Prevalent in many bacterial and fungal pathogens, flavohemoglobins [21] are also known as a “fungal defense” enzymes (Fig. 2). They have been shown to be closely tied to virulence. Upon infectious invasion, macrophages engulf pathogens and produce oxidants like nitric oxide to protect the organism. De Jesus-Berrios et al. [22] have shown that inactivation of flavohemoglobin led to a significant loss of virulence both in vitro and in vivo. Amino acid sequence analyses indicate that bacterial flavohemoglobins have 3/3 Hb fold and are closer to metazoan globins than most of the bacteria hemoglobins, which display a 2/2 Hb fold.

Globin coupled sensors (GCSs) form a large group of chimeric proteins having a N-terminal globin domain acting as a sensor to stimulate a signal controlled pathway [23]. They are one of the four known groups of heme sensors protein, among which one of the most studied was the kinase FixL, a PAS group protein [24]. In these molecules, the heme group, through a conformation change, induced by ligand binding, switch on or off an oxygen dependent metabolic pathway activating some genes. The first GCSs described were regulators of the aerotactic responses in Bacillus subtilis (HemAT-Bs) and Halobacterium salinarum (HemAT-Hs) [25]. Today, more than 30 GCSs are known and the evolution tree of their globin part suggests that they derived from a common ancestor, the “protoglobin”. Such protoglobins have indeed been discovered in two Archaea and two Bacteria [17]. Protoglobin from the strictly anaerobic methanogenic Archaea (Methanosarcina acetivorans) has been recently crystallized and revealed an almost completely buried heme group being only accessible to ligand through apolar tunnels [26]. GCSs could be classified in several categories: the first is made by aerotactics and contains the HemATs. Another group is formed by “second messengers” coupled, for example, to a guanylyl cyclase catalytic domain [23]. Two GCSs were found to be involved in protein-protein interactions through a STAS domain (sulfate transporter and antisigma factor antagonist). Finally, some GCSs were found to be tightly bound to the membrane through a domain which spans the membrane several times.

The evolutionary pathways of GCS and protoglobin are not yet clear. The initial ancestor could be a multidomain protein which split into components that further evolved on their own. Conversely, another hypothesis is that evolution may have started from individual domains forming a multidomain protein in which the two partners evolved together.

The presence in all three kingdoms of life of a large variety of globin molecules belonging to the different lineages raises the question of their functions. It is assumed that the globin ancestor was able to react with diatomic gaseous ligands inducing a change configuration that could be used in many tasks. Hemoglobin is therefore able to detect, scavenge, and detoxify O₂ and O₂-derived species (e.g., NO and CO). The principal role of existent 2/2 Hbs is considered to be NO detoxification [29]. It is also the case for flavohemoglobins acting aerobically as NO oxygenases and anaerobically as NO reductases [30]. This NO detoxification is linked to bacterial resistance: it has been shown in pathogenic microorganisms that flavohemoglobins protect the bacteria from human macrophage NO-mediated killing and promotes the virulence of bacteria.

In plants three types of Hbs have been identified: symbiotic, non-symbiotic and 2/2 Hbs. Two classes of Hbs evolved prior to the divergence of monocots and dicots. Leghemoglobins, which are pentacoordinate O₂ transporters, emerged from the first class, whereas hexacoordinate scavenger/sensor Hbs emerged from both classes [31]. The function of the symbiotic hemoglobins is to establish a low oxygen concentration for the effective functioning of the bacterial nitrogenase necessary for nitrogen fixation. Although the role of 2/2 Hbs in plants is completely unknown, the non-symbiotic hemoglobins, which biosynthesis is induced in plant cells upon exposure to low oxygen concentrations, are thought to play a role in a metabolic pathway which also involves nitric oxide and which provides an alternative type of respiration to the mitochondrial electron transport under hypoxic conditions [32,33].

In the animal kingdom, the 3/3 Hbs and 2/2 Hbs are usually involved in oxygen transport and storage and NO detoxification, respectively. Nevertheless, in some species, hemoglobin molecules developed special evolution pattern allowing their carriers to occupy new ecological niches [29].

In some invertebrates, hemoglobins are found within the circulating cells of the coelomic fluid and/or within muscular type cells or cells of organ that need to be protected from anoxia or oxidant stress. These molecules are usually monomers or made from a small number of subunits.

In other species, such as insects, worms or mollusks, Hbs are freely dissolved in vascular, coelomic, or perienteric body fluids. They are circulating as giant polymers [34,35]. This is observed in the erythrocruorins from several oligochaetes (earthworm), achaetes (leech), vestimentiferans (hydrothermal vent tube worm) and polychaetes. Lumbricus erythrocruorin, shown here as an example, assembles from 144 oxygen-binding hemoglobin subunits and 36 structural “linker” subunits. Four distinct hemoglobin subunits are arranged into 12 dodecamers, which assemble onto a central core formed from linker subunits (Fig. 3) [36]. This supramolecular assembly involving several disulfide bonds raises many questions about the synchronization of the gene expression for the different structural elements and the subsequent folding and assembly mechanism. Such large complexes offer a number of important advantages as oxygen transport vehicles: erythrocruorins are retained in the vascular system, each complex has large oxygen binding capacity and subunits can be arranged to permit cooperative oxygen binding and regulatory features. Lumbricus terrestris erythrocruorin is probably the best studied giant Hb; it has been proposed as a potential model for a blood substitute [37].

The globins have often served as a model for studying structure-function relations in proteins. However, they are not necessarily an easy system to describe for ligand binding, since the crystal structure did not reveal a ligand binding pathway. Rather than a lock and key type system involving a large surface area, ligand binding apparently requires protein fluctuations. The static structure therefore is not sufficient to provide information such as trajectories and affinities [1].

The globins of very high oxygen affinity led to a new description of the critical structural elements. It was originally thought that it was essentially an affair of iron-oxygen binding, in which the heme pocket provides a minor modulation, especially by the distal (E7) residue. It has now been shown that both the E7 and B10 position can participate in H-bonding directly to the oxygen molecule. This places the center of interest on the oxygen ligand itself, capable of forming up to three liaisons. The E7 and B10 positions will provide the bonds for major changes in affinity, which now approach a variation of nearly a factor of a million (see Fig. 4) [38,39].

Globins displaying these extremely high affinities will necessarily have specific residues in these positions capable of forming the H-bonds. Note that the specific residue in the B10 position may be a Tyr or a Phe, depending on the distance requirements. In the E7 position, the classical His residue may form a bond, but the very high affinities seem to employ a Gln residue.

The discovery of neuroglobin has lead to the formulation of an alternate mechanism of oxygen binding. The external ligands (such as oxygen, CO, or NO) may still bind to the iron atom, but in competition with the distal E7 His residue [40,41]. In the absence of external ligands, the ferrous iron may not be in a penta-coordinated state, since the distal His will bind to form a hexa-coordinated system [40]. Other globins such as cytoglobin, and new species of globin also display this mechanism of competitive binding; they are now collectively classed as hexa-coordinated globins to stress the fact that the penta-coordinated state is never more than a transient state.

Note that this mechanism has some interesting physiological consequences. The overall (observed) oxygen affinity is no longer the simple ratio of oxygen binding rates, but rather will be a ratio of the intrinsic oxygen and His binding affinities.

Another direct consequence of the competition is a reduced effect of external factors, such as temperature and pH. While the individual rate coefficients may be sensitive to temperature, the overall affinity is the ratio of affinities. If both the internal (His) and external (O₂) ligands have a similar response to temperature, the overall observed oxygen affinity may be nearly independent of temperature. This has been observed for oxygen binding to various Hex-globins, and even a reverse temperature effect for CO binding [41].

In species where Hbs play only a local role or a function in a signal pathway, oxygen transport is usually assumed by other respiratory pigments such as hemocyanin.

Hemoglobin from Ascaris works probably as an oxygen scavenger. Two Hbs, that have been thoroughly studied, are present in this species. The first one is an extracellular one that circulates as a 328 kDa octamer. Each of the 8 subunits is made from a 39.5 kDa polypeptide chain formed by two covalently bound globin domains (D1 and D2). In the C-terminal part of D2, the sequence “His-Lys-Glu-Glu” is repeated 4 times and is likely to play the role of a zipper for the association into octamer [42]. Electron microscopy pictures show that the octamers are made by two layers formed each by 4 subunits. This Hb displays one of the highest known oxygen affinities: its P₅₀ is of about 0.004 mm Hg, which is 25 000 times higher than that of mammalian Hbs [39,43,44]. Since association rates are limited by diffusion, these very high affinities necessarily display a considerably low value of the dissociation rate, making difficult the release of the bound oxygen. This property will hardly make this molecule suitable for oxygen transport requiring a prompt uptake and release. At the limit it could be a way to store oxygen in order to assume a minimal oxygen supply in case of extreme anaerobic conditions. Conversely, it has also been proposed that this Hb could be an oxygen scavenger allowing its destruction through a pathway involving NO [45]. In fact, nothing precise is known about the physiological role of this Hb, which remains a challenge. In the case of this Hb the only strong data are the molecular mechanism leading to this considerable oxygen affinity which consists in a network of hydrogen bonds extending in the heme pocket between a Gln at position E7 and the Tyr B10 stabilizing the Fe(δ+)–O₂(δ−) dipole binding [39].

In the same species another type of Hb is present in the wall cells of the animal central cavity. It is a homodimer of 40.6 kDa stabilized by SH bridges. Its sequence is 38% similar to that of domain D1 and its P₅₀ is of about 0.1 mm Hg, which could allow a role for oxygen storage at the proximity of muscular cells. In the animal kingdoms, Hbs of nematodes, like neuroglobins and cytoglobins, are the only globins encoded by a gene containing 3 intervening sequences.

In some marine invertebrates, Hb is involved in sulfur transport. For example, the giant extracellular hexagonal bilayer hemoglobin (HBL-Hb) of the deep-sea hydrothermal vent tube worm Riftia pachyptila transports simultaneously O₂ and H₂S in the blood from the gills to a specific organ, the trophosome, that harbors sulfide-oxidizing endosymbionts. Within this organ, symbiotic bacteria use the sulfur as final electron acceptor in a metabolic pathway transforming the dissolved CO₂ into sugars [46]. Three Hbs are present in these animals: the vascular blood contains Hbs V1 and V2, and the coelomic fluid comprises Hb C1. These Hbs are able to bind oxygen and sulfide simultaneously and reversibly at two different sites. The amount of H₂S that they transport is high, varying from 1–10 mmol/L, depending upon its concentration in the surrounding medium. Hb of Riftia pachyptila is a giant polymer made from a basic unit formed by 4 chains with 1 free thiol group, connected by linker polypeptides. The MW of HbV1 is of about 3500 kDa, contains about 150 globin chains and some 40 linkers [47]. Hb V2 and C1 are smaller, with a mw of about 400 kDa, and contains 24 globin chains, many free SH groups are present.

Another example of Hb with high affinity for thiol group is observed in Lucina pectinata. This tropical clam contains two kinds of Hbs: the first one (HbI) is a monomer, 143 residues long, which serves to transport H₂S to autotrophic bacteria whereas two other globins (HbII and Hb III), respectively dimeric and tetrameric in the μM range of [heme], are used for oxygen transport [48,49]. Site directed mutagenesis introducing a Gln residue at position E7 and a Phe at B10 and E11 in sperm whale myoglobin led to a Hb molecule with a high affinity for H₂S but nevertheless with one order of magnitude less than in Lucina pectinata. Even though further modifications of the heme pocket are thus required for the fine tuning mechanism of ligand binding this underlines the presence of key residues located at the distal side of the heme pocket for switching from O₂ to H₂S transport. Same conclusion holds inside the different Lucina pectinata globins. Indeed for Hb II and Hb III while the E7 and E11 residues do not change compared to Hb I the B10 Phe is replaced by a Tyr leading almost to three orders of magnitude decrease of the O₂ dissociation rate. Thus the addition of a single hydroxyl group is able to stabilize the dipolar Fe–O₂ bond through a favorable hydrogen bond, perhaps in association with a network of electrostatic interactions such as those in the Ascaris globin explains the difference of O₂ off-rates between these globins. Note that the B10 residue is also involved in the large decrease of the O₂ off-rates for Hb II and Hb III compared to Hb I and more generally in the ligand accessibility from the solvent to the heme.

As can be expected from the competitive ligand binding, a modification of the His affinity will directly influence the observed oxygen affinity. The effect is exploited by neuroglobin via an internal disulfide bond: when the S–S bond is broken, the His affinity increases by an order of magnitude, thereby decreasing the oxygen affinity [40]. One could therefore imagine a partner molecule which reduces the cysteines of neuroglobin and thereby provokes the release of oxygen. This would lead to a model of Ngb as an allosteric protein providing oxygen delivery on demand of a specific partner.

Neuroglobin and cytoglobin have shed light on an alternate globin mechanism for modulating the ligand binding. These globins belong to the class of hexacoordinated globins: in the absence of external ligands, the distal E7 His forms a stable bond with the heme iron. This bond is reversible so that the external ligand binding requires the replacement of the constitutive residue. The globins are then able to use a mechanism competitive ligand binding to the heme iron. The fine tuning mechanism appears more in the stress induced at a distance by an intra-disulfide bond; note that the location of this disulfide bond is different for Ngb (CD loop) vs Cygb (spanning the E helix with AB turn). Breaking the disulfide bond allows a shift in the orientation of the E-helix, resulting in an enhanced affinity of the distal E7 His for the heme and thereby a release of O₂. The redox state of the cell could act on the state of the cysteines of the neuroglobin and further influence other biochemical reactions involving an electron transfer to the molecular oxygen or another electron acceptor molecule and the NO catabolism.

Hb from the nematode Mermis nigrescens (Mn-GLB-E), a parasite of grasshoppers and other Orthopteran insects, has an optical, light shadowing function [50]. The protein accumulates to high concentration as intracellular crystals in the ocellus of mature phototactic adult females while also being expressed at low concentration in other tissues. From eggs ingested by the host, larvae grow to the adult length inside the body cavity. The last larvae stage break out of the host and burrow into the soil where the final molt takes place and the adult nematode matures. The gravid adult females emerge 1–2 years later from the soil and exhibit a positive phototaxis during the search for suitable egg-laying sites in grass. This positive phototaxis results from a pigmentation of the ocellus by formation of a thick Hb crystal. A sensory nerve ending lies anterior to this area and will point the head away from a light source as long as the ocellus would not be shadowed.

The neural and body wall tissue of a nemertean worm, Cerebratulus lacteus, express a major Hb components for which the amino acid sequences have been deduced from cDNA and genomic DNA [51]. This Hb is 109 residues long and forms the smallest stable Hb known. The globin genes have three exons and two introns with splice sites in the highly conserved positions of globin genes. Alignment of the sequences with those of other globins indicates that helix A is almost missing, helix H is very short and GH corner elongated. Nevertheless the general 3D structure is that of a globin fold. It has been hypothesized that this Hb could be an ancestor of neuroHbs.