Shedding light on ovothiol biosynthesis in marine metazoans

Ovothiol, isolated from marine invertebrate eggs, is considered one of the most powerful antioxidant with potential for drug development. However, its biological functions in marine organisms still represent a matter of debate. In sea urchins, the most accepted view is that ovothiol protects the eggs by the high oxidative burst at fertilization. In this work we address the role of ovothiol during sea urchin development to give new insights on ovothiol biosynthesis in metazoans. The gene involved in ovothiol biosynthesis OvoA was identified in Paracentrotus lividus genome (PlOvoA). PlOvoA embryo expression significantly increased at the pluteus stage and was up-regulated by metals at concentrations mimicking polluted sea-water and by cyclic toxic algal blooms, leading to ovothiol biosynthesis. In silico analyses of the PlOvoA upstream region revealed metal and stress responsive elements. Structural protein models highlighted conserved active site residues likely responsible for ovothiol biosynthesis. Phylogenetic analyses indicated that OvoA evolved in most marine metazoans and was lost in bony vertebrates during the transition from the aquatic to terrestrial environment. These results highlight the crucial role of OvoA in protecting embryos released in seawater from environmental cues, thus allowing the survival under different conditions.

Thanks to its chemical features, ovothiol has been receiving increasing interest for its potential therapeutic use in humans. Some ovothiol analogues were synthetized and their antioxidant properties studied in in vitro experiments 23 . Among these, 1-methyl-2-(3-trifluoromethylphenyl)-4-mercaptoimidazole has been shown to be a potent agent in mammalian cerebral protection 24 . Moreover, ovothiol A isolated from sea urchin eggs has been recently reported to induce autophagy in liver carcinoma cell lines, suggesting a potential role in regulating cancer cell growth 25 . In addition, the enzyme which catalyzes the first step of ovothiol biosynthesis, 5-histidylcysteine sulfoxide synthase (OvoA), recently characterized from the bacterium Erwinia tasmaniensis and the pathogenic protist Trypanosoma cruzi 26 , has been proposed as a target of anti-infective therapy. This enzyme is an iron (II) dependent sulfoxide synthase, which catalyzes the S-conjunction of cysteine with histidine in the presence of molecular oxygen and S-adenosyl methionine. Subsequently, a sulfoxide lyase cleaves the sulfur carbon bond in the cysteine residue leaving 5-thiohistidine, finally methylated at the imidazole ring [27][28][29] . OvoA enzyme is the homologous of an iron dependent sulfoxide synthase involved in ergothioneine biosynthesis (EgtB), a 2-thiohistidine, produced in some actinobacteria, cyanobacteria, and fungi 30 . The key step in the biosynthesis of these thiohistidines is the oxidative insertion of a sulfur atom into the C2 or C5 position of the imidazole ring of L-histidine. This represents an unusual and completely novel reaction with respect to known mechanisms for enzymatic C-S bond formation 31 . In silico analysis has revealed the presence of OvoA mainly in bacteria, protists, fungi, and some animals 26 .
The aim of the present work is to address the biological role of OvoA in marine metazoans. To this aim, sea urchins represent exceptional models for evolutionary, ecotoxicological, and developmental biology studies. We have identified and analyzed the nucleotide sequence of the gene locus for OvoA in the sea urchin P. lividus in comparison with the orthologous one in S. purpuratus. We have investigated the role of OvoA in mediating P. lividus embryos response to environmental stress factors, including heavy metals and natural algal blooms. Evolutionary analysis of OvoA in metazoans has allowed us to hypothesize the involvement of ovothiol in protecting eggs and embryos released in the seawater column from environmental cues.

Results
PlOvoA and SpOvoA gene loci structure. The OvoA genes in the sea urchins P. lividus and S. purpuratus genomes were identified by tblastx using as sequence query the previously characterized OvoA of E. tasmaniensis and T. cruzi. SpOvoA is a multi-exonic gene with 19 exons. PlOvoA shows a similar gene structure (Fig. 1). The in silico analysis of upstream regions of these genes revealed the presence of Metal Responsive Element (MRE) and Stress Responsive Element (SRE). In the 12 kb upstream region of SpOvoA, including the first intron, 9 putative MREs and 14 putative SREs were identified (Fig. 1). In PlOvoA orthologous region, besides 2 MREs and 2 SREs, 1 activator protein 1 (AP-1) and 1 cAMP-Responsive Element (CRE) were found. However, low quality of the assembled sequences of P. lividus genome in that area did not allow to rule out the presence of other responsive cis-elements.
PlOvoA expression profile during embryo development and response to environmental factors. Initial experiments were performed to examine by quantitative RT-PCR PlOvoA transcription during P. lividus development from virgin eggs to pluteus stage larvae ( Fig. 2A), using the zinc-finger transcription factor Pl-Z12-1 as reference gene, whose expression remained constant in all developmental stages. Immediately after fertilization, PlOvoA expression was slightly reduced compared to the highest value in the unfertilized eggs and then strongly decreased at the early and the swimming blastula stages. Thereafter, PlOvoA expression levels significantly increased at the pluteus stage.
In subsequent experiments, sea urchin fertilized eggs were treated with Cd 2+ and Mn 2+ at concentrations mimicking polluted seawater and previously shown to affect development 32 . PlOvoA expression was followed during embryo development until the pluteus stage with respect to untreated embryos (Fig. 2B). A significant increase of PlOvoA mRNA was observed at the swimming blastula stage after exposure to 1 and 5 μ M Cd 2+ , whereas at the early blastula and prism stages, the gene transcription was unaffected. On the contrary, at the pluteus stage PlOvoA mRNA was strongly down-regulated after exposure to 5 and 10 μ M Cd 2+ . In the presence of Mn 2+ , the relative expression ratio of PlOvoA at the swimming blastula stage was close to the minimum value (2) considered significant. No appreciable gene regulation was observed in other developmental stages.
To understand if the metal-induced up-regulation of PlOvoA gene reflected also an increase in ovothiol biosynthesis, the amount of the thiohistidine was determined by HPLC analysis at the swimming blastula stage, when PlOvoA transcription increased, and at the subsequent prism stage, after exposure of fertilized eggs to the highest effective concentration of Cd 2+ (5 μ M). Ovothiol levels significantly increased at the prism stage after exposure to the metal, with respect to the control, whereas no significant increase was detected at the swimming blastula stage (Fig. 2C).
In order to assess PlOvoA gene regulation by other environmental stress factors, we examined its expression in the offspring of sea urchins cyclically exposed to a natural toxic bloom of the dinoflagellate Ostreopsis cf. ovata occurred at the Gaiola Marine Protected Area in the Gulf of Naples. Females were collected in October during Data are reported as a fold difference in PlOvoA expression levels, compared to control (mean ± SEM), embryos developed in seawater without metals. Fold differences greater than ± 2 (see dotted horizontal guidelines at values of 2 and − 2) were considered significant. (C) Ovothiol levels in developing embryos after metal treatment. FE were treated with 5 μ M Cd 2+ and SB and Pr were examined for ovothiol production. Data are reported as μ mol ovothiol/mg of embryos. Results are representative of 3 independent experiments and expressed as means ± SEM, and analyzed by unpaired t-test. *represents the significance respect to the control, **P < 0.01. D. PlOvoA expression analysis in developing embryos after maternal exposure to O. cf ovata bloom. Sea urchin females, exposed to Ostreopsis cf. ovata bloom in July at Gaiola site, were collected in October and fertilized. Different developmental stages (EB, SB, Pr and Pl) were examined for PlOvoA expression. Data (mean ± SEM) are reported as a fold difference in PlOvoA expression levels, compared to control embryos derived from sea urchins collected at control site. Fold differences greater than ± 2 were considered significant. the reproductive season, after O. cf ovata bloom, which occurred in July. The eggs were fertilized in the laboratory and the offspring analyzed for mRNA expression of the PlOvoA gene at the different developmental stages. A significant increase of PlOvoA mRNA was observed at the early and swimming blastula stages with respect to embryos derived from females harvested at the control site, Castel dell'Ovo, which is known to harbor O. cf. ovata at negligible concentrations (Fig. 2D).

Identification and characterization of PlOvoA protein. The open reading frames of PlOvoA and
SpOvoA code for a protein of 763 and 767 amino acids (aa), respectively. The two proteins are highly similar (83% of sequence identity) and share 34-35% of sequence identity with E. tasmaniensis OvoA. Both protein primary structures contain a DinB superfamily domain (36-176 aa, PlOvoA) and a Formylglycine-generating sulfatase (FGE-sulfatase) domain (211-491 aa) in the N-terminal region and a S-adenosylmethionine -methyltransferase (SAM-transferase) domain (572-743) in the C-terminal region (Fig. 3A). The DinB superfamily domain contains the conserved HX3HXE putative iron-binding motif. Moreover, the residues (581-587, 602-603, 661-663, 680) considered to be involved in the formation of SAM-binding site are conserved with the bacterial orthologous gene. OvoA shares two protein domains with EgtB, the DinB superfamily domain and the FGE-sulfatase domain but differs for the additional C-terminal putative SAM-transferase domain 26 (Fig. 3B).
The structure of PlOvoA, excluding the C-terminal additional domain, has been modeled and compared with that of EtgB from Mycobacterium thermoresistibile (MtEgtB, 33% sequence similarity). As expected, the overall structure of PlOvoA is similar to that of MtEgtB (Cα root mean square deviation is 0.77 Å using 397 atoms) (Fig. 4A,B).
The main structural differences are located in the surface loops and in regions corresponding to insertions/ deletions, which adopt different conformations in MtEgtB and PlOvoA. Inspection of the active site regions of the two proteins reveals that, in analogy to what observed for MtEgtB, the active site of PlOvoA is located in a cleft at the bottom of a wide tunnel (Fig. 4C,D). The three conserved histidine residues His74, His169 and His173 (His51, His134 and His138 in MtEtgB) coordinate the catalytic Fe ion; the Tyr residue in position 442 (Tyr377 in MtEtgB) is in close proximity to the metallic center. The putative residues involved in the binding of MtEgtB 30 with its two substrates N-alpha-trimethyl histidine and gamma-glutamyl cysteine (i.e. Arg87, Arg90, Tyr380, Trp415, Asp416, Arg420) are not conserved in the structure of PlOvoA, and this probably accounts for the different products of the reactions catalyzed by the two enzymes. Interestingly, other residues as Met108, Tyr442, Phe445, His480 and Phe481 are easily recognized in the structural model of the PlOvoA active site (Fig. 4D), and are conserved in all marine metazoans OvoAs (see alignments in Fig. 3 and Supplementary Fig. 1).

Evolutionary analysis of OvoA in metazoans.
The evolutionary history of OvoA was surveyed in metazoans. We found OvoA with a high degree of conservation in porifera, placozoa, cnidaria (anthozoa), protostomes (annelida and mollusca) and deuterostomes. Among protostomes, OvoA gene was lost in nematoda (Caenorhabditis elegans) and arthropoda (Drosophila melanogaster), at least according to the genomic and transcriptomic data available at date. Among deuterostomes, we found OvoA in ambulacraria, i.e. echinodermata (P. lividus and S. purpuratus) and hemichordata (Saccoglossus kowalevskii), and in the chordate phylum in cephalochordata (Branchiostoma floridae), urochordata (Ciona intestinalis), and in chondrichthyes (Callorhinchus milii). Interestingly, in S. kowalevskii we found two genes coding for OvoA of 813 aa and 795 aa, respectively. The two proteins share 59% sequence identity. However, we could not identify OvoA orthologous in bony vertebrates, neither in Actinopterygii nor Sarcopterygii ( Fig. 5 and Table 1), suggesting that a second independent event of gene loss took place in the ancestor of Osteichthyes fishes.

Discussion
In this work, we investigated functional and evolutionary aspects related to the gene and the enzyme responsible for the biosynthesis of ovothiol, the methylated thiohistidine, first isolated in sea urchin eggs 5 . Although in the past decades ovothiol has been suggested to be involved in different biological processes 14,15,18,19 , its functions still remain a matter of debate. Here, combining biochemical and molecular studies with bioinformatics tools for available genomes screening, we provide new insights into the role of ovothiol in metazoans.
Since the eighties it has been suggested that ovothiol mainly acted to protect sea urchin eggs from the high oxidative burst at fertilization. Here, we demonstrate that ovothiol plays a role also during development. Indeed, our data showed that, after consumption of the PlOvoA maternal transcript, the levels of PlOvoA mRNA significantly increased at the pluteus stage. This profile overlapped that of ovothiol content in P. lividus early development 10,33 , suggesting a functional role for ovothiol also in the embryo. The PlOvoA gene expression regulated by stress factors, together with in silico analysis of the gene upstream region, demonstrated that in sea urchins OvoA acts as a metal and general stress responsive gene during development. The up-regulation of PlOvoA at the swimming blastula stage after exposure to 1 and 5 μ M Cd 2+ indicated that, under these conditions, the gene was regulated by the metal-induced stress, as suggested by the presence of MREs in the promoter region. At the prism stage, no significant regulation of gene expression was observed. However, the significant increase in ovothiol levels at this stage, suggested that, following PlOvoA mRNA increase at the swimming blastula stage, the synthesis of the metabolite was induced at the later prism stage. Finally, at the pluteus stage, PlOvoA was strongly down-regulated, indicating that the production of high amounts of ovothiol, at 5 μ M Cd 2+ , may exert a negative feedback regulation on its synthesis, as already reported for other metabolic enzymes, including the one catalyzing the first step of glutathione biosynthesis 34 . At higher Cd 2+ concentrations (10 μ M) a negative regulation of PlOvoA occurred probably related to the marked toxicity of the metal. The increase in ovothiol content after Cd 2+ treatment closely resembled the glutathione formation observed in marine organisms in response to metal and natural toxins exposure [35][36][37][38] to counteract the increasing production of ROS. This indicates that ovothiol behaves similarly to the universal cellular antioxidant glutathione. Accordingly, the slight regulation of PlOvoA following Mn 2+ treatment may be due to the essential biological role of the metal, which does not cause an increase of ROS levels in sea urchin embryos 39 . PlOvoA metal-regulation may be mediated by the highly homologous metal responsive transcription factor-1 (MTF-1), which accumulates in the nucleus upon heavy metal exposure. MTF-1    Table 1 for protein sequence ID).
binding to the core consensus TGCRCNC sequence in MRE induces the expression of metallothioneins and other genes involved in metal homeostasis 40 .
As a consequence of metal-dependent ROS formation, Cd 2+ can also modulate antioxidant gene transcription through SREs 38 . However, these elements usually mediate general stress responsive gene activation, including toxin-mediated stress. Our findings indicate that when sea urchins were cyclically exposed to toxic O. cf. ovata bloom, PlOvoA gene was strongly up-regulated during early development of the offspring, thus suggesting a general stress activation of PlOvoA. Interestingly, the promoter region of PlOvoA gene contains, besides SRE, also AP-1 and CRE sequences, mainly involved in the antioxidant response. The coexistence of these 3 regulatory elements further supported the view that PlOvoA is a stress and antioxidant responsive gene 41 .
Previous studies on the characterization of OvoA focused mainly on the kinetics and substrate specificities of the enzyme isolated from some microorganisms 26 . In this work, we identified the intron-exon composition of the gene and the protein primary structure of OvoA in sea urchins, as the first example in metazoans. The active site structural model of PlOvoA in comparison with that of MtEgtB allowed us to identify putative binding residues for the two specific substrates of OvoA, cysteine and histidine. These residues were highly conserved in OvoA from the other metazoans, suggesting their direct involvement in the catalysis of ovothiol production, instead of ergothioneine. These amino acidic substitutions can be considered particularly relevant to the evolutionary divergence of EgtB and OvoA, which likely accounts for adaption to different functional niches for the two enzymes.
Concerning the evolutionary history of OvoA in metazoans, the two independent gene loss events, one in nematodes and arthropods, at least in terrestrial species, and the other one in the ancestor of osteichthyes fishes suggested a crucial function of ovothiol biosynthesis in marine environment. Unfortunately, the lack of available genomes from marine nematodes and arthropods, did not allow to rule out the presence of OvoA in these invertebrates. The occurrence of OvoA in most of broadcast spawners ( Table 1), suggested that this antioxidant gene protects the eggs and the embryos released in seawater from special external factors responsible for ROS production in marine organisms. We propose that invertebrate broadcast spawners (i.e. sea urchins) release gametes in the seawater column, where eggs and embryos are directly exposed to soluble natural toxins or heavy metals. These factors can induce oxidative stress in the embryos, thus activating transcription factors, which bind to MRE and/or SRE to promote OvoA transcription. Finally, OvoA enzyme catalyzes the synthesis of ovothiol to counteract redox unbalance (Fig. 6). However, the occurrence of OvoA also in invertebrates with internal fertilization, e.g. some cephalopods or some vertebrates, such as cartilaginous fishes, which can be oviparous or ovoviviparous (the embryos develop within the mother's body), suggests a broader function for the enzyme. OvoA may play a key role in maintaining cellular redox homeostasis, i.e. in response to the different oxygen pressures. Indeed, in the early stages of development, embryo respiration occurs through a direct exchange of oxygen with the surrounding environment, whereas in the later stages, gas exchanges are guaranteed by fetal membranes (yolk sac, allantoid and placenta), usually highly vascularized in fishes. Therefore, we hypothesize that the loss of OvoA in bony fishes may be related to the appearance and development of the swim bladder and of more efficient systems of respiration and blood circulation, which allow a more controlled oxygen exchange with respect to the cartilaginous fishes and other invertebrates, which have conserved OvoA. Interestingly, the swim bladder, responsible for gas volume regulation and pressure, has a common origin with the lung in dipnoi and tetrapods. Moreover, among bony fishes, coelacanths and lungfishes are considered more closely related to tetrapods 42 than to teleost fishes and represent key species to elucidate the transition from ancestral aquatic vertebrates to terrestrial animals 43 . The transition from the aquatic-to-aerobic environment requires accumulation of pre-adaptive changes in the genomes of ancestral species to adapt to an air-based environment, such as lung development and adaptation to a different oxygen pressure, as well as loss of genes no more necessary to the new conditions. Therefore, OvoA could represent one of the genes lost during the water to land transition, for the concomitance emergence of other mechanisms of oxygen pressure regulation and exchange. In agreement with this hypothesis, the presence in S. kowalevskii of two genes coding for OvoA may suggest a double specialization for the OvoA to adapt to different conditions of oxygen pressure. Indeed, the hemichordate lives in burrows on the sea-bed in anaerobic conditions and during low tide can be exposed to aerobic ones 44 .
In conclusion, this study demonstrates for the first time that OvoA can be regarded as a metal and a general stress responsive gene, and ovothiol a new biomarker of stress conditions in marine organisms. The OvoA evolutionary history in metazoans suggests the involvement of ovothiol in regulating redox homeostasis in organisms that have to survive in a marine habitat, where oxygen pressure and solubility is far from being similar to air-based environment.  Gamete collection. Sea urchins were collected during the breeding season by SCUBA divers in the Gulf of Naples, transported in an insulated box to the laboratory within 1 h after collection, and maintained in tanks with circulating seawater. To perform metal experiments, animals were acclimated for a minimum of 10 days until use. Fertilization was carried out as described in Migliaccio and collaborators (2014) 32 . Animals collected at Gaiola for the in situ study were fertilized with sperm from a pool of males collected in a control area, not affected by toxic microalgae blooms.

Methods
Embryo culture, treatments and morphological analysis. Embryos (150 eggs/ml) were allowed to develop at 18 ± 2 °C in a controlled temperature chamber at 12:12 light:dark cycle. Metal treatment was performed 5 min from fertilization, by adding Cd 2+ or Mn 2+ ions under careful agitation. Nominal concentrations were 1, 5 and 10 μ M for Cd 2+ (cadmium chloride-Sigma-Aldrich) and 36 μ M for Mn 2+ (manganese chloride tetrahydrate-Sigma-Aldrich). Embryos derived from fertilization of females harvested at Gaiola were reared in the field (sampling) water. Experiments were performed in triplicate using the eggs collected from three different females. Gene expression by Real Time qPCR. For real time qPCR experiments the data from each cDNA sample were normalized using the gene encoding for the zinc-finger transcription factor Pl-Z12-1 as well-assessed endogenous control during P. lividus development 32,45 . For Pl-Z12-1, we used primers reported in the previous paper (32). In the case of OvoA, specific primers were designed on the basis of nucleotide sequence using Primer 3: forward primer 5′-AGGTCAGCATGGACATAGCC-3′, reverse primer: 5′-CCTCAGCCGACTTCAAGAAC-3′. The amplified fragment (156 bp in length) using Taq High Fidelity PCR System (Roche) was purified from agarose gel using QIAquick Gel extraction kit (Qiagen) and specificity of PCR product was checked by DNA sequencing. Specificity of the amplification reaction was verified by melting curve analysis. The efficiency of primer pair was calculated according to standard methods curves using the equation E = 10 −1/slope . Five serial dilutions were set up to determine Ct values and reaction efficiencies for the primer pair. Standard curve was generated for oligonucleotide pair using the Ct values versus the logarithm of dilution factor. PCR efficiencies were calculated for control and target gene and were found to be about 2. Diluted cDNA was used as a template in a reaction containing a final concentration of 0.3 μ M for each primer and 1x FastStart SYBR Green master mix (total volume of 10 μ l). PCR amplifications were performed in a ViiA TM 7 Real Time PCR System (Applied Biosystems) thermal cycler using the following thermal profile: 95 °C for 10 min, one cycle for cDNA denaturation; 95 °C for 15 sec and 60 °C for 1 min, 40 cycles for amplification; 72 °C for 5 min, one cycle for final elongation; one cycle for melting curve analysis (from 60 °C to 95 °C) to verify the presence of a single product. Each assay included a no-template control for each primer pair. To capture intra-assay variability all Real Time qPCR reactions were carried out in triplicate. Fluorescence was measured using ViiA TM 7 Software (Applied Biosystems). The expression of OvoA gene was analyzed and internally normalized against Pl-Z12-1 using Relative Expression Software Tool software (REST) based on Pfaffl method (2002) 46 . Relative expression ratios above two cycles were considered significant.

Determination of ovothiol A disulphide concentration in sea urchin embryos. Embryos,
harvested at different developmental stages were homogenized in ethanol-1 M HCl 80:20 v/v (1 mL) and left overnight at room temperature under stirring in the air. After centrifugation at 14000 g for 15 min at 4 °C, the supernatant was recovered. The pellet was washed three times with acidic ethanol and the combined supernatants, concentrated to a small volume, were extracted three times with the same volume of ethyl ether freed from peroxide by passage over alumina column. The aqueous layer was concentrated to small volume and loaded onto a Dowex 50WX2, column (1 cm × 2 cm). Elution was sequentially carried out with water, 0.1 M and 0.5 M HCl. The column was then eluted with 4 M HCl and the collected fractions were monitored spectrophotometrically in the 200-350 nm range. Fractions exhibiting the UV spectrum typical of ovothiol were collected, concentrated to a volume of 100 μ L for HPLC analysis on a column Phenomenex Synergi Sphereclone (25 cm × 0.46 cm, 5 µm particle size) with an isocratic elution in 1% formic acid taken to pH 4.5 with ammonia.
Statistical analysis. Data are presented as means ± SEM and analyzed by One-way ANOVA (P < 0.05) with Turkey's Multiple Comparison Test and Two-way ANOVA (P < 0.05), with Bonferroni post hoc test as reported in figure legends. Statistical analysis was performed using GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego California USA). The results of biochemical experiments were reported as means ± SEM and analyzed by unpaired t-test for comparison between the groups. P < 0.05 was considered statistically significant. For Real Time qPCR analysis, results were reported as means ± SEM and significance was tested using the "Pair Wise Fixed Reallocation Randomisation Test", developed by REST software 47  Bioinformatic analysis. OvoA genomic sequences or transcripts were downloaded from the following sources: S. purpuratus from Spbase (www.spbase.org), P. lividus from BioDev (www.octopus.obs-vlfr.fr), Ciona intestinalis and other invertebrate species from Ensemble (www.ensemble.org), Callorhinchus milii from the Elephant shark Genome Project (www.sharkgenome.imcb.a-star.edu.sg), Platynereis dumerilii was from http:// jekely-lab.tuebingen.mpg.de/blast/ 48 . Amino acid sequence of OvoA from different sources was found by tblastx and Blastp homology search using as template SpOvoA gene sequence. OvoA protein sequence alignments were obtained by ClustalW 49 . 12 kb upstream of S. purpuratus and P. lividus OvoA genes were scanned for searching the following metal responsive DNA cis-elements: Metal Response Element [MRE consensus TGC(A/G) CNC] 50  Structural modelling. The model of PlOvoA structure was obtained using the structure of the Ergothioneine biosynthesis enzyme EgtB from Mycobacterium thermoresistibile as starting model (MtEtgB, PDB code 4 × 8B 30 , and the SWISS-MODEL server (http://swissmodel.expasy.org/). The program DeepView-Swiss-PdbViewer 51 was used to include iron and iron-coordinated water molecules close to the conserved His residues in the active site. The position of the conserved His residues was assumed to be similar to that of MtEgtB. DeepView-Swiss-PdbViewer was also used to minimize the energy of the structure and calculate the root mean square deviation from the starting model. The figures were done with Pymol (www.pymol.org).