heme prosthetic group with the surround- ing aromatic residues and to modifications in the spatial orientation of these amino acids with respect to heme [22]. These modifications affect porphyrin transitions and p–p* transitions in the surrounding aromatic residues. However, the protein-induced heme distortions from planarity and the contributions of polarizable groups (near the heme) have been recently postulated to participate to the ellipticity in the Soret region [23]. According to some authors ([11,23] and references cited therein) a blue shift in the Soret band is a consequence of R fi Ttransitionanditreflects theinteractionbetweena 1 and b 1 subunits. It may be due to tertiary structural changes in regions including aromatic residues and it may be involved in the interactions between these subunits [24]. However, this bandissensitivealsotoliganddetachment(deligation)[17,19]. In the Soret region at pH 7.8, HbIV in the oxygenated- form is characterized by a band at 418 nm, in agreement with our previous data [3]. A similar band is present in other oxy-hemoglobins, such as human HbA [10,25]. A significant red shift (to about 433 nm) and a decreased intensity in this peak was measured in both iron(II)- (Fig. 2A) and met- HbIV (Fig. 2C), as a consequence of pH decrease, although the superimposition of two different bands in met-HbIV at pH 6.0 is evident (Fig. 2C). Compared with the iron(II) form at the same pH, met-HbIV shows a small blue shift in the Soret band (from 418 to 416 nm at pH 7.8, while the position of the peak measured at pH 6.0 is more difficult to calculate, due to its form) and an increased positive ellipticity (Fig. 2A,C). In iron(II)-HbI (Fig. 2B), the Soret band is also localized at 418 nm, in line with previous results [3]. In this protein, the pH lowering induces a slight decrease in the Soret band ellipticity, without wavelength shift (Fig. 2B). Met-HbI (Fig. 2D), compared with the iron(II)- form, shows a small blue shift and an increased ellipticity in this band, at both pH values. The two major peaks in the region 500–600 nm (Q 0 and Q m ) give indications on the constraints at the heme site and reflect the symmetric properties of the heme-iron bound material [11]. In particular, they are correlated to the asymmetry of the proximal bond. In fact, a decreased symmetry leads to an enhanced intensity in the Q 0 band [25,26]. The splitting of these bands has been regarded as a lowering of the heme symmetry in HbA-CO [17,27], due to nonperpendicular iron–ligand bond above the xy plane of the heme group. Fig. 3. Circular dichroism spectra in the visible region of iron(II)-HbIV (A), iron(II)-HbI (B), met-HbIV (C) and met-HbI (D) in 50 m M Tris/ HCl pH 7.8 (solid line) and 50 n M Bis/Tris, pH 6.0 (broken line). Temperature ¼ 6 °C. 1974 R. Gabbianelli et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Both iron(II)-HbI (Fig. 3B) and iron(II)-HbIV (Fig. 3A) at pH 7.8 (oxygenated), show very sharp Q 0 bands with similar intensities, while Q m shows a lower ellipticity in HbI. The spectra obtained are similar to those observed in many vertebrate hemoglobins and myoglobins [10,17,21,25]. The Q bands of iron(II)-HbIV (Fig. 3A) and -HbI (Fig. 3B) at pH 7.8 show some variations in relative intensity, suggesting small differences in the constraints at the heme pocket. In HbI, the spectrum is only slightly modified by pH lowering (Fig. 3B), while HbIV at pH 6.0 (Fig. 3A) shows a band centered around 545 nm and a shoulder around 572 nm (Fig. 3A). This spectrum is quite similar to that of deoxy human HbA [19], suggesting similar changes at the heme site in HbA and trout HbIV. Following oxidation, met-HbIV (Fig. 3C) shows spectra similar to iron(II)-HbIV at both pHs tested. These data indicate only small modifications at the heme site. On the contrary, when met-HbI is compared with iron(II)-HbI, it shows (Fig. 3D) a slightly decreased intensity in the Q 0 band (around 572 nm) with respect to Q m band (around 545 nm), suggesting a modified symmetry in the proximal bond. This behavior is similar at both pHs. Peroxidase activity The peroxidase activity was followed by monitoring the increase in absorbance at 470 nm. Guaiacol is a methoxy- phenol that oxidizes to a radical, followed by dimerization [6]. Figure 4 shows the peroxidase activity of iron(II)- and met-forms of HbI and HbIV. It is evident that the enzymatic activity decreases according to the order iron(II)-HbIV > met-HbIV > met-HbI > iron(II)-HbI. Chemiluminescence measurements Chemiluminescence (CL) measurements were performed by using lucigenin as chemiluminogenic probe for superoxide radical produced by xanthine/xanthine oxidase system. The reaction of lucigenin with superoxide radical gives rise to chemiluminescence whose level indicates the presence of superoxide in the medium under study. Table 1 shows results obtained on met- and iron(II)-Hbs at two different pH values, 7.8 and 6.0. The first parameter reported in Table 1 is the duration of the reaction, which was very different in the two buffers used. At pH 7.8, the reaction time was always about 15 s, while at pH 6 this time was 40 s. The maximum peak value was 1.640 ± 0.001 ( · 10 7 ) c.p.s. for Tris and 2.453 ± 0.0020 ( · 10 5 ) c.p.s. for Bis/Tris buffer, respectively. The presence of 40 lgÆmL )1 of Hb [iron(II)- or met-derivatives of both hemoglobins] does not significantly reduce the CL level when the experiments were performed in 50 m M Tris/HCl pH 7.8. The CL reduction wasnolargerthan5%ineachsample.Whenthe experiments were performed in 50 m M Bis/Tris pH 6, the CL reduction was significantly different for HbI and HbIV: iron(II)-HbI and met-HbI induce a 23.2 and 50.1% reduction of the peak intensity, respectively, when com- pared with the peak of the buffer. Iron(II)-HbIV and met- HbIV induce a 35.8 and 62.1% reduction of the peak intensity, respectively. Discussion In red blood cells, Hb can readily generate or interact with free radicals [28]. In fact, during spontaneous Hb autoxi- dation to met-Hb, the superoxide anion is produced. Most superoxide is reduced by superoxide dismutase to H 2 O 2 . Catalase and glutathione peroxidase eliminate H 2 O 2 . Hemoglobin also presents peroxidative properties (hydro- gen peroxide removal activity), which was recognized more than 30 years ago and could be important for the cellular lifespan [5,6]. The mechanism of hydrogen peroxide removal by iron(II)-Hb (oxyHb and deoxyHb) and met-Hb results in the formation of ferrylhemoglobin (ferrylHb) and oxo- ferrylhemoglobin (oxoferrylHb), respectively. Both are strong oxidizing agents, which can be source of cellular and tissue damage [8]. A recent work [8] suggested that ferryl-Hb takes an electron from a second molecule of H 2 O 2 and is reduced to met-Hb, while the H 2 O 2 is oxidized to super- oxide anion: HbðIIÞþH 2 O 2 ! HbFeðIVÞ¼ O þ H 2 O 2 ! HbFeðIIIÞþO À 2 ! heme degradation Ferryl-Hb [formed by iron(II)-Hb] can be also capable of withdraw an electron from a suitable substrate, resulting in the formation of metHb and a substrate radical [6]. The superoxide anion produced at the heme pocket is thought to easily react with porphyrin molecule, resulting in heme degradation and iron release [8]. However, H 2 O 2 -induced heme degradation products were demonstrated with iron(II)-Hb, but not with met-Hb [8,29]. The inability to produce heme degradation products by met-Hb with H 2 O 2 , was explained by the reaction of oxoferrylHb with H 2 O 2 with the production of met-Hb and oxygen instead of superoxide anion [8]. Our data performed, at pH 5.4, in the presence of H 2 O 2 and guaiacol (as reductant) indicate a larger peroxidase activity of HbIV, with respect to HbI, in Fig. 4. Peroxidase activity of different hemoglobins using guaiacol as substrate, monitored by the increase in optical density at 470 nm. For experimental details see Materials and methods. (j), iron(II)-HbI; (h), met-HbI; (r), iron(II)-HbIV; (e), met-HbIV. Ó FEBS 2004 Structural changes in trout hemoglobins (Eur. J. Biochem. 271) 1975 line with a previous work [5]. However, these results show a decrease of the peroxidase activity in the order: iron(II)- HbIV > met-HbIV > met-HbI > iron(II)-HbI. The larger peroxidative activity of iron(II)-HbIV with respect to met- HbIV is in line with previous data obtained with human HbA derivatives [7]. The unexpected HbI results could be explained suggesting a restricted accessibility to guaiacol for the heme pocket of HbI. In fact, it is known that, although H 2 O 2 binds directly to the heme iron, the guaiacol can have different kind of interactions with Hb, because it could be too voluminous to penetrate in the heme pocket [6]. It is known that one important difference between the two trout Hbs is present on the a-subunit, where the distal Val residue (E11) (present in both a and b pockets of human Hb) is substituted in HbIV by a Thr, and in HbI by an Ile. Because the Val(E11) residue is known to affect the ligand accessi- bility to heme pocket, its substitution with the bulky hydrophobic side chain of Ile could impose a certain steric hindrance to the a-pocket. On the other hand, the polar Thr could likely facilitate the access of guaiacol (o-methoxyphe- nol) to a-pocket by binding it with a hydrogen bond. The b-pocket of HbI is mainly affected by the interaction of Tyr(F1)85b with Ala70b and Ala74b [30]. As these residues are maintained in HbIV, similar b-pocket features are expected in both proteins. However, other amino acid substitutions could affect pockets structure, e.g. the change of Cys(F9)93b of human Hb with Ser in HbIV and Ala in HbI; or the change of His(HC3)146b with Phe in HbI are known to affect the physiological behavior of the proteins [31]. However, these residues may affect also the b-pocket structure [30,31]. According to Perutz and Brunori [31], the polar Ser in HbIV prefers an external position to the pocket. On the contrary, an internal position could by hypothesized for the hydrophobic Ala of HbI. Moreover, the residue b66, present in the b-pocket [32] is a Val for HbI and a Thr for HbIV, so that also in this case the polarity of the pocket is increased for HbIV. Although it is impossible to simply foresee the effect of these changes on the pocket structure, they could suggest a general lower accessibility of a quite bulky and polar molecule a guaiacol to HbI heme pockets. A structural modification of the heme pocket could be hypothesized also to explain the larger peroxidase activity shown by met-HbI, compared with iron(II)-HbI (Fig. 4), in the presence of substrate (guaiacol). The lower reactivity of HbI towards reactive oxygen species is confirmed also by the reaction with the superoxide anion obtained by the xanthine/xanthine oxidase reaction, both at pH 7.8 and pH 6, although a larger activity of both Hbs is evident at the acidic pH, in line with previous data obtained with human HbA [7,33]. Also in this case the larger reactivity shown by HbIV, when compared with HbI in the same iron oxidative state, could suggest a lower accessibility of the HbI heme pocket to superoxide anion. Moreover, a different pocket accessibility to the relatively small and charged superoxide could be related also to the pH-dependent reactivity of the proteins. Differences between HbI and HbIV in the reactions with H 2 O 2 and superoxide anion are likely to be related to possible different features of the heme–globin interaction, which can be modified in iron(II)- and met-Hbs and as consequence of pH changes. As the comparison between the structural characteristics of HbI and HbIV, at two different pH values, had been never studied, we have performed these studies by CD, whose spectra acquired in near UV and visible regions are particularly sensitive to heme surround- ing. In line with a previous work [3], the near-UV and visible regions of HbI and HbIV (oxygenated forms, pH 7.8) are characterized by quite different spectra, more evident at the a 1 b 2 interface, as suggested by the CD spectra in the region 270–290 nm (Fig. 1). Data presented in this work demonstrate that CD spectra (Figs 1–3), acquired under different conditions of pH and oxidation, show different patterns of modifications for the two trout hemoglobin components (HbI and HbIV), which are likely to be related to their different functional properties [2]. HbIV spectral modifications induced by a pH decrease, are quite similar to those reported in the literature for human HbA under deoxygenation [17,18] (Figs 1–3). HbIV deoxygenation is evident by the L-band which responds strongly to spin state [20,21] because it shows a large positive ellipticity for low-spin Hbs and a very reduced contribution for high-spin Hbs. Deoxygenation is evident Table 1. Lucigenin-amplified chemiluminescence of the xanthine/xanthine oxidase reaction in the presence of 40 lgÆmL )1 of oxy and met derivatives of both hemoglobins. The system contained 0.9 U of xanthine oxidase per mL and 150 l M lucigenin; the reaction was started by injecting xanthine at the final concentration of 50 mmol in 50 m M Tris pH 7.8 or 50 m M Bis/Tris pH 6.0. Values obtained are expressed in count per second (c.p.s.). Sample Duration (s) T half rise (s) T max (s) T. half fall (s) Peak value (c.p.s.) CL reduction (%) 50 m M Tris pH 7.8 15 < 2.40 6.60 (1.640 ± 0.001) · 10 7 Fe(II)-HbIV 15 < 2.40 7.20 (1.599 ± 0.006) · 10 7 2.5 Met-HbIV 15 < 2.40 7.20 (1.555 ± 0.005) · 10 7 5.1 Fe(II)-HbI 15 < 2.40 7.20 (1.610 ± 0.005) · 10 7 1.8 Met-HbI 15 < 2.40 7.18 (1.599 ± 0.005) · 10 7 2.5 50 m M Bis/Tris pH 6.0 40 < 3.00 25.80 (2.453 ± 0.002) · 10 5 Fe(II)-HbIV 40 < 1.20 22.80 (1.575 ± 0.010) · 10 5 * 35.8 Met-HbIV 40 < 1.80 24.60 (0.929 ± 0.010) · 10 5 * 62.1 Fe(II)-HbI 40 < 3.00 24.00 (1.884 ± 0.010) · 10 5 * 23.2 Met-HbI 40 < 5.40 26.40 (1.207 ± 0.010) · 10 5 * 50.1 *P<0.05. 1976 R. Gabbianelli et al. (Eur. J. Biochem. 271) Ó FEBS 2004 also by the large red shift of the Soret band (433 nm at pH 6.0 and 418 nm at pH 7.8), in line with the reported spectral characteristics of human deoxy-HbA [17,18]. This red shift is most likely not linked to changes in the quaternary structure, as a similar deoxygenation-induced shift was also observed in the Soret peak of monomeric Lucina pectinata Hb [25]. Moreover, although the data reported here seem not to be in line with previous data by Perutz et al. [18], which show that the Soret band in the T-form is slightly blue shifted and higher in intensity than the R-form, the measured effect in the Soret band is likely due to the superimposition of deoxygenation and the proton-dependent R fi T transition of HbIV. This possi- bility could be consistent with the results indicating that in deoxygenated human Hb, the difference CD spectra shows a maximum in the Soret peak at 437 nm for the R-form, and at 433 nm for the T-form [24]. The near-UV and visible CD spectral features are very similar in HbIV and in human HbA. However, the CD Soret band of human deoxy-HbA shows an increased ellipticity when compared with the oxygenated derivative [17,34]. The spectrum of deoxy-HbIV (pH 6.0) follows an opposite behavior (Fig. 2A). A previous work performed on human pathologic Hbs related the decrease in the Soret band ellipticity to a decreased cooperativity [35]. This possible interpretation of the unexpected decrease in the Soret band ellipticity in deoxy-HbIV is in agreement with data indicating that an acid pH causes a reduction in cooperativity (together with a marked reduction in ligand affinity) in a fish hemoglobin, exhibiting the Root effect [1,2,34]. Comparison between iron(II)-HbIV and met-HbIV at pH 7.8 in the regions of the L-band (about 260 nm), 280– 290 nm and the Soret band indicate quite similar structural features between these two ligated Hbs showing a similar R structure. The small, but evident differences between the spectra of iron(II)-HbIV and met-HbIV pH 7.8 around 285 nm (Fig. 1A,C) may be related to larger values of the equilibrium constant L ¼ [T]/[R] for met-HbIV (in terms of a two-state concerted model). The CD spectra obtained at pH 6.0 for met-HbIV indicate a shift to the T-form, high spin iron, a mixed population of six and five coordinated molecules (as suggested by the Soret band) and a decreased cooperativity (as suggested by the decreased ellipticity of the Soret band). These results on HbIV are in line with a previous work on human HbA by Perutz et al. [14], which suggested that the high-spin ligands as H 2 O (which occupies the sixth coordi- nation state in met-Hbs) favor the transition to the T state more than the low-spin ligands. HbI at pH 7.8 and pH 6.0 is reported to be completely oxygenated and cooperative [2]. As expected, the CD spectra recorded at pH 6.0 show the characteristic bands of a ligated Hb, at each wavelength range tested. The small spectral changes induced by low pH on iron(II)-HbI (decreased intensity, either negative or positive, of each band), in the region 250–470 nm could be due to a possible protein destabilization, which may affect the quaternary structure of the protein and, as a consequence, the a 1 b 2 and a 1 b 1 contacts. Oxidation does not induce important structural modifi- cations in the HbI spectra, although a slightly more negative ellipticity in the 285 nm region, at pH 7.8, can indicate a larger value of the equilibrium constant L ¼ [T]/[R]. However, the modified ratio Q 0 /Qm (visible region) indicates a slightly different symmetry at the axial bond with the proximal histidine, as a consequence of iron oxidation. Comparison with HbIV spectra suggests the possibility that the modified ratio Q 0 /Qm could be due to a partial change to the unligated form following oxidation, at both pH values. At pH 6.0, the low ellipticity value of the L-band (centered around 260 nm) for met-HbI suggests a high spin iron. This is in agreement with the presence, at low pH, of the high-spin ligand, water, bound to iron(III) subunits [36]. According to Perutz et al. [14] in the absence of organic phosphates, the R structure is dominant in all Hbs in which the irons are six-coordinated, but the spin state can modify the allosteric equilibrium between the two structures, shifting the equilibrium constant L ¼ [T]/[R] to a higher value. However, this is not the case of met-HbI at pH 6.0, as indicated by the CD spectra obtained which show no changes indicative of R fi T transition in the region of 285 nm and no blue shift in the Soret band. Conclusions Data presented in this work demonstrate different structural features for the heme–globin interaction in HbIV and HbI, which could be related to their different activities towards ROS. However, CD data did not give any indication about the different reactivity shown by both Hbs towards superoxide anion at pH 7.8 and 6. For this reason it is suggested that the larger reaction with the superoxide at pH 6 could be linked to the pH-induced decrease of negative charge density on dissociable groups on the proteins. Differences in HbI and HbIV behavior are likely related to the known differences [36,37] in number and position of amino acids with charged lateral groups. The larger activity of met-HbI, with respect to iron(II)- HbI, in the presence of the guaiacol (Fig. 4) could be due to the inhibition of R fi T transition in met-HbI. We suggest that the persistence of R form could hinder the structural changes, induced by the low pH, which could affect the interaction of the guaiacol large molecule at the heme pocket. A modification of the heme pocket structure is suggested also by the modified ratio Q 0 /Qm (visible region), indicating a changed symmetry at the axial bond with the proximal histidine. This modified symmetry seems to be related to a partial dissociation of ligand by met-HbI, which could permit a binding of more H 2 O 2 molecules with heme iron and, as a consequence, a higher rate of peroxidative reaction by the R-form of met-HbI pH 5.4 with respect to the R-form of iron(II)-HbI at pH 5.4. This hypothesis is also suggested by the comparison with data obtained for HbIV: at low pH the iron(II) form is completely deoxygenated (as confirmed by CD data, in particular by Soret band), and met-HbIV is only partially delegated (see Soret band). For HbIV the peroxidative reaction follows the order iron(II)-HbIV > met-HbIV, suggesting that in the case of this protein, the deligation also affects the rate of the reaction. We stress that our CD data give new information about ligation features and/or R fi T transition of HbIV and HbI: HbIV appears to dissociate ligands, as a consequence of pH also decreases in the oxidized form, even if the shift to Ó FEBS 2004 Structural changes in trout hemoglobins (Eur. J. Biochem. 271) 1977 the non ligated form is incomplete. Moreover, met-HbIV spectra(atpH6.0)showanRfi T transition, which is related to the change from a low- to high-spin iron. This behavior is in line with data indicating that tetramers with high-spin ligand water of iron(III) subunits, show the most T-state behavior [38]. CD spectra show that iron(II)- and iron(III)-HbI struc- tural features are almost insensitive to pH decrease, as the protein remains in the ligated form. This result is expected in iron(II)-HbI, which remains in the R-form and low-spin iron. However, a particularly high stability of the R-form is evident for met-HbI at pH 6.0, also in the presence of high-spin iron, in contrast with the expected transition to the Tform. In spite of the efficiency of the antioxidant defense system, the trout RBCs can be exposed to a considerable flux of reactive oxygen species, implying a red cell oxidative stress. Our data could suggest that trout HbI, which shows a reduced sensitivity to reactive oxygen species (such as superoxide anion and H 2 O 2 ), with respect to HbIV, could have a role in the maintenance of normal oxygen supply during ROS production. References 1. DeYoung, A., Kwiatkowski, L.D. & Noble, R.W. (1994) Fish hemoglobins. Methods Enzymol. 231, 124–150. 2. Brunori, M. (1975) Molecular adaptation to physiological requirements: the hemoglobin system of trout. Curr. Top. Reg. 9, 1–39. 3. Zolese, G., Gabbianelli, R., Caulini, G.C., Bertoli, E. & Falcioni, G. (1999) Steady-state fluorescence and circular dichroism of trout hemoglobins I and IV interacting with tributyltin. Proteins 34, 443–452. 4. Gabbianelli, R., Falcioni, G., Santroni, A.M., Fiorini, R., Coppa, G.V. & Kantar, A. (1997) Interaction of trout hemoglobin with H 2 O 2 : a chemiluminescence study. J. Biolumin. Chemilumin. 12, 79–85. 5. Fedeli, D., Tiano, L., Gabbianelli, R., Caulini, G.C., Wozniak, M. & Falcioni, G. (2001) Hemoglobin components from trout (Salmo Irideus): determination of their peroxidative activity. Comp. Biochem. Physiol. 130, 559–564. 6. Everse, J., Johnson, M.C. & Marini, M.A. (1994) Peroxidative activities of hemoglobin and hemoglobin derivatives. Methods Enzymol. 231, 547–559. 7. Gabbianelli, R., Santroni, A.M., Fedeli, D., Kantar, A. & Falcioni, G. (1998) Antioxidant activities of different hemoglobin derivatives. Biochem. Biophys. Res. Commun. 242, 560–564. 8. Nagababu, E. & Rifkind, M. (2000) Reaction of hydrogen peroxide with ferrylhemoglobin: superoxide production and heme degradation. Biochemistry 39, 12503–12511. 9. Balagopalakrishna, C., Abugo, O.O., Horsky, J., Manoharan, P.T., Nagababu, E. & Rifkind, J.M. (1998) Superoxide produced in the heme pocket of the beta-chain of hemoglobin reacts with the beta-93 cysteine to produce a thiyl radical. Biochemistry 37, 13194–13202. 10. Geraci, G. & Parkhurst, L.J. (1981) Circular dichroism spectra of hemoglobins. Methods Enzymol. 76, 262–275. 11. Zentz,C.,Pin,S.&Alpert,B.(1994)Stationaryandtime-resolved circular dichroism of hemoglobins. Methods Enzymol. 232, 247–266. 12. Binotti, I., Giovenco, S., Giardina, B., Antonini, E., Brunori, M. & Wyman, J. (1971) Studies on the functional properties of fish hemoglobins. Arch. Biochem. Biophys. 142, 274–280. 13. Antonini, E. & Brunori, M. (1971) Hemoglobin and Myoglobin in Their Reactions with Ligands, p. 10–11. North-Holland, Amster- dam, the Netherlands. 14. Perutz, M.F., Sanders, J.K., Chenery, D.H., Noble, R.W., Pennelly,R.R.,Fung,L.W.,Ho,C.,Giannini,I.,Porschke,D.& Winkler, H. (1978) Interactions between the quaternary structure of the globin and the spin state of the heme in ferric mixed spin derivatives of hemoglobin. Biochemistry 17, 3640–3652. 15. Gabbianelli, R., Santroni, A.M., Kantar, A. & Falcioni, G. (1994) Superoxide anion handling by erythrocyte loaded with alfa and beta hemoglobin chains: a chemiluminescence study. In Biolumi- nescence and Chemiluminescence (Campbell, A.K., Kricka, L.J & Stanley, P.E., eds), pp. 227–230. John Wiley & Sons, New York. 16. Kantar, A., Oggiano, N., Gabbianelli, R., Giorgi, P.L. & Biraghi, M. (1993) Effect of imidazole salicylate on the respiratory burst of polymorphonuclear leucocytes. Curr. Ther. Res. 54, 1–7. 17. Sugita, Y., Nagai, M. & Yoshimasa, Y. (1971) Circular Dichroism of hemoglobin in relation to the structure surrounding the heme. J. Biol. Chem. 246, 383–388. 18. Perutz, M.F., Ladner, J.E., Simon, S.R. & Ho, C. (1974) Influence of globin structure on the state of the heme. I. Human deoxyhemoglobin. Biochemistry 13, 2163–2173. 19. Hamaguchi, H., Isomoto, A. & Nakajima, H. (1969) Circular dichroism of human hemoglobin-haptoglobin complexes. Bio- chem. Biophys. Res. Commun. 35, 6–11. 20. Henry, E.R., Rousseau, D.L., Hopfield, J.J., Noble, R.W. & Simon, S.R. (1985) Spectroscopic studies of protein–heme inter- actions accompanying the allosteric transition in methemoglobins. Biochemistry 24, 5907–5918. 21. Chiancone, E., Vecchini, P., Verzili, D., Ascoli, F. & Antonini, E. (1981) Dimeric and tetrameric hemoglobins from the mollusc Scapharca inaequivalvis: structural and functional properties. J. Mol. Biol. 152, 577–592. 22. Hsu, M.C. & Woody, R.W. (1971) The origin of the heme Cotton effects in myoglobin and hemoglobin. J. Am. Chem. Soc. 93, 3515– 3525. 23. Blauer, G., Sreerama, N. & Woody, R.W. (1993) Optical activity of hemoproteins in the Soret region: circular dichroism of the heme undecapeptide of cytochrome c in aqueous solution. Bio- chemistry 32, 6674–6679. 24. Kawamura-Konishi,Y.&Suzuki,H.(1988)Interactionbetween alpha 1 and beta 1 subunits of human hemoglobin. Biochem. Biophys. Res. Commun. 156, 348–354. 25. Kaca, W., Roth, R.I., Vandegriff, K.D., Chen, G.C., Kuypers, F.A., Winslow, R.M. & Levin, J. (1995) Effects of bacterial endotoxin on human cross-linked and native hemoglobins. Bio- chemistry 34, 11176–11185. 26. Boffi, A., Wittenberg, J.B. & Chiancone, E. (1997) Circular dichroism spectroscopy of Lucina I hemoglobin. FEBS Lett. 411, 335–338. 27. Bolard, J. & Garnier, A. (1972) Circular dichroism studies of myoglobin and cytochrome c derivatives. Biochim. Biophys. Acta 263, 535–549. 28. Winterbourn, C.C. (1990) Oxidative reactions of hemoglobin. Methods Enzymol. 186, 265–272. 29. Nagababu, E. & Rifkind, M. (1998) Formation of fluorescent heme degradation products during the oxidation of hemoglobin by hydrogen peroxide. Biochem. Biophys. Res. Commun. 247, 592– 596. 30. Tame, J.R.H., Wilson, J.C. & Weber, R.E. (1996) The crystal structures of trout Hb I in the deoxy and carbonmonoxy forms. J. Mol. Biol. 259, 749–770. 31. Perutz, M.F. & Brunori, M. (1982) Stereochemistry of cooperative effects in fish and amphibian haemoglobins. Nature 299, 421–426. 1978 R. Gabbianelli et al. (Eur. J. Biochem. 271) Ó FEBS 2004 32. Baldwin, J. & Chothia, C. (1979) Haemoglobin: the structural changes related to ligand binding and its allosteric mechanism. J. Mol. Biol. 129, 175–220. 33. D’Agnillo, F. & Chang, T.M. (1998) Absence of hemoprotein- associated free radical events following oxidant challenge of crosslinked hemoglobin-superoxide dismutase catalase. Free Rad. Biol. Med. 24, 906–912. 34. Mazzarella, L., D’Avino, R., Di Prisco, G., Savino, C., Vitagliano, L., Moody, P.C. & Zagari, A. (1999) Crystal structure of Trematomus newnesi haemoglobin re-opens the Root effect question. J. Mol. Biol. 287, 897–906. 35. Nagai, M., Sugita, Y. & Yoneyama, Y. (1972) Oxygen equilibrium and circular dichroism of hemoglobin-Rainer. J. Biol. Chem. 247, 285–290. 36. Petruzzelli, R., Barra, D., Sensi, L., Bossa, F. & Brunori, M. (1989) Amino acid sequence of a-chain of hemoglobin IV from trout (Salmo irideus). Biochim. Biophys. Acta 995, 255–258. 37. Petruzzelli, R., Barra, D., Goffredo, B.M., Bossa, F., Coletta, M. & Brunori, M. (1984) Aminoacid sequence of b-chain of hemoglobin IV from trout (Salmo irideus). Biochim. Biophys. Acta 789, 69–73. 38. Marden, M.C., Kister, J. & Poyart, C. (1994) Allosteric equili- brium measurements with hemoglobin valency hybrids. Methods Enzymol. 232, 71–86. Ó FEBS 2004 Structural changes in trout hemoglobins (Eur. J. Biochem. 271) 1979 . a decrease of the peroxidase activity in the order: iron(II)- HbIV > met-HbIV > met-HbI > iron(II)-HbI. The larger peroxidative activity of iron(II)-HbIV. activity of iron(II)- and met-forms of HbI and HbIV. It is evident that the enzymatic activity decreases according to the order iron(II)-HbIV > met-HbIV