Hb is known to the reader because this protein is responsible for the red color of vertebrates’ blood ¹⁹ . However, Hbs are widely distributed in living organisms, ranging from bacteria to mammals ^{20– 22} . The tertiary structure of Hbs consists of a specific arrangement of 6 to 8 α-helices (designated with letters A to H) known as the globin-fold. This protein folding forms a hydrophobic pocket where a Fe-heme prosthetic group is located ^{19, 23, 24} . Two structural types of the globin-fold have been identified in Hbs: the 2/2- and 3/3-folding. In the 2/2-Hbs, helices B and E overlap to helices G and H and in the 3/3-Hbs helices A, E and F overlap to helices B, G and H. Likewise, three evolutionary families have been identified in Hbs: the M, S and T Hb families. The M Hbs, which exist in bacteria and eukaryotes, include flavoHbs and single domain globins, the S Hbs, which exist in bacteria and yeasts, include globin-coupled sensors, protoglobins and single domain globin sensors, and the T Hbs, which exist in bacteria, unicellular eukaryotes and plants, include truncated Hbs (tHbs). Canonical T Hbs from bacteria and unicellular eukaryotes are ~100 to 120 amino acids in length, however plant T Hbs are longer than canonical T Hbs because of the existence of extra amino acids at the N- and C-terminal. The M and S Hbs fold into the 3/3-folding whereas the T Hbs fold into the 2/2-folding ^{21, 25– 28} .

A variety of ligands bind to the Fe-heme of Hbs, including O ₂ and NO. Reversible binding of O ₂ is closely associated to the major function of Hbs in organisms, which is the transport of O ₂ ¹⁹ . Binding of NO by Hbs is essential to NO-detoxification via a NO-dioxygenase activity ²⁹ . Several additional functions have been reported for Hbs, including dehaloperoxidase activity and reaction with free radicals, binding and transport of sulfide and lipids, and O ₂-sensing ^{30– 39} . This indicates that in vivo Hbs might be multifunctional proteins.

Land plant Hbs were first identified by Kubo in soybean root nodules ⁴⁰ . Few years after Kubo’s discovery these proteins were named as leghemoglobins (Lbs) by Virtanen and Lane ⁴¹ because they were only found in the symbiotic (N ₂-fixing-) nodules of the leguminous plants. Lbs are the most abundant soluble proteins in nodules ( e.g. in soybean nodules their concentration is as high as 3 mM) ^{16, 42} . The x-ray analysis of lupin Lb revealed that the tertiary structure of Lbs was remarkably similar to that of the sperm whale myoglobin ⁴³ . This evidence demonstrated that Lbs are plant Hbs and indicated that plant and animal Hbs evolved from a common ancestor more than 600 mya ⁶ . Subsequent work led to the identification of Lb-like (or symbiotic) Hbs in nodules of actinorhizal plants ^{44– 48} , purification of an Hb from the root nodules of the dicotyledonous non-legume Parasponia andersonii ⁴⁹ , cloning and sequencing of an hb gene from the non-nodulating dicot Trema tomentosa ^{50, 51} and detection of Hbs in non-symbiotic organs from several land plants, including primitive bryophytes and evolved angiosperms ^{9, 18, 52– 54} . Until now three types of Hbs have been identified in land plants: the symbiotic Hbs, which include Lbs, that are specifically located within nodules of the N ₂-fixing land plants, and the non-symbiotic (nsHbs) and truncated (tHbs) Hbs, that are located within non-symbiotic and symbiotic organs of primitive and evolved land plants ^{9, 18} . Based on sequence similarity the nsHbs are further classified into type 1 and type 2 nsHbs (nsHbs-1 and nsHbs-2, respectively) ^{9, 17, 55} .

Monocots are a large family of flowering plants ^{56– 58} that includes cereals. Cereals, such as rice, maize and wheat, are the main source of food for humans ⁵⁹ . Because of this, during that last decade the genomes of a number of cereals have been sequenced. This allowed the identification of novel cereal Hbs. The search of hb genes in databases by G. Rodríguez-Alonso and R. Arredondo-Peter ^{60, 61} revealed that nsHb and tHb sequences exist in the Brachypodium distachyon, Hordeum vulgare (barley), Oryza glaberrima (rice), O. rufipogon (rice), O. sativa (rice) var. indica, O. sativa (rice) var. japonica, Panicum virgatum (switchgrass), Setaria italica (foxtail millet), Sorghum bicolor (sorghum), Triticum aestivum (wheat) and Zea mays ssp. mays (maize) genomes. The highest number of nsHbs (5) exists in O. sativa var. indica and O. sativa var. japonica, whereas one to three nsHbs exist in barley, Brachypodium, foxtail millet, maize, O. glaberrima, O. rufipogon, sorghum, switchgrass and wheat. Also, with the exception of wheat, which contains two copies of the thb gene, a single copy of thb was identified in the genome of Brachypodium, barley, O. sativa var. indica, O. sativa var. japonica, switchgrass, foxtail millet, sorghum and maize. Little is known about Hbs from non-cultivated monocots. The only Hb reported from a non-cultivated monocot is that of teosinte ( Z. mays ssp. parviglumis) ³ , which is postulated as the ancestor of maize ^{59, 62– 67} . Analysis by Southern blot using the teosinte hb gene as probe showed that apparently a single copy of hb exists in teosinte (J. Sáenz-Rivera and R. Arredondo-Peter, unpublished results). Sequence comparison revealed that maize and teosinte Hb polypeptides are identical ³ .

Monocots were a target for searching Hbs after these proteins were detected in non-symbiotic organs of dicotyledonous plants (see subsection above). At that time, monocot genomes had not been sequenced. Searching approaches consisted in detecting Hb polypeptides and hb genes by spectroscopy and molecular biology methods, respectively. Attempts to detect absorption maxima in the Søret (~410 nm) and Q (~500 to 550 nm) regions, which are characteristic of ferric (Fe ³⁺), ferrous (Fe ²⁺) and liganded Hbs ^{68, 69} , were unsuccessful (R. V. Klucas and C. A. Appleby, unpublished results) mostly due to the very low Hb concentration (~50 to 100 nM) in plant non-symbiotic organs ^{5, 70} . At the molecular level a consensus probe designed from legume and non-legume ( T. tomentosa, P. andersonii and Casuarina glauca) Hb sequences ⁷¹ hybridized with hb-like sequences from rice and other monocot total DNAs ( Figure 1). This observation suggested that hb sequences exist in monocots, however hybridizing fragments were not subsequently cloned and sequenced in order to verify if they actually corresponded to hb genes.

Rice Expressed Sequence Tags (ESTs) were first deposited in databases early in the 1990’s. The first rice Hb (Hb1 and Hb2) sequences were detected from ESTs deposited in the DNA Data Bank of Japan (DDBJ) database ⁷² . Rice Hb1 and Hb2 corresponded to clones C741 and C2576 with DDBJ accession number D15507 and D38931, respectively. Rice hb1 and hb2 genes were subsequently amplified by PCR, cloned and sequenced. Sequence analysis revealed that rice hb1 codes for non-symbiotic Hb1 and that rice hb2 codes for non-symbiotic Hb2 ² . Afterwards, sequencing of the rice ( O. sativa L. ssp. indica) genome more than a decade ago ⁷³ allowed the identification of a family of rice nshb genes and a single copy of the rice thb gene (see subsection below).

The O. sativa var. indica and O. sativa var. japonica genomes are fully sequenced, and the O. glaberrima and O. rufipogon genomes are partially sequenced. Rice genome sequences are mainly available from the GenBank ( www.ncbi.nlm.nih.gov) and Phytozome ( http://www.phytozome.org/) databases. Search of Hb sequences in the above databases showed that a family of the nshb genes, consisting of hb1, hb2, hb3, hb4 and hb5, and a single copy of the thb gene exist in the O. sativa var. indica and O. sativa var. japonica genomes. A single copy of the nshb gene was detected in the O. glaberrima and O. rufipogon genomes, however thb genes have not yet been detected in these plants ⁶¹ . Given that the sequencing of the O. glaberrima and O. rufipogon genomes is in progress the identification of hb genes in these genomes is incomplete. Thus, the following discussion will focus on the O. sativa var. indica and O. sativa var. japonica hbs.

The structure of known rice hb genes corresponds to four exons and three introns, with introns located at similar position as all of the known plant hb genes ⁷⁴ . Canonical TATA boxes and a variety of potential promoters exist upstream of the rice hb genes which suggests that rice hbs are functional and that the regulation of the hb genes in this plant is complex ^{75– 77} . Figure 2 shows the localization of hbs in the O. sativa chromosomes and mapping of hbs in the O. sativa genome. Rice hb1, hb3 and hb4 cluster forming the hb1- hb4 cluster ⁷⁶ which is localized in chromosome 3. Rice hb2 is also localized in chromosome 3 but 467 kb upstream of the hb1- hb4 cluster. In contrast, rice hb5 and thb genes are localized in chromosomes 5 and 6, respectively ( Figure 2A). Rice hbs are flanked by a variety of genes with known and unidentified functions ( Figure 2B). However, with the exception of genes coding for a ternary complex factor macrophage inflammatory protein MIP1 and an ubiquitin fusion protein which are located 239 and 411 nucleotides up- and downstream of the hb1- hb4 cluster, respectively, distance of flanking genes to hbs is >1 kb. This suggests that co-expression of hb and flanking genes is unlikely.

The expression of hb genes and localization of Hb polypeptides have been analyzed in rice growing under normal and stressed conditions. Under normal conditions the expression level of rice nshbs was low ^{2, 75} . However, analysis by RT-PCR revealed that hb1, hb2 and hb5 genes were expressed in embryonic and vegetative organs obtained from rice plants grown under a normal environment ^{2, 75, 78} . Specifically, transcripts for rice Hb1 were detected in embryos, seminal roots, leaves and roots, transcripts for rice Hb2 were detected in embryos, coleoptiles, seminal roots and leaves, and transcripts for rice Hb5 were detected in embryos, coleoptiles, seminal roots, leaves and roots. Likewise, evaluation of the β-glucuronidase (GUS) activity from a construct containing the rice nshb2 gene promoter that is responsive to the cytokinin-regulated ARR1 trans-acting factor showed that this promoter is activated in roots, the vasculature of young leaves, flowers and the pedicel/stem junction of transgenic Arabidopsis ⁷⁷ . In addition, a variety of potential promoters was identified upstream of the rice nshb genes, such as those involved in the ethylene synthesis, photoregulation, heat shock response and plant defense signaling ^{70, 75– 77} . However the activities of these promoters have not been determined.

Transcriptomic analyses revealed that nsHb and tHb transcripts coexist in rice embryonic and vegetative organs ( Table 1). This evidence suggests that nsHb ( i.e. Hb1, Hb2, Hb3, Hb4 and Hb5) and tHb polypeptides coexist and probably function in rice organs. Immunoanalysis by Western blot and confocal microscopy using a polyclonal anti-rice Hb1 antibody revealed that Hb polypeptides exist in rice seeds and in rice leaves and roots from 2 to 14 weeks after seed germination. These analyses also revealed that Hb polypeptides exist in the cytoplasm of differentiating cells of the root cap, schlerenchyma, aleurone, and in the vasculature, principally in the differentiating xylem ^{16, 70, 79} . However, the anti-rice Hb1 antibodies cross-react with different rice Hbs (G. Sarath and E. J. H. Ross, unpublished results) and thus it is not known which Hb polypeptides were detected in the above analyses by the anti-rice Hb1 antibodies.

It is well documented that land plant hb genes are either up- or down-regulated by stress conditions ^{1, 52, 54, 79– 82} . Table 1 shows that Hb transcripts coexist in rice growing under cold, drought and salt stress conditions. Also, Ohwaki and co-workers ⁸³ reported that nshb1 and nshb2 are induced by nitrate, nitrite and NO in cultured rice cells. These observations indicate that rice hb genes response to a variety of stress conditions. However, the detection of Hb polypeptides by Western blot using the anti-rice Hb1 antibodies showed that level of Hbs increased in rice etiolated leaves and flooded roots, but not in rice plants subjected to oxidative (H ₂O ₂), nitrosative (SNP) and hormonal (2,4-D) stresses. These observations suggest that rice Hbs do not appear to be part of a generalized stress response, but may be functional in plant organs subjected to specific stress conditions ⁷⁹ .

Rice hb genes are functional and code for Hb polypeptides with a predicted molecular mass of ~16 to 19 kDa ^{2, 74– 76} . Rice Hb1 was the first monocot nsHb whose crystal structure was elucidated ⁸⁴ . This protein crystalizes as a dimer, thus it is possible that in vivo rice Hb1 forms dimers when its concentration is ≥1 mM ⁸⁵ . After the elucidation of the rice Hb1 structure the tertiary structure of rice Hb2 ⁸⁶ , Hb3, Hb4 and Hb5 ⁷⁵ (CASPUR PMDB ID PM0075009, PM0075873, PM0076005 and PM0075011, respectively) was predicted using computational methods and rice Hb1 (PDB ID 1D8U) as the structural homolog. The crystal structure of rice Hb1 and that of predicted rice Hb2, Hb3 and Hb4 is highly similar. The tertiary structure of these proteins consists of six helices that fold into the 3/3-folding (see subsection on Generalities on hemoglobins). However, the structure of rice Hb1 to 4 is characterized by the existence of a short pre-helix A located at the N-terminal and an extended and poorly ordered CD-loop. The heme pocket in these proteins differs from that in “traditional” Hbs because the proximal and distal His side chains coordinate the Fe-heme forming a hemichrome ( Figure 3), resulting in that Fe-heme from rice Hb1 to 4 is hexacoordinate. Also, the amino acid residues (V50, S53, E123, V124, F127 and A128) located at the monomer-monomer interface of dimeric rice Hb1 ⁸⁴ are highly conserved in rice Hb2 to 4 ⁷⁶ . This suggests that rice Hb1 to 4 can potentially form homo- or hetero-dimers if the hb1 to 4 genes coexpress in rice organs. The tertiary structure of rice Hb5 also consists of six helices that fold into the 3/3-folding. However, rice Hb5 differs from rice Hb1 to 4 in missing 11 amino acids in helix E which results in that the length of the CD-loop and helix E in the predicted Hb5 structure are unusually long and short, respectively. An apparent consequence from this characteristic is that distal His is located far away (13.92 Å, compared to 2.11 Å in rice Hb1) from the Fe-heme within the predicted Hb5 structure, resulting in that Fe-heme from rice Hb5 could be pentacoordinate ⁷⁵ . The amino acid residues located at the monomer-monomer interface of dimeric rice Hb1 ⁸⁴ are poorly conserved in rice Hb5 ⁷⁵ which suggests that rice Hb5 exists in vivo as a monomer.

The folding pathway and kinetics of rice nsHbs were predicted using the Average Distance Map (ADM) method ^{87– 89} . This analysis indicated that rice Hb1 and Hb2 could fold in the C → N direction at a moderate rate, that rice Hb3 could fold in the N → C direction at a fast rate, and that rice Hb4 and Hb5 could fold in the N → C direction at a moderate rate. Thus, it appears that the predicted folding pathway and kinetics among rice nsHbs are diverse. Also, the ADM analysis showed that pre-helix A and CD-loop apparently do not play a role during the folding of rice nsHbs ⁹⁰ . The ADM analysis has not been performed on rice tHb, thus the predicted folding pathway and kinetics for this protein are not known.

Rice ( O. sativa) tHb (GenBank accession number NP_001057972) is 172 amino acids in length, which corresponds to a globin domain (position 26 to 147) flanked by N- and C-terminal extensions. No monocot tHb has been analyzed by x-ray crystallography, however the tertiary structure of a rice tHb was predicted using computational methods ( Figure 3). The predicted structure of rice tHb is highly similar to the crystal structure of an Arabidopsis thaliana tHb ⁹¹ . The globin domain from rice and A. thaliana tHbs folds into the 2/2-folding (see subsection on Generalities on hemoglobins). Similarly to the A. thaliana tHb structure, flanking regions to the globin domain of predicted rice tHb correspond to an N-terminal helical extension and a C-terminal unfolded extension ( Figure 3). The high similarity between the crystal structure of A. thaliana tHb and the predicted structure of rice tHb suggests that the biochemical properties and function of dicot and monocot tHbs are similar.

Visible spectroscopy (see subsection Early search and identification of rice hemoglobins) is a tool to analyze the redox state of and ligand-binding to the Fe-heme of Hbs ^{44, 68, 69, 92, 93} . Rice Hb1 is the only rice nsHb that has been spectroscopically characterized ² . This protein exhibits spectral characteristics that are similar to other Hbs. However, rice Hb1 exhibits distinctive absorption maxima in the deoxyferrous form: the unligated ferrous state exhibits maxima at 526 and 556 nm ² which are characteristic of hexacoordinate Fe-heme ⁹⁴ . This is in contrast to pentacoordinate Hbs which display a broad peak centered at 556 nm in their deoxyferrous form ^{7, 95, 96} . The distal ligand that coordinates the Fe-heme in rice Hb1 was identified as His74 by site directed mutagenesis. Absorbance spectra of the ferric and deoxyferrous forms of an H74L mutant of rice Hb1 showed no evidence of His coordination. Also, the addition of exogenous imidazole to ferric and deoxyferrous H74L mutant resulted in a spectrum identical to that of the wild-type rice Hb1 ² . This evidence indicated that in rice Hb1 the distal ligand to Fe-heme is His74. A similar case can be predicted for rice Hb2 to 4. In contrast, distal His appears to be located far away from the Fe-heme in the predicted structure of deoxyferrous rice Hb5, resulting in that Fe-heme in rice Hb5 could be pentacoordinate ⁷⁵ .

Rice tHb has not been subjected to spectral analysis, however the predicted structure of this protein ( Figure 3) is highly similar to the crystal structure of an A. thaliana tHb ⁹¹ (see subsection on Structure of rice hemoglobins). The absorption spectra of an A. thaliana tHb showed that Fe-heme from this protein is pentacoordinate ^{82, 91} . Thus, it is likely that Fe-heme in rice tHb is pentacoordinate and that the O ₂-binding properties of rice tHb are similar to those of pentacoordante Hbs, i.e. the O ₂-association and -dissociation rate constants are high.

Analysis of ligand-association and -dissociation rate constants of penta- and hexacoordinate Hbs using stopped-flow methods indicated that these proteins exhibit low to moderate and high affinity for O ₂, respectively. Rice Hb1 to 4 are hexacoordinate and apparently rice Hb5 and tHb are pentacoordinate. The O ₂-association rate constants for hexacoordinate rice Hb1 and 2, and possibly for rice Hb3 and 4, are similar to those of other O ₂-transport and -storage proteins, such as the spermwhale myoglobin and soybean Lb a ^{8, 16, 75, 97} . However, in rice Hb1 ² and 2 ¹⁶ , and possibly in rice Hb3 and 4, the bound O ₂ is stabilized by distal His after binding to the Fe-heme, which results in very low O ₂-dissociation rate constants. The O ₂-association and -dissociation rate constants of hexacoordinate rice Hb1 and 2, and possibly of rice Hb3 and 4, result in that the affinity of these proteins for O ₂ is extremely high ( e.g. KO ₂ = 1,800 and 1,316 μM ^-1 for rice Hb1 and 2, respectively ^{2, 16} ). In contrast, the O ₂-association and -dissociation rate constants of pentacoordinate rice Hb5 and tHb could be moderate to high, which could result in a low to moderate affinity for O ₂.

The bis-histidyl hexacoordinated form of rice Hb1 displays a hydrophobic distal cavity which appears to be connected with the external solvent through the position of Phe44 (also known as FB10 because it occupies the tenth position in helix B). It was suggested that this amino acid regulates the migration of small ligands in rice Hb1, for example in ligand binding to the Fe-heme, ligand migration through internal docking sites and ligand release into the external solvent ^{98, 99} . Kinetic analysis after laser flash photolysis of rice Hb1 encapsulated in silica gel combined with computational analysis revealed the existence of two channels in the rice Hb1 CO-bound species. The first channel is located in the distal region of the heme pocket and is connected with a secondary channel that is directly connected with the external solvent. Apparently, the position of FB10 in hexacoordinated rice Hb1 leaves the distal heme pocket accessible to the external solvent, however after the ligand entrance the phenyl ring rotates closing the cavity and thus hindering the exit of the bound ligand ¹⁰⁰ . Thus, together with distal (H74) His (see subsection Spectroscopic characteristics of rice hemoglobins) and aromatic (F40, F41, F55, F57, Y145 and L126) amino acids that are located in the distal region of the heme pocket, FB10 appears to regulate hexacoordination and functioning of rice Hb1.