| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Institute of Protein Biochemistry (IBP), National Research Council (CNR), I-80125 Naples, Italy
2 Department of Chemistry, University of Florence, I-50019 Sesto Fiorentino, Florence, Italy
3 Institute of Biochemistry and Clinical Biochemistry and CNR Institute of Chemistry of Molecular Recognition, Catholic University, I-00168 Rome, Italy
4 Australian Antarctic Division, Kingston, Tasmania 7050, Australia
Reprint requests to: Guido di Prisco, IBP-CNR, Via Marconi 12, I-80125 Naples, Italy; e-mail: g.diprisco{at}ibp.cnr.it; fax: +39-081-593-6689.
(RECEIVED May 13, 2004; FINAL REVISION June 25, 2004; ACCEPTED July 2, 2004)
 |
Abstract |
|---|
 TOP Abstract IntroductionResultsDiscussionMaterials and methodsReferences |
|---|
-chain of Hb 1, although an altered distal histidyl and heme position was identified in the models of the 
-chain, possibly facilitated by a more open heme pocket due to reduced steric constraints on the vinyl substituent groups. This conformation may lead to the hemichrome form identified by spectroscopy in the Met state, which likely fulfils a potentially important physiological role. Keywords: Antarctica; fish; Pseudaphritis urvillii; hemoglobin; Bohr/Root effects; evolution; electronic absorption spectroscopy; Resonance Raman spectroscopy; molecular modeling
Article published online ahead of print. Article and publication date are at http://www.proteinscience.org/cgi/doi/10.1110/ps.04861504.
|
Introduction |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Diversification of the teleost perciform suborder Notothenioidei, largely confined within Antarctic and sub-Antarctic waters, has occurred in parallel with the climatic changes. Notothenioidei are the dominant component of the Southern Ocean fauna; they account for 96 of the 213 benthic species of the continental shelf and upper continental slope, and 122 (this figure includes non-Antarctic species) of the 313 Southern Ocean species described to date (Gon and Heemstra 1990; Eastman 2000). Information on the place of origin and existence of a transition fauna is lacking. Indirect indications suggest that notothenioids appeared in the early Tertiary (Eastman 1993) and began to diversify in isolation on the Antarctic continental shelf in the middle Tertiary, gradually adapting to progressive cooling. The ancestral notothenioid stock probably arose as a sluggish, bottom-dwelling teleost species that evolved some 40–60 million years ago in the shelf waters (temperate at that time) of the Antarctic continent. Molecular phylogeny has recently begun to provide indications about the time of radiation in the Antarctic.
The initial divergence of some of the families took place during the early fragmentation of Gondwana. At this time stocks appear to have become established in brackish water on isolated continental blocks (e.g., Pseudaphritidae in Australia). With the local extinction of most of the temperate Tertiary fish fauna as the Southern Ocean cooled, the suborder experienced extensive radiation, dating from the mid-Miocene, approximately 7–15 million years ago (Bargelloni et al. 1994), that enabled it to exploit the diverse habitats provided by a now chronically freezing marine environment.
According to a recently revised classification (Balushkin 1992; Pisano et al. 1998), Bovichtidae, Pseudaphritidae, Eleginopidae, Nototheniidae, Harpagiferidae, Artedidraconidae, Bathydraconidae, and Channichthyidae are the families of the suborder. Notothenioids are red-blooded except Channichthyidae (Ruud 1954), the only known adult vertebrates whose blood is devoid of hemoglobin (Hb). Some species of Bovichtidae and Nototheniidae also inhabit waters north of the Antarctic and sub-Antarctic, as well as monotypic Pseudaphritidae and Eleginopidae.
The variety of physiological and biochemical adaptations underlying the ability of modern Antarctic fish to survive at the freezing temperatures of the environment represents the extreme of low-temperature adaptations found among vertebrates.
The Hb system of temperate fish is often characterized by multiple, functionally distinct Hbs, probably required to cope with environmental variability. Temperate habitats with constant physicochemical features are inhabited also by species having a single major component (di Prisco and Tamburrini 1992). Notothenioids differ from temperate and tropical species in the reduced Hb concentration and multiplicity; Hb-less Channichthyidae represent the extreme of this trend. Antarctic red-blooded notothenioids are mostly bottom dwellers and have a single major Hb (Hb 1, accounting for 95%–97% of the total), often accompanied by a minor component Hb 2 (di Prisco 1998), both displaying strong Bohr (Riggs 1988) and Root (Brittain 1987) effects, with few exceptions. Hb 1 and Hb 2 always have one of the chains in common. The minor components Hb 2 often differ from Hb 1 precisely in the residues found only in the Antarctic species. On the basis of these observations, and taking into account the fact that in all Antarctic notothenioids Hb 2 is functionally indistinguishable from the major components, we have hypothesized that in adult fish the minor components are vestigial or larval remnants. Hb 2 is indeed one of the dominant components in the juvenile stage (G. di Prisco, unpubl.), still synthesized in limited amounts by adult fish. Hb 1 and Hb 2 would bear very different evolutionary pressures, on one hand reflecting the needs that larvae and juveniles have to survive, and on the other, the much more refined adjustments developed by adult fish.
This communication reports the structure and function of Hb 1 of the non-Antarctic notothenioid Pseudaphritis urvillii (common name, Congolli) of the monotypic family Pseudaphritidae. Comparison between cold-adapted and noncold-adapted species is a useful approach to understanding the fish evolutionary history, as well as the molecular mechanisms of cold adaptation. In fact, P. urvillii is a non-cold-adapted catadromous species from coastal waters, estuaries, and rivers of Southern Australia. Most likely, the ancestors of modern P. urvillii became associated with the Australian component of Gondwana 40–50 million years ago. The fragmentation of Gondwana and separation of Australia from Antarctica is likely to be the vicariant event that led to the speciation of P. urvillii (Bargelloni et al. 2000). It is euryhaline and may migrate upstream as far as 120 km from the sea (Andrews 1980). Because of its geographical origin, this Gondwanian species is especially important for calibrating the evolution molecular clock in Antarctica.
Besides the lifestyle, some features differentiate this species from cold-adapted Antarctic notothenioids, for example, it has glomerular kidneys (Eastman 1993) and has neither antifreeze glycoproteins nor the genes encoding antifreeze glycoproteins (Cheng et al. 2003). Other features are shared with cold-adapted species of the suborder. In freshwater, it is almost neutrally buoyant (Eastman 1993). Similar to most notothenioids, P. urvillii has a major (Hb 1) and a minor (Hb 2) component, in the typical notothenioid proportions and having the -chain in common. The primary structure of Hb 1 has high identity with Antarctic globins; however, the values are at the lower end of the range. Hb 1 shows several nonconservative substitutions, some of which may alter the geometry of the invariably hydrophobic heme pocket.
Hb 1 displays strong, effector-enhanced, Bohr and Root effects. (The oxygen-binding properties of the minor component Hb 2 will not be discussed in detail, in view of the physiologically irrelevant levels of this component, and of the similarity with the functional features of Hb 1, shown by appropriate key experiments.) Some of the functional features of Hb 1, namely the oxygen affinity and subunit cooperativity, appeared exceptional. Molecular modeling, electronic absorption and Raman spectroscopy, powerful techniques that provide information about the overall environment of the heme pocket, were used in an attempt to shed light on possible links between these features and the molecular structure of the heme.
|
Results |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
-chains (indicated as 1 and 2, respectively), a feature shared with only one Antarctic notothenioid, the bathydraconid Cygnodraco mawsoni (Caruso et al. 1991); Hb 1 and Hb 2 of all the other species differ by the -chains.
Primary structure
The sequences of the - and -chains (142 and 146 residues, respectively) which constitute P. urvillii Hb 1 and Hb 2 are reported in Figure 1
. The functionally important residues suggested by Perutz and Brunori (1982) to be involved in the molecular mechanism of the Bohr and Root effects in fish Hbs (Ser F9, Glu FG1, Gln HC1, and His HC3) are all conserved in the -chains. The sequences of the two chains ( and 1) of Hb 1 have higher similarity with major Hbs of Antarctic notothenioids (73%–82%, Table 1) than with minor Hbs of the same species (64%–68%). In contrast, the 2 chain of Hb 2 shows a reverse correlation (70%–77% with major Hbs and 80%–88% with minor Hbs). In comparison with Hb 1 of other notothenioids, the primary structure of Hb 1 reveals several nonconservative substitutions. In the -chain, Glu and Met replace Leu in F7 and FG3, and Thr replaces Lys in E10 and Ala in E6; in the 1 chain Cys replaces Leu in B13. It is worth noting that Cys is also present in B13 of Trematomus bernacchii Hb.
|
|
- and -globin data sets are shown in Figure 2, A and B, respectively. The same topologies were recovered with the two methods used for phylogenetic analysis (i.e., Local Rearrangement and exhaustive search). In both trees, the globins of major and minor Antarctic fish Hbs cluster in two separate, strongly supported groups, with the globins of temperate fish Hbs forming the most ancestral clade. The -chain of P. urvillii Hb 1 and Hb 2 branches off the clade of the major Antarctic Hbs, and the same applies to the -chain of Hb 1. The -chain of Hb 2 places itself in a basal position with respect to the clade of the Antarctic minor Hbs.
|
) in the absence and presence of allosteric effectors such as protons, chloride ions, and organophosphates (ATP and IHP). The effect of IHP, although not physiologically relevant, was investigated because it is known that this heterotropic ligand, possessing a high number of negative charges, may strongly stabilize the low-affinity T state, within the framework of the two-state allosteric model (Kister et al. 1987; Robert et al. 1988).
|
, where oxygen saturation at atmospheric pressure is reported as a function of pH, Hb 1 is characterized by a marked Root effect; at a specific pH value, the oxygen release induced by protons is enhanced by ATP. This is not a novel finding, but once again it highlights the physiological role of anions and in particular of organo-phosphates which, besides modulating the oxygen affinity, appear to finely control the region of pH in which the Root effect is operative. This is even clearer in Figure 3A, where the oxygen-binding experiments, performed in the absence and presence of allosteric effectors, are illustrated in terms of logp50 as a function of pH. The Bohr effect is strongly enhanced by organophosphates; in particular, going from stripped conditions to 3 mM ATP or IHP, the Bohr coefficient (
=
logp50/pH), which denotes the average number of protons bound upon heme oxygenation, is increased, respectively, by about 150% and 80%. The effect of chloride is of the same order of magnitude as that observed in human HbA. However, in Hb 1 from P. urvillii, the presence of chloride influences the pH dependence of both oxygen affinity (Fig. 3A) and cooperativity (Fig. 3B).
In comparison with Antarctic notothenioids, Hb 1 displayed a much higher oxygen affinity. In the absence of effectors, logp50 remained below zero in the pH range 8.7–7.5. In addition to this remarkable difference, the oxygen affinity and cooperativity in the presence of chloride and ATP at pH 6.3 and pH 8.7 call for special attention. At alkaline pH the oxygen affinity was very high (0.90 mmHg) and cooperativity almost absent, whereas at pH 6.3 the oxygen affinity was very low (102 mmHg) and cooperativity completely abolished. This unusual modulation of cooperativity by pH is clearly depicted by the bell-shaped curve of the Hill coefficient (Fig. 3B), in which n50 reaches 2.0–2.5 at intermediate values, approaching one at both extremes of pH.
The stripped hemolysate had oxygen-binding features reflecting those of the major component Hb 1 (Fig. 3 D,E,F). Whole blood and intact erythrocytes were assayed in the absence of effectors, and also showed strong Bohr and Root effects (not shown).
Thermodynamics of oxygen binding
The temperature dependence of oxygen-binding equilibria of stripped hemolysate and Hb 1, was investigated in the range 10°–16°C (Table 2). Under all experimental conditions, that is, in the absence and presence either of chloride alone, or of both chloride and organophosphates, the highest value (in absolute terms) of the enthalpy change of oxygenation was at pH 8.7, where Hb 1 may be considered locked in the high-affinity R state. At lower pH values, the overall enthalpy change progressively decreased as a function of increasing proton concentration, due to the endothermic contribution of the heterotropic effectors, released upon oxygen binding, and of the allosteric transition. At pH 6.3 the overall enthalpy change of the low-affinity T state was measured. Under all conditions, the difference in enthalpy change of oxygenation of the two states was 10–12 kcal mole–1. The value at pH 6.6 in the simultaneous presence of chloride and ATP was of special interest. In this case, the enthalpy change was positive (i.e., endothermic), indicating that oxygen release in the tissues is favored by low temperature. On the grounds of the primary structure analysis, a hypothesis on the binding of chloride will be discussed below.
|
H was also measured in the unstripped hemolysate, containing physiological ATP concentration (not shown). As expected, the values were much less negative than those of mammalian Hbs.
Spectroscopic characterization
Spectroscopy was the first approach aimed at finding a structural explanation to the striking oxygen-binding features. The electronic absorption and resonance R (RR) spectra of Hb 1 are shown in Figures 4 and 5, respectively.
|
|
(C=C) vinyl stretching mode (at 1625 cm–1) in addition to that at 1617 cm–1 (Fig. 5). Accordingly, in the low-frequency region, two 
(CCaCb) bending modes of the vinyl substituents at 405 and 412 cm–1 are observed (not shown). Deoxy Hb 1 displays an RR (Fe-Im) stretching mode at essentially the same frequency (215 cm–1) as in deoxy HbA (Nagai et al. 1980), suggesting a largely unchanged proximal cavity in P. urvillii Hb 1 compared to human HbA.
In oxy Hb 1, the 2-nm blue-shift of the electronic absorption maxima at pH 7.7 (Fig. 4), compared to those reported for oxy HbA at neutral pH (Antonini and Brunori 1971), and the very weak band at 630 nm indicate that a small amount of Met-Hb 1 is present (see below). The RR bands at 1469 and 1478 cm–1, assigned to 5c HS (deoxy form) and hexa-coordinate high spin (6c HS) (Met form), respectively, result from photolysis of the FeO2 bond induced by the laser irradiation. The formation of the Met form at the expense of the oxy upon laser exposure is confirmed by a slight Soret-band blue-shift after the RR experiment. The (Fe-O2) stretching mode was observed (not shown) at a frequency (568 cm–1) very close to that of HbA (Robert et al. 1988).
As the frequencies of the Fe–CO complex are more sensitive to the properties of the heme cavity than those of the oxy form, carbomonoxy Hb 1 was also studied. The electronic absorption spectrum of the CO complex is characterized by bands at 419, 537, and 567 nm (not shown), very similar to human HbA–CO (Antonini and Brunori 1971). The frequencies of the (Fe–C) and (C=O) stretching modes at 502 and 1952 cm–1, respectively (not shown), differ only slightly from those of the corresponding bands of carbomonoxy HbA (507 cm–1 and 1951 cm–1, respectively; Tsubaki et al. 1982). Hence, again any variation in the properties of the heme cavities of the two proteins must be of limited extent.
The electronic absorption spectrum of human Met-HbA at pH 6.5 is characterized by bands at 405, 500, 540, 580, and 631 nm, indicative of a dominant aquo 6c HS state in equilibrium with a hydroxy hexacoordinate low-spin (6c LS) state (Perutz et al. 1974a). In contrast, Met-Hb 1 at pH 6.1 displays a mixture of HS and LS heme spin states, in which the LS form constitutes a significant component (Fig. 4). Furthermore, the ~525-nm (-band) and 560-nm (-band) wavelengths are consistent with an LS species resulting from a hemichrome, rather than a hydroxide complex (Smulevich at al. 1991). In agreement with the principal features of the electronic absorption spectrum, the high-frequency RR spectrum of Met-Hb 1 (Fig. 4) reveals a mixture of 6c HS (1478 [3], 1512 [38], 1560 [2], and 1578 [37] cm–1) and 6c LS (1508 [3], 1578 [2], 1604 [37], and 1638 [10] cm–1) species. As observed for the other P. urvillii Hb forms, two vinyl stretching modes are evident at 1620 and 1627 cm–1.
At pH 9.2 the absorption spectrum of Met-Hb 1 red-shifts compared to that at pH 6.1 (the band maxima are at 409, 536, 570, and 630 nm, not shown). This is consistent with an increased proportion of LS species compared to pH 6.1, as confirmed by the corresponding RR spectrum (not shown). Slight differences are evident particularly in the visible region compared to the spectrum of HbA at alkaline pH (George et al. 1964; Feis et al. 1994), consistent with P. urvillii Hb 1 at pH 9.2 being a mixture of the hydroxide LS complex characteristic of HbA at alkaline pH, a hemichrome LS species, and a small amount of aquo 6c HS heme. The transition to the hydroxide complex could not be followed to completion as the protein was denatured at pH values higher than 9.2.
The complex formed between Met-Hb 1 and fluoride at pH 5.6 gives rise to electronic absorption and RR spectra (Figs. 4, 5) typical of a 6c HS state, indicating that fluoride binds strongly to P. urvillii Hb 1. However, the presence of a weak band at ~1510 cm–1 (3 of LS) suggests that fluoride does not fully bind to the protein. The absorption spectrum is similar to that of HbA (Perutz et al. 1974a) except for a 2-nm Soret blue-shift in the fluoride adduct, consistent with the presence of two vinyl stretching modes in the RR spectrum (at 1620 and 1637 cm–1) and, therefore, a lower degree of vinyl conjugation with the porphyrin macrocycle (Neri et al. 1997). It is worth stressing that only one vinyl band is present in the corresponding spectrum of the HbA-F complex (Choi et al. 1982).
Thus, the spectroscopic analysis of Met-Hb 1 provides evidence for hemichrome formation and supports the presence of two vinyl stretching modes.
Molecular modeling
Another approach aimed at correlating the structure of the heme pocket with the oxygen-binding properties was molecular modeling. Models of - and -chains of P. urvillii Hb 1 were constructed to allow visualization of the heme environment in deoxy, carbomonoxy, and oxy forms. De-oxy and carbomonoxy models were based on the X-ray coordinates of T. bernacchii Hb, whereas the crystallographic structure of human HbA was used as template for the oxy model. Detailed analysis of the final models was performed in comparison to human HbA and T. bernacchii Hb crystallographic structures, including those of HbA under low salt conditions (PDB codes 1bbb
[PDB]
and 1hbb
[PDB]
). T. bernacchii Hb displays the highest sequence identity with P. urvillii Hb 1 (79% and 82% for the -chains and -chains, respectively). Figure 6 reports the sequence alignments.
|
chain of Hb 1 comprises the important replacement of Leu FG3 (strongly conserved in human and fish Hbs) by Met. Modeling indicates that the methyl group (CD1 atom, according to the standard nomenclature for amino acids and heme group used by the Protein Data Bank) of Leu FG3 is at only 3.5 Å from the porphyrin ring of HbA (PDB code 2hhb
[PDB]
). However, the -heme proximal cavity of P. urvillii Hb 1 is able to accommodate the Leu FG3 
met replacement with no significant effect on the heme and on the proximal (F8)-histidyl conformation. Furthermore, the analysis has shown that even the nonconservative replacement of Leu F7 with Glu does not affect the -heme proximal side of P. urvillii Hb 1, probably due to orientation of the carboxyl oxygen of Glu protruding out of the proximal cavity toward the protein surface.
The distal side of the heme in Hb 1 was compared to that of HbA. The most notable change in the model of the -heme distal cavity is Thr replacing Gly E6 (Ala in T. bernacchii Hb) and Lys E10, which might perturb the hydrogen-bonding interaction pattern of the -heme distal His. However, a slight conformational modification of His E7, with respect of HbA, was detected only in the carbomonoxy form. The models of the carbomonoxy and oxy forms of Hb 1 suggest that Thr E10 does not interact with ligands, as its hydroxyl group is oriented away from distal His. Consequently, it is also unlikely that the nonconservative replacements Lys E10 Thr and Gly E6 Thr induce a significantly modified electrostatic field in the vicinity of His E7 in the -chain of P. urvillii Hb 1.
It is worth noting that Thr is also present in E10 of sperm-whale myoglobin; the side chain of Thr E10 is always directed toward the solvent in deoxy, carbomonoxy, and oxy forms of myoglobin (PDB codes 1a6n [PDB] , 1a6g [PDB] and 1a6m [PDB] , respectively; Vojtechovski et al. 1999), hydrogen bonded to water crystallographic molecules. The side chains of E10, E7, and E11, by means of feasible movements, have been indicated (Case and Karplus 1979) as a possible pathway for ligand escape in sperm-whale myoglobin.
A different situation is seen in the -chain. The replacement of the residue interacting with heme vinyl 4 (Leu B13 Cys) reduces the steric hindrance by distal side chains. In particular, the distance of Leu B13 methyl group (CD2 atom) from the CBC atom of heme vinyl 4 is 4.0 Å, 4.8 Å, and 4.5 Å in the three forms of human HbA considered, deoxy (PDB code 2hhb
[PDB]
), carbomonoxy (PDB code 2hco
[PDB]
), and oxy (PDB code 1hho
[PDB]
), respectively. Replacement of Leu B13 with Cys in P. urvillii Hb 1 increases the distance of the B13 side chain from the heme and, consequently, induces a different conformation of -heme vinyl 4. The SG atom of Cys B13 is 6.2 Å, 7.1 Å, and 5.6 Å away from the vinyl CBC atom of deoxy, carbomonoxy, and oxy Hb 1, respectively. Superimposition of -heme environment of the deoxy structures of P. urvillii Hb 1 and human HbA is reported in Figure 7.
|
-heme conformation surrounding the proximal and distal histidines essentially unmodified in comparison with HbA, a different conformation in the -heme vinyl 4 was detected, in agreement with the spectroscopic characterization. |
Discussion |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Three notothenioid families, that is, Bovichtidae (only one out of 10 species is Antarctic), Pseudaphritidae, and Eleginopidae, presumably diverged and became established in waters around areas corresponding to New Zealand, Australia, and South America before Antarctica became isolated. The absence of any detectable antifreeze glycoprotein coding sequence in Bovichtus variegatus, P. urvillii, and Eleginops maclovinus is consistent with this hypothesis (Cheng et al. 2003). However, the three families carry the trypsinogen-like protease gene, in keeping with occurrence of diversification before the evolution of the antifreeze glycoproteins gene from this ancestral gene.
From the evolutionary standpoint, the analysis of the oxygen-transport system of non-Antarcvatic notothenioids is of great interest. The characterization of the Hb system of P. urvillii (among the most primitive notothenioid species) provides useful insights. Taking advantage of the wealth of information available on a high number of Antarctic notothenioid species (e.g., Hb amino acid sequences), this suborder has been considered as a suitable source for the study of the evolution of the oxygen carrier in fish, in a molecular phylogenetic approach. In this study, methods affording the highest level of reliability have been used to build phylogenetic trees of - and -chain amino acid sequences of Hbs of P. urvillii, Antarctic notothenioids, and several non-Antarctic fish species, thus developing and improving previous preliminary results (Stam et el. 1997). The trees are in agreement with those obtained by morphological analysis and sequence studies on mitochondrial RNA (Ritchie et al. 1996), and give strong support to the monophyly of Antarctic notothenioids, with non-Antarctic P. urvillii as their sister taxon.
Although P. urvillii has never developed cold adaptation, the amino acid sequence reveals high identity with the globins of Antarctic notothenioids, and most of the residues that differentiate Antarctic notothenioids from temperate fish are found in Hb 1. However, the identity between Hb 1 of P. urvillii and other Antarctic fish is close to, or lower than, the low extreme of the ensemble of values of Antarctic notothenioids (82%–99% and 77%–93% for the - and -chains, respectively). This argues in favor of a common origin within notothenioids, but also suggests that the primary structure of the major Hb has undergone modifications only to a limited extent. If sequence mutations in Antarctic fish are indeed related to the development of cold adaptation, this may imply divergence during the first stages of the cooling process, and in any case before the event that gave rise to antifreeze glycoproteins.
The relevance of studying Hbs of non-Antarctic species in the evolution of Notothenioidei is highlighted by the analysis of the features of the Hb system of another temperate notothenioid. Recent studies indicate that the genome of the more recent Notothenia angustata (family Nototheniidae), common near the coast of southern New Zealand, does have the antifreeze glycoproteins genes (Cheng et al. 2003), supporting the hypothesis that this species had developed cold adaptation before migration from the Antarctic continental shelf in a relatively recent geological time. The amino acid sequence identity between Hbs of noncold-adapted N. angustata and cold-adapted Notothenia coriiceps (a species of the same family) is the highest ever found among notothenioids (Fago et al. 1992). The presence of antifreeze genes suggests that, unlike P. urvillii, this fish may have migrated from Antarctic to temperate waters in a relatively recent time and was cold adapted prior to radiation. The very high similarity of Hb sequence of N. angustata and the lower one of P. urvillii with Antarctic notothenioids may indeed bear a proportional relationship to cold adaptation, and at the same time appear to reflect the differences in divergence and migration times. On the other hand, both species have high hematocrit, erythrocyte number, Hb content, and cellular concentration. Unlike primary structure, these features might be considered a short-term response to environmental changes. In evolutionary processes, in the same time span changes in amino acid sequence would occur at a much slower overall rate, thus producing fewer variations. We suggest that the common phylogenetic origin remains the prevailing factor in primary structure in the course of readaptation of a notothenioid to a temperate environment. Changes in factors other than sequence, such as (for instance) regulatory processes or enhanced protein synthesis, may well be additional mechanisms of temperature compensation.
Oxygen affinity and subunit cooperativity
P. urvillii Hb 1 has a strong Bohr effect in the presence and absence of the physiological effector ATP. The Root effect is largely ATP-induced. In the temperature dependence of oxygen binding, the positive (endothermic) enthalpy change observed around pH 6.6 may have an important physiological significance. At pH values where the T state becomes preferentially populated, but still in the presence of a significant degree of heterotropic interactions, the contribution of the effectors and of the conformational transition may even overcome that of the T state, thereby producing a change of sign in the overall enthalpy change, which, in fact, becomes endothermic, favoring oxygen release when the organism is challenged by lower temperatures.
The -chain of P. urvillii Hb 1 shows two remarkable replacements with respect to HbA: Lys A5 Arg and Ala E20 Lys. In P. urvillii Hb 1, these residues may represent an "additional" oxygen-linked chloride binding site, as observed in a number of Hbs characterized by a low effect of temperature. It has been shown (De Rosa et al. 2004) that the endothermic contribution by the oxygen-induced release of the "additional" chloride ions (one for each -chain) may significantly contribute to lower the overall heat of oxygenation, thus providing a molecular basis to explain the thermodynamics of the oxygenation–deoxygenation cycle of several mammal Hbs. Similar to human fetal, horse, and bear Hbs, P. urvillii Hb 1 also displays a small oxygenation enthalpy only in the simultaneous presence of both chloride and organophosphate. The -chain of these Hbs has basic residues in A5 and E20; in the mammal Hbs, the interaction with 2,3-DPG in its pocket induces a distortion in helix A, thereby influencing the structure so that Lys A5 (Arg in P. urvillii Hb 1) is able to switch on the additional binding site. Such conclusion, supported by crystallographic data, may apply to P. urvillii Hb 1.
The affinity and cooperativity are modulated by pH and allosteric effectors in a remarkable way. The oxygen affinity of Hb 1 is exceptionally high compared to other notothenioids. Although the Hb multiplicity closely resembles that of sedentary Antarctic bottom dwellers, this high affinity (possibly related to the peculiar habitat constraints) clearly differentiates this species from other notothenioids. A high oxygen affinity may result, at least in part, from an altered intrinsic affinity in the - and -subunits. In the -chain of P. urvillii Hb 1 several substitutions may alter the geometry of the invariably hydrophobic heme pocket. For instance, Glu and Met replace Leu in F7 and FG3; the substitution in 87 (F7) is uncommon, and Glu is indeed found only in this species. Interestingly, all Antarctic Notothenioidei have Gln, whose codon differs from that of Glu of P. urvillii Hb 1 by a single base change in the first position. Moreover, Thr replaces Gly E6 (in HbA) and Ala E6 (in T. bernacchii Hb) and Lys E10 (in both Hbs). As discussed in detail below, the spectroscopic and modeling studies indicate, however, that each of these replacements leave the conformation and electrostatic field surrounding His E7 essentially unmodified with respect to both HbA and T. bernacchii Hb. In fact, although T. bernacchii Hb has lower oxygen affinity, the affinity of P. urvillii Hb 1 and HbA are comparable. In the -chain, Leu B13 is replaced by Cys. This nonconservative substitution, although uncommon in fish Hbs, is found also in T. bernacchii Hb. Although this replacement, therefore, cannot directly explain the high affinity of P. urvillii Hb 1, it does induce a conformational change with significant implications (see below).
The subunit cooperativity, mostly induced by allosteric effectors, is regulated by pH in a very unusual way, in that it not only disappears at low physiological pH (as in all Root-effect Hbs), but is absent also at high pH. Thus, the curve of the Hill coefficient as a function of pH is bell-shaped, and its maximum identifies the pH value where the ligand-induced conformational transition is fully active. Moreover P. urvillii Hb 1 appears to be a rare type-case of Hb, in which the molecule is fully locked in the low-affinity conformation at acidic pH and tends to be locked in the high-affinity state at alkaline pH. In the common sting ray, Dasyatis sabina (Mumm et al. 1978), Hb also exhibits a bell-shaped curve of the Hill coefficient, with maximal value of 2.0 near pH 9.0 but approaching one below 7.0 and above 10, in the absence and the presence of effectors.
Heme pocket properties
The spectroscopic measurements reveal two characteristics which distinguish Hb 1 from HbA, namely the presence of a hemichrome species in the Met form and reduced steric constraints on vinyl group 4. It is noteworthy that the proportion and nature of the LS species observed in P. urvillii Met-Hb 1 are in contrast with those found in HbA. At both pH 6.1 and 9.5 P. urvillii Hb 1 displays a significant proportion of LS heme (compared to HbA) due to a hemichrome conformation, rather than a hydroxy complex as in HbA. Hemichromes have been generally observed in association with Hb-denaturation processes but, unlike hemichromes of other heme proteins that are reduced to hemo-chrome, the oxidized form of P. urvillii Hb 1 is reduced by dithionite to the deoxy form which can then bind oxygen or CO. Although the formation of hemichrome in a native Hb structure is rare, it has recently been reported in the monomeric Hb of the sea cucumber Caudina arenicola (Mitchell et al. 1995), the homodimeric Hb of rice (Hargrove et al. 2000), the tetrameric Hb of the horse (at pH 5.4) (Robinson et al. 2003), and the major Hb 1 of the Antarctic notothenioid Trematomus newnesi (Riccio et al. 2002). Moreover, it has recently been found that in five Hbs from four Antarctic notothenioids the oxidation process, promoted by air or ferricyanide, leads to the reversible formation of hemi-chrome (Vitagliano et al. 2004), in marked contrast with human HbA, and suggesting that this behavior is common in fish. These findings provide new information on tetrameric Hbs in the hemichrome state, and suggest a physiological role of in vivo hemichrome formation in fish, which appears to occur regardless of whether a species is cold adapted or not.
It is worth noting that fish are highly sensitive to a number of environmental stress factors. Although fish erythrocytes may contain low levels of Met-Hb, stress often leads to rapid increase of this form (Graham and Fletcher 1986). It is feasible that fish might use hemichrome formation as a protective response to otherwise irreversible stress-induced Hb oxidation. In other words, spontaneous autoxidation, occurring at higher rates than in mammalian Hbs, might be balanced by hemichrome formation. This in vivo mechanism has been suggested also in other organisms (Robinson et al. 2003).
The formation of a hemichrome (bisHis) in P. urvillii Met-Hb 1 indicates a change in conformation compared to HbA that allows distal His to bind to the iron atom. A slightly modified distal cavity, compared to HbA, is also suggested by retention of an aquo HS species in the Met form at alkaline pH, which indicates that the pKa of the alkaline transition in P. urvillii Hb 1 is higher. Although the plant, invertebrate, and vertebrate Hbs mentioned above differ significantly in their primary structures, all exhibit two hemichrome-induced structural characteristics, namely a shift of the heme group to a more exposed position and a narrowing of the heme pocket. These differences found in P. urvillii Met-Hb 1 compared to HbA do not appear to have significant consequences on the ligand-bound form (oxy, carbomonoxy). In fact, in accordance with the modeling studies, which suggest a slightly modified distal His disposition in the ligand-bound forms but a largely unchanged electrostatic field, the Fe–CO and CO frequencies of the CO complex are not significantly modified. These frequencies are sensitive indicators of modified heme interactions (H-bonding, polarity and steric hindrance) and hence confirm the similarity of the heme cavities of P. urvillii Hb 1 and HbA. The absence of any variation in the Raman frequency of the Fe–O2 stretching mode of P. urvillii Hb 1 compared to HbA is consistent with this conclusion, although it should be noted that the frequency of the Fe–O2 Raman mode is not particularly sensitive to perturbation of the distal cavity (Hirota et al. 1996).
The presence of two vinyl modes in the RR spectra of P. urvillii Hb 1, compared to the single mode observed in HbA, is particularly interesting. A direct relationship between the (C=C) stretching frequency and the orientation of the vinyl groups, as induced by specific protein interactions, has recently been found in heme-containing peroxidases (Smulevich 2001; Marzocchi and Smulevich 2003). In particular, it was found that, when the protein matrix exerts no contraints on the vinyl groups, the latter take the tortional conformation usually found in model compounds in solution (Kalsbeck et al. 1995), and two (C=C) stretching modes should be observed in the RR spectrum. Hence, it is suggested that the presence of two distinct vinyl stretches in the RR spectra of P. urvillii Hb 1 is related to a reduced steric hindrance of vinyl 4 caused by replacement of Leu B13 with Cys in the -chain. This spectroscopic finding is consistent with the reduced steric constraints of the vinyl 4 substituent group of P. urvillii Hb 1 highlighted by the modeling studies. It can be anticipated that the reduced steric hindrance upon replacement of the residue interacting with heme vinyl 4 contributes to the increased mobility of the heme, allowing it to shift to a more exposed position, as required for hemichrome formation.
In conclusion, spectroscopy and molecular modeling were used to investigate correlations between the peculiar oxygen-binding features of P. urvillii Hb 1 and the overall environment of the heme pocket, and showed the unexpected lack of a direct linkage between the nonconservative replacements found in the amino acid sequence and the oxygen affinity of these Hbs on one hand. On the other hand, two closely related and unique features distinguish P. urvillii Hb 1 from HbA, namely (1) the hemichrome species in the Met form, whose formation appears to be mediated by (2) the reduced steric constraints on vinyl group 4. Although their correlation with the oxygen-binding features of P. urvillii Hb 1 remains an open question, the occurrence of a significant amount of Hb in the hemichrome state under physiological conditions in notothenioids (regardless of their temperature adaptation) suggests an important functional role of this form in fish life style.
|
Materials and methods |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Hb purification
Specimens of P. urvillii were captured in North West Bay River, Margate, Tasmania, Australia. Blood samples were drawn from the caudal vein of anesthetized fish with heparinized syringes. Hemolysates were prepared as described (D’Avino and di Prisco 1988). Separation of Hbs was achieved by fast protein liquid chromatography (FPLC) on a DE52 ion-exchange column (1.8 x 30 cm), equilibrated with 10 mM Tris-HCl (pH 7.6), and by stepwise elution with the same buffer, at a flow rate of 1.0 mL/min. The Hb 1-containing fractions were pooled and dialyzed against 10 mM HEPES (pH 7.7). All steps were carried out at 0°–5°C. For oxygen-binding, aliquots of a solution of carbon monoxide Hb 1 were stored at –80°C prior to use within a maximum of 7 days. For each experiment, one aliquot was thawed, converted to the oxy form by exposure to light and oxygen, and immediately used; no oxidation was spectrophotometrically detectable, indicating that final Met-Hb formation was negligible (<2%).
Amino acid sequencing
Purification of globins, fractionation of tryptic and cyanogen-bromide-cleaved peptides, and subsequent amino acid sequencing were carried out as previously described (Verde et al. 2002).
Phylogenetic analysis
Globin amino acid sequences of selected species of temperate and Antarctic fish were aligned using the program CLUSTAL X. Maximum likelihood phylogenetic trees of both -and -globin sequences were inferred by the method implemented in the software package MOLPHY (Adachi and Hasegawa 1996), using the model of amino acid substitution (Jones et al. 1992), adjusted with the empirical amino acid frequencies. The following strategy was applied for inferring the trees. First, a distance matrix and the corresponding neighbor-joining (NJ) tree were estimated. Second, the NJ tree was subjected to a local rearrangement search using the program PROTML contained in the MOLPHY package. An exhaustive topology search of 1000 top-ranking trees was performed on a partially constrained starting tree where the taxa of some undisputed clades were clustered together. The constrained topology contained the major and minor Antarctic globin sequences as fixed groups because their topology was well supported by bootstrap analysis and was congruent with previous assessments (Verde et al. 2003). Finally, the user tree option was used to pick up the best tree among the 1000 trees inferred by exhaustive search. The relative bootstrap probabilities were computed with the ReEstimation of Log Likelihood (RELL) method. Bootstrap values for internal tree branches were calculated as the sum of the bootstrap probabilities of all the trees showing the node being analyzed among the alternatives.
Oxygen binding
Hemolysate stripping was carried out as described (Tamburrini et al. 1994). Oxygen equilibria were measured in 100 mM HEPES/ MES in the pH range 6.0–9.4, at 10°C and 16°C (keeping the pH variation as a function of temperature in due account) at a final Hb concentration of 0.5–1.0 mM on a heme basis. An average standard deviation of ±3% for values of p50 was calculated; experiments were performed in duplicate. To obtain stepwise oxygen saturation, a modified gas diffusion chamber was used, coupled to cascaded Wösthoff pumps for mixing pure nitrogen with air (Weber et al. 1987). In the diffusion-chamber experiments, the oxygenation range was between 10% and 90%. pH was measured with a Radiometer BMS Mk2 thermostatted electrode. Sensitivity to chloride was assessed by adding NaCl to a final concentration of 100 mM. ATP or inositol hexakisphosphate (IHP) were added to a final concentration of 3 mM; hence, a large excess over tetrameric Hb. Oxygen affinity (p50) and cooperativity (n50) were calculated from the linearized Hill plot of logS/(1–S) versus logpO2 at half saturation (S denotes fractional oxygen saturation).
The overall oxygenation enthalpy change H (kcal mole–1; 1 kcal = 4.184 kJ), corrected for the heat of oxygen solubilization (–3 kcal mol–1), was calculated by the integrated van’t Hoff equation H =–4.574[(T1 T2)/(T1 – T2)] logp50/1000.
Molecular modeling
As the crystal structure of P. urvillii Hb 1 has not yet been established, molecular modeling studies were carried out on a Silicon Graphics O2 R12000
[GenBank]
workstation using Insight II (Accelrys Inc.).
Sequence similarity searches within the Protein Data Bank were performed using PSI-BLAST running on the server at NCBI (http://www.ncbi.nlm.nih.gov/BLAST/). The sequence alignments were obtained by CLUSTAL W (Thompson et al. 1994). The PostScript files from CLUSTAL W alignment results were generated with the ESPript program (Gouet et al. 1999).
Crystallographic structures of deoxy and carbomonoxy T. bernacchii Hb (PDB codes 1hbh
[PDB]
and 1pbx
[PDB]
, respectively; Camardella et al. 1992; Ito et al. 1995) were used as templates for building corresponding models of Hb 1 of P. urvillii. As the PDB coordinates of oxy T. bernacchii Hb were not available, a model of oxy Hb 1 of P. urvillii was built based on human HbA (PDB code 1hho
[PDB]
; Shaanan 1983), which displayed 51% and 47% sequence identity with the - and -chains of P. urvillii Hb 1, respectively.
Based on CLUSTAL W alignments, three-dimensional models of - and -chains of deoxy, carbomonoxy and oxy Hb 1 of P. urvillii were then generated by the MODELER program (Sali and Blundell 1993) implemented in Insight II. This program is based on a distance-restraint algorithm using spatial restraints extracted from a multiple alignment of the target sequences with the template structures and from the CHARMM-22 force field (Brooks et al. 1983). Three models, optimized by a short simulated annealing refinement protocol available in MODELER, were generated for each alignment. The model with the lowest value of MODELER objective function (F, molecular probability density function violation) was selected as the representative model for each chain in each structure. To further validate the modeling procedure, models obtained by MODELER were analyzed by the program PROCHECK (Laskowski et al. 1993) to confirm their stereochemical quality.
Spectroscopy
As P. urvillii Hb 1 was isolated in the oxy form, the oxy samples were used without further manipulation. The Met-Hb derivatives were prepared by oxidation of the oxy form using excess potassium ferricyanide followed by gel filtration on a Biogel P-6DG column equilibrated with the appropriate buffer to remove the oxidant. The deoxy samples were prepared by adding 2–3 µL of sodium dithionite (20 mg/mL) to 50 µL of deoxygenated buffered solution of Met-Hb. The carbomonoxy complex was prepared by flushing Met-Hb with nitrogen, then with CO and reducing the protein as described above. Fluoride (F–) adducts were prepared by adding 2 µL of a saturated solution of sodium fluoride, dissolved in 100 mM sodium citrate (pH 5.0), to 40 µL of solution of Met-Hb in 5 mM MES at pH 6.0. The protein concentration of all samples, for both resonance Raman (RR) and electronic absorption spectroscopy, was 30 µM on a heme basis.
Electronic absorption spectra were measured with a Cary 5 spectrophotometer. Infrared spectra were recorded with a Bruker IF120HF FTIR spectrophotometer. Carbomonoxy Hb samples (250 µM) were transferred by a gas-tight syringe flushed with CO into a CaF2 IR cell (path length, 0.1 mm), previously flushed with CO. Tris-HCl at pH 7.6 (10 mM) was used as reference. RR spectra were obtained at room temperature with excitation from the 406.7- and 413.1-nm lines of a Kr+ laser (Coherent, Innova 302C). The back-scattered light from a slowly rotating NMR tube was collected and focused into a computer-controlled double monochromator (Jobin-Yvon HG2S) equipped with a cooled photomultiplier (RCA C310344A) and photon-counting electronics. To minimize local heating by the laser beam, all samples except deoxy Hb were cooled by a gentle flow of nitrogen run through liquid nitrogen. RR spectra were calibrated to an accuracy of 1 cm–1 for intense isolated bands, using indene as standard for the high-frequency region, and indene and CCl4 for the low-frequency region.
The protein sequence data reported in this paper will appear in the Swiss-Prot and TrEMBL knowledgebase under the accession numbers P83623
[GenBank]
(), P83624
[GenBank]
(1), P83625
[GenBank]
(2).
|
Acknowledgments |
|---|
The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.
|
References |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Andrews, A.P. 1980. Family Bovichthydae: Congolli. In Freshwater fishes of south-eastern Australia (ed. R.M. McDowall), pp. 167–168. Reed, Sydney.
Antonini, E. and Brunori, M. 1971. Hemoglobin and myoglobin and their reaction with ligands (eds. A. Neuberger and E.L. Tatum), pp. 17–19. North Holland, Amsterdam.
Balushkin, A.V. 1992. Classification, phylogenetic relationships, and origins of the families of the suborder Notothenioidei (Perciformes). J. Ichthyol. 32: 90–110.
Bargelloni, L., Ritchie, P.A., Patarnello, T., Battaglia, B., Lambert, D.M., and Meyer, A. 1994. Molecular evolution at subzero temperatures: Mitochondrial and nuclear phylogenies of fishes from Antarctica (suborder Notothenioidei), and the evolution of antifreeze glycopeptides. Mol. Biol. Evol. 11: 854–863.[Abstract]
Bargelloni, L., Zane, L., Derome, N., Lecointre, G., and Patarnello, T. 2000. Molecular zoogeography of Antarctic euphausiids and notothenioids: From species phylogenies to intraspecific patterns of genetic variation. Antarctic Sci. 12: 259–268.
Brittain, T. 1987. The Root effect. Comp. Biochem. Physiol. 86B: 473–481.[CrossRef]
Brooks, B.R., Bruccoleri, R.E., Olafson, B.D., States, D.J., Swaminathan, S. and Karplus, M. 1983. A program for macromolecular energy, minimization, and dynamics calculations. J. Comput. Chem. 4: 187–217.
Camardella, L., Caruso, C., D’Avino, R., di Prisco, G., Rutigliano, B., Tamburrini, M., Fermi, G., and Perutz, M. 1992. Haemoglobin of the antarctic fish Pagothenia bernacchii. Amino acid sequence, oxygen equilibria and crystal structure of its carbonmonoxy derivative. J. Mol. Biol. 224: 449–460.[CrossRef][Medline]
Caruso, C., Rutigliano, B., Romano, M., and di Prisco, G. 1991. The hemoglobins of the cold-adapted Antarctic teleost Cygnodraco mawsoni. Biochim. Biophys. Acta 1078: 273–282.[CrossRef][Medline]
Case, D.A. and Karplus, M. 1979. Dynamics of ligand binding to heme proteins. J. Mol. Biol. 132: 343–368.[CrossRef][Medline]
Cheng, C.-H.C., Chen, L., Near, T.J., and Jin, Y. 2003. Functional antifreeze glycoprotein genes in temperate-water New Zealand nototheniid fish infer an Antarctic evolutionary origin. Mol. Biol. Evol. 20: 1897–1908.
Choi, S., Spiro, T.G., Langry, K.G., Smith, K.M., Budd, D.L., and La Mar, G.N. 1982. Structural correlations and vinyl influences in the Resonance Raman spectra of protoheme complexes and proteins. J. Am. Chem. Soc. 104: 4345–4351.[CrossRef]
D’Avino, R. and di Prisco, G. 1988. Antarctic fish hemoglobin: An outline of the molecular structure and oxygen binding properties—I. Molecular structure. Comp. Biochem. Physiol. 90B: 579–584.[CrossRef]
De Rosa, M.C., Castagnola, M., Bertonati, C., Galtieri, A., and Giardina, B. 2004. From the Arctic to the fetal life: Physiological importance and structural basis of an "additional" chloride binding site in hemoglobin. Biochem. J. 380: 889–896.[CrossRef][Medline]
di Prisco, G. 1998. Molecular adaptation of Antarctic fish hemoglobins. In Fishes of Antarctica. A biological overview (eds. G. di Prisco et al.), pp. 339–353. Springer, Milano, Italy.
di Prisco, G. and Tamburrini, M. 1992. The hemoglobins of marine and freshwater fish: The search for correlations with physiological adaptation. Comp. Biochem. Physiol. 102B: 661–671.
di Prisco, G., D’Avino, R., Caruso, C., Tamburrini, M., Camardella, L., Rutigliano, B., Carratore, V., and Romano, M. 1991. The biochemistry of oxygen transport in red-blooded Antarctic fishes. In Biology of Antarctic fish (eds. G. di Prisco et al.), pp. 263–281. Springer-Verlag, Berlin.
Eastman, J.T. 1993. Antarctic fish biology: Evolution in a unique environment. Academic Press, Inc., San Diego, CA.
———. 2000. Antarctic notothenioid fishes as subjects for research in evolutionary biology. Antarctic Sci. 12: 276–287.
Fago, A., D’Avino, R., and di Prisco, G. 1992. The hemoglobins of Notothenia angustata, a temperate fish belonging to a family largely endemic to the Antarctic Ocean. Eur. J. Biochem. 210: 963–970.[Medline]
Feis, A., Marzocchi, M.P., Paoli, M., and Smulevich, G. 1994. Spin state and axial ligand bonding in the hydroxide complexes of metmyoglobin, methemoglobin, and horseradish peroxidase at room and low temperatures. Biochemistry 33: 4577–4583.[CrossRef][Medline]
George, P., Beetlestone, J., and Griffith, J.S. 1964. Ferrihemoprotein hydroxides: A correlation between magnetic and spectroscopic properties. Rev. Mod. Phys. 36: 441–458.
